Datasets:

health360

/

Healix-Shot

like 4

Follow

Health 360 6

C O G N I T I V E T H E R A P Y O F S U B S T A N C E A B U S E -•̂ .̂ '̂ ySi:'yf̂ '̂ ''->->*' 'r... ^'i-'Mt .#''a^K"'-M>' • M ^ i € • ̂ ^. M'^::. A A R O N T . B E C K • ^ • r f - F R E D D . W R I G H T C O R Y F . N E W H A N B R U C E S . L I E S E C O G N I T I V E T H E R A P Y O F S U B S T A N C E A B U S E C o g n i t i v e T h e r a p y o f S u b s t a n c e A b u s e Aaron T. Beck, M.D. Fred D. Wright, Ed.D. Cory F. N e w m a n , Ph.D. Bruce S. Liese, Ph.D. T H E G U I L F O R D PRESS N e w York London ©1993 The Guilford Press A Division of Guilford PubHcations, Inc. 72 Spring Street, New York, N Y 10012 www.guilford.com All rights reserved No part of this book may be reproduced, stored in a retrieval s or transmitted, in any form or by any means, electronic, mechanical, photocopying, microfilming, recording, or otherwise, without written permission from the Publisher. Printed in the United States of America. This book is printed on acid-free paper. Last digit is print number: 9 Library of Congress Cataloging-in-Publication Data Cognitive therapy of substance abuse / Aaron T. Beck . . [et al p. cm. Includes bibliographical references and index. ISBN 0-89862-115-1 (he.) ISBN 1-57230-659-9 (pbk.) 1. Substance Abuse—Treatment. 2. Cognitive therapy. I. Beck, Aaron T. [DNLM: 1. Cognitive therapy—methods. 2. Substance Abuse— therapy. W M 270 C6765 1993] RC564.C623 1993 616.86'0651—dc20 DNLM/DLC for Library of Congress 93-5208 CIP To Phyllis, G w e n , Jane, and Ziana P r e f a c e s L-#ubstance abuse is widely recognized as a serious social and legal problem. In fact, the use of illegal drugs may be responsible for more than 2 5 % of property crimes and 1 5 % of violent crimes. Financial losses related to these crimes have been estimated at $1.7 billion per year. Homicides are also strongly linked to drug dealing. Approximately 1 4 % of homicides per year are causally related to drugs. The costs for criminal justice activities directed against drug trafficking on the federal level were approximately $2.5 billion in 1988, compared to $1.76 billion spent in 1986. There are also many health problems caused by these drugs. Alcohol can damage almost every body organ, including the heart, brain, liver, and stomach. Illegal drugs such as cocaine can have a serious effect on the neurological, cardiovascular, and respiratory systems. Cigarettes can cause cancer, heart disease, and more. The most widely used and abused drug in the world is alcohol. In the United States, two-thirds of the population drink alcohol. About ten out of a hundred people have problems with alcohol so serious that they can be considered "alcoholic" or "alcohol-dependent." (Interestingly, this 1 0 % of Americans buys and drinks more than half of the alcoholic beverages!) At least 14 million Americans take illegal drugs every month. Dur­ ing "peak months" this number climbs to more than 25 million users. Some experts have estimated that approximately 2.3% of Americans over 12 years of age have a problem with illegal drugs serious enough to warrant drug treatment. To a large degree, we have tried to put a halt to drug abuse by making drugs illegal. For example, heroin and cocaine are presently illegal in the United States. Cigarette smoking is becoming increas­ ingly proscribed. At one time we tried to stop alcoholism by legal Vll via Preface mechanisms (i.e., prohibition). Obviously, these methods will never make substances completely unavailable. Not all people who use drugs become addicted to them, although many people have asked themselves, "Am I [or is someone else] an alcoholic [or a substance abuser]?" The American Psychiatric Associ­ ation has defined the addictions very specifically. In fact, the official term for an addiction is "substance dependence." There are some specific signs of substance dependence, including (1) heavy use of the substance, (2) continued use even though it may cause problems to the person, (3) tolerance, and (4) withdrawal symptoms. Cultural and historical factors are implicated in substance use and abuse. The patterns and consequences of drug use have been influ­ enced by historical developments, which have had positive and neg­ ative effects. Two centuries ago, the extraction of pure chemicals from plant materials created more powerful medicinal agents. The inven­ tion of the hypodermic needle in the middle of the nineteenth cen­ tury was also a medical boon, which, on the other hand, allowed drug users to circumvent the body's natural biological controls consisting of bitter taste and slow absorption through the digestive tract. Many synthetic drugs developed in the twentieth century had medical appli­ cation but created further opportunities for abuse and addiction. In short, any activity that affects the reward mechanisms of the brain may lead to compulsive, self-defeating behavior. Social, environmental, and personality factors have affected sub­ stance use and abuse in ways that go far beyond the simple pharma­ cological properties of these agents. Alcoholism, for example, is preva­ lent among certain ethnic groups and practically absent among others, such as the Mormons, who require abstinence for group acceptance. O n the other hand, other social subgroups may condition group accep­ tance on using or drinking. The social milieu may influence using. Soldiers used illegal drugs extensively in Vietnam but, for the most part, relinquished heavy drug use after returning home. Impoverished environments have been shown in both animal experiments and human studies to lead to addiction. As pointed out by Peele, the com­ m o n denominator is the lack of other opportunities for satisfaction. Finally, our clinical experiences have indicated that addicted indi­ viduals have certain clusters of addictive attitudes that make them abusers rather than users. Successful treatment depends on clinicians' effectiveness in deal­ ing with these addictive potentials. And what form will this care take? As pointed out by Marc Galanter, president of the American Academy of Psychiatrists in Alcoholism and Addiction, the long-term efficacy of new pharmacological treatments is open to question. "Tricyclics, Preface ix dopaminergic agents, and carbamazapine for cocaine abusers have yet to be substantiated as a vehicle for continuing care. For opiates, naltrexone and buprenorphine offer only a modest niche in the do­ main that was traditionally occupied by methadone maintenance. Intervention in GABAergic transmission may hold promise for alco- hoHsm, but that promise is far from clinical application" (Galanter, 1993, pp. 1-2). W e have written this book in response to the ever-growing need to formulate and test cost-effective treatments for substance abuse dis­ orders, problems that seem to be multiplying in the population in spite of society's best efforts at international interdiction and domes­ tic control and education. W e believe that cognitive therapy, a well- documented and demonstrably efficacious treatment model, can be a major boon to meeting this pressing need. At one time, "drug abuse rehabilitation counseling" was regarded as a specialty area in the field of psychotherapy—now it is apparent that almost all who engage in clinical practice will encounter patients who use and abuse drugs. Therefore, it would be desirable for all mental health professionals to receive some sort of routine training and education in the social and psychological phenomena that com­ prise the addiction disorders. Our volume is intended to provide a thorough, detailed set of methods that can be of immediate use to therapists and counselors—regardless of the amount of experience they might have had with cognitive therapy, or in the field of addictions. Toward this end, we have strived to make our model and our proce­ dures as specific and complete as possible. W e certainly recommend that those who read this book also read the many valuable sources we have cited in the text. Nevertheless, our intention in writing Cog­ nitive Therapy of Substance Abuse has been to provide a convenient, centralized source that is comprehensive, teachable, and testable. Although advances in the field have been made in the form of pharmacological interventions (e.g., antabuse, methadone, and nal­ trexone), 12-step support groups (e.g.. Alcoholics Anonymous, Nar­ cotics Anonymous, and Cocaine Anonymous), and social-learning models and programs (relapse prevention, rational recovery, etc.), each of these approaches has posed problems that limit its respective poten­ tial efficacy. For example, pharmacological interventions have pro­ duced promising short-term data but are fraught with compliance and long-term maintenance difficultieŝ atients may not take their chem­ ical agonists and antagonists, and they are prone to relapse when the medications are discontinued. Twelve-step programs provide valuable social support and consistent guidance principles for individuals who voluntarily join and faithfully attend the program meetings, but can- X Preface not address the needs of those who will not enter the programs or who drop out. Social-learning approaches provide sophisticated models of substance abuse and relapse, and hold promise to produce and accumulate empirical data, but thus far the resultant treatments (with very few exceptions) have been less well described than the theories that gave rise to them. Although the cognitive approach that we have explicated is most closely related to the social-learning theories of substance abuse, we want to emphasize that we find value in all of the aforementioned treatment modalities. Cognitive therapy is not in "opposition" to 12- step or psychobiological models of substance abuse. W e have found that these alternative treatment systems may be complementary to our procedures. Many of the substance abuse patients that we treat at the Center for Cognitive Therapy concurrently attend Narcotics Anony­ mous and similar 12-step groups. Other patients take the full spec­ trum of pharmacologic agents, from antidepressants to antabuse, under strict medical guidance. The individualized conceptualization of patients' belief systems and the long-term coping skills (to deal with everyday life concerns, as well as to manage cravings and urges spe­ cific to drug use) that cognitive therapy provides for patients can mesh well with medication and 12-step meetings. The main variable that seems to influence whether or not patients avail themselves of all of these treatment opportunities (once they have been presented to the patients in a feasible manner) is not the practical compatibility of the treatments, but rather the attitudes of the treatment providers] At present, an earlier draft of this book is serving as a treatment manual in a National Institute on Drug Abuse collaborative, multisite study on the respective efficacy of cognitive therapy, supportive- expressive therapy, and general drug counseling. Data obtained from this project will help us to answer two important questions: (1) Does Cognitive Therapy of Substance Abuse succeed as a manual for the train­ ing of competent cognitive therapists for patients with addictions? and (2) Do patients who receive the treatment outlined in the text make demonstrable and lasting gains? In order to answer these questions, therapists are provided with intensive supervision (note: the authors of this text serve in that role), complete with competency and adher­ ence ratings on a regular basis; treatment is not confounded with adjunct medications, urinalyses are routinely conducted, and a host of measures other than drug monitoring per se are being administered and evaluated (to examine changes in mood and global adaptational functioning). Drug abuse is a sociological problem as well as a psychological issue. Factors such as poverty and lack of adequate educational and Preface xi vocational opportunities play a role in the epidemic. However, we believe that it is harmful to assume that low socioeconomic status patients cannot be treated as effectively as those of higher socioeco­ nomic status. While social change is desirable, individual change is not necessarily dependent on it. W e are optimistic that cognitive ther­ apy can serve as an important individual-focused treatment in today's society, and that the data will support this. A c k n o

w l e d g m e n t s w w e would like to offer our thanks to our highly esteemed colleagues in the field of substance abuse treatment and research, Drs. Dan Baker, Lino Covi, Tom Horvath, Jerome Piatt, Hal Urschel, David Wilson, and Emmett Velten, for their extremely help­ ful insights and suggestions on earlier versions of this manuscript. Special thanks are due Dr. Kevin Kuehlwein, an important member of our own cognitive therapy team in Philadelphia, for his thorough evaluations and editorial work on many of the chapters in this book. The input of all of the above has been invaluable during the course of this project. W e would also like to offer our thanks and apprecia­ tion to Tina Inforzato, who did yeoman work in typing this volume, and its many revisions. Without her tireless efforts, this volume would still be "on the drawing board." Xll C o n t e n t s CHAPTER 1 Overview of Substance Abuse 1 CHAPTER 2 Cognitive Model of Addiction 22 CHAPTER 3 Theory and Therapy of Addiction 42 CHAPTER 4 The Therapeutic Relationship and Its Problems 54 CHAPTER 5 Formulation of the Case 80 CHAPTER 6 Structure of the Therapy Session 97 CHAPTER 7 Educating Patients in the Cognitive Model 112 CHAPTER 8 Setting Goals 121 CHAPTER 9 Techniques of Cognitive Therapy 135 CHAPTER 10 Dealing with Craving/Urges 157 CHAPTER 11 Focus on Beliefs 169 CHAPTER 12 Managing General Life Problems 187 CHAPTER 13 Crisis Intervention 211 CHAPTER 14 Therapy of Depression in Addicted Individuals 226 CHAPTER 15 Anger and Anxiety 242 CHAPTER 16 Concomitant Personality Disorders 268 CHAPTER 17 Relapse Prevention in the Cognitive Therapy 292 of Substance Abuse Append ixes 311 References 331 Index 347 xm C H A P T E R 1 O v e r v i e w o f S u b s t a n c e A b u s e T .he fabric of America is profoundly affected by problems of subsAtahnc(e abuse. They are problems that directly affect those millions of Americans who suffer from substance abuse and indirectly touch the lives of millions more in the larger social and vocational networks around them. One in every ten adults in this country has a serious alcohol problem (Institute of Medicine [lOM], 1987) and at least one in four is addicted to nicotine (Centers for Disease Control [CDC], 1991a). Approximately 1 in 35 Americans over the age of 12 abuses illicit drugs (lOM, 1990a). This level of substance abuse has profound social, medical, and psychological ramifications on both the individual and the larger societal levels. The C D C (1991b), for example, estimate that approximately 434,000 people in this coun­ try die each year as a result of cigarette smoking, and many thou­ sands also die as a result of alcoholism (lOM, 1987) and/or illicit drug abuse (lOM, 1990a). It must be emphasized, however, that substance abuse spans many more areas and the toll taken is far greater than these simple mortality figures convey. In this introductory chapter we set the stage for the cognitive therapy of substance abuse. W e begin with an overview of psycho-v active substances and substance abuse, we briefly review the history of psychoactive substance use, we describe the most commonly used and abused psychoactive substances, we discuss cognitive models for understanding substance abuse and relapse, and we scan traditional methods for treating substance abuse. 2 COGNITIVE THERAPY OF SUBSTANCE ABUSE BACKGROUND: PSYCHOACTIVE SUBSTANCES AND SUBSTANCE ABUSE Psychoactive substances are chemicals that affect the central nervous system, altering the user's thoughts, moods, and/or behaviors. The revised third edition of the Diagnostic and Statistical Manual of Mental Disorders (DSM-lIl-R; American Psychiatric Associa­ tion [APA], 1987) categorizes psychoactive substances into 10 classes: alcohol; amphetamines or similarly acting sympathomimetics; can­ nabis; cocaine; hallucinogens; inhalants; nicotine; opioids; phencycli- dine (PCP) or similarly acting arylcyclohexylamines; and sedatives, hypnotics, or anxiolytics. Each of these substances has unique prop­ erties and effects. Some substances that are abused have low addic­ tive potential (e.g., hallucinogens), while others have high addictive potential (e.g., crack cocaine). Some are typically smoked (e.g., nico­ tine, cannabis, and crack cocaine); others are ingested orally (e.g., hallucinogens and sedatives); while still others are taken intranasally (e.g., powdered cocaine and inhalants). Some drugs lead the user to feel "up" or energized (e.g., amphetamines and cocaine); some cause the user to feel "down" or relaxed (e.g., sedatives, hypnotics, and anxiolytics); while others (e.g, alcohol and nicotine) simultaneously have both effects on the user. DSM-III-R distinguishes between substance abuse and dependence. Abuse is defined as a maladaptive pattern of psychoactive substance use while dependence (considered more serious than abuse) is defined as "impaired control of use" (i.e., physiological addiction). In this volume, we do not go to great lengths to emphasize this distinction. Instead, we view any pattern of psychoactive substance use as prob­ lematic and requiring intervention if it results in adverse social, voca­ tional, legal, medical, or interpersonal consequences, regardless of whether the abuser experiences physiological tolerance or withdrawal. Further, although we caution against an all-or-none view of addic­ tion and recovery, and although we acknowledge that some patients seem to be more successful at engaging in controlled, moderate sub­ stance use than are others, we advocate a program of treatment that strives for abstinence. In this manner we maximize the patients' chances of maintaining an able and responsible lifestyle, reduce the risk of relapse, and avoid giving patients the false impression that we view a mere reduction in drug use as the optimal outcome. History of Psychoactive Substance Use Psychoactive substances have been used by most cul­ tures since prehistoric times (Westermeyer, 1991). In fact, for centuries Overview 3 psychoactive substances have served many individual and social func tions. O n an individual level, they have provided stimulation, relief from adverse emotional states and uncomfortable physical symptoms, and altered states of consciousness. O n a social level, psychoactive substances have facilitated religious rituals, ceremonies, and medical functions. Egyptian and Chinese opiate use was evident from the earliest writings of these people (Westermeyer, 1991). Marijuana was referenced in India "as far back as the second millennium B.C." (Brecher, 1972, p. 397). Evidence of Mayan, Aztec, and Incan medici­ nal and ritual drug use was evident from their statues and from draw­ ings on their buildings and pottery (Karan, Haller, & SchnoU, 1991; Westermeyer, 1991). Alcohol use goes back to paleolithic times (Good­ win, 1981) and Mesopotamian civilization gave one of the earliest clinical descriptions of intoxication and hangover cures. In modern times the World Health Organization (WHO) has been concerned about drug and alcohol abuse problems on a worldwide scale (Grant, 1986). As early as 1968 the W H O conducted an interna­ tional study of drug use in youth (Cameron, 1968), and in a more recent study (Smart, Murray, & Arif, 1988) drug abuse and preven­ tion programs in 29 countries were reviewed. However, Smart and his colleagues concluded from their review that "the seriousness of the drug problem is well recognized in some countries but not in oth­ ers" (p. 16). Presently the W H O is addressing the issue of alcohol- related problems by developing an international secondary preven­ tion protocol (Babor, Korner, Wilber, & Good, 1987). Drug policies in the United States have been profoundly affected by historical and sociocultural attitudes regarding psychoactive drugs on a spectrum from less restrictive (e.g., libertarian) to more restrictive (i.e., criminal). Between the late 1700s and the late 1800s, for example, psychoactive dnigs (especially narcotics) were widely used in the United States. In fact, Musto (1991) reported that opium and cocaine were legally available during this time from "the local dmggist." A Consumers Union report (Brecher, 1972) described the nineteenth century as "a dope fiend's paradise" due to such minimal restrictions. In the late 1800s and the early 1900s, medical conceptualizations of addiction began to develop, however, influenced to some extent by Dr. Benjamin Rush's (1790) ear­ lier interest in the course of addictions. Magnus Huss, a Swedish physi­ cian, first used the term "alcoholism" in 1849 (lOM, 1990b). At the same time (late 1800s and early 1900s), criminalization of drug use was m- creasingly becoming U.S. policy. In the 1960s and 1970s, however, atti­ tudes about drugs became less restrictive as U.S. sociopolitical attitudes generally became more liberal. Simultaneously, the disease model of addictions was gaining widespread acceptance, partly due to the work ofjellinek(1960). 4 COGNITIVE THERAPY OF SUBSTANCE ABUSE Since the 1980s, the United States has again become less toler­ ant and more restrictive about drugs. At least two explanations can account for this phenomenon: (1) The negative effects of drugs on individuals, families, and society have become more apparent with increased use, and (2) sociopolitical attitudes in the United States generally have become more conservative. At the same time, however, there is increasing controversy about the disease model of addiction ("Current Disease model," 1992; Fingarette, 1988; Peele & Brodsky, with Arnold, 1991) and the criminalization of psychoactive substances (R. L. Miller, 1991). The Most Commonly Used Drugs Alcohol Alcohol is simultaneously a chemical, a beverage, and a drug that "powerfully modifies the functioning of the nervous sys­ tem" (Levin, 1990, p. 1). Approximately 1 0 % of Americans in the United States have a serious drinking problem; 6 0 % are light to mod­ erate drinkers; and the remaining 3 0 % of adults in the United States do not consume any alcohol. Alcohol abuse, however, accounts for approximately 8 1 % of hospitalizations for substance abuse disorders (lOM, 1987). Remarkably, half the alcohol consumed in this country is consumed by the 1 0 % who are heavy drinkers. A larger percentage of men than women drink and a greater percentage of men than women are heavy drinkers. Alcohol initially acts as a general anesthetic, interfering with subtle functions of thought, reason, and judgment (Miller & Munoz, 1976). As blood alcohol concentration (BAC) increases, however, the effects become more intense until gross motor functioning is also affected. At still higher BAC levels, sleep is induced, and ultimately death may occur as a result of respiratory depression. "Alcohol affects almost every organ system in the body either directly or indirectly" (National Institute of Alcohol Abuse and Alco­ holism [NIAAA], 1990, p. 107). Thus with chronic use, alcohol can cause serious multiple medical problems, including damage to the liver, pancreas, gastrointestinal tract, cardiovascular system, immune system, endocrine system, and nervous system. Alcohol has also been strongly linked to the leading causes of accidental death in the United States: motor vehicle accident, falls, and fire-related injuries. Further­ more, approximately 3 0 % of suicides and half of all homicides are alcohol related (lOM, 1987), and estimates of annual deaths related to alcohol use range between 69,000 and 200,000 per year (lOM, Overview 5 1987). In addition, a significant percentage of both violent and non violent crimes are committed under the influence of alcohol (cf. McCord, 1992). Chronic alcohol use can also have other profound negative social consequences, including loss of career, friends, and family. A great deal of physical and sexual abuse, for example, is related to the intoxicated state of the offender (Clayton, 1992; Frances & Miller, 1991; Harstone & Hansen, 1984), and general family dys­ function often is associated with the alcoholism of one or more adult members (Heath & Stanton, 1991). Medical complications can even reach insidiously into the next generation, in that maternal drinking during pregnancy can cause fetal alcohol syndrome and other seri­ ous birth defects. In fact, "prenatal alcohol exposure is one of the leading known causes of mental retardation in the western world" (NIAAA, 1990, p. 139). Illicit Drugs According to the lOM (1990a), at least 14 million persons consume illicit drugs monthly. During peak months this fig­ ures climbs to more than 25 million users. It is estimated that approxi­ mately 2.3% of the U.S. population over 12 years old has an illicit drug problem sufficiently serious to warrant treatment. This statistic is substantially higher, however, for individuals who are incarcerated (33%) or on parole or probation (25%). W h e n these people are included in the epidemiologic data, the estimate of illicit drug use problems in the overall population increases to 2.7%. Regarding

the social costs of illicit drug abuse, it is estimated that more than 2 5 % of property crimes and 1 5 % of violent crimes are related to illicit drug use by the criminal. Financial losses related to these crimes have been estimated at $1.7 billion per year. Homicides are also strongly linked to activities surrounding drug dealing. Approx­ imately 1 4 % of homicides per year are causally related to drugs. The costs for criminal justice activities directed against drug trafficking on the federal level were approximately $2.5 billion in 1988, com­ pared to $1.76 billion spent in 1986. In the following sections we present brief descriptions of the three most commonly used illicit drugs: marijuana, cocaine, and the opioids. In 1972, a Consumers Union report identified marijuana as the fourth most popular psychoactive drug in the world, after caffeine, nicotine, and alcohol (Brecher, 1972, p. 402). Although marijuanas use has declined since its peak in 1979, it still remains the most widely used illicit drug in Western society (APA, 1987; Weiss & Millman, 1991). 6 COGNITIVE THERAPY OF SUBSTANCE ABUSE Marijuana is typically smoked, although it can also be ingested. According to Weiss and Millman (1991), in spite of its generally sedat­ ing effects, marijuana's psychoactive effects in the user are quite varied, "profoundly dependent upon the personality of the user, his or her expectation, and the setting" (p. 160). The health effects of marijuana have been widely debated and remain quite controversial, probably due to the inconsistent effects of the drug on the individual user and across different users. For some time marijuana was considered relatively safe and nonaddictive (Brecher, 1972). Presently, however, it is associated with multiple adverse physical and psychological effects, including labile affect and depression, amotivational syndrome, impaired short-term memory, and pulmonary disease (Weiss & Millman, 1991). According to DSM- III-R, marijuana dependence is characterized by heavy use of the drug (e.g., daily) with substantial impairment. Marijuana dependence also puts one at risk for other psychological problems, as those who are dependent on cannabis are also likely polysubstance abusers or afflicted with other psychiatric disorders (APA, 1987; Weiss & Millman, 1991). Cocaine is a major central nervous system stimulant that produces euphoria, alertness, and a sense of well-being. It may also lower anxi­ ety and social inhibitions while increasing energy, self-esteem, and sexuality. Presently cocaine is among the most widely used illicit drugs. In fact, cocaine use increased in 1991, "despite the Bush adminis­ tration's three-year war against drugs" (Mental Health Report, 1992, p. 5). Clearly, for many people the positive short-term physiological and psychological effects of cocaine maladaptively supersede the dan­ gers associated with acquiring and using the drug. According to Gawin and EUinwood (1988), "The pursuit of this direct, pharmacologically based euphoria becomes so dominant that the user is apt to ignore signs of mounting personal disaster" (p. 1174). Cocaine is an alkaloid (as are caffeine and nicotine) which is extracted from the coca leaf. In its pure form, raw coca leaves can be chewed, although this practice is generally limited to native popula­ tions in the cocaine-producing countries (APA, 1987). In the United States, cocaine is most commonly taken intrana­ sally (i.e., snorted or "tooted") in the powder form of cocaine hydro­ chloride. In this form, the user pours the powder on a hard surface and then arranges it into "lines," one of which is snorted into each nostril (Karan et al., 1991). In powdered form, cocaine hydrochloride can also be mixed with water and administered by intravenous injec­ tion. This process is known as "shooting" or "mainlining" (Karan et al., 1991). Intravenous injection of cocaine results in intense sub­ jective and physiologic effects within 30 seconds Oones, 1987). Overview 7 Cocaine can also be smoked as a paste or in alkaloid form (i.e., "freebased"). In this form it also produces its effects within seconds. Crack cocaine (named for the sound made by the cocaine as it is freebased) is the currently popular form of freebase which is sold in relatively inexpensive, prepackaged, and ready-to-use small doses (Karan et al., 1991). According to Karan et al. (1991), low-cost crack, approximately $2-$ 10 per vial, "has been widely available on the streets in many American cities since 1985" (p. 125), making it easily within the financial grasp of most teenagers and even the impover­ ished. Adding to this high availability is the especially troublesome fact that crack cocaine produces an enormously intense and almost instant high. Crack cocaine is, therefore, extremely addictive, lead­ ing to significant impairment in life functioning after only a few weeks' use on average (Gawin & EUinwood, 1988; Smart, 1991), much faster than, for example, intranasal usage of cocaine. These charac­ teristics of crack cocaine make it especially prone to rapid increase in the prevalence of its abuse. Indeed, many observers suggest that cocaine use has already reached epidemic levels (Weinstein, Gottheil, & Sterling, 1992). In the popular press, for example, a graphic biographical Reader's Digest article describes cocaine as "the devil within" (Ola & D'Aulaire, 1991). This contrasts starkly with the glorification of cocaine in movies and songs of the 1970s and early 1980s, when cocaine was seen as the drug of choice of the affluent and powerful. In the scientific litera­ ture, Gawin and EUinwood (1988) explain that "believing that the drug was safe, millions of people tried cocaine, and cocaine abuse exploded" (p. 1173). These authors report that 1 5 % of Americans have tried cocaine, and 3 million people had abused cocaine regularly by 1986, resulting in "more than five times the number addicted to heroin" (p. 1173). Smart and Adlaf (1990) report also that an increasing num­ ber of cocaine abusers have sought treatment since the 1980s. Cohen (1991) attributes the "cocaine outbreak" to supply factors (e.g., low cost, availability, and high profitability), external factors (e.g., peer pressure and media portrayals of drug usage), internal factors (e.g., hedonism, sociopathy, depression, and life stress), and intrinsic drug factors (e.g., "the pharmacologic imperative"). Strikingly, cocaine abuse occurs and persists in spite of dramatic medical problems that are associated with its use: central nervous system damage, cardiac arrest, stroke, respiratory collapse, severe hypertension, exacerbation of chronic diseases, infection, and psychiatric complications (Estroff, 1987). Because cocaine abuse research has produced fewer pharma­ cological treatment alternatives than has research on some other illicit drugs such as heroin (Alterman, O'Brien, & McLellan, 1991; Covi, Baker, & Hess, 1990; Stine, 1992), and because of the extent and 8 COGNITIVE THERAPY OF SUBSTANCE ABUSE severity of cocaine-related problems, we have placed proportionatel greater emphasis on cocaine and crack cocaine than on other drugs in this treatment manual. The opioids, including heroin, methadone, and codeine, are drugs that pharmacologically resemble morphine. Drugs in this class pro­ duce feelings of euphoria, relaxation, and mood elevation. They also have the potential for reducing pain, anxiety, aggression, and sexual drives (lOM, 1990a), and are considered highly addictive. According to Thomason and Dilts (1991): Opioids have the capacity to commandeer all of an individual's attention, resources, and energy, and to focus these exclusively on obtaining the next dose at any cost. This vicious cycle repeats itself every few hours, 24 hours a day, 365 days a year, for years on end. Comprehending the implications of opioid abuse shocks and stag­ gers the inquiring mind. (p. 103) Although the use of pharmacologic agonists such as methadone (and antagonists such as naltrexone) traditionally has represented an important component of treatment in the heroin abuser, methadone itself is unfortunately subject to various forms of abuse (e.g., black market dealings or use with other drugs). Further, many heroin abusers find methadone to be inferior to the "real stuff," leading to high noncompliance and dropout (Grabowski, Stitzer, & Henningfield, 1984) rates with these programs. Therefore, we posit that pharmaco­ logic approaches (even for heroin) represent an incomplete treatment strategy unless utilized in combination with psychosocial approaches such as support groups and cognitive therapy. Nicotine Cigarette smoking is by far the single most prevent­ able cause of death in the United States. In fact, it has been estimated that 434,000 people died in 1988 due to cigarette smoking (CDC, 1991b). This figure includes those who died of cancer, lung disease, heart disease, house fires caused by careless smoking, and renal and pancreatic disease. Approximately 49.4 million Americans (28.1%) are regular cigarette smokers (CDC, 1991a), despite the fact that cigarette smoking is known to be a leading cause of morbidity and mortality in this country. Since the mid-1970s, however, the number of smokers has admit­ tedly decreased steadily. Historically, more men than women have smoked; however, a higher proportion of men than women have also quit smoking. It has thus been projected that by the year 1995, more Overview 9 women than men will be smokers. Ironically, in spite of cigarettes' historical and advertising linkage with status, wealth, and desirabil­ ity, it is increasingly the case that the socially disadvantaged are over- represented as smokers. The number of minorities, poor, and less educated people who smoke, for example, has been disproportionately higher than those who do not smoke, and this trend is expected to continue (Pierce, Flore, & Novotny, 1989). Nicotine is the psychopharmacologically addictive ingredient in cigarettes. As mentioned earlier, nicotine dependence is included in DSM-III-R, along with the dependence on other psychoactive sub­ stances (alcohol, opiates, cocaine, etc.). Not surprisingly, we have found the addictive process in cigarette smoking to be analogous to the addictive process involved in the other psychoactive substances. Therefore, although nicotine addiction is not associated with the same degree of social, vocational, and legal consequences as is addiction to illicit drugs, its medical hazards and the fact that early-life regular smoking often leads to addiction to "harder" substances (Henning­ field, Clayton, & Pollin, 1990) make it an important area for mental health intervention. Although this volume focuses relatively little on methods specifically geared to smoking cessation, we believe that the same principles of assessment and treatment (e.g., coping with crav­ ings and modifying beliefs) that we outline in this book are highly applicable to the patient addicted to nicotine. Polysubstance Abuse Individuals abusing one psychoactive substance are likely to be simultaneously abusing another substance. In fact, between 2 0 % and 3 0 % of alcoholics in the general public and approxi­ mately 8 0 % in treatment programs are dependent on at least one other drug. A prevalent combination is alcohol, marijuana, and cocaine (N. S. Miller, 1991, p. 198). N S Miller (1991) explains that polysubstance abuse occurs tor multiple reasons. For example, some drugs enhance the effects of other drugs, while some drugs are used to avoid unwanted side effects of other drugs. Some drugs are used to treat drug withdrawal effects of other drugs and, similarly, some drugs are used as substihites for other '^'"^The medical and psychological correlates of polysubstance abus are numerous (N. S. Miller, 1991). They include problems associated with each individual drug (e.g., liver and heart disease associated with alcohol abuse), as well as those more commonly associated with multiple substances (e.g., interaction-induced overdose). 10 COGNITIVE THERAPY OF SUBSTANCE ABUSE Dual Diagnosis: Substance Abuse and Other Psychiatric Disorders The coexistence of substance abuse with other psy­ chiatric disorders is also very common (e.g., Ananth et al., 1989; Brown, Ridgely, Pepper, Levine, & Ryglewicz, 1989; Bunt, Galanter, Lifshutz, & Castaneda, 1990; Davis, 1984; Hesselbrock, Meyer, & Kenner, 1985; Kranzler & Liebowitz, 1988; Nace, Saxon, & Shore, 1986; Nathan, 1988; Penick et al., 1984; Regier et al., 1990; Ross, Glaser, & Germanson, 1988; Schneier & Siris, 1987). In a survey of more than 20,000 Americans conducted by Regier et al. (1990) it was found that individuals with psychiatric disorders were 2.7 times as likely to have alcohol or other drug problems, compared to those without psychi­ atric disorders. In fact, 3 7 % of individuals with substance use disor­ ders had coexisting Axis I mental disorders. From these data it appears that individuals with substance abuse problems should benefit most from therapeutic interventions that simultaneously address their other psychiatric disorders. Cognitive therapy is ideally suited for these individuals, since it has been devel­ oped and tested on patients with depression, anxiety, and personal­ ity disorders (see Hollon & Beck, in press, for a most recent compre­ hensive review).

In fact, an important component of cognitive therapy involves the case conceptualization (Persons, 1989), defined as the evaluation and integration of historical information, psychiatric diag­ nosis, cognitive profile, and other aspects of functioning (see Chap­ ter 5, this volume, for a detailed description of the case conceptual­ ization). When a coexisting psychiatric syndrome is found to exist with a dmg or alcohol abuse patient, for example, the therapist focuses simultaneously on substance abuse and the symptoms of the psychi­ atric syndrome as well as on any factors of interaction (see Chapters 14, 15, and 16, this volume, for more on the treatment of patients with dual diagnoses). RELAPSE P R E V E N T I O N Substance abuse and dependence are characterized both by remission and by relapse. In a classic review by Hunt, Barnett, and Branch (1971) it was found that heroin, nicotine, and alcohol were all associated with similar high rates and patterns of relapse (p. 455; see Figure 1.1). These investigators found that two-thirds of individuals treated had relapsed within 3 months. Many investigators have speculated about the meaning of these findings, most inferring Overview 11 RELAPSE RATE OVER TIME •----•HEROIN ASMOKING OALCOHOL 2weeksJ 6 101112 MONTHS FIGURE 1.1. Relapse rate over time for heroin, smoking, and alcohol addic­ tion. From Hunt, Barnett, and Branch (1971), p. 456. Copyright 1971 by Clinical Psychology Publishing Co., Inc. Reprinted by permission. that they reflect common processes that underlie the addictions. In fact, since the publication of Hunt et al.'s (1971) data, addiction experts have focused on developing and testing comprehensive models of addiction that include all the psychoactive substances, as well as gambling and binge eating. Marlatt and his colleagues (Brownell, Marlatt, Lichtenstein, & Wilson, 1986; Marlatt, 1978; Marlatt, 1982; Marlatt & Gordon, 1985) have made an important contribution to the addiction literature with their cognitive-behavioral model of relapse prevention. According to Marlatt and Gordon's (1985) model (see Figure 1.2), individuals view themselves as having a sense of perceived control or self-efficacy. When they are faced with high-risk situations, this sense is threat­ ened. High-risk situations for the drug abuse patient might include negative or positive emotional or physical states, interpersonal con­ flicts, social pressure, or exposure to drug cues. Individuals faced with high-risk situations must respond with coping responses. Those who have effective coping responses develop increased self-efficacy, result­ ing in a decreased probabiHty of relapse. Those who have relatively fewer coping responses or none at all may experience decreased self- 12 COGNITIVE THERAPY OF SUBSTANCE ABUSE Coping Increased Decreased response self-efficacy probability of relapse Higii-risk situation Decreased Abstinence self-efficacy violation No Initial effect: Increased coping Positive use of probability response outcome substance Dissonance conflict of expectancies and relapse (for initial self-attribution effects of (guilt and perceived substance) loss of control) F I G U R E 1.2. Model of relapse process. From Mariatt and Gordon (1985), p. 38. Copyright 1985 by The Guilford Press. Reprinted by permission. efficacy and increased positive o u t c o m e expectancies about the effects of the drug, followed b y a "lapse" or initial use of a substance. This initial use mi g h t result in w h a t Marlatt calls a n Abstinence Violation Effect (AVE; i.e., perceived loss of control) and a n ultimately increased probability of relapse. T h e w o r k of Marlatt and his colleagues has h a d a profound effect o n knowledge about addictive behaviors. In fact, m o s t current text­ books o n addictions n o w deal with the issue of relapse prevention in some way. Although most of the work on relapse prevention has been generated within the cognitive-behavioral model (e.g., Chiauzzi, 1991), various 12-step programs (e.g.. Alcoholics Anonymous) and other advocates of the disease model have recently also increased their emphasis on relapse prevention (e.g., Gorski & Miller, 1986). MODELS OF ADDICTION Numerous theoretical models have been developed to explain addictive behaviors (see Baker, 1988; Blane & Leonard, 1987, for recent reviews). As previously mentioned, the dominant trend Overview 13 among addiction experts is toward developing comprehensive theo­ retical models that explain all addictions. Cognitive Models of Addiction A variety of related cognitive models of addiction have been developed and evaluated (e.g., Abrams & Niaura, 1987; Marlatt, 1978, 1985; McDermut, Haaga, & Shayne, 1991; Stacy, Newcomb, & Bentler, 1991; Tiffany, 1990; Wilson, 1987a, 1987b) since Bandura's (1969, 1977) classic presentations of cognitive social learning theory. Marlatt (1985) describes four cognitive processes related to addictions that reflect the cognitive models: self-efficacy, outcome expectancies, attributions of causality, and decision-making processes. Self-efficacy refers to one's judgment about one's ability to deal com­ petently with challenging or high-risk situations. Examples of high self-efficacy beliefs include the following: "1 can effectively cope with temptations to use drugs" or "1 can say 'no' to drugs." Examples of low self-efficacy beliefs might include the following: "I'm a slave to drugs," "I can't get through the day without drugs," or "I can't get what I want, so I might as well use drugs." Marlatt (1985) explains that low levels of self-efficacy are associated with relapse and high levels of self-efficacy are associated with abstinence. Marlatt (1985) also explains that self-efficacy increases as a function of success; to the extent that individuals effectively choose not to use drugs, they will experience an increased sense of self-efficacy, for example, believ­ ing that their sense of pride is greater than their need for a "high." Outcome expectancies refer to an individual's anticipation about the effects of an addictive substance or activity. Positive outcome expectancies might include the following beliefs: "It will feel great to party tonight," or "I won't feel so tense if I use." To the extent that one expects a greater positive than negative outcome from using drugs, one is likely to continue using. Attributions of causality refer to an individual's belief that drug use is attributable to internal or external factors. For example, an individual might believe the following: "Anybody who lives in m y neighborhood would be a drug user" (external factor), or "I am physi­ cally addicted to alcohol and m y body can't survive without it" (inter­ nal factor). Marlatt (1985) explains that such beliefs most likely would result in continued substance use, since the individual perceives his/ her use to be predestined and out of control. For example, the AVE is an individual's tendency to believe that he/she is unable to control substance use after an initial lapse. That is, the AVE occurs when an individual has had a "lapse" or "slip" (i.e., has used a drug after being 14 COGNITIVE THERAPY OF SUBSTANCE ABUSE abstinent for some time) and attributes this lapse to a "lack of w power" (i.e., an internal causal factor). Under such circumstances, this individual is likely to continue using, resulting in a full-blown relapse. This is analogous to Beck's (1976) description of all-or-none think­ ing; for example, "I've blown it, so I might as well keep using." Marlatt (1985) also describes substance abuse and relapse as a cognitive decision-making process. He demonstrates (with an amus­ ing example) that substance use is a result of multiple decisions (like forks in the road) which, depending on the decisions, may or may not lead to further substance use. He further explains that some deci­ sions initially appear to be irrelevant to substance use ("apparently irrelevant decisions"); however, these decisions ultimately may result in a greater likelihood of relapse because of their incremental push toward higher-risk situations. In his example, Marlatt "innocently" chooses to sit in the smoking section of an airplane after being absti­ nent from smoking for several months. As a result of this decision he is more vulnerable to relapse (by his exposure to other smokers, their smoke, and their offers of cigarettes to him). W e see this same phenomenon in patients who claim to have had every intention of remaining abstinent from alcohol and illicit drugs, only to bhthely accept an invitation to meet a friend at a local tavern, or to cavalierly choose to drive out of the way in order to go past a street corner where drugs are sold. When such patients lapse into alcohol and drug use, it is striking to see how they fail to realize the ways in which they set themselves up for a fall with their decisions that lead up to the actual using incident. Unfortunately, the cognitive models of substance abuse have not been integrated adequately into many addiction treatment programs (lOM, 1990a; Miller & Hester, 1985). This volume provides a focused, step-by-step treatment based on Beck's (1976) cognitive model. It is our hope that the chapters that follow will stimulate increased appli­ cation of this cognitive model to substance abuse treatment across treatment settings and modalities. The Motivation to Change Efforts to examine the treatment of addictions are incomplete without considering the issue of motivation. Miller and Rollnick (1991) address this issue, explaining that most addicts are genuinely ambivalent about changing (rather than resistant, weak- willed, or characterologically flawed). The authors view motivation as a "state of readiness or eagerness to change, which may fluctuate from one time or situation to another" (p. 14). Overview 15 Prochaska, DiClemente, and Norcross (1992) provide a compre­ hensive model for conceptualizing patients' motivation for change. In their work, Prochaska et al. (1992) identify five stages of change: precontemplation, contemplation, preparation, action, and mainte­ nance. In the precontemplation stage, individuals are least concerned with overcoming their problems and they are least motivated to change problematic behaviors. In the contemplation stage individuals are willing to examine the problems associated with their substance use and consider the implications of change, although they may not take any constructive action. They are also likely to respond more positively to confrontation and education, although they may still be ambivalent. In the preparation stage, patients wish to make actual changes and therefore desire help with their problems, although they may feel at a loss as to how to do what is necessary to become drug free. In the action stage individuals have made a commitment to change and they have begun to actually modify behaviors. Prochaska et al. (1992) point out that this is a particularly stressful stage, which may require considerable therapist support and encouragement. In the maintenance stage individuals attempt to continue the process begun in the contemplation and action stages. In recent years, with so much emphasis placed on relapse prevention, the maintenance stage has received increased attention. Prochaska and DiClemente (1986) caution that the process of change is very complex. They explain that "most individuals do not progress linearly through the stages of change" (p. 5). Alternatively, they offer a "revolving door model" (p. 6), based on the assumption that individuals make multiple revolutions around the circle of stages prior to achieving their long-term goals. Furthermore, they observe that some individuals "get stuck" in the earlier stages of change. In the words of Prochaska and DiClemente (1986), "Therapy with addictive behaviors can progress most smoothly if both the client and the therapist are focusing on the same stage of change" (p. 6). To use nicotine dependence as an example, a smoker in the precontemplation stage will benefit little from advice about specific strategies for quit­ ting smoking. The same smoker, however, might respond well to general questions about health maintenance, which might lead to a discussion of the health effects of smoking, which might lead further to a discussion of the benefits of quitting, which eventually might lead to a discussion of specific strategies. It is clear that the field can benefit from an understanding of what makes a patient ready to seek help (Tucker & Sobell, 1992). The Prochaska et al. (1992) stage model is a useful heuristic. However, it is important to note that patients in a precontemplative 16 COGNITIVE THERAPY OF SUBSTANCE ABUSE Stage of change are not impossible to treat (especially if they ar court order to attend therapy). Conversely, patients in the action phase or maintenance phases are not guaranteed to succeed in treatment. The same degrees of vigilance and commitment are required of the cognitive therapist regardless of the substance abuse patient's stage of change. Treatment Outcome Goals Some models of addiction (e.g., Alcoholics Anony­

mous and other disease-model programs) view total abstinence as the only acceptable goal of treatment. Proponents of these models view addiction as an all-or-nothing phenomenon, with any use seen as pathological and abstinence considered a state of "recovering" (rather than "recovered"). Alternatively, proponents of cognitive-behavioral models are more likely to view light or moderate use (i.e., "controlled drinking") as an acceptable goal of treatment in some cases. At one time controlled drinking was extremely controversial (Marlatt, 1983). Presently, however, it is generally accepted that the goals of treatment should vary according to the patient's needs, prob­ lems, and previous response to treatment. Sobell, Sobell, Bogardis, Leo, and Skinner (1992), for example, surveyed problem drinkers to deter­ mine their preference for self-selected versus therapist-selected treat­ ment goals (e.g., abstinence vs. controlled drinking). They found that most respondents preferred setting their own goals and believed that they would be more likely to achieve them; respondents with more serious drinking problems were even more likely to favor self-set goals. In general, we favor a collaborative approach in setting goals with patients. Therefore, to the extent that allowing severely addicted patients to set the modest goal of substance use reduction succeeds in getting otherwise resistant patients engaged in a more complete course of therapy, we are in favor of a controlled substance use approach. In the long run, however, we strongly advocate assisting patients in becoming drug- and alcohol-free. THE TREATMENT OF SUBSTANCE ABUSE A N D DEPENDENCE In reality, most substance abuse treatment programs are eclectic in theory and practice, and they include varying degrees of inpatient and outpatient services, 12-step program attendance, education, psychotherapy, family therapy, support groups, pharmaco- Overview 17 therapy, and so forth. In our view, cognitive therapy can be compat ible with any of these approaches. In fact, many of our drug and alcohol abuse patients attend support groups, have had inpatient detoxification, and take medication. The special strengths that cog­ nitive therapy adds to this battery of approaches are its emphasis on (1) the identification and modification of beliefs that exacerbate crav­ ings, (2) the amelioration of negative affective states (e.g., anger, anxi­ ety, and hopelessness) that often trigger drug use, (3) teaching patients to apply a battery of cognitive and behavioral skills and techniques, and not just willpower, to become and remain drug-free, and (4) help­ ing patients to go beyond abstinence to make fundamental positive changes in the ways they view themselves, their life, and their future, thus leading to new lifestyles. In the following section we present a brief overview of more tra­ ditional treatments of substance abuse and dependence. Alcoholism Treatment Miller and Hester (1980, 1986) have conducted exhaustive reviews of the alcoholism treatment literature. These authors have examined nine major classes of interventions. The four most common were pharmacotherapy, psychotherapy or counseling. Alcoholics Anonymous, and alcoholism education. The five less com­ monly employed approaches included family therapy, aversion thera­ pies, operant methods, controlled drinking, and broad spectrum treat­ ment. Miller and Hester (1986) conclude from their reviews that alco­ holism treatment is best approached as a two-stage process, requir­ ing different interventions at each stage. The first set of interventions should be focused on changing drinking behaviors to abstinence or moderation (e.g., behavioral self-control training). The second set of interventions should be focused on maintenance of sobriety (e.g., social skills training in order to increase confidence in relating to drug- free people). Miller and Hester (1986) also draw some disturbing conclusions, however, about the poor relationship between empirical research and traditional inpatient treatment approaches. Treatment methods that are supported by controlled research include aversion therapies, behav­ ioral self-control training, community reinforcement, marital and family therapy, social skills training, and stress management, whereas approaches actually currently employed as standard practice in alco­ holism programs include Alcoholics Anonymous, alcoholism educa­ tion, confrontation, disulfiram, group therapy, and individual coun- 18 COGNITIVE THERAPY OF SUBSTANCE ABUSE seling. They point out that there is little apparent overlap betwe these lists: Alcoholism treatment programs in the United States do not tend to use treatment methods that have been validated by controlled outcome studies. Furthermore, Miller and Hester (1986) point out that traditional inpatient treatment programs are very expensive, "despite clear evidence that they offer no advantage in overall effectiveness" (p. 163). Concurring in this, McLellan et al. (1992) note that stan­ dard detoxification and "28-day programs" (in spite of their high costs) are insufficient to deal with long-term issues. Clearly, to help drug and alcohol patients deal with more enduring issues, these treat­ ments need to be supplemented with ongoing outpatient treatment that focuses on attitude change and skills acquisition. The Institute of Medicine recently commissioned a National Acad­ emy of Sciences committee to make an exhaustive critical review of the research literature on treatment for alcohol problems (1990b). The committee discovered that interventions included "a broad range of activities that vary in content, duration, intensity, goals, setting, pro­ vider, and target population" (p. 86). The committee's assessment was that "no single treatment approach or modality has been demonstrated to be superior to all others" (p. 86). Its conclusions, published in Broadening the Base of Treatment for Alcohol Problems (1990a), included the following: 1. There is no single treatment approach that is effective for all persons with alcohol problems. 2. The provision of appropriate, specific treatment modalities can substantially improve outcome. 3. Brief interventions can be quite effective compared with no treatment, and they can be quite cost-effective compared with more intensive treatment. 4. Treatment of other life problems related to drinking can improve outcome in persons with alcohol problems. 5. Therapist characteristics are partial determinants of outcome. 6. Outcomes are determined in part by treatment process factors, posttreatment adjustment factors, the characteristics of indi­ viduals seeking treatment, the characteristics of their problems, and the interactions among these factors. 7. People who are treated for alcohol problems achieve a con­ tinuum of outcomes with respect to drinking behavior and alcohol problems and follow different courses of outcome. 8. Those who significantly reduce their level of alcohol consump­ tion or who become totally abstinent usually enjoy improve­ ment in other life areas, particularly as the period of reduced consumpfion becomes more extended (pp. 147-148). Overview 19 The findings of the Institute of Medicine (1990a) coupled with those of Miller and Hester (1986) make it apparent that there is still a profound need for effective alcoholism treatment interventions. It is hoped that the principles introduced in this text will be integrated into, and evaluated in, traditional treatment programs in order to move toward more effective and appropriate alcoholism treatment programs. Illicit Drug Treatment In addition to its report on alcohol treatment pro­ grams, the Institute of Medicine appointed a separate committee (1990a) to review the treatment of drug problems in the United States. Specifically, the committee divided treatments into four classifications: methadone maintenance, therapeutic communities, outpatient non- methadone programs, and chemical dependency programs. These findings (1990a) were similar to those of Miller and Hester (1986). The most empirically validated programs have been metha­ done maintenance clinics for opioid dependency. Some evidence also supported the efficacy of therapeutic communities and outpatient nonmethadone treatment. Nonetheless, "Chemical dependency is the treatment with the highest revenues, probably the second largest number of clients, and the smallest scientific basis for assessing its effectiveness" (lOM, 1990a, p. 18). The Institute of Medicine acknowl­ edges that most of the studies on methadone maintenance were con­ ducted in the 1970s and early 1980s, however. As a result, research has insufficiently addressed the growing cocaine problems in this country. By contrast, this volume will focus heavily on the cognitive therapy of cocaine and crack cocaine addiction. Smoking Cessation Interventions In a report published by the National Cancer Insti­ tute, Schwartz (1987) critically reviewed the literature on smoking cessation interventions. He divided the various methods into 10 cate­ gories: (1) self-care, (2) educational approaches/groups, (3) medica- fion, (4) nicotine chewing gum, (5) hypnosis, (6) acupuncture, (7) physician counseling, (8) risk factor preventive trials, (9) mass media and community programs, and (10) behavioral methods. Schwartz (1987) found considerable variability in cessation rates among these methods. Approximately 1 million Americans per year quit smoking, and most do so on their own through "self-care." In fact, three-fifths of all smokers would prefer to quit on their own, rather than seek group 20 COGNITIVE THERAPY OF SUBSTANCE ABUSE quit-smoking programs (Schwartz, 1987). There are many self-help aids for those wishing to quit smoking, including books, pamphlets, audio cassettes, drug store preparations, correspondence courses, and so forth. Almost all self-care efforts and aids involve some cognitive techniques. In fact, those who successfully quit on their own have higher levels of success expectancy and self-efficacy (areas strongly affected by cognitive interventions) than those who are unsuccess­ ful. Approximately 16%-20% of smokers who quit on their own are abstinent at 1 year (Schwartz, 1987). For those who wish to receive assistance with smoking cessation, there are nonprofit and commercial clinics and groups available. Most of these utilize cognitive methods, including education, self-monitor­ ing, and modifying attitudes about smoking. In a review of 46 group smoking cessation programs, Schwartz (1987) found median cessation rates ranging from 2 1 % to 36%, depending on the length of follow- up and the time the study was conducted. A number of medications have also been tried as aids to smok­ ing cessation over the years. These have included lobeline, mepro- bamate, amphetamines, anticholinergics, sedatives, tranquilizers, sym­ pathomimetics, anticonvulsants, buspirone, propranolol, clonidine, nicotine polacrilex, and most recently transdermal nicotine. Of these, the most promising medications have been those that replace the nicotine from cigarettes with prescription nicotine (i.e., nicotine gum and transdermal nicotine). In fact, the median cessation rates for nico­ tine gum at 6-month and 1-year follow-ups were 2 3 % and 11%. These rates were substantially higher when the gum was used in conjunc­ tion with cognitive-behavioral smoking cessation programs: 3 5 % and 2 9 % (Schwartz, 1987). At the time this book was being written, transdermal nicotine delivery systems had just been approved by the Food and Drug Administration. Hence, substantial field trials of these "patches" have not been conducted. Both hypnosis and acupuncture have been of interest to the gen­ eral public as smoking cessation techniques. However, empirical vali­ dation of these methods has been weak and hirther controlled studies are necessary prior to assuming their efficacy (Schwartz, 1987). SUMMARY Huge numbers of people in the United States are affected by substance abuse. Thousands of books and articles have been written and millions of dollars have been spent on research on the addictions. Nonetheless, there is a noticeable paucity of reliably effec- Overview 21 five substance abuse treatment strategies. For years, however, it h been noted that there are underlying cognitive processes common to the addictions. (Even Alcoholics Anonymous warns alcoholics about "stinkin' thinkin.'") W e believe strongly that understanding and work­ ing with these cognitive aspects more explicitly will help to resolve some of the uncertainty plaguing the field of substance use treatment. In the chapters that follow we strive for a high degree of speci­ ficity in describing the procedures that comprise this approach. A preliminary version of this book currently serves as a therapist manual in an ongoing National Institute on Drug Abuse pilot study compar­ ing cognitive therapy, supportive-expressive therapy, and general drug counseling treatment outcomes for cocaine abusers. Our hope is that Cognitive Therapy of Substance Abuse will continue to serve as a train­ ing guide for further clinical and empirical tests. C H A P T E R 2 C o g n i t i v e M o d e l o f A d d i c t i o n W R Y DO PEOPLE USE DRUGS (AND/OR ALCOHOL)? Some individuals are "generalists" and may use a wide variety of addictive substances almost randomly or depending on their availability. Others are "specialists" and their drug of choice may depend on its specific pharmacological properties as well as its social meanings (e.g., alcohol is often viewed as manly and associated with sports, whereas cocaine is associated with group acceptance and sexual activity). Cocaine may be used because of its stimulant properties- producing a rapid "high," for example. Similarly, amphetamines may be chosen as psychic energizers. In contrast,

barbiturates, benzodiaz­ epines, and alcohol may be preferred because of their relaxing effect and, perhaps, their presumed relief of inhibitions. Hallucinogens are attractive to some to relieve boredom and "expand consciousness." Most people addicted to cocaine have also abused other drugs and/or alcohol (N. S. Miller, 1991; Regier et al., 1990; Stimmel, 1991). There are numerous explanations for why people use—and become addicted to-psychotropic substances. In general, the process of addiction can be understood in terms of a few simple, perhaps obvious, formulas. A basic reason for starting on drugs or alcohol is to get pleasure, to experience the exhilaration of being high, and to share the excitement with one's companions who are also using (Stim­ mel, 1991). Further, there is the expectation that the drug cocaine, for example, will increase efficiency, improve fluency, and enhance creativity. 22 Cognitive Model of Addiction 23 How do people progress from recreational or casual use to regu­ lar use? In time, additional factors may contribute to becoming depen­ dent on the drug. Some people find that drug taking-for example, heroin, benzodiazepines (such as Valium), or barbiturates^rovides temporary relief from anxiety, tension, sadness, or boredom. These individuals soon develop the belief that they can weather the frustra­ tions and stresses of life better if they can turn to drugs and/or alco­ hol for a period of escape or oblivion. People with adverse life cir­ cumstances are more likely to become addicted than are those with more sources of satisfaction (Peele, 1985). For a while, real-life prob­ lems fade into insignificance and life itself seems more attractive. As one patient put it, "If I take coke, m y bad thoughts go away." Fur­ ther, people whose self-confidence is low may find that the drug or alcohol boosts their morale—in the short run. Finally, many individu­ als discover that using drugs provides new social groups in which the only requirement for admission and acceptance is that they are users. If drug using has so many advantages, why should we be con­ cerned with getting people off the "drug habit"? The profound impli­ cations of breaking the law by using illegal drugs (and selling them in order to support their habit) are so obvious that they do not need further elaboration. Regardless of whether the drugs are legal, such as alcohol, or illegal, substance abuse creates serious personal, social, and medical problems (Frances & Miller, 1991; Kosten & Kleber, 1992). A major problem is that the drug seems to take control of addicted individuals. Their goals, values, and attachments become subordinate to the drug using. They cannot manage their lives effectively. They become subject to a vicious cycle of craving, precipitous drops in mood, and greater distress that can be relieved immediately only by using drugs again. The web of external and internal problems leading to and, later, maintaining compulsive drug use is a defining characteristic of addic­ tion. Far from soothing life's pains, the drugs create a new set of prob­ lems-enormous financial outiays (for illegal drugs), threat of or achial loss of employment, and difficulties in important personal relation­ ships, such as marriage. The individual also becomes stigmatized by society-as a "lush" or a "junkie." Finally, of course, chronic use may cause serious medical problems and even death. As pointed out by Peele (1989), the compulsive use of psycho­ tropic agents depends on a wide variety of personal and social fac­ tors. If the environment is malevolent and there is group support for drug use-as in the case of U.S. soldiers in Vietnam-widespread drug use is more likely. W h e n the environment is comparatively less stress­ ful (as when veterans retiirn to civilian life), individuals do not con- 24 COGNITIVE THERAPY OF SUBSTANCE ABUSE tinue excessive use—except for those who had been heavy users prio to military service (Robins, Davis, & Goodwin, 1974). A number of characteristics distinguish addicted individuals from casual users. A major difference, as pointed out by Peele (1985), is that while addicted individuals subordinate important values to drug using, casual users prize other values more highly: family, friends, occupation, recreation, and economic security, to name a few. In addition, drug users may have certain characteristics, such as low frus­ tration tolerance, nonassertiveness, or poor impulse control, that make them more susceptible. Thus, psychological and social factors may be the determinative factors—rather than the pharmacological prop­ erties per se—in converting a drug user into a drug abuser. Support­ ing this hypothesis is the commonly encountered phenomenon in hospital settings where "patients who take opioids for acute pain or cancer pain rarely experience euphoria and even more rarely develop psychic dependence or addiction to the mood-altering effects of nar­ cotics" {Medical Letter on Drugs and Therapeutics, 1993, p. 5). If drug addiction were merely a biological process, we would not expect this to be the case. The sequence of using or drinking is illustrated in Figure 2.1. An addicted individual who is feeling anxious or low decides to have a smoke or a snort. The short-term relief is followed by delayed, longer- term negative consequences: problems about breaking the law, seri­ ous financial problems, family difficulties, and possibly medical prob­ lems. These problems lead to realistic fears of being apprehended, becoming bankrupt, losing a job, disrupting close relationships, and becoming ill. These fears generate more anxiety and lead to craving and further using or drinking to neutralize the anxiety. Thus, a vicious cycle is established. Many other kinds of vicious cycles, which are described in Chap­ ter 3 (this volume), may be created. These involve a number of psy­ chological factors such as low self-esteem, emotional distress, and hopelessness. W H Y NOT STOP IF DRUGS OR ALCOHOL CREATE PROBLEMS? By definition, addicts are people who have difficulty in stopping permanently. They may have started to use voluntarily, but they either do not believe that they can stop or they do not choose to stop voluntarily. At the first sign of medical, financial, or interper­ sonal problems, many users ignore, minimize, or deny the problems Cognitive Model of Addiction 25 Anxiety/Low Mood Using ^k- Flnanclal, Social, Medical Problems FIGURE 2.1. Simple model of vicious cycle. or attribute them to something other than drugs (e.g., they m a y blame their spouse for domestic problems). Others m a y be aware of the problems, but they evaluate the advantages of using as greater than the disadvantages. M u c h of this evaluation is based on avoiding a true assessment of the disadvantages (Gawin & EUinwood, 1988; Gawin & Kleber, 1988). As the problems increase, m a n y users become more ambivalent and begin to vacillate in their decision to use. One factor in maintaining drug use is the c o m m o n belief that withdrawing from the drug will produce intolerable side effects (Horvath, 1988, in press). However, these effects vary enormously from person to person—and from substance to substance—and the impact is greatly enhanced by the psychological meaning attached to the withdrawal symptoms. These meanings are often more salient than the actual adverse physiological sensations in determining the inten­ sity of withdrawal symptoms. Most cocaine abusers participating in detoxification programs, for example, feel better in the early stages after they stop using (Ziedonis, 1992). A major obstacle to eliminating using or drinking is the network of dysfunctional beliefs that center around the drugs or alcohol. Exam­ ples of these beliefs are: "I can't be happy unless I can use," and "I a m more in control w h e n I've had a few drinks." A n individual w h o is contemplating eliminating the use of drugs or alcohol m a y feel sad or anxious. Termination of reliance on drugs or alcohol is seen as a deprivation of satisfaction and solace or a threat to well-being and functioning Qennings, 1991). Stopping m a y mean, for some, remov­ ing the "security blanket" used to cushion dysphoria. Addicted individuals often try on their o w n to stop using or drink­ ing. However, w h e n they experience the craving (often stimulated by low m o o d or exposure to the drugs or related stimuli), they feel dis­ appointed if they restrain themselves from using or drinking. They perceive their feelings of disappointment and distress as intolerable; the thought, "1 can't stand this feeling," upsets them even more. Hence, they feel driven to yield to the craving in order to dispel the 26 COGNITFVE THERAPY OF SUBSTANCE ABUSE sense of loss and relieve their distress. Patients often have a cl of beliefs that seem to become stronger when they decide to stop using. These center around the anticipated deprivation: "If I can't use, I won't be able to bear the pain (or boredom)," "There is nothing left in life for me," "I will be unhappy", or "1 will lose m y friends." These beliefs are elaborated more in the section on low frustration toler­ ance (Chapter 15, this volume). Another set of beliefs centers around the addicted individual's sense of helplessness in controlling the craving: "The craving is too strong," "1 don't have the power to stop," or "Even if I do stop—1 will only start up again." These beliefs become self-fulfilling prophecies. Since the patients believe they are incapable of controlling their urges, they are less likely to try to control them and, thus, confirm their belief in their helplessness in overcoming their addiction. W H Y D O PEOPLE W A N T HELP? There are roughly five stages people go through in seeking help (Prochaska et al., 1992). In the precontemplative stage, they do not even acknowledge to themselves that they have a problem (or else they consider using more important than the problems it causes). In the contemplative stage, they are willing to consider their problems, but are still unlikely to stop using on their own. Individuals in the preparation stage intend to take action to cease their drug and alcohol use, but are uncertain about being able to follow through. In the action stage, patients behaviorally demonstrate a decrease in their drug-taking behaviors and a therapeutic modification in their drug-taking beliefs. Those who are successful enough to reach the maintenance stage have already taken great strides toward a drug-free and alcohol-free life, and are actively working to maintain consistency in this endeavor over a period of months and years. People come to therapy for a variety of reasons. Some users have been arrested for "dealing" or possession and are referred by the courts. Others see their lives deteriorating as a result of the financial, psychological, and interpersonal consequences of using or drinking. Still others are pressured by friends or family. By the time these patients are labeled drug abusers, addicts, or alcoholics, they have often hit a low point in terms of any combination of the following: health, social adjustment, employment and economic status, and psychologi­ cal well-being. Many people with drug and alcohol problems have tried repeat­ edly to "break the habit," only to relapse eventually. Others suffer from Cognitive Model of Addiction 27 a personality disorder (e.g., Mirin & Weiss, 1991; Nace, Davis, & Gaspari, 1991; Regier et al., 1990) and/or a psychiatric syndrome such as chronic anxiety (e.g., Kranzler & Liebowitz, 1988; LaBounty, Hat- sukami, Morgan, & Nelson, 1992; Walfish, Massey, & Krone, 1990) or depression (e.g., Hatsukami & Pickens, 1982; Rounsaville & Kleber, 1986). For some, drug use is simply a manifestation of their mani­ fold difficulties. For others, drugs represent a form of self-medication (Castaneda, Galanter, & Franco, 1989; Khantzian, 1985) to relieve their feelings of distress, sadness, or anxiety. Given the consequences of sustained drug use, it is important to consider the problem in terms of its sociological, interpersonal, and psychological dimensions, in addition to the strictly pharmacologi­ cal properties of drugs. In fact, substance abuse or addiction could be defined as compulsive use leading to a web of entanglement involv­ ing social, economic, and legal problems over which the patient no longer has control. Given their acknowledgment that they are addicted, many of these individuals come to the conclusion that the only way they can manage or even salvage their lives is to receive assistance, professional or otherwise. H O W C A N COGNITIVE THERAPY HELP? Cognitive therapy is

a system of psychotherapy that attempts to reduce excessive emotional reactions and self-defeating behavior by modifying the faulty or erroneous thinking and maladap­ tive beliefs that underlie these reactions (Beck, 1976; Beck, Rush, Shaw, & Emery, 1979). The approach to a particular patient is derived from a thorough concephialization of the particular case. The specific case formulation, in turn, is based on the cognitive model of that disorder. The thor­ ough case conceptualization, including the relationship of early life patterns to current problems, at the beginning stages of treatment differentiates cognitive therapy from some of the other forms of therapy. The approach is (1) collaborative (builds trust), (2) active, (3) based on open-ended questioning to a large degree, and (4) highly structured and focused. As applied to substance abuse, the cognitive approach helps indi­ viduals to come to grips with the problems leading to emotional dis­ tress and to gain a broader perspective on their reliance on drugs for pleasure and/or relief from discomfort. In addition, specific cognitive strategies help to reduce their urges and, at the same time, establish a stronger system of internal controls. Moreover, cognitive therapy 28 COGNITIVE THERAPY OF SUBSTANCE ABUSE can help patients to combat their depression, anxiety, or anger, w frequently fuels addictive behaviors. A major thrust of cognitive therapy of substance abuse is to help the patient in two ways: (1) to reduce the intensity and frequency of the urges by undermining the underlying beliefs, and (2) to teach the patient specific techniques for controlling or managing their urges. In a nutshell, the aim is to reduce the pressure and increase control. When the patient's addiction is related to a coexisting psychiatric disorder, that condition also needs to be addressed by the cognitive therapist. Cognitive therapy is carried out in several ways. The therapist helps the patient to examine the sequence of events leading to drug use and then to explore the patient's basic beliefs about the value of drugs, alcohol, and nicotine. At the same time, the therapist trains the patient to evaluate and consider the ways in which faulty think­ ing produces stress and distress. Therapists help patients to modify their thinking so that they can gain a better grasp of their realistic problems and can disregard pseudo-problems derived from their faulty thinking. In addition, through rehearsal and practice, patients are trained to build up a system of controls to apply when confronted with strong urges. The techniques the therapist uses include a painstaking evalua­ tion of the short-term and long-term benefits and disadvantages of using: the cost-benefits analysis (also called the advantages-disadvan­ tages analysis; see Chapters 9 and 10, this volume). The therapist also helps the patient to find more satisfactory ways of coping with real­ istic problems and unpleasant feelings without turning to drugs or alcohol for relief. They also work together to structure the patient's life so that other sources of pleasure are made available (cf. Havassy, Hall, & Wasserman, 1991). Since many patients have a low frustra­ tion tolerance (Ellis, Mclnerney, DiGiuseppe, & Yeager, 1988), they are shown how their self-defeating attitudes about themselves and their capabilities lead to overreacting when they encounter obstacles, delays, or thwarting (Chapter 15, this volume). The therapist also demonstrates how patients can approach these obstacles as problems to be solved rather than as barriers to their goals. Many patients who suffer from difficulties in asserting themselves in an appropriate way are likely to be dominated and even exploited by other people, and thus are prone to experience frequent impatience, anger, and disappointment. By learning new interpersonal skills, the patients are able to assert their rights more effectively. The same type of assertion can help them to refuse when others coax them to start Cognitive Model of Addiction 29 using. Refusal can take on a new meaning for them^standing up for themselves, putting long-term interests before short-term gains, and becoming desensitized to derogatory or profane epithets. One of the main features of cognitive therapy is the use of "Socratic questioning." By skillfully asking questions, the therapist leads the patient to examine areas that the patient has closed off from scrutiny, for example, the true frequency and quantity of drug use, the actual losses from the addiction, and the quality and effects on interpersonal relations. Also, questioning leads patients to generate options and solutions that they have not considered. Finally, this approach puts patients in the "questioning mode" (as opposed to the "automatic impulse" mode) so that they will start to evaluate more objectively their various attitudes and beliefs. In a sense, stopping drug use or drinking is a technical problem. The patients coming for help would like to stop using but they do not know how. Many of them have tried to stop many times but have been unsuccessful. Cognitive therapy provides them with tools that will enable them to stop and maintain the abstinence from drugs or to moderate their drinking and smoking. Moreover, they can apply these same useful techniques to their daily problems and thus have a more enjoyable, more fulfilling life. D O SUBSTANCE ABUSERS H A V E ADDITIONAL PSYCHIATRIC PROBLEMS? Many of the patients seeking—or referred for—treat­ ment of addictions have a "dual diagnosis" (Mirin & Weiss, 1991; Regier et al., 1990). By this we mean that in addition to their diagno­ sis of addiction, they also have a syndromal diagnosis (Axis I), such as depression, or a diagnosis of personality disorder (Axis II), or a com­ bination of both. A good conceptualization takes into account the vari­ ous ways in which the patients' psychological problems play them­ selves out. For example, a patient with a dependent personality disorder centered around a poor self-concept may become depressed following a rejection and seek to counteract the depressed feelings through using and/or drinking. Linking these behaviors may be a common thread, such as "I am too weak or fragile to make it on m y own." This belief may lead to clinical depression when interpersonal supports are removed. The same belief promotes using or drinking when the patient is confronted with a difficult problem or a stressful situation: "I can't handle this without a drink (or drug)." 30 COGNITIVE THERAPY OF SUBSTANCE ABUSE WHY DO PEOPLE RELAPSE AFTER NOT USING FOR A SUBSTANTIAL PERIOD? Many individuals handle the withdrawal symptoms, if present, and go for significant periods without using but then relapse—sometimes, for no apparently compelling reason (Carroll, Rounsaville, & Keller, 1991; Tiffany, 1990). The problem seems to lie in the fact that these individuals have not become "inoculated" to the external or internal conditions that can trigger the craving and undermine the control. These circumstances fall into the category of "people, places, and things," which is described in 12-step programs. This category includes situations such as associating with compan­ ions or sex partners who urge one to have a "hit" or drink, visiting a place where one has previously used or drunk, seeing drug parapher­ nalia, or receiving one's paycheck. These individuals also may expe­ rience a craving for the substance if they are feeling sad, bored, or anxious. Some individuals have a lapse when an unusual stressful situation occurs: death of a friend or relative, serious argument with a spouse, or loss of a job. One of the underlying reasons why recovering addicts are still prone to react with powerful urges to various stimulus (high-risk) situations is that their basic beliefs regarding the relative advantages and disadvantages of drug taking have not changed substantially. They may have acquired a number of strategies for controlling their drug- taking behavior, but they have not significantly modified the attitudes that help to fuel the craving. Consequently, when their controls are weakened, perhaps as a result of stress, and their urges are stimulated, for example by exposure to a high-risk situation (a situation that activates their drug-using beliefs), they are vulnerable to lapse by using or drinking a minimum or moderate amount. This lapse is accentu­ ated by a sense of helplessness or hopelessness: "It proves I can't control m y urges"; "I will never be able to beat this problem." As they are swept back into the drug-using cycle, the lapse becomes a relapse. Sometimes, patients may lapse for no discernible reason—that is, they have not been exposed to a high-risk situation (Tiffany, 1990). The probability of such a lapse is increased any time the ratio of the perception of control to the intensity of craving is decreased; that is, when control is weakened by fatigue and a gradual slippage of the constructive beliefs (anti-indulgence behefs) and/or an increase in the desire to use or drink, based, for example, on tran­ sient unpleasant feelings. The degree of commitment to abstinence may simply decrease with the passage of time-perhaps because of fading of memories of the bad effects of using or drinking (Gawin & Cognitive Model of Addiction 31 EUinwood, 1988; Velten, 1986). At this time, a "normal" degree of craving may lead to a lapse. If the patient's reaction is, "My control must be pretty poor if I give in to such mild craving," he/she may progress into a relapse. The basic beliefs that have been dormant but become stimulated by exposure to the stimulus (high-risk) situations include notions such as "If I use, I can handle m y problems better," "Having a smoke or hit will make life more enjoyable," or "I need a drink to overcome my anxiety." As soon as these beliefs are activated, the individual experiences an exacerbation of craving. The patient's attempts at self- confrol are undermined by permission-related thoughts (stemming from the beliefs) such as "I can do it this once and stop," or "There's no reason why I should continue to deprive myself." There is, thus, a continuing conflict between the attitudes concerned with control­ ling the urge and those attitudes favoring yielding to the temptation (or, more strictiy, initiating the behavior that would satisfy the urge). P H E N O M E N A OF ADDICTION Cravings and Urges In helping patients deal with their substance use prob­ lems, it is crucial to have a full understanding of the phenomena associated with drug use. Craving refers to a desire for the drug, whereas the term urge is applied to the internal pressure or mobiliza­ tion to act on the craving (Marlatt, 1985, and Horvath, 1988, use the terms in a similar way). In short, a craving is associated with wanting and an urge with doing. The two terms are often used interchange­ ably, but it is useful to separate them. Cravings represent a strong desire for a particular type of experi­ ence, for example, the pleasure from eating, relaxation from smok­ ing, or the gratification from sex. The fulfillment of the wish may be labeled the consummation and the means, the consummatory act. When one form of consummation is not available, an individual may turn to another form. For example, if there is no satisfaction in sight for yearning for affection, an individual may reach for a sweet or a beer instead. An urge is the instrumental sequel to a craving. A person desires to experience a "high" or relief from discomfort and feels a pressure to act to obtain this experience. Marlatt and Gordon (1985) define an urge as a behavioral intention to engage in a specific consumma­ tory behavior. Urges may be regarded as compulsions when the indi­ vidual feels incapable of resisting them. Thus, an urge may be insti- 32 COGNITIVE THERAPY OF SUBSTANCE ABUSE gated by an unpleasant feeling state (such as anger or anxiety) or anticipation of an unpleasant stressful event. The ultimate goal of consummating the urge is a reduction of the instigating state, whether it be a craving for excitement or a desire to relax. The delay between the experience of craving and implementa­ tion of the urge does provide an interval for a therapeutic interven­ tion—for the technical application of control or what is called, in common parlance, "will power," which we define as an active pro­ cess of applying self-help techniques, not simply a passive enduring of discomfort. Additionally, fostering a delay between the craving and the use of drugs allows for the natural diminishing of the acute crav­ ing episode (Horvath, 1988), thus lowering the chances that the

patient will act on the craving (Carroll, Rounsaville, & Keller, 1991). Urges are governed by the anticipated consequences, for example, reward for doing something or pain for not doing it. The urge may be accompanied by a positive feeling when it is driven by a positive expectation or a negative feeling when it is driven by expectation of unpleasantness unless the urge is consummated. Some people con­ fuse urge with "need." They will say "I need a smoke" or "I need a drink" as though they cannot survive, or at least function, without it. Such a belief is, of course, spurious and becomes a focus for thera­ peutic interventions. Cravings and urges tend to be automatic and may become "auton­ omous"; that is, they can continue even though the individual tries to suppress or abolish them. They may become imperative and are not easily dissipated even if blocked from being carried out. At this point, the word "compulsion" seems most appropriate to describe cravings and urges. W e see compulsions most clearly in obsessive- compulsive disorder, in which the individual experiences strong pres­ sure to engage in a repetitive act in order to ward off some feared event. Addictive behaviors incorporate some of the same characteris­ tics. The Role of Beliefs Dysfunctional beliefs play a role in the generation of urges. The beliefs help to form the expectation, which then molds the urge. For example, a patient with a serious drinking problem had the following beliefs: "If I am 'amusing and friendly' I will receive lots of praise" and "If I have a drink I will be more entertaining." He translated these beliefs into a specific expectation for receiving praise when an opportunity arose for entertaining people. The expectation, then, led to the urge to "show off." However, he was uncertain of his Cognitive Model of Addiction 33 success unless he had crack cocaine first. His expectation of succe was enhanced by his belief in the stimulating or disinhibiting effect of cocaine. As it happened, he would usually "overshoot the mark" and become so excited that people considered him "pathetic." Following Bandura (1982), Mariatt and Gordon (1985) have refined the concept "beliefs about the positive effect of using" into "positive outcome expectancies." Research by Brown, Goldman, Inn, and Anderson (1980) has shown that the expectancies of alcoholics fall into six factors: that drinking will (1) transform experiences in a positive way, (2) enhance social and physical pleasure, (3) increase sexual performance and satisfaction, (4) increase power and aggres­ sion, (5) increase social assertiveness, and (6) decrease tension. A simi­ lar set of expectations is associated with drug use (see Drug Belief Questionnaire in Appendix, this volume). The "Drug Habit" The habit of taking substances for relief or pleasure differs from the way the term "a bad habit" is generally understood. A particular "habit," such as grimacing when frustrated or leaving clothes on the floor, is a repetitive pattern—but it is not experienced as a craving or a need. For the drug abuser the immediate response to a relevant situation is subjective, namely, a craving or an urge. There is a delay between the stimulus and the consummatory act, such as preparing the syringe or the powder. What are chained to the stimu­ lus, thus, are the cravings and urges. Through continual repetition, the chain becomes stronger. In contrast to the habits involved in skilled acts such as driving, the pattern of drug taking is compulsive and dysfunctional. In addition, the skilled acts are based on volun­ tary decisions, whereas drug-taking cravings are involuntary (even though the control of the urges is voluntary). Because of the differ­ ence between using and the habits of everyday life, the term "drug habit" is probably a misnomer. Through a process of "stimulus generalization," the addicted individual is likely to respond with craving to an increasingly broader range of stimulus situations. Whereas originally the individual might have felt the craving for a drink or smoke only in a group, he or she now may experience it when upset, bored, or lonely. With the bind­ ing of the craving to more and more stimuli, there is a concomitant expansion of the dysfrinctional beliefs about drug use. Whereas ini­ tially the belief might be "I should take a smoke to be part of the group," the beliefs may build up to "I need a smoke to be accepted" and later to "I have to take a snort to relieve m y loneliness and dis- 34 COGNITIVE THERAPY OF SUBSTANCE ABUSE tress." The urges, thus, become more generalized and more impera­ tive in keeping with the broadening content of the beliefs. Furthermore, the rebound dysphoria experienced particularly after a "cocaine crash" (Karan et al., 1991; Ziedonis, 1992), for example, leads to a renewal of the craving in order to counteract this low feel­ ing. The consequence of the repetition of emotional distress leading to craving to indulgence to temporary relief of dysphoria is the develop­ ment of beliefs such as "I need a hit in order to feel better." When a drug or alcohol is taken to relieve stress-related or naturally occur­ ring tension, anxiety, or sadness, it tends to reinforce the belief "1 need the drug," as well as "I can't tolerate unpleasant feelings." The Control/Urge Equation There is a common belief that addicted individuals have little or no control over their urges and behavior or that the craving is irresistible. O n the surface, this seems to be true because these people seem to be driven by such a powerful force that they engage in addictive behavior even though they recognize its destruc- tiveness; many make repetitive abortive attempts to control their behavior and will say that they know they want to confrol their behav­ ior but simply carmot. This common observation of their cravings and urges overwhelming any resistance has led to the principle expressed by Alcoholics Anonymous: "I recognize that I am powerless." Their perception of "being out of control" has the positive bene­ fit of inducing addicted people to seek professional help rather than continuing to waste energy in futile attempts to exercise confrol—often followed by self-castigation for not successfully counteracting the urge. Developing control is a technical problem to a large extent. Learning specialized techniques for reducing craving and establishing some measure of control is generally necessary for those who are truly addicted. On the one hand, the sources of craving need to be explored. On the other hand, the notion of total loss of control is simplistic and does an injustice to the potential internal resources available to the individual. In actuality, most people who abuse drugs do exer­ cise control most of the time. When the urge is not strong or the substance is not currently available, they are able to abstain. They do not necessarily go off in wild pursuit of the drug at the first sensa­ tion of craving. There is a qualitative difference between the wish to use (to experience "benefits" of the drug) and the wish to control the urge. The craving activates a drug-taking routine: The individuals' sources for consummating the urge are scanned, a plan emerges, the body becomes mobilized to act, and the physiology shifts to a recep- Cognitive Model of Addiction 35 five state (e.g., the parasympathetic nervous system goes into an a vated state). Since craving is an "appetitive state," it is accompanied by a variety of bodily sensations somewhat akin to hunger or an unpleasant yearning for someone or something. This kind of appe­ tite operates according to the pleasure principle, in contrast to the wish to control the urge, which operates according to the reality principle. The wish not to use, thus, to control, is not expressed in visceral terms (as is craving) but is experienced as a sort of mental state. It has a sttong cognitive component, specifically, decision-making. What powers the decision-making is a sense of resolution or commitment that is felt in the musculature (in contrast to craving, which is more visceral). Thus, the two opposing motivations—craving and self-control (or will power)—are qualitatively different. Parallel to the decision not to use (refusal state) is the decision to indulge (permission giving). Permission giving and permission refusal are akin to gatekeepers. Their relative'strength determines whether the gates will open or close. There is more conscious (vol­ untary) participation in the gatekeeping than in the craving; there­ fore, the individual can reflect and decide whether or not to indulge. If the craving is strong, the decision to refuse/abstain may be too weak to control it. If the balance favors refusal, the using does not occur. Even when the urge is strong, addicted individuals can abstain at times, particularly if the drug is not immediately available. It is important to recognize that addictive behavior is related to the bal­ ance of control versus urge. Put in more abstract terms, the ratio of the strength of the control to the strength of the urge influences whether the individual will abstain or use. The formula or ratio power ofcontiol/power of urge may be used as a guide for intervention. Treat­ ment is focused on increasing this ratio. It does not require a super­ human effort to change the relative strengths. It may simply involve reducing the denominator (urge) or increasing the numerator (con­ trol) or, preferably, doing both. Beck et al. (1979) have used the analogy of the votes in Congress for a declaration of war to illustrate how suicidal behavior may be modified. A somewhat similar analogy may be applied to a decision to use. To declare war requires a simple plurality, a margin of one vote of the yeas over the nays. However, just as in the case of sui­ cide, if the decision is postponed or the relationship of yea to nay votes is changed in favor of the nays, the progression to action is arrested. In the case of declaration of war, lobbying for a few votes for peace may forestall the fateful action; in the case of addiction, strengthening the votes for abstinence can reverse the tendency to use. In the long run, however, it is necessary to build a solid "major­ ity" to forestall relapse. 36 COGNITFVE THERAPY OF SUBSTANCE ABUSE The point to this analogy is that it is not necessary to eliminate cravings totally or to institute absolute control. It is sufficient to change the relative strengths of the two parts of the equation. A change involving reduction of craving or increase in control may interrupt the drug-using progression in the short run. Since the goal is usually permanent abstinence, a durable improvement requires enough last­ ing change in the ratio to provide a sufficient margin of safety. Treat­ ment, thus, is directed toward both halves of the equation: increas­ ing control and reducing craving. Increasing Control Many addicted individuals simply have not developed the skills to control temptation. If such a skill deficit exists, one part of the therapy is directed toward increasing self-control skills. A vari­ ety of methods can be used to increase control. These techniques can be practiced in the therapist's office. The basic procedure is to repro­ duce stimulus conditions that will elicit craving and then to rehearse control behaviors as the craving is stimulated. For example, the indi­ vidual is asked to imagine a situation in which she is offered crack cocaine. She then imagines ways in which to refuse the offer. Or she might imagine feeling blue or anxious and then desiring relief from the discomfort. She then pictures what she will do when craving occurs: divert herself by calling a friend, become engaged in some pleasant activity, or read a flashcard detailing rational responses to cognitions related to craving. Another approach involves dealing directly with permission- giving thoughts. This exercise is carried out in the form of a debate. The patient mentally verbalizes or rehearses reasons for giving per­ mission to indulge and, at the same time, presents a rebuttal to this argument. At some point, however, it is necessary to identify and evaluate the underlying beliefs regarding permission giving and per­ mission refusing. Ultimately, of course, the therapist needs to help the patient reduce the craving by dealing with its various psychological and social sources. These sources may cover very broad domains of

the patient's life ranging from low frustration tolerance to marital problems. "Will Power" In the context of drug using, "will power" refers to a deliberate conscious decision (plus sufficient drive and technical self- help know-how to enforce it) to halt or delay the implementation of Cognitive Model of Addiction 37 an urge. When the urge to use is low or absent, the individual's dr to abstain from further use may appear to be quite strong. However, when the temptation is strong, the will power may become attenu­ ated. Marlatt and Gordon (1985) consider will power in terms of the strength of the commitment not to use or drink. Commitment means attaching a value to a particular goal so that it supersedes other con­ tradictory goals. Thus, the allocation of importance to abstinence can power the resolve to resist cravings. The successful application of will power when cravings and urges are aroused depends on a number of factors. An individual may make a serious commitment to stop smoking, drinking, or using but may not have the technical skills to fulfill the commitment. The applica­ tion of this technical knowledge can greatly increase the amount of leverage when the resolve to abstain is opposed to cravings and urges. Further, core beliefs about oneself (e.g., whether one is effective or helpless) may affect one's capacity to apply will power to controlling urges. W e must caution that patients tend to misconstrue the mean­ ing of will power, seeing it as an almost masochistic battle to main­ tain an unceasing state of discomfort in the face of drug urges (Tiffany, 1990). Clinicians must emphasize to patients that they will be taught to modify their beliefs and behaviors (cf. Washton, 1988) so that positive self-image and lifestyle changes will take place. This, along with the natural dissipation of cravings over time (Horvath, 1988), will help patients to feel good about resisting drug use in the long mn, as opposed to feeling deprived and in pain. According to the myth of the "rational man" (e.g., in jurispru­ dence or economics), an individual weighs the risks and benefits of a given action and makes a rational decision. In the case of the addicted individual, however, the objective cost-benefit analyses, or advantages-disadvantages calculations, are thrown off by the momen­ tary appeal of using, drinking, or smoking. The immediacy and reli­ ability of the effect of the drug and the subjective certainty that some desired effect will be achieved right away contrasts with an uncer­ tain, possibly undesirable consequence in the future. Some individu­ als become oblivious to the negative consequences when they expe­ rience the craving (Gawin & EUinwood, 1988). Others simply shrug off the long-range effects with the attifride "I'll take m y chances," or rationalize, "It won't hurt if I give in this one time." O n the other hand, a number of individuals are able to summon up, on their own, arguments and unpleasant memories that deter them from yielding to the temptation. In any event, there is always a con­ flict when individuals try to utilize will power to forestall yielding to their urges. O n the one hand, for example, an individual experiences 38 COGNITIVE THERAPY OF SUBSTANCE ABUSE the craving (and the anticipated relief or pleasure) and, on the o the voice of reason and restraint (and the anticipated deprivation and distress). After many unpleasant experiences, one may be able to issue oneself warnings of the dangers of indulgence when exposed to a high- risk situation or when aware of the lowering of one's resistance. Whether one will be able to heed these warnings to oneself depends to a large extent on one's access to techniques to implement them. Addictive Beliefs In our work, we have been impressed by the common­ ality of certain beliefs across various types of addictions (cocaine, opiates, alcohol, nicotine, and prescription drugs) and various addicted individuals. Even individuals susceptible to binge eating or general­ ized overeating show these types of dysfunctional beliefs (Heatherton & Baumeister, 1991; Lingswiler, Crowther, & Stephens, 1989; Zotter & Crowther, 1991). The addictive beliefs characterize those individu­ als after they have become addicted (i.e., they are characteristic of the disorder), however, and cannot in themselves be considered predis- positional to addiction. Nonetheless, the addictive beliefs do contrib­ ute to maintaining the addiction and provide the groundwork for relapse. Addictive beliefs may be considered in terms of a cluster of ideas centering around pleasure seeking, problem solving, relief, and escape. The specific items will vary depending on the type of preferred sub­ stance. Among the dysfunctional ideas are (1) the belief that one needs the substance if one is to maintain psychological and emotional bal­ ance; (2) the expectation that the substance will improve social and intellectual functioning; (3) the expectation that one will find plea­ sure and excitement from using; (4) the belief that the drug will energize the individual and provide increased power; (5) the expecta­ tion that the drug will have a soothing effect; (6) the assumption that the drug will relieve boredom, anxiety, tension, and depression; and (7) the conviction that unless something is done to satisfy the crav­ ing or to neutralize the distress, it will continue indefinitely and, possibly, get worse. In addition to these expectations/beliefs, the patients have a variety of beliefs relevant to justification, risk taking, and entitlement. These attitudes fall into one category of "permission-giving beliefs," such as "Since I'm feeling bad, it's O K to use," "I've been having a hard time; therefore, I'm entitied to relief," "If I take a hit, I can get away with it," "The satisfaction I get is worth the risk of relapsing," or "If I give in this time, I will resolve to resist the temptation next time." Cognitive Model of Addiction 39 Predispositional Characteristics A number of characteristics of the drug abuser, how­ ever, may have existed prior to drug use and thus may be considered predispositional. These characteristics center around (1) general sen­ sitivity to their unpleasant feelings or emotions—for example, they have a low tolerance for the normal cyclical changes in mood; (2) deficient motivation to control behavior^hus, instant satisfaction is more highly valued than control; (3) inadequate techniques for con­ trolling behavior and coping with problems—therefore, even when motivated to exert restraint, they do not have the technical knowl­ edge to follow through with it; (4) a partem of automatic, nonreflective yielding to impulses; (5) excitement seeking and low tolerance for boredom; (6) low tolerance for frustration (low frustration tolerance in itself rests on a complex set of beliefs and cognitive distortions); and (7) relatively diminished future time perspectives, such that the individual's attention is focused on here-and-now emotional states, cravings, and urges and on the actions for relieving or satisfying them. None of the attentional resources are devoted to the consequences of these actions. Low frustration tolerance (LET) seems to be an important precur­ sor to drug using (Chapter 15, this volume). Specifically, a number of dysfunctional attitudes magnifying the usual everyday sources of fmstration lead to excessive disappointment and anger. Among the components of this belief complex are attitudes such as (1) things should always go smoothly for m e or things should not go wrong; (2) when I am blocked in what I am doing, it is awful; (3) I cannot stand being frustrated; (4) other people are to blame for m y being thwarted, and they should be punished; and (5) people deliberately give m e a hard time. When individuals with LFT find that their activity is blocked or their expectations are thwarted, they are likely to (1) greafly exagger­ ate the degree of loss resulting from thwarting, (2) exaggerate the long- range consequences of this loss, (3) blame whomever they think might be responsible for thwarting, (4) experience excessive anger, (5) have a sfrong desire to punish the offender, and (6) importantly, overtook other ways of achieving their goal, such as problem solving. The result of this sequence of events is that an individual becomes overmobilized to attiack the offender. Since there is rarely a legitimate avenue for expressing the hostile impulses, the individual is left in a highly energized state, full of tension and anger. At some point, such individuals find that drug taking may reduce the highly volatile state and relieve the pent-up tension. Of course, the use of drugs for this 40 COGNITIVE THERAPY OF SUBSTANCE ABUSE purpose is at best only a temporary remedy and in the long run is self-defeating because the individual never learns ways of coping directly with frustration and solving the contributing problems. Con­ sequently, LFT is perpetuated, as are the beliefs regarding helpless­ ness. S U M M A R Y Many addicted individuals have characteristics that predispose them to drug abuse. These predispositional factors include (1) general exaggerated sensitivity to unpleasant feelings, (2) deficient motivation to control behavior, (3) impulsivity, (4) excitement seek­ ing and low tolerance for boredom, (5) low tolerance for frustration, and (6) in many cases, insufficient prosocial alternatives for gaining pleasurable feelings, and a sense of hopelessness in ever achieving this goal. LFT is characterized by exaggeration of the degree of loss result­ ing from thwarting, blaming other people for any frustration, a strong desire to punish the offender, and overlooking other ways of prob­ lem solving. Each of these predispositional factors is addressed in the course of cognitive therapy. The sequence of addiction often follows a vicious cycle proceed­ ing from anxiety or low mood to self-medication by using or drink­ ing. This behavior, in turn, produces and/or exacerbates financial, social, and/or medical problems, which lead to further anxiety and low mood. Patients often ascribe their drug and alcohol use to "uncon­ trollable cravings and urges." However, certain dysfunctional beliefs tend to fuel these cravings. Abusers tend to ignore, minimize, or deny the problems resulting from their drug use or attribute these prob­ lems to something other than the drugs or alcohol. An important factor in maintaining psychological dependency is the belief that withdrawal from the drug will produce intolerable side effects. In actuality, through careful clinical management these side effects gen­ erally turn out to be tolerable. Another important set of core beliefs centers around the addicted individual's sense of helplessness in con­ trolling the craving. Cravings are associated with wanting gratification or relief, whereas urges are concerned with doing something to provide a grati­ fication or relief. The delay between the experience of craving and the implementation of the urge provides an interval for therapeutic intervention. Cravings and urges tend to be automatic and may become autonomous; the thrust of therapy is to provide voluntary methods for managing them. Patients tend to equate the strong crav- Cognitive Model of Addiction 41 ing with an imperative "need" and an uncontrollable urge. Although the craving leading to drinking and using is involuntary, controlling the urge is voluntary and can be adopted even though the patient may feel helpless. Increasing the ratio of the subjective power of control to the subjective power of the urge may be used as a guide for inter­ vention. Cognitive therapy is a system of psychotherapy that attempts to reduce self-defeating behavior by modifying erroneous thinking and maladaptive beliefs and teaching techniques of control. In the cogni­ tive therapy of drug abuse, the specific case formulation forms the basis for the therapeutic regimen. This formulation, in turn, is based on the cognitive model of addictions. The therapeutic approach consists of undermining the urge by weakening the beliefs that feed into the urge and, at the same time, demonstrating to the patient various ways of controlling and modi­ fying their behavior. Cognitive therapy of substance abuse is charac­ terized by the following: (1) It is collaborative (builds trust), (2) it is active, (3) it is based, to a large degree on guided discovery and empir­ ical testing of beliefs, (4) it is highly structured and focused, and (5) it attempts to view the drug or drinking problem as a technical prob­ lem for which there is a technical solution. C H A P T E R 3 T h e o r y a n d T h e r a p y o f A d d i c t i o n A b-ccording to the

cognitive perspective, the way people interpret specific situations influences their feelings, motiva­ tions, and actions. Their interpretations, in turn, are shaped in many instances by the relevant beliefs that become activated in these situ­ ations. A social situation, for example, m a y activate an idiosyncratic belief such as "Cocaine makes m e more sociable" or "1 can be more relaxed if I have a beer (or a cigarette)," and lead to a desire to use, drink, or smoke. Specific beliefs such as these constitute a vulnerability to substance abuse. Activated under particular predictable circum­ stances, the beliefs increase the likelihood of continued drug or alco­ hol use (i.e., they stimulate craving). Beliefs also shape the individual's reactions to the physiological sensations associated with anxiety and craving (Beck, Emery, with Greenberg, 1985). Beliefs such as "1 cannot tolerate anxiety" or "I must give in to this hunger" will influence the person's reactions to these sensations. Individuals with such beliefs are likely to be hyperattentive to these sensations. Even a low-level degree of anxiety or craving can elicit a substance-using belief such as "I must take a hit (or drink) to relieve m y anxiety (or satisfy m y craving)." The activation of substance-using beliefs is illustrated in the expe­ rience of Les, a chronic cocaine user, w h o experienced a sudden crav­ ing for cocaine while attending a party. In this scenario, his acute urge to use was related to his sense of social isolation within a group. His underlying belief, "I can't stand it without cocaine," was activated by his sad feelings at seeing other people having a good time using drugs. Les lived in a rundown neighborhood in which there was a great deal 42 Theory and Therapy 43 of drug fraffic. He had a longstanding belief, "I'll never get out this awful environment." This belief (not the environment per se) led to chronic feelings of sadness and hopelessness. The belief underly­ ing his chronic urge to use cocaine was "I need some coke to get through the day." This case illustrates the coexistence of acute cravings and urges related to a specific situation with more chronic urges related to the patient's general life situation. The combination of these beliefs made Les prone to addiction. LAYERS O F BELIEFS There were several levels of beliefs underlying Les's addictive behavior: (1) his more general basic belief that he was "trapped" in a noxious environment; (2) his belief that the only way he could escape from his environment and his unpleasant feelings was to take drugs; and (3) the belief that he "needed" drugs to relieve any unpleasant feelings. Added to these drug-related beliefs was a basic belief that he did not belong and was not accepted as a member of his peer group. This cluster of beliefs made Les vulnerable to addic­ tive behavior; that is, they fed into a compulsive urge to relieve his distress through drug taking. The essence of a large proportion of addictive behaviors, consist­ ing of the types of general and specific beliefs held by this patient, are illustrated in Figure 3.1. The addictive beliefs (Chapter 2, this volume) seem to derive from either one or a combination of core beliefs (sometimes referred to as "core schemas"). The first set of dysfunctional core beliefs has to do with personal survival, achievement, freedom, and autonomy. Depend­ ing on the precise nature of the patient's vulnerability, the core belief that is expressed m a y have a content such as any of the following: "I am helpless, trapped, defeated, inferior, weak, inept, useless, or a fail­ ure." The second set of dysfunctional core behefs is concerned with Core Beliefs Emotions Addictive Beliefs "I am trapped/ Sad or "Drugs are an escape." alone." angry "Drugs make me more sociable." Addictive Behavior Go to a crack house FIGURE 3.1. Sequence of core beliefs and addictive beliefs. 44 COGNITIVE THERAPY OF SUBSTANCE ABUSE bonding with other individuals or to a group. This set of beliefs is concerned with lovability or acceptability. The various permutations of the core belief m a y take the following form: "I a m unloved, unde­ sirable, unwanted, repulsive, rejected, different, socially defective." Such core beliefs constitute a specific sensitivity or vulnerability: W h e n circumstances (e.g., social rejection) that are relevant to the core belief arise, they trigger the belief (e.g., "I a m defective") and lead to distress. Les had a double set of core beliefs revolving around the notions "I a m helpless" and "I a m undesirable." W h e n he noted the difficult conditions in his neighborhood, the first belief was triggered and took the form "I a m trapped." O n c e this notion took hold, he believed himself incapable of improving his lot, saw the future as hopeless, and felt frustrated and sad. The specific addictive belief-was then trig­ gered: "The only w a y to get relief is to take a hit." In a group situation his automatic thought was "I don't belong." This thought stemmed from his other core belief, "I a m unaccept­ able." These beliefs converged o n the addictive belief: "The only way to get accepted is to use coke." T h e relation between his two core beliefs, his automatic thought, his addictive belief, and his craving is illustrated in Figure 3.2. The same sort of constellation of core belief, addictive belief, and craving m a y apply whatever the instigating factor and whether the form of relief is alcohol, illegal drugs, legal drugs, or tobacco. The sequence generally proceeds from (1) a core belief, such as a nega­ tive view of the self (helpless, undesirable) and/or a negative view of the environment (noxious, oppressive), and/or a negative view of the future (hopeless), to (2) unpleasant feelings, such as dysphoria or anxi­ ety. From there, the addiction-prone individual experiences (3) crav­ ing and psychological dependency o n drugs (e.g., "1 need cocaine to m a k e m e feel better). Core Belief Automatic Thought Emotions "I am "I am trapped in this Sad or helpless." bad environment." frustrated .r Core Belief Addictive Belief Conclusion "I am "I don't belong unless I "1 need the drug" undesirable." use." \f Craving FIGURE 3.2. Interaction of multiple beliefs. Theory and Therapy 45 It is important to note that the perception of a noxious environ­ ment is not limited to inner-city individuals. Privileged individuals who perceive their job, family, or marital situation as inimical, who experience the same sequence of discouragement over life circum­ stances and have negative views of themselves and their future, may turn to drugs as a form of escape. In depression (Chapter 14, this volume), the negative view of the self, the current circumstances, and the future often is exaggerated. After patients modify their depressive thinking, the therapist often finds that compulsive drug use is dimin­ ished (Woody et al., 1983). Individuals like Les become habitual users because they regard using as a way of gaining or maintaining social acceptance. They have addictive beliefs such as "I can't let m y friends down ... they will reject me if I don't use." (This fear, of course, may be realistic and one of the goals of therapy may be to help the patient to develop friendships with nonusers.) One patient greatly admired his cousin who was addicted to cocaine. The patient constantly used crack when he was with his cousin. The instigating factor each time was a desire to please his cousin. Eventually, using became embedded in his sys­ tem of coping with his fear of becoming socially ostracized. S E Q U E N C E O F BELIEFS Although the core beliefs represent the background of the addictive beliefs, they are not immediately apparent unless the patient is depressed (Chapter 14, this volume). The addictive beliefs may be more accessible. These addictive beliefs are activated in a specific sequence. First in the sequence are anticipatory beliefs. Initially these take a form such as "It will be fun to do this .. . It's okay to try it occasionally." As the patient gains satisfaction from using, he/she often develops romanticized beliefs predictive of gratification or escape: "1 will have an hour or so of sheer pleasure ... I wiU feel less sad/anxious ... It wiU be a sweet oblivion." Some beliefs are predic­ tive of increased efficacy or socialization: "I will perform better ... I will be more entertaining and will be accepted into the group." As the individuals start to rely on the drug to counteract feelings of disfress, they develop relief-oriented beliefs, such as "I need cocaine in order to function ... I can't continue without it... I will feel well again if I use ... I need the drug ... I can't control the craving ... I must have it or I'll fall apart." Note the imperative quality of these beliefs: "I must have a smoke to make it through the day." The acti­ vation of these beliefs then leads to cravings. 46 COGNITFVE THERAPY OF SUBSTANCE ABUSE Activating Anticipatory Craving situation beliefs Drug-seeking Permissive < plan of action beliefs FIGURE 3.3. Sequence of anticipatory and permissive beliefs. Since addiction-prone individuals may have some conflict about using (e.g., medical, financial, social, or legal consequences of using), they generally develop a facilitating or permissive belief, such as "1 deserve it" or "It's all right, I can handle it... Since I'm feeling bad, it's all right to use . .. Nothing else is going right; this is the only right thing in m y life." The relation of these beliefs is illustrated in Fig­ ure 3.3. The sequence of these beliefs formed by Les is illustrated in Fig­ ure 3.4. His uneasiness in a social situation triggers the anticipatory belief "I will feel better if I use," which is immediately followed by a craving and then the plan of action to call his cousin for a "hit." CONFLICTING BELIEFS In the various stages of cocaine use the patient can have conflicting sets of beliefs, such as "I should not use cocaine" versus "It's O K to use this one time." Each behef can be activated under different circumstances or even at the same time. The balance between the relative strength of each belief at a given time will influence whether the patient uses or abstains. (Of course, the availability of the drug will also be a determining factor.) Sometimes the individual experiences a conflict between the desire to use and the desire to be free of drugs. This ambivalence may be formulated as a conflict between two beliefs: "It's O K " (permis­ sive) versus "It's not O K " (abstinent). The conflict between these beliefs results in discomfort or m a y increase the individual's current discomfort. Paradoxically, the individual m a y experience an even greater pull toward using in order to relieve the uneasiness produced by the conflict. The belief "I need relief from this feeling" becomes more potent and may tip the scales in favor of using. In therapy, patients learn skills to cope with the discomfort and to test out and restructure their belief that using or drinking is the most usefial way of dealing with discomfort. Theory and Therapy 4 7 Social 1 will leei situation Craving better If..." Call cousin "It's OK." , _ FIGURE 3.4. Simple model of Les's substance use (maps onto Figure 3.3). A C T I V A T I O N O F BELIEFS IN S T I M U L U S SITUATIONS Drug-using beliefs and desires typically are activated in specific, often predictable, circumstances, which we term "stimu­ lus situations." These are also labeled "cues" (Moorey, 1989). How­ ever, depending on the patient's current mood and self-control, the degree of riskiness of a situation may vary considerably from time to time. That is, a situation that is manageable at one time may be stimu­ lating enough to promote drug use at another time. These circum­ stances, which can be external or internal, correspond to what Marlatt and Gordon (1985) term "high-risk situations." These situations stimu­ late the craving to "smoke, shoot, snort, or swallow drugs." Examples of external stimulus situations are a gathering of friends using cocaine, contact with a

drug dealer, or receiving a weekly pay­ check. Internal circumstances (or cues) include various emotional states such as depression, anxiety, or boredom, which can trigger drug- using beliefs and, consequently, craving for the drug. As shown in Figure 3.5, drug use may be regarded as represent­ ing the final common pathway of the activation of the cluster of the aforementioned beliefs. Cognitive therapy is aimed at modifying each of the categories of beliefs: anticipatory a n d permissive, as well as the underlying core beliefs (e.g., "I a m frapped") that potentiate these drug- Activating Beliefs Automatic > Craving/ stimulus: activated thoughts urges • Internal cues • External cues t \f Continued Focus on Instrumental Facilitating use or strateoies beliefs relapse (action) (Pe rmi 3sion) F I G U R E 3.5. Complete model of substance use. 48 COGNITIVE THERAPY OF SUBSTANCE ABUSE related beliefs. The therapist attempts to introduce or reinforce more adaptive beliefs relevant to each of the classes of beliefs. Other tech­ niques are concerned with dealing with major life problems (see Chapters 12 and 13, this volume) or personality difficulties or disor­ ders (see Chapters 14, 15, and 16, this volume) leading to drug use. As shown earlier, craving is aroused in a specific situation and seems to arise as a reflex reaction to the stimulus. However, the situ­ ation does not directly "cause" the craving: Interposed between the stimulus and the craving is a drug-related belief that is activated by the situation and an automatic thought derived from this belief. For example, w h e n he was feeling sad, Les would get the thought "If I take a hit now, I will feel better." His underlying belief was: "I can't stand discomfort... I need a fix to make the discomfort go away." The sequence then proceeded to craving, to facilitating beliefs ("It's O K this time"), to an actual plan for obtaining the drug, and finally to using. These beliefs can be ascertained by direct questioning and the use of inventories (see Appendix, this volume). The sequence proceeds so rapidly that it is often viewed as a "conditioned reflex" (O'Brien, 1992). The automatic thought, in par­ ticular, seems to be almost instantaneous and can be captured only if the patient learns to focus on the chain of events. Figure 3.6 illustrates the sequence from the activating stimulus to the implementation of the plan to get the drug. It should be noted that each step offers an opportunity for a cognitive intervention. Using the method of guided discovery (Beck et al., 1979), for example, the cognitive therapist questions the meaning attached to the activating stimulus, the relief-oriented belief that taking a fix is the most desir­ able solution, the permission-giving belief ("I can do it without harm"), and the implementation plan (the decision to look around for money). Les had a very low tolerance for unpleasant feelings, whether sadness, anxiety, or sheer boredom. His belief regarding the neces­ sity for alleviating feelings of distress was activated w h e n he attended a party. His drug-taking beliefs centered on the anticipation of reUef Feeling sad "If I take a fix I will feel better.' "What the hell." Craving Purchase and Look around to "I can do It this use drugs get the money. time without any harm." F I G U R E 3.6. Example of Les's drug-using sequence (maps onto Figure 3.5). Theory and Therapy 49 from any negative feelings. Other examples of his anticipatory rel oriented beliefs were: "There is only one way for m e to have fun," "I can't stand the withdrawal symptoms," "I feel better knowing it's there," and "If I don't take a hit regularly, I will feel much worse." It should be noted that the patient's permissive thoughts about the harmlessness of drug taking stemmed from a simplistic (and deceptive) set of beliefs. He believed that since he only snorted cocaine, he could not be addicted: He saw himself as being safe from addic­ tion provided he did not smoke crack. In fact, one of his typical per­ missive thoughts was, "I'm O K since I don't smoke crack." "Spontaneous craving" (i.e., craving in the absence of an obvi­ ous external stimulus) is also often observed. For example, a patient with a 5-year history of cocaine use reported having a dream about using cocaine. Upon awakening, he "felt high." Next, he started day­ dreaming about the last time he had used cocaine. This imagined scenario in turn activated the belief, "Life is more fun when I use," and was followed by the automatic thought, "I love this stuff." A permission-giving belief was also activated, "There is no harm in this." His attention then focused on checking to see whether he had enough money to buy cocaine. Although the craving appeared to be sponta­ neous in this case, the patient's mental state during the dream and upon awakening set the stage for daydreaming about using. This imagery served as a catalyst for the permission-giving thoughts. His attention then focused on implementing his craving and shut out any consideration of the ill effects of using. INFORMATION PROCESSING: MEANING, SYMBOLISM, AND RULES In referring to the kinds of circumstances (external or internal) that excite the craving-using cycle, we generally use the term "stimulus situations" or "triggers" rather than "high-risk sifrra- tions" introduced by Mariatt and Gordon (1985). Although many situ­ ations have a high probability of setting the craving-using pattern into motion, their effect varies from person to person and even for the same person over time. By conceptualizing these situations in terms of their stimulus properties and meanings, we can align our concept of drug use and abuse with concepts regarding stress (Beck, 1993), syndromal disorders (e.g., depression; Beck et al., 1979), and person­ ality disorders (Beck, Freeman, & Associates, 1990). While the term "high-risk situation" fits nicely into a descriptive model, the formulation in more cognitive terms can fit our observa- 50 COGNITIVE THERAPY OF SUBSTANCE ABUSE tion into an explanatory model. This model, encompassing concepts of the activation of beliefs, symbols, information processing, and motivation, provides a broader framework for understanding and psychological intervention. Although we use the terms "stimulus situations" and "stimulus properties," it should be noted that the actual situation is neutral. It becomes a stimulus if a person attaches a special meaning to it. For example, an addicted individual looks at a cocaine pipe and other paraphernalia and becomes excited and experiences craving. Another person, indifferent to drugs or not knowledgeable about the parapher­ nalia, simply sees a pipe. For the first person, the pipe is a symbol, a coded message, packed with meaning. The meaning is not inherent in the pipe but in the individual's personal symbolic code (embed­ ded in his or her information or cognitive processing system). The individual automatically applies this code when he/she perceives the paraphernalia, for example, and consequently experiences pleasure and craving. The therapist's task is to help the patient to decode the symbol. If one "unpacks" its meaning, it would read something like this: "The pipe means taking a hit, which will give me pleasure." The pipe and the concept of pleasure have become fused so that the expectation of pleasure in the future gives pleasure now and leads to craving. The drug abuser may seem to be stimulus bound. Any depiction of or reference to drugs on television, radio, or magazines, for exam­ ple, may be sufficient to excite the individual. The addicted person is actually "schema driven"; that is, Les's reactions are produced by inter­ nal cognitive structures, labeled schemas, that contain the code, for­ mulas, or beliefs that attach meaning to the situation (see Beck, 1967, for a complete description of schemas). Thus, a schema containing the belief "Using is necessary for m y happiness" will be primed when the person is exposed to a relevant situation. Similarly, a schema containing the belief "I cannot be happy unless I am loved" will be activated if the person perceives that he or she has been rejected by a lover and, thus, will feel sad. The experi­ ence of the sad affect will, in turn, be processed cognitively by behefs such as "I can't stand sadness," "I need relief by using." The indi­ vidual then experiences craving. The therapeutic application of this explanatory model involves attaching more importance to modifying the individual's belief sys­ tem than to simply getting him to avoid or cope with high-risk situ­ ations. Since some "situations" (such as internal states) are unavoid­ able and other sihiations (e.g., exposure to drug-related situations) may be inevitable (Childress, Hole, & DePhilippis, 1990; Moorey, 1989; Theory and Therapy 51 O'Brien, McLellan, Alterman, & Childress, 1992; Shulman, 1989), the best outcome can be derived from changing the beliefs that make these situations risky. Les, for example, often compared himself with other people more successful than he. W h e n he saw such a person, his negative beliefs ("I'm inferior" and "I'm no good") were activated. Thus, the mean­ ing attached to the perception of the other person was a self-devalu­ ation, leading to sadness. He also attached a meaning to sadness: "My life is intolerable ... I can't stand the pain." Following the activation of the belief "I need dope to ease the pain," Les experienced craving. In therapy, each of these beliefs was explored. The proposed mechanism for therapeutic change consists of align­ ing the belief system more closely with reality. Since the beliefs are maladaptive (e.g., "I need the drug [or alcohol] in order to function"), it is necessary either to modify these beliefs or to substitute more functional beliefs (or both). The process of change, however, involves more than simple modification of the beliefs. The therapist and patient need to work together to improve the patient's system of controls (e.g., by practic­ ing delay of gratification) and to learn coping techniques such as anticipating and solving problems. Thus, the therapeutic goals are (1) conceptual change and (2) tech­ nical development of proficiency in coping. Cognitive Blockade W h e n they are not experiencing craving, patients are generally able to recognize the disruptive effects of the drug on their lives. However, once the drug-taking beliefs are activated, a "cogni­ tive blockade" inhibits awareness of or attention to the delayed long- term consequences of drug use (Gawin & EUinwood, 1988; Velten, 1986) and increases the focus on immediate instrumental strategies, such as searching for money to buy drugs. As these beliefs become hyperactive, recognition of the drawbacks of drug use become attenu­ ated. W h e n Les was not feeling sad, he was convinced that using cocaine was bad for him, but once his craving was stimulated, he had difficulty in remembering his reasons for not using. His attentional processes were predominantly allocated to using. The immediacy of the stimulus and the activated meanings shut out serious consider­ ation of long-range consequences. This kind of "tunnel vision," in which the individual's attentional resources are devoted almost totally to the immediate situation, has been demonsfrated in cognitive psychology experiments (Beck, 1991). 52 COGNITIVE THERAPY OF SUBSTANCE ABUSE For example, very hungry individuals will be hypersensitive to sti relevant to food or eating and will be relatively insensitive to other signals. The introduction of danger stimuli, however, will shift the attention to the danger stimuli and away from food stimuli. Clinical states show the same type of phenomenon. Information that is con­ gruent with the clinical condition will be processed very rapidly and memories congruent with the state can be rapidly recalled, but the patient has trouble gaining access to stored information that is not congruent with the clinical condition. Depressed patients, for example, quickly assimilate negative information about themselves but block out positive information. Also, they recall negative information much better than positive information (Beck, 1991). Similarly, patients hav­ ing a panic attack readily respond to suggestions that they are expe­ riencing a serious condition but have problems in recalling benign explanations for their attacks or even in applying reason to counter­ act the catastrophic interpretations they are making (Beck, 1986). A somewhat similar phenomenon may be observed among many individuals addicted to drugs, alcohol, or nicotine. Although when sober they may be quite adept at reeling off (with sincerity and

con­ viction) the reasons for not using, drug users have difficulty in recall­ ing or attaching the same significance to these reasons once they are in the throes of a specific drug-using episode. Since all their atten­ tion is focused on the mechanics of obtaining the drug, the reasons for using at that time become very salient and the contradictory rea­ sons become inaccessible or insignificant. This phenomenon is termed the "cognitive blockade" because of blocking out the incongruent (i.e., the corrective, realistic) informa­ tion. The therapeutic task is to lift the blockade, as it were, through a variety of tasks. One approach is to deliberately activate the crav­ ing cycle in the office (e.g., through imagery) and, while the craving is strong, review the reasons for not using. Of course, sufficient time must be allotted for this maneuver to preclude the craving's being maintained following the session (Childress et al., 1990). A similar strategy involves the preparation of flashcards (listing reasons for not using) which patients will read when they experience craving in the natural environment. SUMMARY At the core of the problem of the addicted individual is a set of addictive beliefs which appear to be derived from core beliefs such as "1 am helpless," "I am unlovable," or "I am vulnerable." These Theory and Therapy 53 core beliefs interact with life stressors to produce excessive anxiety, dysphoria, or anger. These stressful or stimulus situations do not directly "cause" craving, but they activate the drug-related beliefs that lead to the craving. Although w e use the term "stimulus situation," it should be noted that the situation itself is neutral. The meanings, derived from the beliefs, that are attached to a situation are what cause the individual's craving. Individuals with beliefs that they cannot tolerate anxiety, dysphoria, or frustration, for example, will tend to be hyperattentive to these sensations and m a y build up expectations that they can relieve the sensations only through using or drinking. Thus, w h e n an unpleasant affect arises, the individual attempts to neutralize it by using or drinking. A specific sequence of drug-related beliefs leading to drinking or using m a y be delineated. First is the activation of anticipatory beliefs relevant to obtaining pleasure from using or drinking. These antici­ patory beliefs usually progress to relief-oriented imperative beliefs, which define using or drinking as a dire necessity and stipulate that the craving is uncontrollable and must be satisfied. The anticipation of pleasure or relief leads to the activation of craving and facilitating 01 permissive beliefs, such as "I deserve it" or "It's O K this time," which legitimize using or drinking. Finally, the instrumental plans, which have to do with plans or strategies for obtaining drugs or alcohol, are propelled by the imperative craving. M a n y individuals have conflicting beliefs regarding the pros and cons of using. At times they are locked in such an unpleasant struggle between these opposing beliefs that, paradoxically, they m a y seek dmgs simply to relieve the tension generated by the conflict. The therapeutic application of this model, consisting of modify­ ing the individual's belief system, goes beyond teaching the individual to avoid or cope with "high-risk situations." C H A P T E R 4 T h e T h e r a p e u t i c R e l a t i o n s h i p a n d I t s P r o b l e m s A , , collaborative relationship between the therapist and the patient is a vital component of any successful therapy. The most brilliantly conceived interventions will be reduced in effective­ ness if the patient is not engaged in the process of treatment. All the support and effort that the therapist m a y put forth in an effort to help the patient will make little impact if the therapist has not gained some measure of the patient's trust. While this seems to be relevant to almost every type of patient, it is most especially true of the substance-abusing population. Numer­ ous potential factors interact to create an almost adversarial relation­ ship between the therapist and the drug-abusing patient at the begin­ ning of therapy and during the course of treatment. These factors include: 1. Drug-abusing patients often do not enter treatment on a vol­ untary basis. 2. Patients often maintain highly dysfunctional presuppositions about therapy. 3. Patients often are not very open and honest, at least at the start of therapy. 4. Patients may be currently involved in felonious activities, thus presenting confidentiality dilemmas. 5. Patients view their therapist as part of the "system," and not as an ally. 54 The Therapeutic Relationship 55 6. Patients have a difficult time believing that their therapist r cares about their problems. 7. Patients look askance at therapists w h o m they perceive to differ from them markedly in terms of demographics and attitudes. 8. Therapists may maintain negative presuppositions about drug- abusing patients. Many of these patients do not come into therapy of their own volition. Some are given an ultimatum by their significant others (e.g., spouse, children, or parents) or employers, while others are remanded by the courts following criminal legal proceedings (Frances & Miller, 1991). Consequently, the drug-abusing patient may enter the ther­ apist's office with any number of counterproductive automatic thoughts, such as "I don't want to be here; I'm only here so m y wife will get off m y back," "I'll just tell this doctor what he wants to hear, and then I'll blow out of here," "This whole therapy thing is like doing real easy time compared to prison; I'll just go along with this and do what I want to do anyway," "I don't really have a problem; maybe I'll show up for therapy, and maybe I won't," and "I'm not going to tell this shrink anything that can be used against me; m y life is nobody's business but m y own." The list could go on and on. To complicate matters further, drug-abusing patients typically enter therapy playing their cards close to their vests, and therefore conceal the kinds of automatic thoughts listed above. The therapist must actively probe for them, as the patients often will not divulge them in an unsolicited fashion (Covi et al., 1990). Another factor that militates against the ready formation of a positive therapeutic relationship is that substance abuse often repre­ sents felonious behavior. As such, patients are highly motivated to be dishonest in self-reporting their substance abuse activities. Although the vast majority of therapeutic interactions represent privi­ leged communications between therapist and patient, drug-abusing patients are typically well schooled in covering their tracks. As the stakes are high, such patients may simply decide it is best to take no chances, and therefore will not readily admit to drug-abuse-related behaviors. Furthermore, some actions of the patients may be serious enough threats to themselves or to the general public that the thera­ pist may legally and ethically be required to contact the authorities (e.g., when the patient admits that a drug-related murder has been arranged, or when the patient calls the therapist and claims to have taken a drug overdose in order to attempt suicide). Since therapists should inform their patients about the limits of confidentiality at the outset, drug-abusing patients will come to know 56 COGNTTIVE THERAPY OF SUBSTANCE ABUSE what information they cannot safely reveal. If they do come forth such sensitive material, the therapist is placed in the uncomfortable position of having to serve as society's watchdog, and may in the process completely discourage the patient from continuing with much needed treatment. This point highlights another more general factor that adds to the difficulty in forming a working alliance with drug-abusing patients-̂ iamely, that such patients often view the therapist as an agent of the police, the courts, "the system," or a more privileged socio­ economic class. Such patients find it hard to believe that their thera­ pists will sincerely try to help them with their problems, or will treat them with honesty, respect, and positive regard. As a result, the patients tend to dread and avoid therapy sessions. They may take con­ frontational statements from the therapist as confirmation that the therapist is working against them, while positive statements from the therapist may be seen as naive, manipulative, insincere, or patroniz­ ing. This places the therapist in a "damned if I don't, damned if I do" position, which, left unaddressed, may sabotage therapy before it gets started. Yet another stumbling block to the formation of a healthy thera­ peutic relationship is the perceived demographic and/or attitudinal differences between therapist and patient. For instance, the patient might think, "This doctor is probably rich and has everything she wants in life. There's no way that she could possibly understand what it's like to struggle every day of your life like I do. H o w in the worid can she help me? The rules of her world just don't apply to mine. Whatever she says is just bullshit." Another thought might be, "I wonder if this therapist ever used drugs. If he did, then he's no bet­ ter than me, so why should I listen to him? If he didn't, then how can he know what it's like to be hooked? Only someone who's been there could know what he's talking about." Similarly, the therapist may have maladaptive beliefs about the patient, such as "This guy is a low-life. At best he's going to waste m y valuable time, and at worst he's going to be a liability or a threat to m y personal safety," or "These types of patients are beyond help. They have a chronic disease for which there is little hope for cure or rehabilitation. Therefore, there's not much point in investing too much of m y time or energy," or "I can't relate to this patient at all. 1 wouldn't associate with him (or her) in 'real life' so I don't think I'll be able to form a working bond with this patient here in the office either." Admittedly, working with drug-using patients can be highly stress­ ful; therefore we strongly encourage therapists to engage in regular The Therapeutic Relationship 57 peer supervision with colleagues in order to receive professional support and objective advice. Such consultations can help therapists to avert burnout, and to combat their own dysfunctional beliefs pertinent to working with the drug-abusing populations (cf. Weiner & Fox, 1982). W e acknowledge that the obstacles are formidable. However, based on extensive clinical experience, we believe it is possible to establish a positive, collaborative therapeutic relationship with the substance abuser. W e consider this to be the case even when such patients exhibit severe concomitant Axis II disorders, such as para­ noid, narcissistic, and/or antisocial personalities (see Chapter 16, this volume). To be sure, the task is difficult, and frequently trying. At the same time, freating the substance-abusing patient can be reframed as represeriting a growth-enhancing challenge for the therapist. The skills of developing therapeutic alliances with difficult populations (e.g., substance abusers and borderline patients), comprise the "art" of therapy, and as such are very much a measure of the competency of the mental health professional. This chapter presents guidelines for facilitating the formation and maintenance of an adaptive and functional therapeutic relationship with the drug-abusing patient. Case illustrations are provided in order to highlight various techniques and strategies, as well as to demon­ strate how things can go awry. The central messages of this chapter are that (1) a positive therapeutic relationship does not occur by chance—it can be actively constructed, (2) treating the drug-abusing patient requires careful and vigilant attention to the vicissitudes of the interactions between the therapist and the patient, and (3) the management of the therapeutic relationship with the drug-abusing patient is neither a straightforward nor an overwhelming task. ESTABLISHING R A P P O R T The initial interactions between the patient who is just entering therapy and the therapist are extremely important. Even when dmg-abusing patients are self-referred, they often have a great deal of ambivalence about seeking ongoing contact with a therapist (Carroll, Rounsaville, & Gawin, 1991; CarroU, Rounsaville, & Keller, 1991; Havassy et al., 1991; Institute

the patient's thoughts, followed by sincere involvement by way of questioning and direct, honest, humble feedback, will be a boon to the establishing of rapport. As patients attempt to engage in the process of treatment, thera­ pists can help facilitate the establishment of rapport by giving posi­ tive verbal reinforcement for the patients' pro-therapy behaviors and attitudes. For example, therapists can provide encouragement and praise to patients for demonstrating good attendance, promptness, active participation in sessions, and cooperation with therapeutic homework assignments (e.g., writing down the disadvantages of using drugs each time the patient experiences a strong urge to go out to make a "score"). Such positive feedback from therapists helps patients to feel supported, to understand their role in therapy, and to decrease their anxieties and negative expectations about the process of work­ ing with mental health professionals. BUILDING T R U S T Trust does not develop immediately. It cannot be asked for, and it cannot be artificially rushed. Only through the therapist's consistent professionalism, honesty, and well-meaning actions over a period of time can trust enter fully into the therapeu- The Therapeutic Relationship 63 tic relationship. It does no good for the therapist to say merely, worry, you can trust me." It is far more realistic to admit that there is little reason for the patient to trust the therapist in the beginning, but that "I hope that in time you will decide for yourself whether or not I can be believed and trusted." Unfortunately, trust can be impaired or lost relatively quickly, and therefore it must be nurtured and managed in a delicate, painstaking fashion. In short, therapeutic trust with the substance-abusing popu­ lation is difficult to establish, and may be more difficult to maintain. Furthermore, even if the patient learns to trust the therapist, there may be little reason for the therapist to trust the patient. Inaccurate and/or incomplete reporting by patients is a frequent phenomenon with this population, a situation to which the therapist must remain sensitive. Nevertheless, since the professional is held to a higher stan­ dard of behavior than is the patient, the therapist must be willing to continue benevolently to assist the substance-abusing patient, even if that patient has been untruthful. Later, we discuss ways in which the therapist can confront such dishonesty on the part of the patient, yet continue to strengthen the therapeutic relationship and work toward greater progress in treatment. The following suggestions and illustrations are offered to assist the cognitive therapist in achieving and holding on to this most valu­ able therapeutic asset. The basic elements of trust building are very simple and undra- matic. They include behaviors that consistently demonstrate the ther­ apist's genuine involvement in the therapeutic process, and com­ mitment to being available to the patient. Such behaviors include (1) being available for therapy sessions on a regular basis, (2) being on time for sessions (even if the patient is not), (3) returning patient telephone calls in a prompt manner, (4) being available for emergency intervention (e.g., by giving the patient a telephone number where the therapist can be reached in case of the need for crisis interven­ tion), (5) showing concern and being willing to try to contact the patient if he or she fails to keep an appointment, and (6) remaining in touch with the patient (and available for the resumption of outpa­ tient cognitive therapy) if inpatient hospitalization, detoxification freatment, halfway house rehabilitation, or reincarceration takes place during the course of the therapeutic relationship. Therapists foster trust when they assiduously avoid making dis­ paraging comments about the patient, the patient's family members, other substance abusers with w h o m the therapist has had contact, or any socioeconomic, ethnic, or gender group. Even if the therapist makes the derogatory comment about someone else, the patient may 64 COGNITIVE THERAPY OF SUBSTANCE ABUSE think that this is how the therapist truly thinks of him or her wh not working in the role of "therapist," and such a remark may foster the patient's possible belief that the therapist is insincere in his or her show of respect. Trust is also built when therapists serve as role models who have "clean" lifestyles and attitudes. Offhanded comments by therapists about their own "partying" or "getting buzzed" clearly are contra- indicated. Such statements give drug-abusing patients a confusing mixed message. This message may lead the patient to perceive the therapist as a hypocrite who operates on a "Do as I say, not as I do" policy. Related to this issue is the situation that arises when patients ask therapists about their own experiences with drug use. Certainly, thera­ pists are under no obligation to answer this type of question. A typi­ cal appropriate response would be, "I know you're curious about it, but I'm going to have to decline to answer your question. W e really have to stick to talking about issues that are relevant to you." At the same time, therapists may use their discretion in choos­ ing whether to answer. A brief, honest reply may go a long way toward fostering the patient's sense of trust for the therapist. For example, the therapist might answer, "No, I've never used any drugs on more than a try-and-see basis, and even that was fifteen years ago. I was playing with fire, and I guess I'm lucky it never progressed. But I've seen enough misery in the lives of those who've gotten into more regular drug use to know that I'd be a damned fool to ever try any­ thing again." Another honest answer could be, "No, I've never used drugs. I was always too afraid that I might like them. But really, we need to focus back on you because this is your session." Those thera­ pists who have used drugs in the past may choose to be silent about this matter or may use the experience to make rare but relevant self- disclosures as a way of keeping a patient engaged in treatment or to drive home an important point. The goal here again is to nurture the therapeutic relationship, not to get sidetracked from the work of therapy. SETTING LIMITS While it is crucial that therapists strive to work in a collaborative fashion with their drug-abusing patients, they must take care not to become oversolicitous to the point that patients know they can take advantage of their therapist. Limits must be set (Moorey, 1989)-for example, that a therapy session will not be held if it is de­ termined that the patient is in an inebriated or drug-intoxicated state. The Therapeutic Relatioruhip 65 Another such limit might be that the therapist will not condone "a little bit" of drug use. Therapists can establish such ground rules without sabotaging the therapeutic relationship if they take care to maintain a respectful tone, and reiterate their commitment to act in ways that are in the best therapeutic interest of their patients (Newman, 1988, 1990). When one of our patients arrived drunk to a session, the following dialogue took place: TH: Walt, pardon me for asking this ... and if I'm mistaken please accept m y apology... but have you had something to drink before coming to this session? PT: I had a few. N o big deal {belches to be humorously obnoxious). TH: H o w many is "a few"? PT: You know, a few. TH: Walt, I think you're intoxicated. PT: I'm fine. I can hold m y beer pretty good. TH: Walt, we've discussed this before. If you're in an altered state of mind ... and believe me, drinking "a few" means that you're in an altered state of mind ... there's no point in going through with this session. I have no reason to believe that you'll be able to pay serious enough attention to what we do here to warrant continuing with this session. PT: Shit man, you're making a big deal out of nothing. TH: Walt... PT: I'm fine I tell you. TH: Walt... PT: I shouldn't have said anything. TH: Walt... I'm glad you were up front with me. I respect you for it. I'm depending on you to be a man and tell m e the real story to m y face. It's just that we can't go through with this session. That was our agreement, and I think we should stick to our agreements. PT: Shit, man. TH: Did you drive here? PT: No, I was beamed down {sneers). TH: I have something important to ask you. I need to ask you to hang out in the waiting room for a couple hours until you're sober enough to drive safely. 66 COGNITIVE THERAPY OF SUBSTANCE ABUSE PT: Doc, I don't got time for this shit. I got here fine, and I'll home fine. TH: Walt, you've worked too hard to get to this point to mess up now. If you get pulled over, or worse, you're risking going back to jail. I don't want to see that happen to you. What's a couple of hours to ensure your freedom? You can have m y newspaper to keep you occupied for awhile. The patient ultimately complied with this therapist's request. The limit was clearly set, but the tone of the communication was neither critical nor controlling. The therapist emphasized that he was look­ ing out for the patient's welfare, and this had a lot to do with the patient's compliance and willingness to continue actively with cog­ nitive therapy. When the therapist sets a limit, sticks to it, and does so in a respectful way, trust is fostered and the patient learns to have respect for the therapist as well. Parenthetical to the above, it is necessary that the therapist be amenable to continuing with therapy once the patient is in compli­ ance again after a slip (Mackay & Mariatt, 1991). Since many drug- abusing patients frequentiy test limits, no gains will be made if thera­ pists are disinclined to go forward with therapy when their patients engage in defiant and/or manipulative behaviors. Therapists serve their drug-abusing patients best when they follow through on predeter­ mined agreements on how to deal with counterproductive patient behavior but also show genuine support and encourage the patients to "get with the program" again. The above vignette brings up the issue of the role of alcohol in the illicit drug-abusing patient's life and therapy. While we believe that it is theoretically possible for illicit-drug-abusing patients to con­ tinue to drink alcoholic beverages on a casual basis during treatment, in practice our experience tells us that the use of alcohol undermines their abstinence from drugs such as cocaine and heroin. One reason is that the use of alcohol lowers patients' inhibitions. Patients have reported that when they are drinking they are less likely to think about the compelling reasons for staying free of drugs. Even when they can stay focused on the disadvantages of drug use, patients report that they are less apt to care about the long-term consequences of their behavior than when they are sober. Thus, they are more likely to resume the use of harder substances. Further, when patients use alco­ hol as a "substitute" for drugs such as cocaine or heroine, their con­ sumption quickly escalates to levels indicative of abuse and depen­ dence. Therefore, we discourage the use of alcohol durtng patients' treatment and recovery from illicit drugs. The Therapeutic Relationship 67 PROTECTING CONFIDENTIALITY As alluded to previously, there are limitations to con­ fidentiality. Therapists should spell this out to their patients from the very start. The following monologue may serve as a model: "Mr. A, I want you to know that almost everything we discuss here will be kept just between you and me, unless you want rne to talk to someone else about your situation or you otherwise give m e permission. So, for the most part, things that you tell m e here will be kept confidential. But I want to inform you that there are certain exceptions to this rule. If you tell m e something that indicates that you or someone else is in danger, and you're not willing to help me fix the situation so that everyone is safe, then I am legally obligated to contact the authorities and anybody else who

their home tele­ phone numbers for use in emergency situations, we realize that some therapists may prefer instead to make use of an intermediary such as an answering service. In either case, we believe that it is necessary for patients to be able to make contact with their therapist after hours in the event of critical situations.) APPEALING T O PATIENTS' POSITIVE SELF-ESTEEM As many substance abusers evidence defiant attitudes and/or pathological levels of self-importance, it is often necessary for the therapist to appeal to patients' narcissism in order to elicit col­ laboration from them. This does not have to entail gross hyperbole on the part of the therapist. If fact, such an approach is contraindicated as the intelligent patient will rightly see it as an insincere, manipula­ tive ploy. Rather, the therapist needs to focus on some of the patient's actual strengths and positive points, and express appreciation for these qualities. This approach serves to strengthen rapport and to elicit greater cooperation. The following clinical vignette demonstrates an appeal to the patient's sense of entitlement in order to defuse his anger toward the therapist. The problem arose when the patient did not show up for his session, and instead called 5 hours later to say that he had gotten a fiat tire on the way to the therapist's office. The dialogue (a con­ densed version of the actual interchange) proceeded in the following manner: TH: Walt, we've talked about how important it is for you to get to sessions on time, and to keep me informed of your whereabouts. The fact that you waited five hours to call me concerns me. PT: {Exasperated) I was on the road. I couldn't get to a phone. I didn't have a spare tire so I had to wait to get help. There was no way 1 could call any sooner. TH: Walt, ninety percent of m e wants very much to believe you, but I have to be honest with you—ten percent of m e has m y doubts. I can't help but wonder whether your lateness in getting in touch with me is drug related. [The patient responded very angrily, vilifying the therapist for being "such a hard-ass" and for insulting the patient by "calling me The Therapeutic Relationship 73 a liar." The therapist answered with a reply that was geared to use Walt's narcissism in the service of repairing the therapeutic relation­ ship.] TH: Walt, I'd like to believe everything you say to me. But you and I both know that you have a lot of skill and experience in cov­ ering your tracks. You could easily outsmart m e if I'm not care­ ful. If I just blindly believe everything you tell me, then I'm a fool, and frankly, I think you deserve better than to have a fool for a therapist. This latter statement achieved its intended effect of disarming th patient's hostility long enough to get him to agree to come in for a session early the next day. Later in treatment, the therapist and Walt were discussing Walt's unsafe sexual habits. Walt noted that he did use condoms when he had sex with prostitutes, but refused to wear one with his many "girl friends," stating facetiously that it was against his religion. Therapist and patient discussed these practices at length, trying to get a handle on the automatic thoughts and beliefs that led him to act so reck­ lessly in this era of the AIDS epidemic. Additionally, the therapist attempted to focus Walt's attention on the dangers involved in his sexual behavior by noting the pros and cons of wearing condoms. Finally, when it seemed that these tactics were falling on deaf ears, the therapist resorted to making an appeal to Walt's intelligence by saying: TH: The fact that you wear condoms with hookers is a smart move. I wouldn't expect anything less than a smart move where you're concerned. You're very good at taking care of number one. So it confuses m e how you would stop short of doing the smart thing with your girl friends as well. It just doesn't seem like you, Walt. It's out of character for you to leave any loose ends like that [no pun intended]. You normally have all your bases cov­ ered [again, no pun intended]. This approach effectively pitted Walt's desire to be seen as an intelligent person against the "macho" rules that governed his unsafe sex practices. It allowed the therapist to be confrontive without dam­ aging rapport or collaboration. In other cases, we have helped bring our patients back into a collaborative mode by appealing to their sense of justice, their posi­ tive feelings for involved significant others, their survival skills, their integrity, their potential abilities to be positive role models for others, and other personal attributes. 74 COGNITIVE THERAPY OF SUBSTANCE ABUSE MANAGING POWER STRUGGLES In spite of the therapist's best efforts to maintain an ongoing positive therapeutic relationship with the drug-abusing patient, there will almost certainly be times when therapist and patient are at odds, and when negative feelings will be rather intense on one or both sides. However, this does not have to spell doom for the working alliance. W e rely on the following guidelines for managing such power struggles: 1. Don't fight fire with fire. 2. Maintain honesty. 3. Remain focused on the goals of treatment. 4. Remain focused on the patient's redeeming qualities. 5. Disarm the patient with genuine humility and empathy. 6. Confront, but use diplomacy. 1. Don't fight fire with fire. When a patient becomes hostile, lou intransigent, and/or verbally abusive, it does little good for the thera­ pist to respond in kind. In fact, such a reaction on the part of the therapist could potentially lead to a dangerous escalation of the con­ flict. Instead, therapists must show confidence and conviction in their position in a matter-of-fact way. Concern and strong feelings can be expressed (e.g., "Ms. G, I urge you to reconsider your intentions in this matter. I am greatly concerned that you are headed for a big-time fall if you go ahead with your plans to attend that dealer's party!"); however, it is advisable that such sentiments be expressed in a way that communicates a genuine concern for the patient's well-being and best interests. A controlling or disrespectful response (e.g., "You're dead wrong! If you go to that party you're an idiot! I simply can't allow you to do it.") will undermine the therapeutic alliance and probably will not effectively control the patient's behavior anyway. Instead, the strategy advocated here is more akin to the philosophy espoused in Asian martial arts that states that a strong opponent must not be fought head on but rather through leaning back and allowing the adversary's misguided brute force to carry him past you, to stag­ ger, and to fall. 2. Maintain honesty. During times of conflict with a drug-abus­ ing patient, there is often a great temptation to try to appease the patient artificially through reassurances that are less than completely truthhil (e.g., getting a patient "oft your back" by telling him that it won't really matter too much if he continues to be late for therapy The Therapeutic Relationship 75 sessions). Not only is it unwise to reinforce patients' maladaptive interpersonal behavior by capitulating to them, it also sets up the therapist to look like a liar if the therapist later reverses his/her posi­ tion or otherwise reneges on the reassurances. Instead, the therapist must be willing to "take the heat," and not simply say things that the patient wants to hear in order avoid the unpleasantness of a power stmggle. 3. Remain focused on the goals of treatment. W h e n therapist and patient are at odds, it is extremely helpful if the therapist calls atten­ tion to mutually set goals. In effect, therapists can remind both them­ selves and their patients that a disagreement in one area does not alter the fact that there are fundamental areas of agreement and collabora­ tion in other areas. One therapist diffused a heated exchange by tell­ ing his ex-football player patient, "We may not agree on whether we should run the ball, or pass, but we have to remember that we're on the same team and we both want to get into the end zone." 4. Remain focused on the patient's redeeming qualities, as well as your own {as therapist). Power struggles are often fueled in part by the therapist's cognitive biases. This happens when the therapist reacts to an aversive power struggle by focusing only on the patient's irri­ tating qualities, and glossing over his/her strengths. Similarly, the therapist may lapse into dysfunctional self-blame (regarding the lack of the patient's therapeutic cooperation and progress), thus engen­ dering more ill feelings. In such instances, it is extremely helpful for therapists to use cognitive therapy procedures on themselves in order to notice and modify the following types of automatic thoughts that might be exacerbating negative interactions with patients: • "This patient is a loser. He'll never listen to me." • "This patient is so dense. I'm going to have to beat this guy over the head with m y point of view until he agrees with me." • "Why can't I reach this patient? What am I doing wrong? I'm ready to give up on working with this patient." • "Maybe I'm not cut out to work with such a patient. I don't like being reminded of m y shortcomings, so this patient is really on m y shit Ust." • "You just can't compromise and be reasonable with these people. If you give them an inch, they take a light-year. There­ fore, I will not budge from m y position one iota." • "Why did I ever take on the responsibility of treating this patient in the first place? I must have been an idiot. I almost wish this patient would get arrested so I can be rid of this case." 76 COGNITIVE THERAPY OF SUBSTANCE ABUSE Obviously, the aforementioned automatic thoughts are very del­ eterious to the therapist, the patient, and the prospects for the con­ tinuation of treatment. Therapists would do well to focus on their own idiosyncratic automatic thoughts, to produce the kinds of rational responses that would diminish the anger, frustration, and exaspera­ tion that escalate power struggles and undermine problem solving and therapeutic collaboration (Weiner & Fox, 1982). Examples of such rational responses might be: • "There have been a number of sessions in which the patient and I worked very well together. Those were very rewarding experiences that 1 must not forget. • "This patient is not dumb. He's convinced he has his reasons for defying the therapeutic plan the way he's doing. Let me try to understand his resistant automatic thoughts and beliefs, rather than simply label him a troublemaker." • "My worth as a therapist does not hinge on m y patient believ­ ing everything 1 say, doing everything I suggest that she do, and staying free of drugs for the rest of her life. I'd like for her to be compliant and to make progress, but the fact that she sometimes thwarts this doesn't prove that she can't succeed in therapy with me, and it certainly doesn't prove that 1 should throw in the towel with all drug-abusing patients." • "If 1 keep m y cool, present m y point of view resolutely, and also show that I'm willing to be flexible within reason, I'll probably get a lot more therapeutic mileage out of this con­ flict than I will if I become strident or stubborn." • "This power struggle is a great opportunity to get at some really hot interpersonal cognitions!" 5. Disarm the patient with genuine humility and empathy. Fre­ quently, drug-abusing patients will become angry if they perceive the therapist to be flaunting their authority over the patient or acting with a holier-than-thou air. This perception can lead the patient to fight against the therapist's position in order to reassert some measure of control. This implies that it is important for therapists to be aware of the possibility that the patient is viewing him or her in this negative way, and to respond with behavior that gives the patient evidence to

process that is associated with their emotional life. The patient's behaviors are the end products of the vulnerable situations, and the activation of beliefs, automatic thoughts, and emotions. C o m m o n dysfunctional behaviors include actively seeking drugs, using drugs, engaging in irresponsible activities (e.g., unpro­ tected sex), abusive interpersonal confrontations, avoidance of self- help activities, and others. The integration of the above data is the most challenging and most important step in the ongoing process of conceptualizing the patient's life and problems. Here, therapists piece together all the information into a "story of the patient's life" that provides plausible explanations for the patient's difficulties and suggests treatment recommendations that may break into the patient's self-defeating patterns and vicious cycles. For example, the therapist may posit the following: "The patient was subject to a frequent barrage of harsh disap­ proval in childhood, and came to believe that he was inadequate and unlovable. These core beliefs have been carried into adulthood, where the patient experiences chronic discomfort in social situations where he believes that he will not measure up. The patient took to using cocaine in the belief that it would make him feel confident enough to make positive impressions on others. Unfortunately, this dysfunc­ tional compensatory strategy has led to compulsive use of cocaine, leading further to a depletion of his money and endangering his marriage. These life problems have fed back into the patient's cycle of anxiety, sadness, low self-worth, and renewed belief that the only way to be accepted is to become outgoing through the use of cocaine. As a result, his drug, financial, and marital problems have worsened, and his sense of helplessness and hopelessness have increased." CASE STUDY The following case illustrates in more detail the ten essential components of a case formulation. The patient (described in the integrative example above), David, is a 40-year-old white male. He has been married for eight years and has one child. His complaints at intake evaluation included high anxiety and a long history of alcohol and cocaine abuse. He reported recently feeling more anx- 84 COGNITFVE THERAPY OF SUBSTANCE ABUSE ious in social situations, and he feared that if his anxiety got w he might have a relapse and start using cocaine and alcohol. In other areas of his life, he was working and received a good salary; however, he was $40,000 in debt as a result of his cocaine "habit." His mar­ riage was "on the rocks," and he also suspected that his wife was an alcoholic. Upon completion of his intake evaluation, David met the DSM- III-R criteria for polysubstance use, with cocaine as his preferred drug, social phobia, and generalized anxiety disorder. The social phobia was the area on which David wanted to work first. David felt an urgency in this regard because it was the holiday season and he had several business obligations that required his attendance at social functions such as dinner-dances and parties. David felt that if he did not learn to cope with his anxiety in these situations, he would lose control and start using cocaine again. David grew up in a household where his father was seen as a workaholic and someone who "drank too much." David stated that his mother was nurturing but somewhat timid around her husband. In school, David did not do well. He received mostly C's and D's and only stayed in college for one year. David first began drinking at about age 9. Because of his father's business, the family was involved in many social activities and David would often go around drinking out of the glasses of some of the guests. By age 13, David already had experimented with alcohol, marijuana, speed, and diet pills. As a child David was often humiliated and degraded by his father, usually at social events, after his father had been drinking heavily. O n one occasion after his father called him "stupid," he ran outside and sat under a tree and felt humiliated, worthless, and helpless. On another occasion, he ran out onto a pier and sat there feeling ashamed and helpless. These were significant childhood events for David, and they served as the foundation for some of his core beliefs and com­ pensatory strategies. His typical style was to run away and avoid unpleasant situations. Later in life, David realized that alcohol and drugs helped him to cope with unpleasant emotions. Since drugs and alcohol worked so well, David did not develop many other strategies for coping with unpleasant emotions. Under relevant childhood data, we can see that there were several incidents when David was shamed by his father. These incidents helped to form his core belief, "I am unloved, unwanted." He later developed a conditional assumption for coping, "If I do everything perfectly, then people will like me." David had several compensatory strategies, such as always to strive to do things perfectly or to avoid doing things that were unpleasant. Other strategies included avoid- Formulation of the Case 85 ing showing others how he really felt, and using alcohol and drugs. After using cocaine, David felt especially confident that he could do "everything perfectly," which in turn led him to believe that he was loved and wanted by others. A vulnerable situation occurred when David was invited to a din­ ner party at a friend's house. Prior to going to the friend's house, David was acutely aware of the fact that he was becoming anxious and ner­ vous. He was also aware of the fact that his automatic thoughts cen­ tered on such ideas as "I'll screw up," and "They will see m e trem­ bling," and he imagined himself being "overcome with anxiety" and eventually running out of the house. This, in turn, led to taking a drink and snorting a line of coke before going to the party. From this example, the clinician can see the relationship between the development of David's core belief, his conditional assumptions, and his compensatory strategies, as well as their cumulative impact on a vulnerable situation—being invited to a dinner party at his friend's house (see Figure 5.1). Once again, we need to remember that these compensatory strategies tend to be rather compulsive, inflexible, inappropriate at times, energy depleting, and not balanced by other adaptive strategies. In addition, in spite of the compensatory strate­ gies, the patient still tends to have hidden doubts and secret fears about coping. GATHERING DATA FOR CASE FORMULATION The Case Summary and Cognitive Conceptualization Worksheet is an excellent form for compiling data that will be used in the case formulation. This worksheet is divided into eight main sections: I. Demographic Information II. Diagnosis III. Inventory Scores IV. Presenting Problem and Current Functioning V. Developmental Profile VI. Cognitive Profile VII. Integration and Conceptualization of Cognitive and Develop­ ment Profiles VIII. Implications for Therapy The demographic information section is where the therapist collect such information as the patient's age, sex, race, religion, employment 86 COGNITIVE THERAPY OF SUBSTANCE ABUSE Relevant Childhood Data Father was alcoholic. H/lother took verbal abuse from father. David often humiliated by father. Core Belief(s) I'm unloved, unwanted. Conditional Assumptions/Beliefs/Rules If I do everything perfectly, then people will like m e and I'll feel com­ fortable. If I show others how I really feel (anger), they will abandon me. Compensatory Strategy(ies) Use drugs If you are uncomfortable. Don't do anything unless you are 1 0 0 % sure you can accomplish it. Don't show others how you really feel. Situation # 1 Situation # 2 Situation # 3 Invited to a dinner party. Automatic Thought Automatic Automatic "I'll screw up." Thought Thought 'They will see m e trembling." Meaning of the Meaning of Meaning of Automatic Thought the Automatic the Automatic "People won't want Thought Thought to be around me." Emotion Emotion Emotion Fear; apprehension T Behavior Behavior Behavior Took a drink. Did a line of coke F I G U R E 5.1. Cognitive conceptualization diagram developed by Judith S. Beck. From J. S. Beck (in press). Copyright Guilford Press. Reprinted by per­ mission. Formulation of the Case 87 status, marital status, and other pertinent identifying characteris This is standard information that would be a part of any psychologi­ cal evaluation. In the diagnosis section, it is advantageous to formulate a diag­ nosis on all five axes of DSM-III-R. Clinical syndromes are designated on Axis I. O n Axis II, developmental disorders and personality disor­ ders are noted. Physical disorders and conditions pertinent to the patient's psychological difficulties are presented on Axis III. Severity of psychosocial stressors is identified on Axis IV. The level of sever­ ity of psychosocial stressors ranges from code 1 (none) to code 6 (cata- sttophic), such as the death of a child, the suicide of a spouse, or a devastating natural disaster. Axis V can be determined from the Global Assessment of Functioning Scale, which has a code number descend­ ing from 90 to 1, with 90 signifying abstinence and ideal coping, and progressively lower numbers indicating an increasing severity of drug use or deficits in coping and functioning. All the data on the five axes will have an impact on the clinician's understanding of the patient, and in the subsequent designing of the treatment plan. Inventory scores, such as the Beck Depression Inventory, Beck Anxiety Inventory, and Beck Hopelessness Scale (discussed later), are listed in this section. Intake scores plus scores from the first six ses­ sions are reported. There is also an extra column in order to note the scores of the most recent session. These inventory scores are extremely important because the therapist can quickly see general trends and patterns—changes for the better or for the worse. The presenting problem and current functioning section describes the patient's current difficulties and focuses on such areas as employment, concurrent psychiatric disorders, nature of drug use, criminal activ­ ity, interpersonal problems, and other data. This is a cross-sectional analysis of the patient's current functioning. The developmental profile examines the patient's social history, educational history, medical history, psychiatric history, and voca­ tional history. In addition, relationships with parents, siblings, peers, authority figures, and significant others over the life span are also noted. It is also important to ascertain any significant events or trau­ mas in the patient's formative years or recent past. This longitudinal analysis is akin to paging through a family photo album. The thera­ pist can see the patient in different stages of development. This retro­ spective analysis also includes an evaluation of the patient's introduc­ tion to psychoactive substances and how the problem became a full-blown addiction. The cognitive profile section addresses the manner in which the patients process information. The patients' typical problem situations 88 COGNITIVE THERAPY OF SUBSTANCE ABUSE are noted, and the corresponding automatic thoughts, feelings, and behaviors in these situations are outlined. In addition, possible core beliefs, conditional beliefs, and drug-related beliefs are described in this section. The integration of cognitive and developmental profiles takes into consideration the patient's self-concept and concept of others. It also focuses on the interaction of life events with cognitive vulnerabili­ ties, as well as compensatory and coping strategies. An important part of this section is a description of how self-concept and concept of others might have played roles in the onset and progression of sub­ stance abuse. This section is illustrated by the case of Mike, a 31-year- old cocaine addict. The patient lives with his parents in a rundown neighborhood where there is high unemployment and high crime rates. He grew up around drugs and alcohol, and his mother and all his siblings have had problems with drugs and alcohol. Mike is now serving five years probation for insurance fraud. The patient's work history is poor. After dropping out of high school, Mike found only unskilled labor jobs, and he is currently unemployed. There is lots of dealing in his neighborhood; Mike has sold drugs in the past and knows there is lots of fast money in deal­ ing. Mike sees himself as a loner and does not have any real friends. His problematic beliefs are "I'm no better than the rest of my fam­ ily," "I'll never get a job," "I can't get away from it (drugs)," "Deal­ ing is the only way out of here (urban ghetto)," and "Using

is the only way to cure the boredom." His circumstances and beliefs have led to the following behavioral patterns: brief periods of abstinence from drug and alcohol followed by solitary use of alcohol with Valium, which, in turn, leads to intermit­ tent use of crack cocaine, and then to daily use of crack cocaine. Mike's case shows many of the cognitive factors (among other factors) that can influence drug use (see Figure 5.2). The section on implications for therapy examines some of the fol­ lowing areas: 1. The patient's initial "aptitude" for cognitive interventions; 2. The patient's personality characteristics, such as "sociotropic" or "autonomous"; 3. The patient's motivations, goals, and expectations for therapy; 4. The therapist's goals; and 5. Anticipated difficulties in treating the patient. In this section, the clinician hypothesizes how psychologically minded the patient is. For example, is the patient aware of the nature and severity of his or her problems, and does he or she have the ability Formulation of the Case 89 Unemployed—Lots of free time Til never get a job." Long family Drugs easily Poor work history history of drug available Quick money dealing and alcohol use X "I'm no better "I can't get "Dealing Is the only than the rest away from it.' way out of here." of my family." Loner; no non- social support "Using Is the only way to cure the boredom." Mike's Drug Use FIGURE 5.2. Beliefs and other factors contributing to Mike's d m g use. to self-monitor automatic thoughts? The patient's capacity for access­ ing automatic thoughts and beliefs certainly has important implica­ tions for the pace of treatment. In addition, personality characteris­ tics—for example, sociotropic or autonomous—can provide the therapist with s o m e indication as to the conditions under which the patient might relapse, for example, the sociotropic person in situa­ tions in which there is social pressure, and the autonomous individual w h e n blocked from reaching his/her achievement goals. The patient's motivation, goals, and expectations for therapy are noted and dis­ cussed o n the conceptualization form. Also noted are the therapist's goals for treatment, and h o w compatible these are with the patient's goals. Finally, it is also useful to anticipate difficulties that might arise during the course of therapy that might warrant special attention. For example, if the patient has a history of periodic homelessness, this will need to be addressed early in treatment lest the patient suddenly "disappear" from treatment and be unreachable by mail or telephone. A case study illusfration of the Cognitive Conceptualization Work­ sheet (for patient "D.D.") follows. 90 COGNTTIVE THERAPY OF SUBSTANCE ABUSE CASE SUMMARY AND COGNITIVE CONCEPTUALIZATION WORKSHEET Dr. R. D.D. 1/10/92 Therapist' Name Patient's Initials Date Session # I. Demographic Information Ms. D is a 38-year-old, white, single female who is currently unemployed. She lives alone and has recently broken up with her boyfriend. She was in school but stopped going to class this month. II. Diagnoses Axis I: Cocaine abuse Alcohol abuse Major depression, recurrent Axis II: Avoidant personality disorder Histrionic personality disorder Axis III: N o physical illness reported Axis IV: Code 2 (Mild)—Not attending her college classes Axis V: GAP Code 60—Moderate depressive symptoms; not function­ ing well in school; difficulties eating, sleeping, and concen­ trating III. Inventory Scores Sess. Sess. Sess. Sess. Sess. Sess. Latest Intake # 1 # 2 # 3 # 4 # 5 # 6 Sess. BDI 25 26 20 21 17 17 16 16 BAI 2 1 1 12 11 12 13 13 BHS 11 11 12 12 12 10 7 7 Other General Trend of Scores: BDI and BHS scores are improving; how­ ever, her BAI scores have worsened over the past six sessions. (Higher scores indicate greater symptomatology.) IV. Presenting Problem and Current Functioning There were three presenting problems: (1) history of alcohol and cocaine use (though she reported that she had not used in the past 30 days), (2) moderate number of depressive symptoms and related fear that the symptoms would get worse, and (3) worry and fear that she might start using drugs and alcohol, progressing toward total relapse. {continued) Formulation of the Case 91 V. Developmental Profile A. History (family, social, educational, medical, psychiatric, voca­ tional) The patient was born on the East Coast and spent most of her life in an urban setting. Her father had numerous businesses and her mother was a homemaker. W h e n D was 13, her mother and father divorced. The patient liked elementary school but was anxious. She "hated" high school. D went through inpatient detox in 1985, and previously suffered from major depression in 1987. B. Relationships (parents, siblings, peers, authority figures, signifi­ cant others) The patient described her father as a loving, dedicated man. She stated that he was a heavy drinker^unpredictable when he was dmnk. Her mother was the "perfect mother," always there for the family. D was the oldest of five children. She stated her relation­ ship with all of them was "great." C. Significant events and traumas The patient described three significant events: (1) her parents' divorce, (2) her unhappiness throughout high school, and (3) her realization in 1985 that she had a d m g and alcohol problem. VI. Cognitive Profile A. The cognitive model as applied to this patient 1. Typical current problems/problematic situations: Situations that can lead to her taking drugs: (a) W h e n she is around former drug friends. (b) W h e n there is a breakup in a significant relationship. (c) W h e n bored and alone. 2. Typical ATs, affect, and behaviors in these situations: Automatic Thoughts: "They still use." "I'll never find someone who really cares." "I can't stand the boredom." Affect: Angry Depressed Hopeless Behavior. Uses alcohol first, then cocaine B. Core beliefs (e.g., "I a m unlovable") "I a m unlovable." "I a m undesirable." "I a m powerless." "I a m weak." {continued) 92 COGNITIVE THERAPY OF SUBSTANCE ABUSE C. Conditional beliefs (e.g., "If I fail, I a m worthless") "If I do what is expected of me, then people will accept me." "If I do things perfectly, then I feel competent." D. Rules (shoulds/musts applied to self/others) "I must be accepted by others or I'm worthless." "I must be emotionally in charge or there is something wrong with me." VII. Integration and Conceptualization of Cognitive and Develop­ mental Profiles A. Formulation of self-concept and concepts of others The patient believed she had to do the "right" thing as a child or her father would push her away. The patient could not always please her father predictably because his drinking led to erratic and "fickle" treatment of his children. Father left mother without any explanation; led to D's belief that men are not to be trusted. B. Interaction of life events and cognitive vulnerabilities Boyfriend Beliefs: "I a m D feels sad, leaves her unlovable" and depressed, "Men can't be hopeless tmsted" get activated C. Compensatory and coping strategies Avoids doing things that she feels she cannot do perfectly. Tendency to use drugs and alcohol when she is upset. D. Development and maintenance of current disorder Low fmstration tolerance for anxiety, boredom, and depression. Drug-related beliefs such as "Cocaine is the only way to relieve the boredom." VIII. Implications for Therapy A. Aptitude for cognitive interventions (rate low, medium, or high, and add comments, if applicable): 1. Psychological mindedness: very good 2. Objectivity: good 3. Self-awareness: very good 4. Comprehension of cognitive model: very good 5. Accessibility and flexibility of automatic thoughts and beliefs: very good General adaptiveness: very good 7. Humor: Excellent (continued) Formulation of the Case 93 B. Personality characteristics: sociotropic vs. autonomous Sociotropic: Strong need for attachment to others. Achievement needs are far less pronounced. C. Patient's motivation, goals, and expectations for therapy Strong motivation to stop using dmgs and alcohol. Other goals are realistic (i.e., would like to finish college). D. Therapist's goals 1. Teach patient to monitor, examine, and respond to negative automatic thoughts. 2. Help her to acquire skills for coping with d m g and alcohol urges and cravings. 3. Improve D's problem-solving skills. E. Predicted difficulties in therapy 1. Lapse could turn into a relapse if patient does not contact the therapist as soon as possible, which she might not do if she feels ashamed of her behavior and expects the therapist to criti­ cize her. 2. Patient will tend to minimize her problems (e.g., as she did when she described her relationships with family members in glow­ ing terms in spite of serious conflicts). 3. Patient still maintains contact with drug friends who may try to sabotage her abstinence goals. ADDITIONAL DATA FOR THE CASE CONCEPTUALIZATION In addition to the clinical interview, self-report inven­ tories provide data that are important in the conceptualization of the case. The following is a list of such questionnaires and scales. 1. Beck Depression Inventory 2. Beck Anxiety Inventory 3. Beck Hopelessness Scale 4. Dysfunctional Attitude Scale 5. Beliefs about Substance Use Scale 6. Relapse Prediction Scale 7. Craving Beliefs Questionnaire 8. Sociotropy-Autonomy Scale Beck Depression Inventory The Beck Depression Inventory (BDI) (Beck, Ward, Mendelson, Mock, & Erbaugh, 1961) is a self-report scale composed 94 COGNITIVE THERAPY OF SUBSTANCE ABUSE of 21 items, each comprising four statements reflecting gradations the intensity of a particular depressive symptom. The respondent chooses the statement that best corresponds to the way he or she has felt for the past week. The scale is intended for use within psychiat­ ric populations as a measure of the symptom severity of depressed mood and as a screening instrument for use with nonpsychiatric populations. Beck Anxiety Inventory The Beck Anxiety Inventory (BAI) (Beck, Epstein, Brown, & Steer, 1988) is a 21-item, self-report instrument designed to measure the severity of anxious symptoms. The BAI overlaps only minimally with the BDI and other measures of depression while measuring anxiety. Beck Hopelessness Scale The Beck Hopelessness Scale (BHS) was developed by Beck, Weissman, Lester, and Trexler (1974) to measure negative expectancy regarding the future. The BHS is composed of 20 true-false items assessing the expectation that one will not be able to overcome an unpleasant life situation or attain the things that one values in life. The BHS has demonstrated predictive validity for completed sui­ cide (Beck, Steer, Kovacs, & Garrison, 1985). Dysfunctional Attitude Scale The Dysfunctional Attitude Scale (DAS) (Weissman & Beck, 1978) is a self-report scale composed of 100 items. It was developed to assess underlying assumptions and beliefs that consti­ tute schemas by which individuals construe their life experiences. Beliefs about Substance Use The Beliefs About Substance Use inventory is a self- report scale composed of 20 items that can be scored on a range from 1 to 7. A "1" indicates that the person totally disagrees with the state­ ment. A score of "7" means that the person totally agrees with the statement. This scale measures many of the commonly held beliefs about drug use (see Appendix 1, page 311). Formulation of the Case 95 Relapse Prediction Scale The Relapse Prediction Scale (RPS) is a 50 item self- report scale. Each item is composed of situations that typically are reported to trigger urges for cocaine or crack. Each situation is rated on two dimensions: "Strength of Urges" and "Likelihood of Using," with all situations being rated on a 0-5 scale (0 = none to 5 = very high) (See Appendix 1, page 313). Craving Beliefs Questionnaire The Craving Beliefs Questionnaire (CBQ) is a self-report scale that measures beliefs about the craving phenomenon as it pertains to cocaine and crack. Each of 28 items is rated on a 1-7 scale (e.g., 1 = totally disagree and 7 = totally agree) (See Appendix 1, page 312). Sociotropy-Autonomy Scale The Sociotropy-Autonomy Scale (Beck, Epstein, & Harrison, 1983) is a measure of two broad personality dimensions that are associated with depression (Beck, 1967). One is "sociotropy," which refers to the degree of importance a person places on interper­ sonal affiliation in order to be happy. The second is "autonomy," referring to the degree to which a person believes he or she must achieve and attain success in order to be happy. Each of the 60 questions asks the respondent to rate the percent­ age of time that a

statement applies to himself or herself. Half of the questions indicate a sociotropic personality style, and the other half indicate an autonomous personality style. One example of a socio­ tropic item is, "I find it difficult to say no to people." There are five possible responses-this appUes to m e (1) 0%, (2) 25%, (3) 50%, (4) 75%, (5) 100% of the time. Sociotropy and autonomy subscores are summed separately; therefore, a person may be high on both scales or low on both scales. S U M M A R Y The purpose of this chapter has been to provide the therapist with a comprehensive methodology for achieving a sound case formulation. W e began with a rationale supporting the importance of a good case formulation, and suggested key questions for the clinician to ask 96 COGNITIVE THERAPY OF SUBSTANCE ABUSE when treating substance abuse patients. Next, we introduced what w consider the ten essential components of the formulation: (1) relevant childhood data, (2) current life problems, (3) core beliefs, (4) condi­ tional assumptions/beliefs/rules, (5) compensatory strategies, (6) vul­ nerable situations, (7) ATs, (8) emotions, (9) behaviors, and (10) inte­ gration of the data. Each component was described in detail and pertinent examples presented, followed by an illustrative case study. Next, we reviewed methods for gathering and organizing data for the case formulation. W e introduced the Cognitive Conceptualization Worksheet, provided an explanation of its use, and noted examples to aid in the understanding of the use of this form. This chapter con­ cluded with a description of eight self-report inventories used in pro­ ducing information toward a sound case formulation. C H A P T E R 6 S t r u c t u r e o f t h e T h e r a p y S e s s i o n T , he structure of the therapy session is one of the more noticeable Xahnd< essential characteristics of cognitive therapy. Stmcture is important for the following reasons: (1) Within a typical 50-minute session, substance abusers often present a large amount of material to discuss, either longstanding or acute crises, yet there is a limited amount of time. Structuring the session provides the oppor­ tunity to make maximum use of time. Patient and therapist collaborate to most effectively handle problems in the time allowed. (2) Struc­ ture assists in focusing on the most important current problems. (3) Learning new skills, such as better problem solving, requires hard work. Structuring the therapy session sets the tone for a working atmo­ sphere. (4) Structured sessions fight against therapy drift, whereby continuity from session to session is lost. Knowing the elements of the structured session facilitates adherence to the cognitive model and minimizes the chances that drift will occur. This chapter focuses on eight important elements of the struc­ ture of a session: 1. Setting the agenda 2. Mood check 3. Bridge from last session 4. Discussion of today's agenda items 5. Socratic questioning 6. Capsule summaries 7. Homework assignments 8. Feedback in the therapy session 97 98 COGNITFVE THERAPY OF SUBSTANCE ABUSE SETTING THE AGENDA Time is precious. Setting an agenda helps to make efficient use of time and provides a focus for the therapy session. It also teaches patients to set priorities, usually a skill deficiency in impulsive drug-addicted individuals. Because they spend a consider­ able amount of time seeking, using, or recovering from their drug use, patients often spend little time focusing on solving the other problems that are plaguing their lives. Setting agendas has a positive effect on the therapeutic alliance as well. It reinforces the collaborative agreement between patient and therapist as each party has an opportunity to contribute to the pro­ cess of therapy. It allows patient and therapist to target specific goals (see Chapter 8, this volume) for the session and to discuss the appro­ priateness of focusing on specific topics. It also sets the stage for modeling better ways of resolving conflicts, for example, when the patient's agenda item seems incompatible with what the therapist wants to discuss. This is illustrated in the case in which the patient says, "I want to give you all the details ... I just want to get it off my chest... It makes me feel better." In order to preserve collaboration, the therapist might reach an agreement with such a patient that a certain portion of the session can be used to "let off steam" but may also suggest that other topics will also need to be covered, such as ambivalence about abstinence, continued drug use, and triggers to using. Some patients have a low tolerance for anxiety and therefore avoid bringing up topics that provoke discomfort. W h e n therapists provide a rationale for putting such topics on the agenda in spite of the dis­ comfort the topics evoke, they help to avoid power struggles between themselves and their patients. Therapists also make good, collabora­ tive use of the agenda by demonstrating empathy for their patients' reluctance to discuss certain hot topics, such as their spouses' sub­ stance abuse problems. At times, therapists need to be flexible in setting agendas. Some­ times patients come to a session in crisis, such as after being fired from a job or being left by a spouse. These types of problems may require immediate attention, superseding ongoing issues. Likewise, a lapse or relapse should be dealt with immediately because patients who have used often feel hopeless about their ability to stay off drugs, and thus are at increased risk for a full relapse. This in turn often leads to feeling hopeless about therapy and may precipitate a prema­ ture flight from treatment. A key point to remember is the importance of the therapist's not structure of Therapy Session 99 being rigid or dictatorial in setting and following agendas. For ex ple, when it becomes clear that a high-priority agenda item will re­ quire most of the session to addresss adequately, the therapist needs to be willing to shelve less important topics. Also, therapists can modify agendas by periodically checking the number of agenda items to be covered and the amount of time left in the therapy session. If there is insufficient time, patient and therapist then collaborate on deciding which agenda items might need to be postponed. The following is a transcript of the beginning of a therapy ses­ sion in which the therapist and patient are setting an agenda collab­ oratively. In reading this transcript, keep in mind that the therapist is working with the patient to set an appropriate agenda with a spe­ cific target problem, to keep the agenda suitable for the amount of time available in the therapy session, and to prioritize the topics. TH: Well, what are we going to focus on today? PT: Some things, but you know ... m y burden now ... m y thing is that I need a job. TH: That is something important to put on the agenda for today. Are there other things that we need to talk about? For example, your current frequency of drug use? PT: No, I'm doing all right as far as m y drug things are concerned. TH: H o w much of anything have you used since the last time I saw you? PT: Nothing. TH Nothing? PT: Nothing. I go to m y meetings now. TH No drinking? No alcohol? PT: Nothing. TH OK. PT: When I wake up in the mornings now, I really don't have that craving for drugs. Do you know what I mean? So, now I have to just put this energy into getting up in the morning and get­ ting out and getting a job. That's a real problem; let's make sure we get to that. TH: OK. What I will do is write on the board a list of things that we need to cover. W e might not be able to get to everything today. First we have "finding a job"-difficulty getting up in the morn­ ing to go looking for one. The next thing I was going to put there had to do with "cravings." So, at some time today, I would like to talk with you about what happened that led to your last 100 COGNTTIVE THERAPY OF SUBSTANCE ABUSE slip. The whole point of doing this is for you to get more familiar with what happened so you can recognize when you are feel­ ing bad, what that usually leads to, and to try to come up with some ways to keep it from going all the way down to your using again. To the alcohol, then to the coke, then to the heroin. Oh, we also need to go over the homework from last week. PT: Finding a job is the most important. TH: OK. Maybe we can start with that. Is there anj^hing else you want us to focus on? PT: No, this is good. As this brief transcript illustrates, the therapist set the stage focusing on two primary goals for treatment: reducing drug use, "What happened that led to your slip?" and doing problem solving, "Find­ ing a job ... difficulty getting up." The therapist asked two of three important questions that should be asked at every session: (1) Have you used since the last session? (2) Have you had any urges/cravings to use? and (3) Are there any situations coming up before our next session where you might be at risk to use? (This third question was not asked by the therapist.) M O O D CHECK Since depression, anxiety, and hopelessness are inter­ nal stimuli that have the potential to trigger continued use and/or relapse, it is important to monitor these (and other) states. Therapists should pay special attention to feelings of hopelessness as it has been shown that a chronic, marked negative view of the future is one of the best predictors of suicide (Beck, Steer, et al., 1985). It is desirable to have the patient complete the BDI, BAI, and BHS at every session (see Chapter 5, this volume, for descriptions of these instruments). Scores and their meanings should be discussed with the patient, especially if there are substantial changes in scores. Sometimes there can be a change in mood as measured by these instruments, but the patient seems unaware of the change. Therefore, the thera­ pist should ask the patient if he or she is aware of changes in his or her mood. The therapist might say, "Your score on the BDI is higher this week, which may indicate that you have been feeling more depressed. Do you agree with that?" Important points to remember are (1) mood is an important vari­ able with regard to drug use and relapse, (2) hopelessness is one of structure of Therapy Session 101 the best predictors of suicide, (3) mood levels should be measured each session, and (4) therapists should discuss scores obtained from the BDI, BAI, and BHS with their patients. B R I D G E F R O M LAST SESSION Drug and alcohol abusers often have chaotic lives; therefore it is easy for therapists to get drawn into a pattern of jump­ ing from one topic to another in a disjointed fashion. Therapists should think carefully as to how they will stay focused and maintain continuity across therapy sessions. They should ask themselves, "How do the present agenda items relate to what was discussed in the pre­ vious session, and how do these items relate to the overall goals of treatment?" The therapist also reviews the patient's feedback about the pre­ vious session. There are two ways to accomplish this. First, the thera­ pist asks the patient if there is any unfinished business from the most recent session, including any negative reactions he or she might have had. Second, the therapist may reflect on the Patient's Report of Therapy Session (see Appendix 5, page 324), which patients are asked to complete after each session. Usually this is brief; however, some responses might require considerably more attention and time to address. For example, one patient reported after the last session that he did not expect to make any progress in that session, that he did not in fact make any progress, and that he did not

expect to make progress in future sessions. The therapist, recognizing that this feed­ back indicated that the patient held very negative views about therapy, suggested that this be discussed at some length in the current ses­ sion. To get a sense of the patient's world, it is helpful to review briefly the patient's life during the past week. Therapists can use activity schedules to structure this review. Therapists must encourage their patients to keep this review as brief as possible, so that it does not deteriorate into idle chit-chat about the patient's general goings-on that takes up valuable time in the session. DISCUSSION OF TODAY'S A G E N D A ITEMS When therapists and patients proceed to discuss the agreed-on agenda items for the session, they must bear in mind the following points. First, it is important to prioritize the list of topics. It is not always possible to discuss every item within the time con- 102 COGNITIVE THERAPY OF SUBSTANCE ABUSE straints of a given session. Some topics will need to be shelved u the following session. Therefore, it makes the most sense to deter­ mine which topic or topics are essential to discuss in the present session, and to discuss these topics first. By doing so, therapists can avert unfortunate problems such as an entire session being used to discuss a patient's complaints about his car troubles, only to find out as the sessions ends that the patient went on a drinking binge after his wife walked out on him. Clearly, this latter topic needed to be discussed first and foremost, not just for 2 minutes at the tail end of the session. Second, therapists must be alert to patients' tendencies to stray from agenda items and to go off on irrelevant tangents. A polite but prompt statement, such as "I don't mean to interrupt, but I think we should refocus on the topic we started talking about," usually is suf­ ficient. At times, when patients seem to "stray" to even more impor­ tant issues (e.g., a discussion of the patient's marriage leads into hints that the patient is contemplating suicide), it is advisable for thera­ pists to switch gears to accommodate and follow up on these impor­ tant topics by revising the agenda. In general, topics such as the patient's active drug use, suicidality, or hopelessness about therapy will supersede most other agenda items. Third, therapists need to be somewhat conscious of time in the session so that the various topics are covered in sufficient breadth and depth, and so that transitions from one agenda item to the next can be made in a timely manner. At times, therapists may choose to inter­ ject the following question in order to facilitate this process: "We're about halfway through the session, and we have a decision to make. Should we keep talking about our current topic a while longer, or would it make sense to wrap this up and go on to our next item?" This is a collaborative, flexible way to stay focused on meaningful therapeutic material, and to be as efficient as possible in making the best use of valuable therapy time. Fourth, therapists need not be stymied by patients who say "1 don't know" when asked what topics should be discussed as part of the agenda. (In fact, good cognitive therapists almost never take "I don't know" for an answer They persist nicely, find alternative ways to ask the question, or ask the patient to deliberate further.) Instead, thera­ pists can explain that one of the patient's responsibilities in therapy is to think about what he or she would like to talk about in session. At first, the therapist may assist the patient by suggesting some agenda items, asking, "Which of these is most important to you?" The thera­ pist may also ask, "What has been on your mind lately? What's on your mind right now?" Later, if the patient continues to be unwill- Structure of Therapy Session 103 ing or unable to generate topics for discussion, this problem in a itself can become an important agenda item. For example, the therapist may say, "Let's discuss your difficulty in thinking of things to talk about in session. Let's try to understand where the problem is, and how to overcome it." In doing so, the therapist avoids falling into the trap of accepting the patient's helplessness or resistance as an unchangeable fact. In addition, the patient learns that saying "I don't know" will not be reinforced, and that this strategy will fail as an intended means of escape from the work of therapy. S O C R A T I C Q U E S T I O N I N G Overholser (1987, 1988) defines Socratic questioning as a method of intervening that encourages the patient to contem­ plate, evaluate, and synthesize diverse sources of information. This type of questioning, also referred to as "guided discovery," is utilized over the entire span of the session. In contrast to questions typically designed for the therapist to gather information regarding the frequency, intensity, and duration of the substance abuse problem, Socratic questioning is used to bring information into the awareness of the patient. Therefore, Socratic questions are designed to promote insight and better rational decision­ making. Questions should be phrased in such a way that they stimu­ late thought and increase awareness, rather than requiring a correct answer. The proper choice, phrasing, and ordering of questions has a strong impact on the organization of thought in the patient. Further, we have found that most of our drug-abusing patients respond more favorably to exploratory questioning than to didactic "lecturing." Socratic questioning is a powerful technique to use while discuss­ ing the various agenda items. Therapist asks questions in such a way as to help patients to examine their thinking, to reflect on erroneous conclusions, and, at times, to come up with better solutions to prob­ lems. This often leads to patients' questioning, and thereby gaining greater objectivity from, their own thoughts, motives, and behaviors. Also, Socratic questioning establishes a nonjudgmental atmosphere and thus facilitates collaboration between patients and therapists. This can help patients come to their own conclusions about the serious­ ness of their drug abuse problem. As a rule of thumb, therapists should start utilizing Socratic ques­ tioning from the beginning of treatment. This helps to orient patients to an active thinking mode. If therapists find that Socratic question­ ing appears to be overwhelming patients more than helping, then the 104 COGNITIVE THERAPY OF SUBSTANCE ABUSE therapists may choose to be more direct, such as pointing out inco sistencies and errors in thinking and asking if the patients agree with and follow this logic. While it is important to use questioning to explore problems and to help patients draw their own conclusions, there should be a bal­ ance between questioning and other more direct modes of interven­ tion, such as reflection, clarification, giving feedback, and educating the patient. The following dialogue illustrates such a balance, with the therapist starting with some basic assessment questions: TH: Charleen, have you used any drugs or alcohol this week? PT: No, none. It's been over a month now. TH: What about your pain medication from the dentist? PT: What about it? TH: Well, I have a number of questions. First, are you taking the amount that you're supposed to take, and not more than that? Are you taking it when you're supposed to take it, and not more often than that? PT: I'm doing just what I'm supposed to do, so don't worry. TH: Do you know why I'm asking? Do you know why it matters? PT: Yeh, because you told m e that pain medication is like a dmg... TH: Not like a drug. It is a drug. It's a mild narcotic. PT: And I could get addicted to it. TH: Right. And do you know why I'm concerned about the amount and the frequency with which you're taking it? PT: No. TH: Think about it for a minute. W h y do you think we should be concerned about it? PT: I don't know. TH: Well, I realize that you might not know exactly why, but could you try to guess some possible reasons. I'll be happy to tell you m y reasons after you give m e your theory. Note that the therapist is asking a number of open-ended questions in the hope that the patient will begin to do some active thinking in the session. The fact that the patient does not respond to the latter question does not deter the therapist. Instead, he finds a tactftil, col­ laborative way to encourage the patient to apply some cognitive effort. Later, he plans to "reward" the patient for her effort by giving her some additional information in order to educate her about the dangers structure of Therapy Session 105 of pain medications in the hands of a recovering addict. First, how­ ever, he continues with some Socratic questioning. PT: I guess if I took more than I'm supposed to, I could get addicted faster. TH: That's right. What else? PT: I could get high on the pain medication and lose control and go out and use other drugs. TH: Absolutely right. Excellent answer. See, you do understand. Any­ thing else you can think of? PT: No. Not really. TH: Well, consider this. What would happen if you ran out of the pain medication before you were supposed to run out? PT: I'd have to get more. TH: Yes, but if the dentist knows that you're supposed to stiU have some medication left, and you're already asking for more, what would happen? PT: He might say no. TH: What might you do then? PT: I might have to find some other way to kill the pain. TH: Such as? PT: Such as whiskey, {laughs) TH: W h y would that be a problem? PT: Because then I would blow m y streak of staying off stuff I shouldn't take. TH: And would you just drink whiskey? PT: I might also use crack if I had the chance. TH: Right. Now, you've worked very hard to get to this point. It would be a crying shame if you set yourself back by taking too much pain medication. PT: I agree. TH: So, Charieen, have you been taking the medication as pre­ scribed? PT: Yes, but I still have pain, so I've been taking the Advils and the Tylenols too. At this point, the therapist is satisfied that the patient has arrived at her own conclusion that she could be at risk for a lapse or a full­ blown relapse if she misuses her prescribed medications in any way. 106 COGNITFVE THERAPY OF SUBSTANCE ABUSE At the same time, he has just heard something a bit disturbing; the fore he will ask for clarification before proceeding with some non- Socratic, didactic education. TH: Uh oh. You're taking more medications? H o w much? PT: {Getting a little annoyed) Until I feel better, that's how much! TH: Do you read the instructions before taking the over-the-counter medications? PT: No, I just take it until I feel better. TH: Charleen, please bear with me for a few minutes. I can tell you're getting a little ticked off right now, and I don't mean to get you angry but this is important. Can you hear m e out? PT: Do I have a choice? TH: Well, yes. You could ignore m e if you wanted to, but I'm hop­ ing you'll give me a chance to make m y point before you decide whether to disregard it or not. PT: Go ahead. TH: Thanks, I appreciate your being a good sport. Charleen, there are good reasons why medications have instructions. If people ignore the instructions, they can overdose. Or, they can cause something called "interactions" with other drugs. In your case, the over-the-counter medications could combine with the den­ tist's medication to create an effect in your body that's equal to many, many, many medications, which could be dangerous. Also—and I'm not sure that you knew this—every time a person takes a pain medication he lowers his body's own natural abil­ ity to kill pain. So, if you take too much of

anything, it can suppress your ability to feel well after you stop taking the medi­ cation. You see, if the medication runs out, and you've sup­ pressed your body's own natural abilities to kill pain, you're going to go into withdrawal and be in a lot of discomfort. Then, you won't be able to get a refill of the dentist's medication and you'll probably think that you have no choice but to drink whiskey or get some crack. That's why it's so important for you to take only what is prescribed, and nothing more, not even over- the-counter stuff. Do you get m y point? PT: You mean if I take these medications, m y body will never be able to kill pain by itself? TH: Not "never." It will just be suppressed for a few days. That's the withdrawal phase, just like for any drugs. But can you hold off on using drugs for a few days when you're in pain? structure of Therapy Session 107 PT: No way. TH: That's m y point. If you go on the way you're going on right now, you'll be in danger of using alcohol and crack, especially when you run out of the prescription. PT: I see. What should I do? TH: That's an excellent question. Can I turn it back to you? What do you think you should do? Now, the therapist shifts back into the mode of Socratic questioni PT: I guess I have to stop taking the Advils and the Tylenols. TH: And how about the dentist's medication? PT: I guess I have to make sure I read the instructions. TH: But what if you do exactly what you're supposed to do, and you're still in pain? PT: I don't know. TH What do you think I would do in your situation? PT You would call the dentist. TH I might do that, yes. What else? PT: You would try distracting yourself with activities, right? TH: Correct! Could you try that? PT: I could try. TH: What kinds of things could you do? The therapist continues to ask open-ended questions so that Charleen can generate her own interventions, the likes of which she is more likely to follow between sessions than those interventions simply directed toward her. Thus, the dialogue has demonstrated that a mix­ ture of interventions, including education, clarification, and Socratic questioning, can help patients to do meaningful work in session, and to elicit the maximum amount of information and cooperation. CAPSULE SUMMARIES Capsule summaries are an important part of the learn­ ing process in therapy sessions. As a general rule, therapists and patients should summarize what has been discussed in a session a minimum of three times. This provides opportunities to adjust agen­ das and to maintain the focus of the therapy session. The first cap­ sule summary typically is done after the agenda has been established. 108 COGNTTIVE THERAPY OF SUBSTANCE ABUSE the second one approximately halfway through the therapy session, and last, toward the conclusion of the therapy session. The first summary helps patients make a connection between the agenda of the present session and the long-term goals of therapy. The following represents a typical first capsule summary: TH: OK, let's summarize what we are going to focus on today. One thing is the situation when you had the strong urge to pick up on some crack. Second is your situation at work. You are anx­ ious about the fact that you might be laid off. Was there any­ thing I missed? PT: No, that's it. TH: Both of these issues fit very nicely with your long-term goals of treatment, one being coping with urges to use crack, finding other methods for coping with anxiety, and, last, your concerns about employment and saving money. Do you see how they connect? PT: Yeh, it all makes sense. The second summary helps the therapist to collect his own thoughts, to decide what to do next (such as advancing to the next item on the agenda), to convey understanding of the patient and provide an opportunity to correct any misunderstanding, and to make the therapy process more understandable to the patient. Initially in treatment the final summary is done by the therapist. However, as therapy progresses the therapist should move very quickly to get the patient to do end-of-session summaries. When patients summarize, it gives them responsibility for processing the session, and it lets therapists check on patients' understanding of what went on in the session. Further, patients improve their retention for the contents of the session when they actively review what has been dis­ cussed. H O M E W O R K ASSIGNMENTS The homework assignment is a collaborative enter­ prise generated and agreed on by the therapist and patient as a team. Its two main functions are to serve as a bridge between sessions, ensuring that the patient continues to work on his or her problems, and to provide an opportunity for the patient to collect information to test erroneous beliefs and to try new behaviors (Blackburn & David­ son, 1990). Structure of Therapy Session 109 Patients are encouraged to view homework as an integral and vital component of treatment (Burns & Auerbach, 1992; Burns & Nolen- Hoeksema, 1991; Persons, Burns, & Perioff, 1988). Since the therapy session is time-limited, normally less than an hour, homework assign­ ments become extremely important as they offer patients ongoing opportunities to practice various skills that they have been taught in the therapy session. It is best to assign homeworks that draw from the therapy ses­ sion, as homework is most effective when it is a logical extension of the therapy session (Newman, 1993). This can be done by reviewing what has happened in the therapy session and then focusing on how these points or lessons can be continued and reinforced outside treat­ ment. Ideally, such assignments ultimately lead to the continued use of new skills, even after the termination of formal treatment. It is generally advisable to review the previous week's homework as an early agenda item in each therapy session. By doing so, thera­ pists convey to patients that homework is an important part of the therapy process (Burns & Auerbach, 1992). Also, by reviewing home­ work from previous sessions, therapists can correct patients' mistakes early in freatment—for example, in completing a Daily Thought Record (DTR) (see Chapter 9, this volume). By making sure that the home­ work assignment is reviewed, therapists can make certain that patients are practicing new cognitive and behavioral skills correctly. Therapists who neglect to review the homework in each session create three problems. First, the patients usually begin to think that the homework is not important and, therefore, that treatment is some­ thing done to them rather than something they actively work on even in the absence of the therapist. Second, the therapists miss oppor­ tunities to correct mistakes such as the patients' inadequately respond­ ing rationally to their automatic thoughts. Third, the therapists lose the chance to draw helpful lessons from the homework and to rein­ force these lessons. The therapist can minimize patient noncompliance by being sure to explain the rationale for the assignment and by discussing with the patient any possible or expected difficulties (Newman, 1993). For example, the therapist might ask: "What are some things that could happen that might get in the way of completing the assignment?" and/or "What are the odds of your completing the assignment?" In addition, if the therapist has some doubt about the patient's under­ standing of the task, he or she should, if possible, rehearse the assign­ ment before the patient leaves the session. If a homework assignment is not carried out, therapists should address this issue. One method is to use the "Possible Reasons for Not 110 COGNTTIVE THERAPY OF SUBSTANCE ABUSE Doing Self-Help Assignment" Checklist (see Appendix 6, page 327). This checklist helps to identify those reasons why patients often do not do homework assignments. The therapist can pull out a copy of this list and ask the patient to select those items that apply to his/her non­ compliance. The following are some examples of items on this check­ list: "I don't have enough time, I'm too busy," "I feel helpless and 1 don't really believe that I can do anything that I choose to do," and "It seems that nothing can help m e so there is no point in trying." These beliefs become new targets for examination and testing. In summary, the homework assignment functions as a bridge between therapy sessions and provides an opportunity to test beliefs and prac­ tice skills learned in the session. The task should be a logical exten­ sion of the session and be relevant to the goals of therapy. The thera­ pist can minimize noncompliance by giving rationales for assignments and discussing possible difficulties with the patient. To facilitate patients' understanding, homework assignments can be rehearsed in session. Therapists should explain the importance of homework, and are advised to review assignments at each session. Incomplete assign­ ments should be discussed as an agenda item in the session. The rea­ sons that patients cite for not doing the homework can be ascertained through questioning or a checklist, and these reasons are treated as beliefs to be tested. F E E D B A C K IN T H E T H E R A P Y SESSION Therapists and patients regularly exchange feedback during therapy sessions. Throughout the session, the therapist asks questions to be sure that the patient understands what the therapist has said and where the therapist is heading. For example, the thera­ pist might ask, "Can you tell me what point I'm trying to make with these questions?" Sometimes patients misunderstand what therapists are trying to accomplish. Asking questions at these points gives patient and therapist an opportunity to clarify miscommunications in the therapy session. At the end of the session, the therapist should try to get feedback from the patient regarding (1) what was learned in the session, (2) how the patient felt during the therapy session, and (3) how the patient feels about the therapy in general. For example, the therapist might ask the following questions: "What did you get out of today's session?" "Was there anything that I said or did that rubbed you the wrong way during today's session?" "Do you feel we are accomplishing something useful?" Other ways of eliciting feedback include responding to nonverbal Stmcture of Therapy Session 111 behavior in the therapy session. For example, if the therapist not that the patient is frowning, the therapist might say, "I noticed you just had a frown on your face. What thoughts were going through your mind right then?" This will often result in eliciting valuable feed­ back. The key points to remember are that the therapist should endeavor to become adept at eliciting and responding to verbal and nonverbal feedback throughout the therapy session, that the thera­ pist should regularly check for the patient's understanding of what is going on in the therapy session, and that key points should be sum­ marized periodically throughout the therapy session. This, in turn, helps to build a strong collaborative relationship. SUMMARY In this chapter the importance of session structure and its eight components are discussed. Setting agendas helps to make maximum use of time, keeps the sessions focused, sets the tone for a working atmosphere, and counters therapist drift. Repeated mood checks identify changes in mood that might lead to relapse. Bridging sessions provides continuity across sessions and keeps therapy ses­ sions focused on goals of treatment. In discussing the list of agenda items, therapists help their patients to prioritize the list, to stay focused on important material, to make the most efficient use of time, and to contribute actively to the discussion. Also, therapists use skillful Socratic questioning as often as possible, which helps patients make their own discoveries. Capsule summaries should occur at least three times in a session. The importance of homework must be conveyed to the patient, and the appropriate steps for minimizing noncompli­ ance should be taken. Therapists provide and elicit feedback to clear up possible misunderstanding and/or misinterpretation of what is happening in the session. C H A P T E R 7 E d u c

a t i n g P a t i e n t s i n the Cognitive Model A J. ̂ s the cognitive therapy of substance abuse is a collaborative enterprise between therapist and patient, it is essential that patients gain a conceptual grasp of the key components in the model, such as understanding the associations and causal relation­ ships between cognition, affect, behavior, craving, and using. Patients need to learn about the phenomenon of automatic thoughts and the key elements for testing hypotheses. Some therapists start educating drug abuse patients in the treat­ ment model before they themselves have gained an adequate under­ standing about their patients' formulations of the various problems. In doing so, therapists may miss an opportunity to foster an atmo­ sphere of teamwork that is so important to nurture early in freatment. Asking patients for their views helps to nurture the collaboration. In some cases, patients are quite aware of the specifics of their problems; they just feel stuck and are not sure what steps they need to take to arrive at functional solutions. While gathering the patients' formula­ tions, therapists can begin to educate patients in the cognitive therapy model by focusing on the beliefs that are inherent in their interpreta­ tions of their drug problems. ELICIT PATIENTS' F O R M U L A T I O N S OF THE PROBLEM Substance abuse patients generally have explanations for their drug problems, such as "I have a high-stress job," "Today everybody uses drugs," "It's this marriage; if she would only change," and so on. By beginning to explore these "reasons," the therapist starts to understand the patient's "internal reality" and to establish a col­ laborative set for therapy. 112 Educating Patients 113 Therapists ask patients how they believe their drug problem devel­ oped, and how they would explain their current difficulties with work, relationships, the law, and other important life areas. In addition, therapists inquire about what the patients think they must do to solve their drug problem. Similarly, clinicians ask why patients believe that they have not been willing or able to solve their drug and general life problems on their own to this point. Although patients may present what appear to be understandable reasons for their drug abuse, they usually have some degree of doubt (as do their skeptical therapists!). H o w much they believe their own explanation can be assessed by asking them to rate it on a scale of 0 to 100, with 0 meaning they do not believe it at all and 100 meaning they believe it completely. This subtie tactic begins to teach patients that their beliefs are not the same as facts and will be subject to evaluation. For example, one patient explained to the therapist during the ini­ tial therapy session that he believed his alcohol and cocaine use were the direct result of where he lived. He stated that the conditions were miserable: "There's high unemployment, poor housing, and drugs all over the place." He believed that anyone with these types of hard­ ships would also be drinking and using cocaine. When the therapist asked how much he believed in his own explanation, the patient re­ plied "85%." The patient stated that there were indeed some people in the area who did not use drugs, but they were "religious." However, on closer review, the patient noted that there were two members of his own family who were using neither drugs nor alcohol—his sister and father. The patient was given an assignment to list all of the people on his street that he believed used and those who he believed did not use. Although there were a large number of users, a clear majority of people on his street did not use drugs or abuse alcohol. The patient was quite surprised with these findings. Therapist and patient then agreed that looking at other explanations for his cocaine and alcohol abuse would be worth pursuing. DEMONSTRATE THE RELATIONSHIP BETWEEN SITUATIONS, COGNITIONS, AFFECT, CRAVING, BEHAVIOR, AND DRUG USE The following is an example of a didactic presenta­ tion: "An automatic thought is a spontaneous thought or picture in your mind. Right now you might not pay much attention to these 114 COGNITIVE THERAPY OF SUBSTANCE ABUSE thoughts or pictures or make any connection between them and how you feel, but they do in fact affect your emotions and cravings for drugs. Automatic thoughts can be related to past, present, or future events, such as 'I knew I should not have gone in that bar,' 'I can't stand this craving,' and 'Oh hell, I'm never going to kick this prob­ lem.' Furthermore, these thoughts seem completely believable when they occur; therefore, they are accepted as fact without question. Further, they seem to make sense in spite of evidence to the contrary." A powerful method for teaching patients how to recognize auto­ matic thoughts is to have them relate their ongoing thoughts live in the therapy sessions. Therapists might say, "I want you to remember when you were in the waiting room just prior to this session. How were you feeling?" Patients sometimes will respond by saying "anx­ ious," "nervous," "unsure of myself," "bored," "angry," and other emotions. Therapists can then ask, "What was going through your mind right then as you were sitting out there?" Some typical responses are, "I hope no one sees me here," "1 wonder what this is going to be like," "Is the therapist going to like me?" "Am 1 going to be able to do this?" "What is this therapy all about?" "What am I doing here?" "I won't be able to stop using," and "I don't need to see a shrink." Patients then are told that these are examples of automatic thoughts and that they have direct bearing on the aforementioned emotions. For home­ work, patients may be instructed to self-monitor some of their auto­ matic thoughts between therapy sessions. They are asked, for example, to write down their thoughts while feeling depressed, bored, anxious, angry, and especially when having cravings or urges to use drugs. The next phase entails demonstrating the relationship between situations, emotions, cognitions, behaviors, and cravings. This may be accomplished by using examples patients bring to the therapy sessions. Therapists use the patients' examples to show how the patients' thoughts played a role in their negative feelings, their urges to use drugs, and their resultant drug-related actions. For example, "Walter" stated that he was extremely angry because the therapist had implied that the patient's failure to attend a recent therapy session was drug-related. He added, "1 was so pissed off that 1 thought about going out and getting bombed." The therapist, rather than becoming defensive, seized this opportunity to teach Walter about the connections between situations, thoughts, emotions, and drug urges. Specifically, he helped Walter to realize that his thoughts about "getting bombed" did not arise spontaneously, nor did his urges to use result purely from having a chemical addiction. Rather, Walter's thought that the therapist distrusted him triggered an angry reaction Educating Patients 115 (a thought led to an emotion), and the anger in turn sparked though about reasserting control in the therapeutic relationship through using drags (an emotion led to a drug-related belief), and this finally stimu­ lated an urge to use (a drug-related belief led to a drug urge). Thus, a potentially destructive interaction between patient and therapist was turned into an opportunity to learn about the cognitive model of drug abuse. As patients become more skillful at making their own connec­ tions between situations, affect, cognitions, and craving, the thera­ pist can begin to discuss with them the concept of beliefs. The therapist explains that the way we interpret events is largely determined by our belief systems. Beliefs tend to lie in a dormant state out of awareness until they are activated by specific situations. Patients who have difficulty understanding the notion of beliefs within a drug use context may be given the following nonclinical example: "A person has the belief that 'all people are created equal.' This is a belief that usually is dormant; it is not a statement that he goes through life thinking to himself. However, under certain circum­ stances, this belief is activated, such as when the person sees an injustice occur (e.g., someone who is guilty of a serious crime is set free because he is wealthy and influential). In these examples, the belief 'all people are created equal' is activated because circumstances occur that have to do with the belief. This, in turn, leads to a series of automatic thoughts, such as 'This shouldn't happen,' 'This is unfair,' and 'Why is this happening?' At this point, when the belief is triggered, the automatic thoughts are brought to awareness for that person and he becomes righteously angry." Therapists explain that while the above belief is adaptive, many drag-related beliefs are not, and the thoughts and feelings they lead to may make matters worse by triggering drag cravings and urges. Dysfunctional beliefs lead people to misinterpret situations, to over- generalize, to exaggerate, to see things in all-or-none terms, and to engage in other errors of thinking. Drug-abusing patients often make the mistake of assuming that their dysfunctional beliefs are valid and, therefore, that their interpretations are correct. One way to illustrate the notion of accurate/inaccurate beliefs is to remind patients of the story of Christopher Columbus. At one time, many people in the Western world held the belief that the world was flat. However, this did not make it so. It took a bold expedition to test the accuracy of this belief. One might say that when Columbus proved that the worid was not flat, many people updated their information or belief about the surface of the worid. However, there were still a few skeptics who 116 COGNTTIVE THERAPY OF SUBSTANCE ABUSE held on rigidly to the idea that the world was flat. In the same wa people may believe that drags are the only way to feel good, even though drags have time and time again been proven to cause more misery than joy. As pointed out earlier (Chapter 2, this volume), addictive beliefs are an essential component in the sequence leading to compulsive using. Patients often start off with the belief that their drag of choice is not harmful. They may also have the belief that they function bet­ ter with other people when under the influence of drugs such as cocaine. This belief can lie dormant until the patient faces a situa­ tion such as a social event. Being informed of an upcoming party may activate this belief and thus lead to craving, followed by certain auto­ matic thoughts that give permission to use the cocaine. This in tum leads to a series of habitual behaviors that facilitate finding the cocaine. Patients also have beliefs that develop over time. For example, a patient may start off with the belief that it is O K to use cocaine, that "it's not addictive." He may also entertain the belief that using cocaine makes him more sociable. However, over time, he might also develop a new set of beliefs regarding the cocaine, such as "I can't be social without it" and "I must have the cocaine in order to function." In this scenario the patient has moved from being a recreational user to substance dependence. In educating drag-abusing patients about the cognitive therapy model, it is helpful to teach them to identify drag-related beliefs. As stated earlier, drag-related beliefs can take different forms. They may be (1) beliefs about the drug itself, such as "Cocaine is not addictive," (2) beliefs about what is expected from the drug, such as "Coke will help to chill me out," and (3) permission-giving beliefs, such as "1 deserve to feel good." Mr. C, a 34-year-old polysubstance abuser, had the initial belief that he could not become addicted to using cocaine. He started off using cocaine socially-"only" at parties. Later he developed the belief that he could work better using cocaine. This, in turn, led to his use of cocaine at work when he was under pressure to meet deadlines. He had the illusion that he was much more productive at work when using coke. However, he overlooked the large amount

of money he was spending on cocaine. Furthermore, he was not as productive because he began to miss days at work after cocaine binges. Later, when he tried to stop and he began to experience strong urges and cravings, Mr. C's beliefs centered on the cravings themselves. Some of these were, "I can't stand the craving," "These feelings won't go away," and "The urges make m e use." Educating Patients 117 In order to assist therapists in the process of identifying patien beliefs about drags, we have developed the Beliefs About Substance Use questionnaire (see Appendix 1, page 311). This questionnaire, clinically generated, lists 29 common beliefs that patients report about substance use. W e have found that patients who are otherwise un­ skilled in reporting their thoughts can recognize beliefs that they main­ tain regarding drag use when they utilize this inventory. A noteworthy feature of the Beliefs About Substance Use questionnaire is that it allows patients to endorse beliefs that are contradictory to one another. This can be important in helping patients understand that they may hold conflicting beliefs about their drag-taking behavior. Thus, both their ambivalence and the dysfunctional nature of their thinking styles may be highlighted. Another useful method for identifying beliefs is via inductive questioning (also known as the "downward arrow technique"; cf. Burns, 1980). The substance-abusing patient first recognizes an auto­ matic thought; then the patient and therapist attempt to understand underlying meanings of the thoughts. This technique is illustrated by the following example: Mr. C. reported that during the week he was feeling extremely angry and anxious. The therapist then asked the patient to describe the specific situation. Mr. C. reported that while at a party, he had seen other people using and also saw some of his drug buddies, and he started having urges to use. TH: What thoughts were going through your mind right then? PT: It's not fair; they can use and I can't. TH: Let's presume for the moment this thought is accurate. What about it is important—what does it mean to you? PT: I'll never be able to use again. TH: And if you'll never be able to use again, what will be the sig­ nificance of thatl PT: I'll always have these urges and feel anxious. TH: And what are the implications for your life? PT: There's no escape. I'm trapped and helpless. As one can see, an important core belief ultimately was uncov­ ered through using the downward arrow technique (see Figure 7.1), and the patient came to understand the role of beliefs in his prob­ lems a bit more clearly. 118 COGNTTIVE THERAPY OF SUBSTANCE ABUSE "It's not fair. They can use and I can't." Automatic Thoughts I'll never be able to use again." "I'll always have these urges and feel anxious." Deeper core beliefs "There's no escape. I'm trapped and helpless." FIGURE 7.1. The downward arrow technique. THE "CRAVING SCENARIO" Therapists can teach their patients a great deal about the patients' substance abuse as seen within a cognitive therapy per­ spective by diagramming a "craving scenario." Essentially, this entails spelling out the cognitive m o d e l of substance abuse in the form of a flow chart, complete with examples that pertain directly to the patient's subjective experiences with drags. For example, M s . L. reported a cocaine lapse to her therapist, w h o proceeded to m a p out the patient's "craving scenario" (see Figure 7.2), w h i c h highlighted the sequence of events a n d beliefs that led u p to the actual episode of drag use. PROBLEMS IN EDUCATING PATIENTS IN THE COGNITIVE MODEL Sometimes patients initially fail to identify their auto­ matic thoughts. They report, "I don't have any thoughts." In order to overcome this problem, therapists may wait for the patients to dem­ onstrate affective shifts in the therapy session and then ask, "What is going through your mind right now?" W h e n aroused in this man­ ner, patients often have access to their thoughts. Patients are also asked, "If you don't have any thoughts, can you report what you are feeling?" Oftentimes, they report feelings in the form of cognitions; for example, "I feel like I don't want to be here today." Initially, thera­ pists may choose to accept these responses as feelings, but, at a later point, it will be important to educate patients to make a distinction Educating Patients 119 between cognitions, such as "I don't feel like being here today," and feelings, such as anxiety, anger, sadness, shame, and guilt. S o m e patients have difficulty labeling the particular feelings that they are experiencing. They say such things as "I feel like shit" or "I feel awful." O n e m e t h o d to help patients label feelings is to encour­ age them to use the simplest terms possible in describing their feel­ ings, for example, m a d , sad, or glad. Also, if a patient says, "I'm upset," the therapist could ask, "And where in your body d o you experience this feeling of being upset?" T h e patients m a y then report s o m e type of physiological indicator such as tightness in the stomach, tightness in the chest, stiff neck, and so on. Patients can be taught to use these bodily sensation cues to ask themselves the important question: "What's going through m y m i n d right n o w ? " With repetition, patients eventually c o m e to understand that it is important to notice and to modify their thoughts and beliefs. stimulus Circumstances—External Ms. L was at a party, met up with some of her former drug buddies. Stimulus Circumstances—Internal Had a sudden feeling of euphoria. Later became tired; felt bored, sad. Drug Beliefs Activated "Using Is a lot of fun." "I can't stand the boredom." "My life is screwed up anyway so why not use?" Core Beliefs "I'm a failure." "My life is hopeless.' Automatic Thoughts "Go for It." "This will be great." "They're having a blast." Focus on Short-Term Solution Ms. L joins the group and snorts a few lines. FIGURE 7.2. The craving scenario. 120 COGNITIVE THERAPY OF SUBSTANCE ABUSE SUMMARY An important part of the early stages of therapy involves educating patients about the cognitive model of drag addic­ tion and its treatment. Most fundamental to this process is the therapist's explication of the causal and correlational connections between stimulus situations, thoughts, beliefs, emotions, drag urges, and drag taking. Therapists can achieve this important goal by highlighting natu­ rally occurring sequences of events in the patients' lives, as well as in session. In this manner, patients learn that therapy entails much more than simply venting about problems and/or being persuaded to give up their drag use. Rather, patients learn that their drag prob­ lems involve an understandable series of external and internal events that, left undiscovered and unmanaged, "automatically" lead to drag use. They learn that these same events, once understood, offer a num­ ber of choice points for patients to minimize the chance that they will experience opportunities, urges, and actions that will perpetuate drug use. Most important in this process is the patients' understand­ ing that their automatic thoughts, triggered by core beliefs and beliefs about drugs, play an important role in their addiction. As a result, they learn that by modifying these thoughts and beliefs, they assist themselves in the process of recovery. The "craving scenario" is one useful method for illustrating the series of external and internal events outlined above. Another useful tool is questionnaires, such as the Beliefs About Substance Use inven­ tory, which helps patients to recognize some of the implicit beliefs that fuel their drag use. In addition, skillful questioning by therapists can help patients to illuminate the cenfral role that their thinking plays in any situation that is pertinent to their risk for using drags. Ulti­ mately, the patients themselves become adept at modifying their thinking styles, a vital therapeutic step. C H A P T E R 8 S e t t i n g G o a l s T X here is an old saying that maintains, "If you don't know where you're going, you won't know when you get there." This statement rings especially trae for therapy with addicted indi­ viduals. In this chapter, we examine the reasons it is important to establish goals for therapy. Setting goals creates a therapy map that helps to give a sense of direction to patients and therapists. Generally, patients enter treatment wanting to feel better—to get rid of the depression, anxiety, panic, and other negative affect states. W h e n the patient and therapist agree on a set of goals, they collaboratively focus on change in the patients' behavior, for example, being abstinent from cocaine and finding bet­ ter ways of solving real-life problems. In the absence of clearly defined goals, therapy sessions often appear fragmented or disjointed. In one case, a patient came in with a long histoty of cocaine use. However, her presenting problem was wanting to "feel better" about herself. Through careful questioning, the therapist was able to conclude that this might be achieved by her going back to school, finishing school, and pursuing a career. Since her use of cocaine was interfering with these goals to a large degree, the therapist assumed that one of her goals was to become totally abstinent. It was not until they were halfway through the first therapy session that it became apparent that the patient planned to continue her cocaine use, "but only on the weekends," and wanted to work only on the goal of going back to school. She saw no contradiction between these goals. Formulating goals tends to make explicit what patients can expect from treatment. Sometimes patients' expectations of therapy are unrealistic. By initially discussing the expected outcome of therapy and defining it in concrete behavioral terms, patients know where therapy is headed and, through understanding the cognitive therapy 121 122 COGNITIVE THERAPY OF SUBSTANCE ABUSE model, have an idea as to how the goals will be obtained. In sum, a alluded to at the beginning of the chapter, patients will know where they are going and how they are going to get there. Focusing on the expected outcome of therapy tends to make patients feel hopeful about the possibility of change. Many substance abusers have made numerous unsuccessful attempts in the past to stop on their own and feel hopeless about kicking their habit. Clear goals direct their attention to the possibilities for change. For example, "Jake" entered treatment and, at the vety first therapy session, stated that he could not imagine himself being off of cocaine, even though he had recently finished a detoxification program and had not used cocaine for several weeks. It was a difficult (but not impossible) task to get this patient to focus on the possibility of change and to try to imagine himself being cocaine free and coping with many of the other concerns that he brought into therapy—excessive debt and family dis­ cord. Setting goals helps to prevent therapy drift. Many substance abus­ ers enter therapy only after accumulating many problems such as the loss of a job, marital problems, poor health, depression, and anxiety. With so many presenting concerns, it is quite easy to shift haphaz­ ardly from one topic to another in each session. Knowing specifically the long-term goals of therapy and the priority order in which they will be addressed helps to prevent drift from taking place. The thera­ pist and patient can focus on one or two of the most immediate and pressing problems, yet still fully realize that there are additional issues that will be dealt with as therapy progresses. Specific goals tend to act as anchors and thus make it more obvious to the therapist when therapy has taken a turn in a new direction. For example, consider the case of a patient who came into treatment for help with his anxiety about abstinence from cocaine. He had not used cocaine for over a year, but in the past few weeks he noticed that he was becoming more and more anxious about the possibility of using again.

The initial goal in therapy was to help the patient develop better ways of coping with his anxiety about some of the urges and cravings that were reappearing. Although this was the primaty focus of therapy, it became obvious after several sessions (when the patient was less anxious about relapse) that he was also experiencing severe marital discord. He disclosed that his wife was also abusing diet pills and alcohol and that this was causing a great deal of strife between them. Knowing the original goal of therapy, which was to help him deal with the anxiety about a possible relapse, the therapist and patient were able to see that focusing on marital concerns was going to be a change in the original treatment plan. It Setting Goals 123 would have been vety easy to drift automatically into working on t marital discord at the expense of a discussion about managing anxiety and cravings for drags. Instead, patient and therapist put both topics on the therapeutic agenda, allotted a certain proportion of session time for each, and decided that future sessions would explore the causal connection between his renewed drag cravings and his wife's substance abuse. The setting of mutually established goals reinforces the therapeu­ tic alliance and the spirit of collaboration between patient and thera­ pist. It also gives the patient a sense of active participation in his or her treatment. This is especially important for substance abusers who often see their lives in disarray and feel out of control and at the mercy of their dependency. Collaboratively setting goals aids in fostering the patient's sense of efficacy and confidence to overcome drag depen­ dence and other problems. For example, a young cocaine addict reported the following after a goal-setting session: For the first time since he had tried stopping on his own he had some sense of control over his life and it was clear what he wanted to get from therapy—to learn techniques for coping with cravings and to learn better ways of finding a job. He felt as if he was part of the therapeutic process, and that therapy was not something that was "being done" to him. To define positive therapeutic outcomes in concrete terms is an important part of the structure of the cognitive therapy session. Goal setting, along with other elements of the structure of the therapy session, such as agenda setting, helps to avert the common trap whereby each session is reduced to a series of crisis interventions. Therefore, patient and therapist gain a sense of the long-term goals of therapy along with the short-term goals of the session at hand. Understanding the goals of therapy also is important for evaluat­ ing therapeutic progress and outcome. Oftentimes, patients become dis­ couraged in freatment because of a lack of progress or setbacks, such as lapses in drag use. This, in turn, often stimulates black-and-white thinking about therapy: "Therapy is not working at all." By referring back to the original goals of treatment, and by reviewing the patient's progress throughout, the therapist can undermine some of the patient's hopelessness about treatment. For example, one patient and his therapist had documented the following goals at the beginning of treatment: to abstain from cocaine and alcohol, to gain more confidence in social situations, and to obtain and maintain a steady job. Over a period of about 6 months, the patient did remain abstinent from cocaine and alcohol. However, he was unable to obtain any type of employment. This was vety dis­ couraging for the patient and he began reporting negative automatic 124 COGNITIVE THERAPY OF SUBSTANCE ABUSE thoughts about the therapy. The therapist then was able to point o that, although he had not found employment, the patient was feel­ ing much better and had a strong sense of pride about being able to stay away from drags and alcohol. Thus, an important goal of treat­ ment was being met. In addition, the therapist questioned the patient about the specifics of the original goal of employment. In order to achieve this goal, the patient and therapist had agreed on the short- term goal of developing job-seeking skills, such as how to conduct himself at an interview, where to look for a job, and how to prepare a resume. Indeed, the patient had made gains in these areas. After their discussion, the patient was able to see the progress he had made and was able to challenge his negatively biased thinking about the thera­ peutic "failure" of not finding a job. The patient felt a bit more hopeful about therapy and more motivated to continue in treatment. Conversely, a patient who is discouraged because she has not completely quit her smoking and drinking may be cheered somewhat on realizing that some of her more general goals for therapy are being met, thus giving her momentum to tackle further her alcohol and nicotine addictions. This brings up an important point; namely, that goals for therapy do not simply entail cessation of problematic drag use. Criteria for success in therapy must be assessed across a number of important life concerns, including family relationships, social func­ tioning, and work productivity, to name but a few (Covi et al., 1990; McLellan et al., 1992). There are numerous issues that can be brought up and discussed in any one particular session. With clear, concrete goals, the thera­ pist and patient can make maximum use of the therapy time, and can address problems in an organized and systematic fashion. G E N E R A L RULES F O R SETTING G O A L S Therapists should collaborate with their patients in establishing goals for treatment. When patients enter treatment they frequently are ambivalent about abstinence from drugs and alcohol (Carroll, Rounsaville, & Keller, 1991; Havassy et al., 1991; Miller & Rollnick, 1991). It is inadvisable for therapists simply to proclaim that their patients must strive for abstinence as a condition of being in treatment. Instead, it is important for the therapist to explore collaboratively with the patient the benefits of total abstinence from drags such as cocaine. The act of collaboration will help the patient to feel that he or she traly is a part of the process of change. In setting goals, therapists try to highlight the relationship Setting Goals 125 between abstinence and problem-solving. For example, therapists dis cuss with patients how being drug free can contribute to keeping a job and to having better relationships and more money for other things such as clothes, vacations, and a car. Nevertheless, it must be explained that while abstinence increases the chances of obtaining the desired outcome, it does not guarantee it. As a case in point, Ms. F. presented as her primaty goal the wish to stop using crack. The thera­ pist then asked her what the benefits of not using crack would be. She stated that by not using she would be able to save money, that she would be able to pay her bills, and also that she would feel like going to work each day. The therapist then summarized, "Being drag- free is your primaty goal and other goals will be to save money, to be able to pay your bills, and to be able to keep your job. It is impor­ tant to note that abstinence alone will not insure achieving these other goals. It will help you to be in a better position to learn how to get what you want in these other areas of your life. Perhaps we'll work on these skills as well." Goals are best stated in concrete, specific terms. Often, at the beginning of therapy, patients present vague, nonspecific goals for treatment, such as "I just want to get m y life in order," "I just want to be m y old self again," or "I just want this anxiety to go away." Therapists assist patients in defining treatment goals in more circum­ scribed behavioral terms, such as finding a job; staying away from people, places, and things associated with drugs; being able to go out and have a good time without using drags; or reestablishing a bro­ ken relationship. For example, Mr. R., a 42-year-old cocaine addict, stated that he wanted "his world to stop falling apart." In order to concretize the goals, the therapist asked what he would like to be doing differently at the end of treatment. Mr. R. then presented more focused objectives: "First of all, I would like to get a chance to see my children more often. I'm separated from m y wife right now and she doesn't allow m e to see the kids. I would like to stop using coke. 1 want to get involved with the church again. I used to be really into it. I would like to have a regular job. I'm tired of doing odd jobs. I want more excitement in m y life. I'm bored most of the time except when using drags." The therapist facilitated this process by periodi­ cally asking, "And what else would you like to be doing differently?" As a result, Mr. R.'s goals were translated from a vague statement of his "world to stop falling apart" to much more concrete, behavioral, measurable events. The therapist summarized Mr. R.'s goals as follows: 1. See children more often. 2. Stop using coke. 126 COGNITIVE THERAPY OF SUBSTANCE ABUSE 3. Get involved with the church. 4. Have a regular job. 5. Have more excitement, but stay drag free. Once the patient's goals are concretized, the therapist can start to think of the necessaty operations to achieve them, and the criteria on which to assess treatment outcome. An example of goal attainment was Mr. R.'s finding a permanent job with a construction company. Another goal was reached when, after getting a job, his wife allowed him to see the children each weekend. Further, through urine test­ ing, the therapist was able to establish that Mr. R. was not using cocaine. Finally, therapist and patient worked on developing sources of nondrug positive reinforcement, such as hobbies and physical rec­ reation (cf. Stitzer, Grabowski, & Henningfield, 1984). Having these specific goals in the forefront helped to keep the therapist and patient from drifting in each therapy session. Also, hav­ ing these goals written down at the beginning of treatment proved to be a powerful motivator for Mr. R., as the patient was able to com­ pare his situation at the beginning of treatment with his functioning at later stages in treatment, and thereby to recognize his progress in therapy. In setting goals, it is important to remember the following: 1. Be collaborative in setting goals. 2. Establish goals in positive terms as they relate to abstinence. 3. Be concrete and define goals in measurable behavioral terms. STANDARD GOALS OF TREATMENT Two standard goals of treatment are (1) to reduce drag dependency, with the cornerstone being to help the patient develop techniques for coping with urges and cravings, and (2) to help patients learn more adaptive methods for coping with life problems. When substance abusers enter therapy, they often are ambivalent about their desire to stop using. Increasing their motivation to reduce drag dependency becomes an important early focus in therapy. In the first phase of treatment, it is imperative to facilitate the patient's understanding of the various advantages and disadvantages of using or not using drugs and alcohol. The following is a transcript of a therapist discussing with his patient the advantages and disadvantages of using cocaine. Setting Goals 127 TH: One thing that we said that we would go through are the ad­ vantages and disadvantages of using drags. What are some of the advantages to using cocaine? W h y is it good to use? PT: It ain't good. It ain't good but it makes m e feel good. "TH: OK. It makes you feel good. PT: Yeah, thaf s an advantage of it. It makes m e feel good for awhile. TH: What is another advantage? PT: I see evetybody else doing it, so I want to do it too. TH: So, are you saying that it makes you fit in? PT: Yeah, it makes m e fit in the crowd. W h e n I see somebody else doing it, I want

so I can't possibly have a drinking problem." "My life's a mess anyway, so drags couldn't make it worse." "If I stop using, I'll get depressed." Individuals who abuse drugs are typically not attentive to their drag-related beliefs, often viewing their drug use as a function of extrinsic factors. For example, they attribute cigarette smoking to "stress" rather than to their attitudes about smoking. Therefore, it is essential to help patients monitor and identify their beliefs about drags and drag use. In cognitive therapy the drag user is taught about these beliefs as they apply to his or her drag-use patterns. Specifically, the thera­ pist explains and illustrates the "cognitive model of substance abuse" (see Figure 3.5, page 47) to the patient, and together they fill in each box of the flow chart with examples of basic beliefs, automatic thoughts, feelings, and behaviors that are pertinent to the patient's life. For example, consider the case of "Mack," who reported to his therapist, "I went on a binge on Saturday night." The following dia­ logue ensued. TH: Tell me about your binge. PT: What's there to tell? I just felt pretty good and I decided to drink. TH: Let's look at the cognitive model together. [The therapist took out a printed copy of the cognitive model of substance abuse 140 COGNITIVE THERAPY OF SUBSTANCE ABUSE and showed it to the patient.] So the initial stimulus was inter­ nal; you "felt pretty good." PT: Yeh. TH: In response to feeling good, what was going through your head that contributed to your drinking? PT: I guess I was thinking "Gee, a drink would make m e feel even better." TH: And what automatic thought went through your mind? PT: "What the hell! I might as well." TH: And by that time you were beginning to crave? PT: Big time! When patients are systematically taught to monitor their basic drag-related beliefs and automatic thoughts, and when these cogni­ tive processes are shown to be related to their subsequent drag use, patients tend to report an increased understanding of why they use, and a better sense about how better not to use. D o w n w a r d Arrow Technique It is common for drag-using patients to have cata­ strophic (all-or-none or overgeneralized) thoughts not only about their substance use but about themselves, their life, and their future (the cog­ nitive friad; Beck, 1976). Such tiioughts might include tiie following: "My life is going to cramble." "I'll fall apart if I can't get m y fix." "I can't do anything right." "Nobody gives a damn about me." "I am totally out of control." The downward arrow technique is quite useful for addressing such beliefs as these. Many patients are unable to articulate these underly­ ing beliefs until they have been asked to consider the personal mean­ ing thdt their more manifest thoughts have for them. Therefore, when patients exhibit strong negative emotions that seem to be far more intense than their automatic thoughts alone would cause, therapists can ask patients to probe a bit deeper by asking successive variations of the question "What does that mean to you?" Oftentimes, the end result of the question is the elicitation of an underlying or core belief. The following is an illustration: Techniques of Cognitive Therapy 141 TH: Phil, you seem to be having a vety strong reaction against the idea of ttying to steer clear of drinking alcohol at your upcom­ ing office party. What is your concern about being sober at the party? PT: I wouldn't be any fun at parties if I didn't drink. TH: And if you weren't fun at parties, what would the implica­ tions fee? PT: People wouldn't hang around me. TH: And if people didn't hang around you, what would that mean? PT: It would mean they didn't like me. TH: And assuming that all of the above is trae, what would the con­ sequences be? PT: It might mean that m y career would suffer since I am a sales­ man, and I depend on people liking m e for m e to succeed. TH: And if your career suffered, what would the ultimate conse­ quences be? PT: I could lose m y house and m y family and evetything I've worked for! TH: And all this would happen because you weren't drinking alco­ hol at social events? PT: Well, when you put it that way, I guess it's pretty unlikely. TH: I agree, but do you see how a chain of progressively more prob­ lematic beliefs leads you to assume that catastrophe would result if you followed through with the assignment of being "dty" at the party? PT: Yes, I never realized that before. It does seem like it's do or die, but maybe it isn't. TH: We'll have to stay alert to similar chains of beliefs. For now, let's write down what we've just learned on paper [See Figure 9.2.] The downward arrow technique (so-called because of the way it is illusfrated on paper, with each successive belief pointing an arrow downward to the next underlying belief) is effective because it helps a patient to "decatastrophize" (i.e., to reevaluate and modify cata­ strophic thoughts). In the above example, the downward arrow tech­ nique helped Phil to articulate the catastrophic thought that he would lose his career, family, and house if he stopped drinking. O n doing so, Phil was able to see the distortion in his thinking. As a result, he 142 COGNITIVE THERAPY OF SUBSTANCE ABUSE Automatic Thought = "I can't not drinic at this party!" "I wouldn't be any fun." i "People wouldn't hang around me." •X' "They wouldn't like me." I "My career would suffer." Underlying Beliefs = "I could lose everything." FIGURE 9.2. Downward arrow illustration. was in a better position to modify this thought. In fact, the irony of Phil's catastrophic thinking traly came to light w h e n he did an A-D analysis and determined that he could "lose evetything" if he did not stop drinking. Reattribution of Responsibility Those who use drags will often attribute their use to extrinsic factors. For example, the alcoholic m a y attribute drinking to a bad marriage, a stressful job, and drinking buddies exerting pres­ sure to be "one of the guys." The therapist can help individuals to reattribute responsibility for their drag use to themselves so that they m a y take initiative to modify drag-using behaviors. Reattribution of responsibility requires the skillful application of the Socratic method, so that patients do not feel that their therapists are being judgmental or accusatoty. The following is an example of this process: TH: You say that you've been drinking again. What are your thoughts about w h y this is happening? PT: Well, m y wife has really been hassling m e lately. I might be able to stay sober if she would get off m y back once in a while. TH So you see your wife as causing your drinking? PT: In a way, yes ... but it's not that simple. TH OK. Tell m e more about some of the complexities involved here. Exactly what role does your wife play, and what role your beliefs and actions play. PT: She criticizes m e , and I start thinking that I can't get any peace and that I'm trapped. Techniques of Cognitive Therapy 143 TH: And then? PT: And then I think that m y only escape in life is to get bombed and forget about m y miserable marriage. TH: So when you focus on the misety in your marriage, you deter­ mine that there is only one remedy—to drink until you go "blotto." A m I on target? PT: Yes. TH: Would it be fair to say that you decide to drink, rather than choosing some other way of resolving your problems with her? PT: Yes. I guess it's m y decision, but it's easier to blame her. {chuck­ les) TH: It may be easier to blame her, but does that help you to reach your goal of dealing with your problem with alcohol? FT: No. TH: What would help? PT: If I take charge of m y own life and m y own decisions, regard­ less of how pissed off I am at m y wife. TH: Not easy, I admit, but a worthy goal. In this example, the patient's focus shifted from external to inter­ nal factors. Initially, he blamed his wife for his drinking. At the end of this dialogue he had begun to see that he had some responsibility for his drinking. As a result, he may begin to take some responsible actions toward changing his drinking. Daily Thought Record The Daily Thought Record (DTR) is a ftindamental strategy in cognitive therapy, which has been useful in the freatment of depression, anxiety, and other problematic mood states (Beck et al., 1979; Burns, 1980). The standard DTR is a five-column form that is completed by the patient (see Figure 9.3). Those who abuse substances tend to do so as a result of their beliefs (often maladaptive) about drags. For example, the alcohol abuser, prior to going to the corner tavern, may have the belief "I need a drink." By using the DTR, the patient is able to examine this belief and consider its validity in a more systematic, objective fash­ ion. In addition, the DTR provides a time lag after the initial urge during which the patient may choose not to drink (or use other drags), and may experience a natural diminution of the craving. Addition- 144 COGNITIVE THERAPY OF SUBSTANCE ABUSE ally, the DTR provides a method for coping with negative mood stat so that they are not as likely to trigger drag use. Consider the case of Mack, who reported that he was most tempted to drink when alone on weekends. For homework his thera­ pist asked him to complete a DTR based on the experience of being alone on a Saturday night (see Figure 9.3). Mack initially described himself as being lonely and angty as a result of not having a date. He rated these emotions as quite high: 80 and 75, respectively. These feelings were based on his beliefs: "No one cares about me" and "It's unbearable to be alone." Using the DTR as a guide. Mack's therapist helped him to consider alternative rational responses, such as "My friends and family care about me" and "I CAN bear to be alone." As a result of doing so. Mack reported feeling sufficient relief that he held off on having a drink. This result boosted his sense of self-efficacy markedly. A useful method for teaching patients to generate objective rational responses involves the application of a series of open-ended questions (see the bottom of Figure 9.3). These questions include the following: 1. What concrete, factual evidence supports or refutes my auto­ matic thoughts and beliefs? 2. Are there other ways I could view this situation? Is there a blessing in disguise here? 3. What is the worst thing that could happen? What is the best thing? What is most likely to realistically happen? 4. What constractive action can I take to deal with the situation? 5. What are the pros and cons of m y changing the way I view this situation? 6. What helpful advice would I give m y best friend if he/she were in this situation? Any or all of these questions can help stimulate patients to think of rational responses. W e have found that the regular application of these questions makes for an excellent ongoing homework assignment and helps patients learn to use DTRs to their maximum benefit. Imagery Imagety techniques can be used with dmg users to help them visualize "self-control" and avoid drag use. Imagety can be a useful technique for focusing patients on drag-related beliefs and automatic thoughts, or distracting them from their cravings and urges. Directions: W h e n you notice your mood getting worse, ask yourself, "What's going through m y mind right now?" and as soon as possible jot down the thought or mental image in the Automatic Thought column. SITUATION AUTOMATIC THOUGHT(S) EMOTION(S) RATIONAL RESPONSE O U T C O M E Describe: 1. Write automatic thought(s) 1. Specify

sad, 1. Write rational response 1. Re-rate belief 1. Actual event leading to preceded emotion(s) anxious/angry automatic thought(s) in automatic unpleasant emotion, or 2. Rate tielief in automatic etc. 2. Rate belief in rational thought(s) 2. Stream of thoughts, 0-100%. 2. Rate degree 0-100%. 0-100%. daydreams or recollection, of emotion 2. Specify and leading to unpleasant 0-100% rate subsequent DATE/ emotion, or emotions TIME 3. Distressing physical 0-100% sensations 4x Questions to help formulate the rational response: (1) What is the evidence that the automatic thought is true? Not true? (2) Is there an alternative explanation? (3) What's the worst that could happen?Could I live through it? What's the best that could happen? What's the most realistic outcome? (4) What should I do about it? (5) What's the effect of my believing the automatic thought? What couW be the effect of changing my thinking? (6) If was in this situation and had this thought, what would I tell him/her? (friend's name) F I G U R E 9.3. Daily T h o u g h t Record (blank). 146 COGNITIVE THERAPY OF SUBSTANCE ABUSE It can also serve as a method for changing drag-related beliefs an thoughts. Examples of imagety used in this fashion include imagin­ ing assertive, direct methods for "saying no" to others who offer drags; imagining positive, enjoyable activities as alternatives to drag use; and imagining a healthy, productive life as a result of freedom from drags. The following example demonstrates the use of imagety in smoking cessation. PT: I don't know how I'll survive without my first cigarette of th TH: What are your thoughts just before you smoke that cigarette? PT: I usually don't think. I just go to m y pack, take one out, and light up. TH: Perhaps your thoughts are so automatic that you don't pay at­ tention to them. I'd like to help you recall thoughts that lead to smoking that first cigarette of the day. Close your eyes for a moment and imagine what happens when you first wake up in the morning. What do you see? Smell? Hear? H o w do you feel? Now, what thoughts are going through your mind? What are you telling yourself? PT: Well, first I'm just kind of groggy. When m y mind starts to clear, I lie in bed thinking about what I have to do that day. I feel myself getting a little nervous and I think "Oh just relax. Go to the kitchen for a cup of coffee and a cigarette." TH: What happens next? PT: Well, to be honest, I usually go the bathroom first, but after that, I go to the kitchen and have a cigarette and coffee. TH: As you head to the kitchen can you recall what you are think­ ing? PT: Sure! I am thinking about the lift I'll soon get from the ciga­ rettes and coffee. TH: So you think "I'm going to get a lift from that first cigarette. Quick, go smoke!" What do you feel with that image and what happens next? PT: I feel an urge to smoke and 1 indulge myself. TH: OK, let's tty something. The images of smoking and drinking coffee are quite positive. Can you replace these with an alterna­ tive image of what could happen in the morning? PT: 1 guess so. Last time I quit smoking I would wake up and go jogging. TH: What were your thoughts prior to going jogging? Techniques of Cognitive Therapy 147 PT: They were sort of tike the smoking thoughts, but instead I thought that the run would give m e a lift. TH: OK, let's tty to imagine jogging in the morning. What do you see? Smell? Hear? Feel? Make it as attractive and tempting as possible. PT: I see the soft light of the morning sun as it shines on the trees, the house, and the landscape. The morning air smells fresh and clean. I hear lots of birds chirping because it's early morning. 1 feel healthy and alive. TH: What happens when you produce that new image? PT: 1 lose some of m y interest in the cigarette. TH: Great! So now you have an alternative image to the smoking image when you wake up in the morning. Can you tty to vol­ untarily "call up" this image to your awareness in the morn­ ing? PT: I'll fry. TH: Let's also make sure your family physician agrees that it is safe for you to take up a regimen of jogging again before you start, OK? BEHAVIORAL TECHNIQUES Activity Monitoring and Scheduling Patients who abuse drags tend to engage in activities and behaviors that support their drag abuse and may concurrently fail to take part in activities that promote prosocial life goals, such as work, hobbies, community service, and stable relationships. Activity monitoring and scheduling can be useful basic strategies for under­ standing and modifying drug-related behaviors and for increasing productive behaviors. The process of activity monitoring and scheduling is simple and straightforward. The patient receives a blank grid (the Daily Activity Schedule) which contains the 7 days of the week divided into 1-hour blocks (see Figure 9.4). For a period of 1 week the patient records daily activities and the degree to which he or she felt a sense of pleasure or mastety from participating in each activity. Pleasure and mastety, recorded on a scale from zero (none) to ten (extreme), provide an indication of the patient's mood and the level of reward or satisfac­ tion derived from each activity. The Daily Activity Schedule can be used for at least three pur- NOTE: Grade Activities M for Mastery and P for Pleasure 0-10 M T W Th F S Su 6-7 7-8 8-9 9-10 10-11 11-12 12-1 1-2 2-3 3-4 00 4-5 5-6 6-7 7-8 8-9 9-10 10-11 11-12 12-6 REIVIARKS: FIGURE 9.4. Daily Activity Schedule (blank). Techniques of Cognitive Therapy 149 poses. First, it serves as a journal of present activities. By revi completed schedule, therapist and patient gain a baseline understand­ ing of the patient's activities and how they relate to drag use. Sec­ ond, the Daily Activity Schedule can serve as a prospective guide for future activities. That is, patient and therapist can use a blank form to schedule alternative activities that are less conducive to drug use. Furthermore, to the degree that the patient lacks satisfaction and a sense of accomplishment in life, the therapist may choose to exam­ ine the patient's core beliefs about his or her lovability and adequacy, respectively. And finally, the Daily Activity Schedule can be used to evaluate the extent to which the patient has been following his or her proposed schedule successfully. That is, after a weekly plan has been completed, patients may take home a blank form to monitor actual behaviors. Frequently, a failure to follow through with planned activities comes about as a result of drag-related behaviors, along with their concomitant beliefs, such as "I can't do anything right" or "I never reach m y goals." W h e n this happens, therapists must remain upbeat, helping patients to see that useful information has been obtained, and that goals can still be achieved in spite of early setbacks. When patients succeed in planning and completing nondrug activities that give them satisfaction and build their self-efficacy, they begin to view themselves as less helpless, less out of control, and less dependent on chemical "fixes." Behavioral Experiments Behavioral experiments are used to test the validity of patients' drag-related beliefs and core beliefs. For example, con­ sider the patient who believes "I would lose all m y friends if I didn't smoke pot." A behavioral experiment might involve having this patient participate in "usual activities" with friends, without using marijuana. (The patient would be encouraged to fully participate in all non-drag- related activities.) Thus, the "independent variable" in this behavioral experiment would be the patient's use of marijuana. The "dependent variable" would be maintenance of friendships. The patient would be encouraged to avoid any "extraneous variance" by maintaining con­ sistency in all other aspects of his or her behavior. Regardless of the results of this experiment, the patient is likely to learn some impor­ tant lessons. Specifically, if he loses friends, he will be encouraged by the therapist to examine the meaning of his pre-abstinence friend­ ships. If he maintains his friendships, he will, it is hoped, modify his original distorted belief: "I will lose all m y friends if I don't smoke marijuana." 150 COGNTTIVE THERAPY OF SUBSTANCE ABUSE Another form of behavioral experimentation is the "as if" tech­ nique. Using this technique, the therapist encourages the patient to act as if a desired behavior or set of circumstances were trae for him. For example, the patient who wishes to quit smoking might spend a week acting "as if" he were a nonsmoker. For example, he might ask others not to smoke around him, he might exercise, or he might sit in the nonsmoking sections of restaurants. Such activities are designed to modify the patient's drag-related beliefs as well as behaviors. Behavioral Rehearsal (Role Play and Reverse Role Play) Many patients with substance abuse problems have concurrent problems with interpersonal communication (e.g., assertiveness, self-disclosure, and active listening) (Piatt & Hermalin, 1989). As a result, they often feel frastrated and overwhelmed in inter­ personal situations, resulting in a vulnerability to drag use. Hence, the therapist may initiate role-playing to teach the patient effective interpersonal skills. The following is an example: PT: I feel like getting wasted evety time my wife nags me. TH: What do you mean when you say she "nags" you? PT: I mean when she asks me to do more than m y share of the work. TH: What do you do when you think she's nagging? PT: Well, typically I go to another room or I just leave. That's when I'm most likely to drink. TH: So you withdraw. PT Yeh. TH How else could you handle the situation? PT I have no idea. TH OK, let's tty some different options. I will play the role of your wife, and I would like you to discuss your feelings with me as honestiy as possible, without attacking or being aggressive. PT: I don't understand. What should I do? TH: Simply pretend that I am your wife and practice talking to me in a constractive fashion about m y nagging. PT: OK, who starts? TH: 1 will, {acting as patient's wife, sounding annoyed at patient) I want you to take care of the yard this weekend. It's been three weeks since you promised to plant grass and soon it will be too late to even tty. I'm sick and tired of waiting for you to do things Techniques of Cognitive Therapy 151 around here! {silence; patient looks puzzled) [The therapist momentarily stopped role play to encourage patient.] N o w you respond to m e as your wife. PT: Man, that sounded too much like her! TH: OK, now talk to m e as you would to her. PT: {Another pause, then role play continues) Maty, I'm pretty sick of you nagging me! TH: It's the only way anything gets done around here! PT: I do a lot around here! {in raised voice) You're so busy complain­ ing that you never notice! TH: [Again, role play stopped. Therapist talked to patient.] Now, let's look at what has happened. What are your thoughts about this role play? It became apparent that this patient lacked effective conflict reso lution skills. Thus, the therapist must begin by teaching the behav­ ioral skills of active listening, assertiveness (vs. aggressiveness), and compromise. The teaching process involves some didactic training, along with frequent role-playing, to learn and practice more effective interpersonal behaviors. One way to gain a reluctant patient's active participation in role playing is to volunteer to take the patient's role while the patient portrays the "problematic other" (e.g., an employer, a spouse, or an old drag associate who is ttying to convince the patient to get high). In this manner, the patient can show the therapist just how difficult the situation is to manage, while the therapist can model some responses that the patient might not have thought of before. A more demanding variation of this procedure

has been described by Moorey (1989), who suggests that drag abusers can learn to empa­ thize with important people in their lives by role-playing the part of a significant other who has been hurt by the patients' drag use. This exercise also serves to highlight the destruction that the drag abuse has wrought on the patients' personal lives. On the plus side, repeated role playing helps patients develop new, mature, effective repertoires of social behavior in a safe envi­ ronment where errors can be corrected without actual consequences. Relaxation Training There is often a component of anxiety to drag use (see Chapter 15, this volume). For example, cigarette smokers and heavy alcohol users often report that they smoke or drink to relax. Hence, drag use may be a form of self-medication for people who have dif- 152 COGNTTIVE THERAPY OF SUBSTANCE ABUSE ficulty relaxing (Castaneda et al., 1989; Khantzian, 1985). Even co users, who consume cocaine for its stimulating effects, may feel anx­ ious or tense in anticipation of using, especially if there is some delay between the time they crave and when they actually use. Thus, relaxation training may be a useful technique in that it provides the patient with a safe (drag-free) method of relaxing. Sec­ ond, it provides the patient with a time lag after the initial craving experience, during which the craving may subside (Carroll, Roun­ saville, & Keller, 1991; Horvath, 1988). Ultimately, relaxation train­ ing may be useful in building the patient's new belief that he or she is in control of and responsible for his or her coping responses (see Chapter 15, this volume, for additional information). Graded Task Assignments Quite often patients must make dramatic behavioral changes in order to facilitate a drag-free (or drag-reduced) life. For example, the patient whose only friends are other crack users will have to restracture his or her social life almost entirely in order to mini­ mize the chances of relapse. C o m m o n sense and experience tell us that this is no easy task. Hence, the patient is encouraged to engage in approximations of the desired behaviors in order to build toward the end goals. For example, the patient who needs to modify her friendships might be encouraged to begin by spending drag-free time with a non-drag-using acquaintance (e.g., a new friend from a sup­ port group meeting), such as going to lunch or to a movie. On suc­ cessful completion of this exercise, the patient would choose another (more challenging) assignment until he or she had built up a new drug-free network of cohorts. Problem Solving Drag users who frequently demonstrate impulsivity often are vety poor problem solvers. In fact, in advanced stages of drag abuse, many patients either ignore their problems (denial, avoid­ ance) and/or respond to their problems by anesthetizing themselves with drags. For those patients with a long histoty of drag use, it is strikingly apparent that they have precious littie accumulated experi­ ence in recognizing and solving life's problems constractively. For example, one of our drag-abusing patients was troubled by her husband's ongoing drag abuse. One day she found a stash of his crack cocaine. Rather than confront him with this finding, or flush the drags down the toilet, she smoked evety cap until the stash was Techniques of Cognitive Therapy 153 used up. When she reported this to her therapist, he shook his head in disbelief and asked her what her rationale was for such a self- defeating act. She replied, "I didn't want him to use, so I figured if I smoked it all, he wouldn't be able to use." The therapist noted that this was an example of a "permission-giving" belief, and that it reflected a failure to think the problem through carefully (to say the least). Another patient reported a scheduling conflict between his therapy sessions and his job. His immediate response was to quit the job, rather than wait a week to work out a new schedule with the therapist. The upshot of these illustrations is that drag-abusing patients must be taught the principles of problem-solving (Nezu, Nezu, & Perri, 1989). The steps of problem solving include the following: 1. Defining the problem in clear specific terms 2. Brainstorming a number of possible solutions 3. Examining the pros and cons of each brainstormed solution (for the present, for the future, and for significant others as well) 4. Choosing the best hypothesized solution 5. Implementing the behavior after some planning, preparation, and practice 6. Evaluating the outcome and assessing for more problems to solve This is a long, gradual process that is fraught with frustration along the way. Therapists must remain supportive, patient, and encouraging if patients are to persevere in learning these skills. Exercise Most substance abuse is incompatible with physical health and sustained exercise. For example, cigarette smoking and regular aerobic exercise (e.g., jogging) would seem to be incongra- ous. Hence, regular aerobic activity is likely to heighten a person's awareness of the disadvantages of substance use and the advantages of quitting. The therapist may encourage the patient to engage in physical exercise as part of the treatment program (of course, only with a physician's medical approval). Such activity may help the patient to redefine him- or herself as a healthy, physically fit person. This image should cause cognitive dissonance for the patient and may motivate him or her to modify the pattern of substance abuse. [A notable exception to the above involves athletes who abuse 154 COGNFTTVE THERAPY OF SUBSTANCE ABUSE anabolic steroids, thinking that it is very much compatible with a letic achievement. In such cases, therapists would not want to focus on exercise as an intervention. Instead, anger control, interpersonal conflict resolution, and focusing on the unappealing side effects of steroids (e.g., balding, acne, reduced sexual responsivity, and myriad serious medical consequences) should be the central focus.] Stimulus Control At first glance it would seem that an effective strat­ egy for reducing substance abuse would be to eliminate all stimuli that trigger drag use. However, it quickly becomes apparent that doing so is not practically possible: all people will have episodes of feeling sad, lonely, anxious, bored, frastrated, and other internal sources of high risk. Ex-smokers inevitably find themselves in places where others are smoking, alcohol-dependent individuals eventually face enticing advertisements, and many users of illicit drags have to come in contact with relatives who abuse such drags. In order to minimize contact with drug triggers, patients are encouraged to identify those stimuli (internal and external) that put them at high risk for the activation of drag-related beliefs that trig­ ger drag use (Carroll, Rounsaville, & Keller, 1991). For example, some people are vulnerable to negative moods (boredom, anxiety, sadness, etc.) while others are vulnerable to positive moods (joy, happiness, excitement, etc.). Still others are vulnerable to extrinsic cues (meals, other users, time of day, geographic location, etc.). Patients are encouraged to plan ways to avoid these cues whenever possible. How­ ever, more importantly, patients are encouraged to prepare methods for dealing with these cues when the cues are encountered (see Chap­ ter 10, this volume). C O M M O N OBSTACLES IN TREATMENT When a patient does not appear to be responding favorably to cognitive therapy, it is possible that there is a problem in the therapeutic relationship. The patient who does not trast the therapist or feel comfortable in therapy is likely to be guarded in sessions; therefore, he or she might avoid important self-disclosures for fear of judgment or retribution. Without these admissions it is unlikely that the patient will objectively examine or acknowledge biased thinking patterns. Such examination is central to the success Techniques of Cognitive Therapy 155 of cognitive therapy, and can be fostered via an accepting, warm, collaborative therapeutic approach. When the therapeutic relationship is judged to be sound, but the techniques appear to be ineffective, it is important to review the case conceptualization for missed diagnoses, overlooked beliefs, and impor­ tant unassessed historical events. For example, many substance-abus­ ing patients have a coexisting personality disorder (borderline, anti­ social, avoidant personality disorders, etc.). These patients require careful attention simultaneously to their chronic maladaptive person­ ality (i.e., belief) patterns and to their substance abuse problems. Patients may also have problems with anxiety and depression, and they may be attempting to treat these problems with psychoactive sub­ stances (i.e., to self-medicate). Therapists must keep in mind that such a patient's "resistance" to treatment simply may reflect an unspoken fear that without drags his or her anxieties or dysphoria will become overwhelming. Also, taking an updated review of the patient's his­ toty can stimulate a breakthrough in treatment. For example, one therapist did not realize until many months into treatment that his patient had gone through humiliating failure experiences in grade school and that memories of this had led the patient to resist any­ thing called "homework," for fear that it would make him look stu­ pid. Once this was understood, the therapist was able to help the patient to generate rational responses to combat these negative expec­ tations and fears. The topic of homework is important enough to merit its own separate discussion. Patient nonparticipation in homework assign­ ments can hinder the therapeutic learning process. It is tempting in any medical or psychological intervention to simply tell the patient to "go home and do. ..." However, for homework to be maximally effective it should have certain qualities. First, it should be collabo­ ratively determined. Second, the therapist should check for the patient's understanding of the exact nature of the assignment, as well as the patient's understanding of the underlying rationale for the assignment. Third, the therapist and patient should consider any potential barriers to completing homework assignments and prepare contingency plans. Finally, the patient should have an opportunity to practice the homework with the therapist in order to test his or her understanding of and ability to do the assignment. A final problem in applying the techniques of cognitive therapy to substance abuse might be in the therapist's therapeutic style. It is essential that the therapist stimulate the patient's thinking process with sensitively worded open-ended questions, rather than by lectur- 156 COGNTTIVE THERAPY OF SUBSTANCE ABUSE ing or preaching to the patient. When the cognitive therapy techni described here are presented in a dictatorial fashion, the therapist's words are likely to "go in one ear and out the other." Generous appli­ cation of the Socratic method often is an antidote to this problem. SUMMARY In this chapter, we presented many of the most widely used techniques of the cognitive therapy of substance abuse. Most of the techniques that have been successfully applied in the treatment of other psychiatric syndromes are useful in the freatment of substance abuse. It is important that these techniques be applied with careful attention paid to the therapeutic relationship, the patient's individual case conceptualization, the patient's application of these techniques in the form of homework assignments, and the therapist's use of open- ended questioning. The techniques that we described do not represent an exhaus­ tive list. In fact, we encourage therapists to make use of their creativ- ify and their patients' unique individual needs and strengths to devise new variations of cognitive therapy techniques and assignments. As long as the technique serves a logical purpose, fits within the case conceptualization, adheres to ethical guidelines, and focuses on changes in beliefs as well as drag-related behaviors per se, there is no limit to what may be applied successfully. C H A P T E R 1 0 D e a l i n g w / i t h C r a v i n g / U r g e s B 'ecause of their resurgence during and after treat­ ment, uncontrolled cravings and urges to use are major factors con- ttibuting to treatment dropout and often lead to relapse even after long periods of abstinence. Teaching patients to cope with craving is therefore one of the most important goals of treatment (Annis, 1986; Carroll, Rounsaville, & Keller, 1991; Childress et al., 1990; Covi et al., 1990; Horvath, 1988, in press; Shulman, 1989; Tiffany, 1990; Washton, 1988). The therapist initially assesses the patient's idiosyncratic per­ ception of craving. Then, over the course of treatment, the therapist helps the

patient to understand the various factors that contribute to craving, to reframe the experience, and to develop better ways to deal with this problem. Horvath (1988) has distinguished the phenomenon of cravings from urges, describing the former as the subjective sense (e.g., physi­ cal arousal, emotional arousal, "need," and "desire") of wishing to attain the psychological state induced by drags. In contrast, urges are described as the behavioral impulse to seek and use the drugs. Although these are useful theoretical distinctions, in practice we have used the terms "cravings" and "urges" interchangeably, as we will in the remainder of this chapter. TYPES OF CRAVING We have identified four major types of craving, each with its own unique characteristics (although there is some overlap from one type to another). 157 158 COGNTTIVE THERAPY OF SUBSTANCE ABUSE 1. Response to withdrawal symptoms. Heavy users of drags such as cocaine and heroin often come to experience a diminishing sense of gratification from the use of the drag, but an increasing sense of inter­ nal discomfort on cessation of the use of the drag. In such cases, the craving takes on the form of a "need to feel well again." This is espe­ cially trae for the heroin user who experiences severe, flu-like symp­ toms during withdrawal, and the cocaine user who becomes deeply depressed during a "crash." Therapists who treat patients who are going through this type of craving will need to be empathic to the patients' acute sense of pain and suffering as a result of abstinence. It is most important to inform such patients that this discomfort is tem- poraty (although in extreme cases medical supervision may be nec­ essaty). 2. Response to lack of pleasure. Another type of craving involves patients' attempts to improve their moods in the quickest and most extreme way possible. This phenomenon is most likely to occur when patients are bored, are unskilled in finding prosocial means of enjoy­ ment, and wish to "self-medicate" (Castaneda et al., 1989; Khantzian, 1988) in order to blot out unpleasant thoughts or feelings. Here, thera­ pists must be aware that the therapy sessions themselves may be suf­ ficiently upsetting to patients that they may experience a craving for drags in order to forget their froubles. Therefore, therapists are advised to assess their patients' moods and degrees of craving at the end of particularly stressful (i.e., productive and meaningful) sessions. 3. "Conditioned" response to drug cues. This type of craving requires no particular dysphoric mood, stressor, or hedonic urge on the part of the patient. Patients who have abused drags have learned to asso­ ciate many otherwise neutral stimuli (a particular sfreet comer, a given person, a telephone number, a certain time of day, etc.) with the acute gratification obtained from the use of drags. These neufral stimuli therefore become "charged" with meaning and can induce automatic cravings even in the absence of stressors. Therapists must help their patients to become aware of, and cope with, the cravings that arise simply as a result of their association with these evetyday stimuli. 4. Response to hedonic desires. Patients sometimes experience the onset of drag cravings when they wish to enhance a positive experi­ ence. For example, some patients have made a habit of combining drags and sex as a way to magnify the sexual experience. Others, for example, seek drags as a way to make their social interactions more "enjoyable and spontaneous." Unfortunately, the high that is achieved by such practices is difficult to match (in the short term) in a drug- free life. Therefore, these are particularly difficult types of cravings to combat in therapy. Therapists have the most leverage when patients' Craving/Urges 159 drag use has progressed to such a degree that their life problems overwhelmed their isolated moments of drag-induced hedonic joy. Under these conditions, patients typically are more willing to work to find other means of achieving gratification in life. Still, there will be an ongoing battle with these cravings whenever the patient expe­ riences a natural good mood. CRAVINGS: UNDERSTANDING THE PATIENT'S EXPERIENCE A therapist can begin to understand the patient's crav­ ing experience first by identifying automatic thoughts (ATs) associ­ ated with the experience. Cravings can be triggered in the therapist's office by having patients simply describe the last time they used drags. Induced imagety is a more powerful method for evoking these crav­ ings. As a word of caution, it is important that the patient understand the rationale for this induction, namely, to learn to identify and cope with cravings and cues to craving. Also, therapists must be prepared to help reduce the strong feelings of craving prior to the end of a session (Childress et al., 1990). If not, patients may leave the therapy session highly aroused and without the skills necessaty to cope with their cravings. Thus, a lapse may ensue. Craving induction techniques should not be implemented until after patients have had practice with general cognitive therapy skills and coping techniques (see Chapter 9, this volume). In a typical craving induction, patients are asked to imagine the last time they used cocaine, and then to describe the image. They are instracted to give as much detail as possible. The aim is to have patients relive the experience as vividly as possible, and therefore gain access to the "hot" cognitions that accompany the craving. A typical way to introduce a craving experience is the following: "Jim, today I want to tty to understand your experience of craving so that I can help you to develop better strategies for coping. Therefore, I'm going to ask you to do a short exercise with me." "I want you to think back to the last time that you used cocaine [or another drag relevant to the patient's problem]. [Wait a minute or two for the patient to get into the image.] N o w I want you to think about the events that led up to your using. I would like you to tty to picture it in your mind, describe the setting, tell m e the sequence of events, and relate what you are feeling. As soon as you notice that you are beginning to have a craving, please indicate this by lifting your hand." {Patient indicates he's having an image) "Describe what 160 COGNITIVE THERAPY OF SUBSTANCE ABUSE you are feeling and what thoughts are going through your mind just now." Therapists should ask patients to compare how similar the crav­ ings experienced in the office are to those they experienced outside. In addition, patients can be asked to come up with their own meth­ ods to help induce the craving. For example, some cocaine abusers report that one of the strongest cues for craving is remembering a particular sex partner that they had while using cocaine. The thought of that person and the image of having sex serve as powerful cues for eliciting strong desires to use cocaine. COPING WITH CRAVINGS Patients can be helped to reduce the aroused cravings by a number of techniques, including (1) distraction, (2) flashcards, (3) imagety, (4) rational responding to urge-related automatic thoughts, (5) activity scheduling, and (6) relaxation training. These techniques should be demonsfrated and taught early in treatment. To develop a durable sfrategy for handling cravings, of course, patients must also learn to deal with the dysfunctional beliefs that facilitate using. Distraction Techniques The key goal of distraction techniques is to get patients to change their focus of attention from internal (e.g., auto­ matic thoughts, memories, physical sensations) to external. Although some of these techniques seem quite simple, they do help to dimin­ ish strong cravings. The following are brief descriptions of commonly used distrac­ tion techniques: 1. Instruct patients to concentrate their attention on describing their surroundings, such as cars, people, trees, and storefronts. Ini­ tially, patients can practice in the office. The more that they can focus and give details about these external events, the more likely they are to focus less on the internal cravings. 2. Use talking to disfract. This can involve starting a conversation with a friend, a relative, a support group sponsor, or the therapist. 3. Patients can remove themselves from the cue-laden environ­ ment. They can take a brisk walk, visit a friend, or go for a drive. One of our patients found the public libraty an excellent place to escape in order to reduce cravings. Craving/Urges 161 4. Perform household chores as a positive distraction. If patients are at home and they notice these cravings and urges, something as simple as beginning to clean the house can distract them from the craving. In addition, this goal-directed activity also helps increase their self-esteem because they have accomplished something useful. (As a caveat, this activity may be ill-advised if drags and paraphernalia are scattered in hiding places throughout the house.) 5. Encourage patients to recite a favorite poem or prayer. For some patients, it is more helpful actually to write down the poem or prayer on a piece of paper. 6. Suggest that patients spend time involved in games, such as cards, video games, board games, and puzzles. These activities can be quite challenging and therefore require focused concentration. Fur­ ther, patients can do some of these activities even if they are alone. Flashcards W h e n cravings are strong, patients seem to lose the ability to reason objectively. Generating coping statements can be helpful in getting patients through this critical period. The usefulness of coping statements can be enhanced by asking patients to write these statements on flashcards (e.g., 3" x 5" index cards). Some examples include a flashcard with the list of advantages for not using drags and a list of things that could be bought with the money intended for cocaine. The following are examples of statements that one patient wrote on his flashcards: 1. You feel more sane when you don't use. 2. Things are going great with m y wife; keep it that way! 3. You look good physically; keep it that way! 4. Get the hell out of this situation nowl Imagery Techniques These techniques include (1) image refocusing, (2) negative image replacement, (3) positive image replacement, (4) image rehearsal, and (5) image mastety. Refocusing is essentially a disfraction technique. Patients direct their attention away from internal cravings by imagining external events. Refocusing can begin first by saying "Stop!" In order to accen- hiate this thought-stopping technique, patients may interject a visual image of a stop sign, a police officer, or a brick wall, to name a few. 162 COGNITIVE THERAPY OF SUBSTANCE ABUSE They then begin to describe to themselves what they see going on around them. For example, one patient was at a picnic where many people were drinking and having a good time. He began to have spontaneous memories of the last time he used cocaine, which led to a desire to use in the present. He said "Stop" to himself, pictured a stop sign, and then began focusing his attention on the people around him who were not drinking. He identified them by name, what they were wear­ ing, and what they were doing. In doing so, he was able to focus his attention away from the cocaine memories and he experienced sub­ sequent reduction in the craving. Another imagety technique is negative image replacement. Often­ times during the first few weeks of abstinence, patients report pictur­ ing themselves using, sometimes even having dreams about using. In these images they see the use of drags as a method for coping with their current distress, and the image takes on a positive glow. In response, it is helpful for patients to substitute a negative image regard­ ing the many unfortunate consequences of taking the drug, such as feeling helpless and hopeless (especially after a period of abstinence) or losing money, jobs, and relationships. For example, one patient, while at a nightclub, became quite angty that he could no longer drink alcohol. He saw other people around him drinking and this brought back nostalgic memories of some of his drag and alcohol days. In response, he replaced the image with one pertaining to the unpleas­ ant physical experiences that he would have when he crashed the day after using alcohol and cocaine. This image was strong enough to dissuade

him from taking the first drink. Positive image replacement is a related technique to help cope with cravings and urges. For example, one patient experienced vety strong negative images about his current situation, that almost his entire family was Strang out on drags. His father was about to lose the house where the patient was living. Consequently, the patient had images of losing his children, having to put them in a foster home, and having to live in a shelter. At that moment, in his sense of hope­ lessness, the patient began to have thoughts of giving up his absti­ nence from cocaine. However, he instead referred to a flashcard that described a positive scene. He imagined himself being back at work again, being in his own home, and able to take care of his chil­ dren. Also, a concomitant of this positive image was the self-satisfac­ tion that he would have after a long day of working. This technique diminished his hopelessness somewhat and, along with it, his crav­ ing. Imagety rehearsals should be used to prepare patients when it is Craving/Urges 163 known that they are going to be in cue-laden situations. One patien who had been abstinent from cocaine and alcohol for about one year, planned to go to a formal banquet where alcohol would be served. In the image rehearsal the therapist asked the patient to imagine going to the banquet and saying, "No thanks, I'll have a club soda," when he was offered a drink. The therapist told the patient to repeat the image several times and monitor his thoughts and feelings associated with the imagety rehearsal. The patient initially was quite anxious while doing this, but later developed a sense of mastety or confidence in being able to go to the banquet and still deal with alcohol being served all around him. Some patients fear that while experiencing vety strong cravings and urges to use drags they will not be able to tolerate the negative feelings that they are experiencing without giving in to the urge to use (Horvath, 1988; Washton, 1988). W e have found that it is help­ ful to teach these patients mastety imaging, seeing themselves as a vety strong and powerful person who is overcoming cravings and urges. One patient who was an intravenous cocaine user reported having repeated images of what he described as "the cocaine lady." In this image, he would be have the sensation of strong cravings for cocaine as he would picture a beautiful woman who was going to offer him cocaine. The patient was taught by his therapist to change the image so that he would have more control. A metaphor that was used was that of a director of a play; that is, the cocaine lady was one character in the play and the patient was the director, able to decide what she would look like and how she would act. He subsequently redirected the image so that "the cocaine lady" was a grotesque-looking person who was vety small, while the patient was a heavyweight boxer able to fight off the urges that he was experiencing. Rational Responding to Urge-Related Automatic Thoughts Therapists start by training patients to self-monitor automatic thoughts when they are having unpleasant emotions such as anger, anxiety, sadness, or boredom. Later, the patients are in­ stracted how to assess their automatic thoughts while experiencing cravings and urges. It is helpful to have patients carty a "therapy notepad" and a pen in order to write down these thoughts. Patients are told that anytime they experience strong cravings or unpleasant emotions, they should ask themselves "What thoughts are going through my mind right now?" They are also instructed to note any physiological distress and then 164 COGNTTIVE THERAPY OF SUBSTANCE ABUSE to ask themselves, "What am I feeling?" and "What thoughts are goi through m y mind?" They then write down the answers and bring their notepads to the next therapy sessions. The Daily Thought Record (DTR) is used to help patients exam­ ine negative automatic thoughts and to generate adaptive responses. DTRs can be completed before, during, and after episodes of craving. The use of DTRs can demonstrate to patients that they are not help­ less in the face of their cravings/urges, and a review of old DTRs can serve as a reminder of this key fact. Figure 10.1 depicts "Jim's" DTR. As can be seen under the Situa­ tion column, Jim was sitting at home. He had recently had an acci­ dent at a constraction site, resulting in a broken hand and wrist. Under the heading "situation" he wrote: "Sitting at home, m y hand is bro­ ken, and I can't go to work. There is plenty of money in m y pocket." Also, he started thinking about some of his old drag buddies and the last time that he used cocaine. Jim was also aware of having a strong craving for cocaine. He rated at 9 5 % the amount of boredom and anxiety that he was experiencing—an indication of sfrong negative feel­ ings. Some of the automatic thoughts he identified were "There is nothing to do" and "I can't stand this boredom." In examining their automatic thoughts, patients ask themselves five basic questions (see Chapter 9, this volume). The first question is "What is the evidence for and against m y automatic thoughts?" The second question is "Are there other ways of looking at this situation?" The third question is "If it is trae, what are the realistic consequences?" The fourth question is "What are the drawbacks to m y continuing to dwell on these thoughts?" The final question asks, "What construc­ tive action can I take to solve this problem?" Jim began examining the automatic thought, "There is nothing to do," by asking himself, "What's the evidence against this?" He responded to this question by saying, "Actually, there are plenty of things that I could do; for example, go to a meeting, watch a game, or call m y therapist." Next, he asked himself, "Are there other ways of looking at this? Do 1 really mean there is nothing to do?" His response was, "No, it is not trae that there is nothing to do, but the pain of boredom makes it difficult for m e to see other things that 1 might be able to do." He then asked, "If trae, what are the realistic consequences?" He responded, "Well, if it is trae that there is noth­ ing to do, then the consequences will be that I will feel bored, and, although the boredom is painftil, it's not the end of the world. The consequences are that I will feel bored and even that will go away." Jim's next question was, "If trae, what are the drawbacks to my con­ tinuing to dwell on these thoughts?" He responded, "The disadvan- DAIL Y THOUGHT RECORD Directions: W h e n you notice your mood getting worse, as(< yourself, "T/Vhat's going through m y mind right now" and as soon as possible jot down the thought or mental image in the Automatic Thought Column. SITUATION AUTOMATIC THOUGHT(S) EMOTION(S) RATIONAL RESPONSE OUTCOME Describe: 1 Write automatic thought(s) 1. Specify sad. 1. Write rational response to automatic ttiought(s). 1. Re-rate belief 1. Actual event leading to that preceded emotion(s). anxious/ 2. Rate belief in rational response 0-100%. In automatic unpleasant emotion, or 2. Rate belief in automatic angry, etc. thought(s) 2. Stream of thoughts, day­ thought(s) 0-100% 2. Rate degree 0-100% dreams or recollection. of emotion 2. Specify and leading to unpleasant 0-100%. rate emotion, or subsequent DATE/ 3. Distressing physical emotions TIME sensations 0-100%. sitting at home,my "There is nothing to do" Bored 'Actually, there are plenty of things I 10% hand broken, and I could do; for example, go to a meeting. can't go to work. "I can't stand this Anxious watch a game, or call my therapist." Bored There is money in boredom." (30%) my pocket. (95%) "It is not true that there is nothing to Thinking about some dfoi cublutt ftohre mpaei nt oo sfe eb oroetdhoemr mtahkiensg s itth adti f­ Anxious (20%) old drug buddies and I might be able to do." the last time I used cocaine. "If it is true that there is nothing to do, Start craving for btohreend ,t hae ndc onasletqhuoeungch est hwei lblo rebdeo mt haits pIa ifnefeull . cocaine. it's not the end of the world. The conse­ quences are that I will feel bored and even that will go away." "The disadvantage of continuing to dwell on these thoughts is that I feel helpless, which in tum, leads to the desire for cocaine." "I can go get a newspaper and read the sports page until it's time for a meeting." (90)% Questions to help formulate ttie rational response: (1) What is the evidence ttiat the automatic thought is true? Not true? (2) Is there an alternative explanation? (3) Whafs the worst that could happen? Could I live through it? What's the best that could happen? What's the most realistic outcome? (4) What should I do about it? (5) Whars the effect of my believing ttie automatic thought? What could be the effect of changing my thinking? (6) If (friend's name) was in this situation and had this thought, what would I tell him/her? FIGURE 10.1. Jim's Daily Thought Record. 166 COGNITIVE THERAPY OF SUBSTANCE ABUSE tage is that I feel helpless, which, in turn, leads to the desire cocaine." The patient's last question, "What constractive action can I take?" led to the response, "I can go get a newspaper and read the sports page until it's time for a meeting." Jim's belief in his rational response was 90%. In the last column, we see that Jim has re-rated his belief in the automatic thoughts at 10%. This indicates that the rational response has had a significant impact on modifying Jim's belief in that automatic thought. W e also see that his level of bore­ dom and anxiety has diminished, going from 9 5 % for each to 30% for boredom and 2 0 % for anxiety. This level of diminishment indi­ cates a reduction in the degree of intensity for these emotions, which, in turn, may help to reduce the likelihood of Jim's using cocaine to cope with the boredom and anxiety. In addition, by using a 0-100 rating scale on the standard DTR form (regarding the level of confidence about the thoughts), the patient understands that the automatic thoughts are not necessarily objective realities. Patients can learn to use the rating scale to gauge changes in their perceptions as they apply cognitive and behavioral techniques. The 0-100 rating scale serves therefore as a useful cogni­ tive barometer. Craving is an idiosyncratic experience made up of cognitive, affective, behavioral, and physiological components. Therefore, by identifying its various components patients can be more objective about the craving experience, and thus diminish the subjective in­ tensity. Later, through the use of subsequent DTRs, the therapist was able to help Jim to be more objective about craving. The trigger (e.g., a sense of extreme boredom or anxiety) activated the beliefs, "I can't cope without cocaine" and "The craving makes me do it," represented by the automatic thought, "I can't stand this." His physiological responses, tension, excess sweating, and urge to seek relief represented the sequence of events that Jim originally labeled "craving." The thera­ pist then helped Jim see that interventions could be made, for exam­ ple, becoming aware of the automatic thought and being able to respond rationally to it, thus diminishing the urge to use cocaine. In addition, it is important to help patients to cope with the crav­ ing by teaching them to test their idiosyncratic predictions about the duration and intensity, as well as the patients' mastety, of the crav­ ing phenomenon. This result can be achieved by teaching them to monitor the intensity, frequency, and duration of craving in order to attack their dichotomous or catastrophic view of the craving phenom­ enon (Horvath, 1988, in press; Tiffany, 1990). Examples of such catastrophic predictions are the following: Craving/Urges 167 "If I don't use something right now, I won't be able to face going to work; I'll lose m y job." "Without the drug,

I'll be a nervous wreck all day. Evetyone will think I'm having a breakdown." "I'll never have a normal life again. I'm a slave to the drag. I simply have to have it to get through the day." "If I resist taking the drag now, I'll just need twice as much later to feel normal later. I might even overdose if that hap­ pens." Activity Scheduling Patients who have a long histoty of drag abuse often engage only in activities that center around the use or the procure­ ment of drags. Oftentimes, their entire social network is drag related. When patients are ttying to control their substance use they often find that initially it is beneficial to stay away from the people, places, and things associated with their former lifestyle (O'Brien et al., 1992). As a result, recovering patients may be faced with a great deal of idle time on their hands. The boredom that accompanies this state can spell frouble for patients' abstinence unless new activities are substituted. Activity schedules are helpful in this regard. First, the activity schedule is used to gather baseline data on how patients actually spend their time. Activity schedules can also be used proactively to strac- ture the patients' day in a constructive way. Many drug-abusing patients have forgotten some of the activities they once enjoyed prior to their drag-using days. The scheduling of activities can revive some of these old, enjoyable, prosocial activities, the likes of which may assist the patient in rebuilding a drag-free life (Hall, Havassy, & Was­ serman, 1991). Such activities serve two purposes. Some of them help patients in the short term to deal with the immediate crisis of coping with urges. Other acti'̂ dties on this list are long-term alternatives to patients' previous drag-related behaviors. In general, the purpose of these sub­ stitute activities is to give patients something to do other than the short-lived and deceptively positive experience of using drags. This is not to say that these activities would be equal to the immediate, intense pleasure received from using such drags as cocaine; however, these activities do have many long-term advantages. While it is vital to teach drag-abusing patients to find alterna­ tive, nondrug sources of reinforcement, therapists must bear in mind that this may require a great deal of training and practice. The diffi­ culty lies in the fact that taking drags requires no particular skills. 168 COGNITIVE THERAPY OF SUBSTANCE ABUSE but the alternative nondrag activities may require considerable sk (Stitzer et al., 1984). Therefore, therapists must not take for granted that patients have the know-how to schedule positive activities and must be prepared to deal with patients' low self-confidence, high fras- tration, marked hopelessness, and passive avoidance surrounding this technique. Relaxation Training Another technique that we have also found to be use­ ful is relaxation training (Bernstein & Borkovec, 1973). Relaxation training gives patients a tool that they can use to help cope with such feelings as anxiety and anger, which, for some patients, can be trig­ gers for cravings (see Chapter 15, this volume). When introducing relaxation techniques to patients, it is impor­ tant to offer them a rationale for the use of this intervention. For example, relaxation training is a method for reducing tension that, left unchecked, might trigger cravings. Relaxation training also helps one to develop an improved general sense of well-being, and to lower one's sense of stress in (fey-to-day life. In addition, a relaxed individual is less likely to act in*pulsively and out of a sense of desperation. It is important that patients understand that relaxation is a skill that can be learned like any other skill. The more they practice their relaxation training, the better they will become at evoking a deep state of relaxation. W e recommend that the first relaxation exercise take place in the therapy session, under the therapist's supervision, in order to ensure that the patients are doing it properly. SUMMARY The three main goals of this chapter were: (1) to focus on the importance of dealing with urges and cravings, (2) to understand better the patient's subjective experience, and (3) to describe techniques that can help patients to cope better with urges and cravings. Uncontrolled urges and cravings are a major factor contributing to treatment failure. Therefore, it is imperative to teach patients early in treatment how to monitor and deal with urges and cravings. C H A P T E R 1 1 F o c u s o n B e l i e f s B eliefs are relatively rigid, enduring cognitive stractures that are not easily modified by experience. In cognitive therapy it is generally maintained that beliefs have a profound impact on feelings and behaviors. For example, depressed patients have global, negative views about themselves, the world, and the future that con­ tribute to their feelings of despair, guilt, and sadness (Beck et al., 1979). Negative beliefs also contribute to such depressive behaviors as isola­ tion and withdrawal. In cases of anxiety disorders, patients have nega­ tive, apprehensive beliefs about some future threat that contribute to avoidance, anxiety, and perhaps panic (Beck et al., 1985). There are at least three types of beliefs pertinent to the patients' addiction to drags: anticipatoty, relief-oriented, and facilitative or permissive. As described in Chapters 2 and 3 (this volume), anticipa­ toty beliefs involve some expectation of reward, such as "The party tonight will be great. I can't wait to go get high!" Relief-oriented beliefs are those that assume that using drags will remove an uncom­ fortable state, for example, "I can't stand withdrawal. I need a hit." And finally, facilitative or permissive beliefs are those that consider drag use acceptable, in spite of the potential consequences, for exam­ ple, "Only weak people have problems with drags. It won't happen to me." Permissive beliefs also have much in common with what are more commonly known as "rationalizations." Patients have thoughts that seem to "justify" their drag-using, such as "I have to use cocaine or I won't be able to concentrate on m y work." Such thinking is tanta­ mount to self-deception. The following examples illustrate the addictive beliefs of two individuals who are cocaine-dependent. (These case examples will be used throughout this chapter to illustrate important points.) 169 170 COGNTTIVE THERAPY OF SUBSTANCE ABUSE "Louise" is a 21-year-old unemployed, single parent who has been using cocaine for two years. Most people who encounter Louise can see that she has had a "rough life." W h e n Louise cannot afford to purchase cocaine, she turns to prostitution to acquire money. She holds the following addictive beliefs about cocaine and herself: "I need drags to numb the pain." "I might as well do drags since m y life won't ever improve any­ way." "Getting high is the only thing I look forward to." "I'm tough; I can handle drags." In contrast to Louise, "Bill" is a 39-year-oId successful sales execu­ tive who has been using cocaine for the past 3 years. He "loves to party" with his friends and coworkers. Bill holds the following addic­ tive beliefs: "I work like hell all week. I deserve to get high on the weekends." "I can't keep up the pace without an occasional 'pick-me-up' [cocaine]." "I've never failed at anything, so drags won't hurt me." "I'm basically a pretty decent guy." At first glance. Bill and Louise appear to be quite different from each other. However, as we examine them more closely we see that they each fall prey to anticipatoty, relief-oriented, and permissive beliefs that perpetuate their addictions. In conttast to addictive beliefs, individuals may have control behefs. Conttol beliefs are defined as beliefs that decrease the likelihood of dmg use and abuse. The following are examples of conttol beliefe: "Drags are dangerous to m y well-being." "I am capable of withstanding the urges." "If I tolerate this craving for awhile, it will go away." "If I resist the urges, 1 will feel stronger." "It is in m y best interest to stay drag-free." An addicted person may maintain contradictoty beliefs. For exam­ ple, a drag-dependent person may hold the addictive belief "I love the feeling of being high." Simultaneously, the same person may hold the control belief "This is killing me." As a result of these contradic­ toty beliefs, drag-dependent persons often feel a great deal of ambiva­ lence about their habit. They may find themselves, for example, awak­ ening in the morning and "swearing off" drugs. By noon they may Belief 171 be seeking treatment for their addiction, although by evening they may be using again. (Confradictoty beliefs are also common, of course, among people who have tried to diet or quit smoking.) When addictive beliefs are more salient than control beliefs, a drug-dependent person is more likely to use drugs. Of course, the opposite is trae: when a person's control beliefs predominate over addictive beliefs, that person is more likely to abstain from drags. An ideal goal for the cognitive therapist is to identify and eliminate the patient's addictive beliefs, replacing them with more adaptive con­ trol beliefs. More realistically, the goal of cognitive therapy is to facilitate a process whereby the patient's control beliefs become more salient than his or her addictive beliefs. The result, of course, will be that the patient abstains from, or at least diminishes, drag use. The following is a list of specific methods to address the drug-abusing patients' problematic belief systems: 1. Assess beliefs. 2. Orient the patient to the cognitive therapy model. 3. Examine and test addictive beliefs. 4. Develop control beliefs. 5. Practice activation of control beliefs. 6. Assign homework that addresses beliefs. In the remainder of this chapter, these methods are described in detail. ASSESS BELIEFS In order to modify addictive beliefs, the cognitive therapist must first have an accurate understanding of the role of these beliefs in the patient's life. Thus, a careful assessment must take place for each patient. This assessment may be accomplished in two ways: through therapist^atient interaction during psychotherapy sessions and standardized questionnaires. The following open-ended questions are examples of those use­ ful for eliciting information about patients' beliefs: "What are your thoughts about ?" "What was going on in your head when happened?" "How do you explain ?" "How do you interpret ?" "What does mean to you?" 172 COGNITIVE THERAPY OF SUBSTANCE ABUSE "What's your 'rale of thumb' here?" "How did you size up the situation?" As the patient responds to these questions, the therapist verbally reflects what the patient has said, with particular emphasis on the beliefs expressed by the patient. At various points in the interview, the therapist provides "capsule summaries" of what has been dis­ cussed, with strong emphasis placed on the patient's thoughts. To illustrate this process we present the following dialogue between a patient and her therapist, taken from their first session. Louise was referred for therapy by a primaty care physician who treated her at a county health clinic for gonorrhea. TH: Hello, Louise. What would you like to talk about today? PT: I don't really even know. I was sent here by that doctor at the clinic. I figured I had to be here. TH: You must have some concerns ... some things that are bother­ ing you. PT: Yeh, I guess. But I don't know what good it will do to talk to you. TH You doubt that this will be helpful. PT Yeh, that's right. TH What other thoughts do you have about being here? PT Well I've been in treatment before, but as you can probably guess, I'm on the shit again. TH: What do you mean by "on the shit"? PT: O h come on, man! You know what that means! I'm doing drags! TH: Does "on the shit" mean that you are doing drags daily? weekly? monthly? PT: To me it means doing any drags at all! TH So any slip and you consider yourself "on the shit"? PT: Yeh! I was clean for a month and then last week 1 had a really bad time with my old man. I went right out on a two-day binge. TH: And what have you done since then? PT: If you mean drags, I have been clean since then, but I don't suppose it

She is calling m e all kinds of names: "lazy," "worthless," you know the list. Anyway, I feel this urge to ran, but I know there's nowhere to go. I want to hit her but of course I don't. I want to cty, but I wouldn't give her the satisfaction of seeing that she has hurt me. And finally I start thinking about going to Michelle's house. She always knows were to find some shit. I think: "That's the one thing that will get m e feeling better." I know that I can leave my daughter with m y mother while I go out to get high. I also think about what I'll do if I don't find Michelle. If she's not around, I'll just go downtown and tum a quick ttick. And then I get more intense feelings of wanting to get high ... {attention returns to therapist)... Well, you've succeeded in making m e crave the shit. N o w what are you going to do? TH You are craving cocaine now? PT: Damn straight! TH OK, now start reviewing your control beliefs out loud; you know what they are. PT: The disadvantages? TH Yes, but say it with feeling! Take out your flashcards if you like. PT: No, I don't need the flashcards. {she begins tentatively at first, but then builds her enthusiasm) I am not going to get high; my baby needs me. The stuff is killing me. It really doesn't make anything better. I'm better off without it. I can stay clean if I want to. I can make a life for myself and m y baby if I stay cleanl {she smiles) TH: What are you thinking right now as you smile? PT: That I can do it! TH: Great! Beliefs 185 In this dialogue, Louise succeeded in creating a strong urge to use cocaine, but she successfully countered that urge with control beliefs. This process was repeated several times over the next few ses­ sions with Louise. For homework, she was encouraged to practice this exercise in vivo, as drag temptations naturally arise. ASSIGN H O M E W O R K THAT ADDRESSES BELIEFS As in all other applications of cognitive therapy, homework involves applying the skills learned in therapy sessions. Thus, homework is a vital extension of therapy (Burns & Auerbach, 1992; Persons et al., 1988). As a long-term goal of homework in cog­ nitive therapy, patients should learn to use self-guided Socratic ques­ tioning spontaneously in their lives; for example, "What evidence do I have for that belief?" "How else can I look at this situation?" "What are the consequences of m y beliefs?" Homework is an opportunity to practice applying control beliefs in the "real world." It may involve having patients practice activat­ ing confrol beliefs in the face of tempting high-risk stimuli, since they will never succeed in avoiding all tempting stimuli. Louise, for exam­ ple, is unlikely to change her mother's behavior, although she can learn to cope more effectively with her. A specific homework assign­ ment given to Louise, therefore, might be to practice control beliefs in response to her addictive beliefs and automatic thoughts when in the presence of her mother. Homework may also involve testing addictive beliefs to evaluate their validity. For example. Bill might be challenged to tty various methods for having fun, in order to test his belief that "there is noth­ ing more fun than using cocaine." As explained in Chapter 6 (this volume), homework is assigned at the end of each session and it is reviewed at the beginning of each follow-up session. Initially, homework is quite stractured. For example, patients are instracted to complete DTRs on a daily basis. Later, how­ ever, homework can be less formal, as the patient develops new, more adaptive, patterns of thinking. SUMMARY The basic beliefs and automatic thoughts about drags may account for much of their use. There are at least three types of 186 COGNTTIVE THERAPY OF SUBSTANCE ABUSE drag-related beliefs that contribute to urges, craving, and ultimate use of drags: anticipatoty beliefs, relief-oriented beliefs, and permissive beliefs. The role of the cognitive therapist is to assess, examine, and test these beliefs with the patient, in order to ultimately replace them with control beliefs. There are m a n y cognitive strategies that facili­ tate this process, many of which were presented in this chapter. C H A P T E R 1 2 M a n a g i n g G e n e r a l L i f e P r o b l e m s P Aa.t ait ients rarely enter treatment for drag addiction or dependence on their own accord in the absence of general life prob­ lems (Carey, 1991). W h e n patients are in an early phase of their drug use, they are typically quite pleased with the effects of the drugs. This is so either because the drags produce a state of unmatched excite­ ment and euphoria or because they offer the abuser an artificial respite from the demands, pressures, ennui, and emotional pain that they may be suffering. At such times, patients operate under the assump­ tion that the use of mood-altering chemicals offers a workable, viable option to functioning in a drug-free state of mind. However, as the drag user becomes more regularly active in seek­ ing, achieving, and repeating the "high" experience, a number of problems surface. Such problems include, but are not limited to, (1) the realization that drug use does not help the actual demands, responsibilities, and troubles of evetyday life magically to go away; (2) the development of a physiological tolerance to the drags, and therefore the need to expend more time, energy, and money in the search for the ever-elusive "high"; (3) the exacerbation of neglected life concerns, thus increasing stress and the desire to escape through the use of psychoactive substances; and (4) fallout from worsening habits, in terms of damaged relationships, vocational and/or academic failure, and serious medical and legal complications. It is in this advanced state of psychosocial difficulties and life crisis that the individual may be regarded as a drug abuser and will most likely appear for psychological treatment. 187 188 COGNITIVE THERAPY OF SUBSTANCE ABUSE The therapist is faced with the daunting task of helping such patients not only to arrest and ameliorate their drag addictions, but also to deal with many serious real-life difficulties. While many of these problems are similar to those of any other diagnostic categoty of patient, there are many that are particularly salient to the drag- abusing population. It is imperative that therapists be aware of the typical life prob­ lems that the drug-abusing patient presents at the start (and during the course) of treatment. It is equally important that these problems be given adequate attention in therapy, in spite of the temptation to focus solely on the substance abuse disorder. With this in mind, the current chapter presents an overview of the most common life issues that the cognitive therapist will need to address in the treatment of this challenging population. "CHICKEN A N D EGG" C O N U N D R U M : W H I C H COMES FIRST, SUBSTANCE ABUSE OR ASSOCIATED LIFE PROBLEMS? An important assessment question concerns whether the patient's major life problems precede or postdate the onset of the substance abuse disorder. (Similarly, it is crucial to note whether the patient's antisocial behaviors are primaty or secondaty to the drug addiction. See Chapter 16, this volume, for more details.) Informa­ tion regarding the chronology of life problems and drag abuse pat­ terns can shed light on the "function" of the patient's use of psychoactive substances, as well as elucidating the factors that serve as triggers for the abuse of drags. For example, "Maria" presented with a crack addiction that had originally begun two years previously, in the aftermath of the vio­ lent death of her younger sister. After being arrested for possession, Maria spent the next three months in various inpatient rehabilitation facilities. When she was discharged, Maria was convinced that she would never "pick up" (resume using drugs) again, and that she was on the right track. However, shortly thereafter she was strack another blow when one of her best friends died. Almost immediately, she resumed her use of crack and once again found herself back in court. After a brief stay in jail she was released on parole, on the condition that she receive ongoing treatment. It was at this time that Maria entered cognitive therapy. In Maria's case, grief and loss were powerful triggers for drug abuse. Since she had no histoty of drug abuse prior to the death of Managing Life Problems 189 her sister, her prognosis was quite hopeful. It was clear that she had many years of experience as a well-functioning, responsible per­ son. Furthermore, her life was fairly well ordered, and her stressors ,/ere not out of the ordinaty realm of evetyday life. However, in order for therapy to be complete, it would be necessaty to help Maria come to terms with the deaths of her sister and friend and her concomi­ tant belief "I'm all alone." Additionally, the therapist would need to teach the patient to use cognitive therapy skills in a preparatoty fashion in anticipation of episodes associated with loss. For example, the illness of a significant other might be sufficient to induce Maria to have catastrophic expec­ tations of that person dying. Such an extreme worty would put Maria at risk for relapse, as her anxiety at the thought of losing someone (and feeling alone) might induce her to self-medicate with crack cocaine. Similarly, the calendar became a source of negative cues that had to be anticipated, and that served as a therapeutic call to action. Major holidays, birthdays, and anniversaries of the deaths of loved ones had become capable of arousing upsetting automatic thoughts and emotions, the likes of which could trigger a resumption of drug abuse. As it turned out, Maria learned the skills of rational responding, scheduling activities, and problem-solving quite well, and she survived the "anniversaty phenomenon" without incident. However, when she was "blind-sided" by her boyfriend's decision to leave her, she suf­ fered a temporaty drag lapse (a single binge episode). Another inter­ personal loss had triggered a need to blot out her emotional pain through the use of crack cocaine. Maria's case seems straightforward—life problems preceded the onset of substance abuse. O n the other hand, we see cases where the onset of serious life difficulties seems to occur as a result of the sub­ stance abuse. Until the time that "Roland" began using drugs in high school he seemed to have a fairly unremarkable life. He was an aver­ age student, came from an intact family free of substance abuse, and he had a circle of regular friends. After being introduced to marijuana, Roland began to skip classes and to disengage himself from his fam­ ily, many of his friends, and most of his normal recreational activi­ ties. As his drug use came to involve "harder" substances such as quaaludes and cocaine, he incurred more and more serious concomi­ tant life problems. For example, although he somehow managed to graduate from high school, Roland did not pursue further education or vocational training, nor did he seek employment. His sole source of income was through petty drag trafficking. His parents, recogniz- 190 COGNITIVE THERAPY OF SUBSTANCE ABUSE ing that something was dreadfully wrong, pressured Roland to "clea up his act" and to get a job. This frequently led to screaming and shoving matches between son and father. Finally, after Roland's par­ ents were frightened by two incidents when drag associates came to the door in the middle of the night, Roland was told to leave the house. Shortly thereafter, Roland was arrested for driving while intoxicated, as well as for possession of a controlled substance. Roland's steadily increasing problems with substance abuse led to academic and vocational stagnation, family conflicts, loss of his primaty domicile, and legal troubles. His main interests and activi­ ties were reduced to procuring, using, and selling illicit substances. His chief associates were drag users and dealers, and he had no friends who

led a drug-free lifestyle. Roland's drag use had led to numerous negative consequences in his life, yet his sense of helplessness in dealing with these mounting difficulties, coupled with his desire to "forget all his troubles" through the use of drags such as cocaine, perpetuated his drag-related lifestyle. These were problems that had to be addressed when Roland ultimately entered therapy, in addition to his problematic abuse of drags per se. Although the cases of Maria and Roland seem quite distinct in that the former suffered identifiable life troubles as a precipitating factor to the onset of her drag abuse, while the latter produced the majority of his own life problems as the result of his apparently "unprovoked" recreational use of drags, the two cases have at least one vety important similarity. Both patients eventually became trapped in a vicious cycle, where the dual problems of drug abuse and general life problems began to exacerbate each other. For example, Maria's use of crack cocaine, originally the result of her grief, became the cause of her arrest and incarceration. Aside from the obvious negative impact that this had on her life in its own right, Maria further suffered from an inability to find employment as a result of her criminal record. In her anger and frastration, she once again resorted to the use of crack cocaine. Renewed legal troubles soon ensued, and the vicious cycle was com­ plete. Roland's vicious cycle of substance abuse and life crises was even more entrenched and self-perpetuating than Maria's. His "druggie lifestyle" led to numerous personal, financial, and legal complications. He had no stable relationships, no sense of trast for others, no sense of self-esteem when he came off his high, no legal source of income, no healthy sources of stimulation and pleasure, no treatment for numerous medical problems, and a constant nagging fear of crimi- Managing Life Problems 191 nal apprehension. All these concerns were quickly "fixed" by using drags such as cocaine, thus keeping him firmly ensconced in his drag lifestyle. Drags were a major source of his chaos and crises, yet he continued to view them as his only "solution" to deal with these problems. Later, when Roland began treatment, he would not only have to overcome his addictions, he literally would have to "get a life" as well. (Roland's personality disorders also played an important role with regard to his drag use and general life problems. For more on this, see Chapter 16, this volume, on the interaction between sub­ stance abuse and personality disorders.) In either of these prototypical cases, the life problems represent broad "stimulus situations" that both trigger and are worsened by the patient's continuing abuse of substances. Figure 12.1 illustrates the beliefs and automatic thoughts that feed into this process. As depicted by Figure 12.1, the patient's chronic unemployment serves as a stimulus situation that sets off a chain of beliefs, automatic thoughts, and actions that lead to drug use. The patient's drag use then contributes to his or her becoming less employable (less moti­ vated, less reliable, etc.), and the cycle continues. This chapter focuses on the major areas of chronic life stress that plague so many substance abuse patients. (See Chapter 13, this vol­ ume, for examples of common acute crises and emergencies, many of which are related to the issues described later.) Some of these are more likely to be precursors to a drag problem, while some are more clearly the sequelae to drag abuse. In either case, it is important to bear in mind the way that these trouble spots become part of a vicious Life Beliefs Automatic Urges/ Problems Thoughts cravings "I'm a failure. (Chronic My life is "Might as unemploy­ hopeless, so well get ment) drugs can't high now." hurt me." :^ Lapse Instnjmentai Facilitating Beliefs episode drug-seeking (permission) behaviors "Using will make m e (action) feel better." F I G U R E 12.1. The vicious cycle of life problems and drag abuse. 192 COGNITIVE THERAPY OF SUBSTANCE ABUSE cycle that has spiraled out of control. The following life problem described in detail here: marital and family problems, socioeconomic problems, daily stressors, legal problems, and medical problems. MARITAL AND FAMILY PROBLEMS As noted earlier, it is vital for patients to be aware of the "people, places, and things" that they associate with the using of illicit substances. In the early stages of treatment, it is wise for the substance abuser who is ttying to remain drag-free to avoid needlessly coming into contact with stimuli that might provoke strong cravings or facilitate the obtaining of drags. (Later in treatment, however, it is important to teach patients to apply and practice self-help techniques in the face of those kinds of drag-related stimuli that may be unavoid­ able in their everyday lives.) W h e n the "people" and "places" are impersonal acquaintances and out-of-the-way locations, it is reason­ ably straightforward (under the proper motivational and coping con­ ditions) for the patient to steer clear of these stimuli. The picture is decidedly more difficult when the people and places are the patient's own family members, in the patient's own residence. Substance abuse amongst one's relatives poses a serious problem for the patient. In terms of etiology, a person whose parents or siblings abuse drags is at risk for modeling these behaviors (Gomberg & Nirenberg, 1991; Lang, 1992). The patient may have easy access to drag paraphernalia, and may have knowledge of family members' secret stashes. In terms of maintenance and relapse, the patient may have a particularly difficult time turning away from the continual lure of drugs when family members are high, when these family members encourage the patient to get high as well, and when drags and money are so blatantly available at all hours of the day. Furthermore, there may be strong social pressures to conform to the "drug subculture" in the household. One of our patients. Dee, was taunted unmercifully by her two substance-abusing brothers for her attempts to stay away from drugs and to attend therapy and support group meetings. They attempted to "bring her down" by ripping into her for her "holier- than-thou" attitude, hoping that she would succumb to their barbs by joining them in their smoking binges. Luckily for Dee, these broth­ ers lived a block away, so she did not have to deal with them on a constant basis. As it was, however, a great deal of time and energy in therapy was spent in learning how to assert herself effectively to her brothers, and to respond rationally to her own initial automatic thoughts. These thoughts included "If you can't beat 'em, join 'em," Managing Life Problems 193 "I have to get them off my back," and "I have to prove that I ain't stuck up." Rational responses that were generated in response to these automatic thoughts included the following: 1. "If they [the brothers] want to be losers, that's their business I'm doing the right thing by staying straight." 2. "I'd rather that m y brothers hate m e than go back to using and have m y kids hate me." M y kids are much more important." 3. "I'm not stuck up. I'm tiying to be humble and know m y limi­ tations. M y brothers are the ones who are being stuck up, only they're so stoned they don't see it." 4. "I can't change m y brothers, so why should I let them change me?" and, 5. "I've dealt with worse hassles before. I'll just remind myself that evetything they say is just bullshit, and I'll get back to m y own business." Dee was able to defend herself from her brothers' attempts at heavy-handed peer pressure, but the fact that her two "favorite" sib­ lings had drag problems was a considerable source of stress for her. Dee believed that her brothers' drag problems signified that she came from a family of "losers," each member of which (including herself) inevitably would have a life filled with troubles. Furthermore, she believed that she would need to stay away from her brothers as much as possible, thus depriving her of what had once been enjoyable and companionable sibling relationships. This exacerbated her sense of loneliness. It is generally unrealistic for patients to expect that they can solve the substance abuse problems of family members. It is generally wis­ est to focus on cognitive coping skills that enable patients to gain some distance from the problem, and to stay focused on their own recovety. Toward this end, support group meetings can be recom­ mended as an adjunct to cognitive therapy, especially when patients feel overly guilty for family members' drag problems. The negative experiences that these patients have had during their upbringing in alcohol and drag-using households can be discussed as part of treat­ ment as well. This builds rapport between patient and therapist and helps to elucidate implicit dysfunctional rales about substance abuse and family relationships that the patients may have learned. To have a mate who is self-absorbed in drag abuse entails tre­ mendous stress for a patient. Problems include the partner's (1) hav­ ing erratic mood swings, (2) being financially and sexually reckless 194 COGNTTIVE THERAPY OF SUBSTANCE ABUSE and irresponsible, (3) risking major medical and legal consequence (4) endangering the patient with regard to AIDS and other sexually transmitted diseases, and (5) seducing the patient into relapsing into substance abuse once again. When a drag-addicted mate refuses to acknowledge a problem or to get help, it is often necessaty for the patient either to leave the household or to induce the mate to leave. Such a strategy may be the best way to maintain a home environ­ ment that is conducive to recovety. In order to maximize the chances that patients will be willing be go along with such a strategy, therapists can emphasize that it may not be necessaty to leave the relationship permanently, but that such a move may be needed at present. The drawbacks to the above strategy, in the mind of the patient, may be compelling. The patient may be loath to confront a potentially violent partner, and may not have the resources to make a clean start in another domicile. Furthermore, the patient is forced to face the loss of a significant love relationship—a daunting realization, especially if the patient has dependent personality characteristics (see Chapter 16, this volume). When this is the case, the therapist may notice that the patient seems excessively emotionally attached to the substance-abus­ ing partner, no matter what the personal cost. In short, the patient may love the partner deeply, in spite of this person's drug abuse. The patient may go to great lengths to excuse or cover up the partner's transgressions, because the prospect of losing the partner (or even of hearing negative feedback about the partner) is too much to bear. Even worse, the patient may collude with the partner to use drags together in secret and to provide cover for each other. Such phenomena have been written about extensively in recent times under the rubric of "co-dependency" (cf. Lyon & Greenberg, 1991). In such cases, cognitive interventions must focus heavily on adaptive problem-solving (Piatt & Hermalin, 1989), and on the patient's exaggerated guift and fears of loneliness. For example, the patient may believe erroneously that the significant other would use more drugs if left alone, when in reality the significant other is dem­ onstrating an escalating drug problem whether the patient is there or not. Similariy, the patient may believe that he or she could not bear the loneliness of being without the drug-using partner, when in actijality the patient already is coping with being functionally alone, as the drag-using partner is emotionally absent. Additionally, it is vital to teach the patient to expand his or her social network of nonusers, so that the substance-abusing mate is no longer the dominant source of interpersonal reinforcement. At the same time, patient and therapist must be ever vigilant to the patient's own vulnerabilities to using drags. It is important to pay Managing Life Problems 195 close attention to the way that the patient's chronic life problem substance abuse feed into each other, rather than looking at each problem in isolation. For example, in the case of the

patient whose significant other is abusing drags, the belief that "if I stay off drugs I will have to give up the person I love and be all alone" may prime the patient to resume using drags. "Walter" had told his therapist many times that the love of his children was one of the only things he cared about in his life. There­ fore, the fact that he was now not permitted to see them was a source of fremendous distress. As a result, he escalated his drag bingeing— partly out of a belief that "a greater amount of drags is needed to blot out a greater amount of suffering," and partly out of a desire to make a suicidal gesture that would attract attention and care-and soon wound up in the hospital due to overdose-induced convulsions. His failure to handle his distress without the use of chemical substances led to an accentuation of his sense of loss (his children), which in turn brought about a life-threatening exacerbation of his drug abuse. Following his release from the hospital, and his resumption of outpatient cognitive therapy, a great deal of time in session was spent in reviewing the pros and cons of various strategies for coping with interpersonal loss. Among the most favored therapeutic suggestions were seeking the support of others, such as family, friends, and thera­ pist, and focusing on work-related activities, especially those that required a great deal of strenuous physical labor. Additionally, Walter was instructed to respond to his urges to buy and use drags by con­ juring up images of his children, the people he most wanted to be proud of him. These images were to serve as strong deterrents to his initiating the search for drags (however, they were less effective in stopping him from using once he had the drags and paraphernalia in his hands). Ongoing marital discord (or discord in any love relationship) can serve as a powerful stimulus for drag cravings and drug abuse. W e have seen patients use drugs in such cases for a variety of purposes. Such "functions" of the drag use include (1) reviving a chemically based, false and temporaty boost in self-esteem; (2) exerting control in the relationship by defying the spouse, and by inducing a self-pro­ tective state of apathy or invulnerability; (3) "soothing" the anger that is felt toward the spouse, especially if the patient has the belief that not using something will lead to an escalating physical confrontation; and (4) finding an emotional escape from marital unhappiness, espe­ cially if the drag-abusing spouse believes that there is no way to solve the marital problem or to leave the marriage. In general, unhappi­ ness and anger in a relationship feed into patients' feelings of help- 196 COGNITIVE THERAPY OF SUBSTANCE ABUSE lessness and loneliness, which in turn may lead to a wish for imme ate relief. The result may be a craving for drags, followed by their use. It is not surprising to note that the use of drags within the con­ text of a marriage will likely affect the relationship adversely. If the marriage is already disturbed, the substance abuse will exacerbate matters, thus contributing to yet another vicious cycle. The situation is perhaps most dire when the abuse of substances leads to physical, psychological, and/or sexual assault. In contrast to the oft-held belief (alluded to above) that the use of drags can "soothe" the angty spouse, drug use actually more often than not serves to disinhibit and to activate the potential for violence (Amaro, Fried, Cabral, & Zuckerman, 1990). Therefore, the therapist must be ever vigilant to possible signs of domestic violence in the life of the substance-abusing patient. Along with this vigilance, it is wise for the therapist to have ready access to information about crisis hotlines and shelters for victims of physical and sexual abuse. If the patient is primarily a perpetrator of such violence, the thera­ pist must do all that is possible to contract with the patient to deal with this problem as a number one priority in treatment. As discussed previously in Chapter 4 (this volume), the limits of confidentiality that exist when the patient is a threat to the well-being of others need to be spelled out. At the same time, great therapeutic skill must be employed to keep the patient positively engaged in the process of therapy and to keep the patient's trast. An examination of the patient's personal and family histoty may demonstrate that physical, psychological, and sexual abuse during the patient's childhood and adolescence played a significant role in the development of the patient's drag problem. (Note: In order to high­ light patterns and relational configurations in the patient's family, it is sometimes useful to map out a family tree on paper. Such a tree can indicate separations, divorces, deaths, abusive relationships, per­ sons who abused alcohol and drags, and the like.) This may be trae for any number of reasons, such as (1) the aversive conditions at home force patients to spend more and more time on the streets, thus exposing them more to the drug culture; (2) the perpetrators of vio­ lence in the family may be drug abusers themselves, thus modeling the behavior for the victims to emulate; (3) the use of drugs may be viewed as the only means of escape from an intolerable situation at home; and (4) the loss of self-esteem that is incurred over years of being abused creates an increased addictive vulnerability in victims, as they search for quick ways to produce good feelings and to obtain a crowd of associates who will validate them and do the same. These background factors produce such beliefs as "Using drugs Managing Life Problems 197 is a natural way of life," "Violence-particularly when under the influence-is acceptable," "Using is the only way of blotting out unpleasant feelings," and "I am weak and inferior and can feel good (and accepted) only by using." In treating such patients, self-esteem issues become a central component to the therapy. Therapists can help their patients begin to develop a stronger internal deterrent to drag abuse if they find (and work to build) evidence that dispels the patients' notions that they are worthless, deserving of a froubled life, and incapable of meeting their own needs through more socially acceptable and healthier means. SOCIOECONOMIC PROBLEMS Substance abuse disorders frequently occur against the backdrop of socioeconomic problems. For example, in the last decade we have witnessed a significant increase in the prevalence of cocaine abuse among the impoverished of our inner cities (Closser, 1992; O'Brien et al., 1992). Crack cocaine in particular has become more and more frequently abused within this milieu (Gawin & EUinwood, 1988; Smart, 1991). As the people of this segment of society grow more and more disillusioned with their chances of improving their lots in life through standard long-term means (e.g., finding quality educa­ tion, staying in school, and working one's way up a ladder of voca­ tional success), they may be more inclined to turn to nonstandard, short-term fixes (e.g., using drags for a boost in mood and selling drags for an increase in income). It has been argued that it is quite difficult to steer children, ado­ lescents, and adults alike away from drags when so few alternatives for finding enjoyment exist in the immediate environment. Addition­ ally, it is equally difficult to stay focused on the distant rewards of schooling, or the modest remuneration of low-level employment, when the distribution of drugs can produce staggering material rewards in a relatively short space of time. Furthermore, the risks of arrest and incarceration, overdose and poisoning, and violence are less of a deterrent against using drugs when there is a prevailing sense that the future holds no promise anyway. This is especially so for vulnerable individuals who have developed core beliefs of hopeless­ ness and helplessness. For these reasons and others, therapy with lower socioeconomic status substance abusers poses considerable challenges to the thera­ pist. Walter typified this type of patient. When the therapist made an attempt to appeal to Walter's sense of pride and self-worth (in order 198 COGNTTIVE THERAPY OF SUBSTANCE ABUSE to counteract his belief that he was a helpless victim) by asking could be "strong enough to be a man and walk away from his deal­ ing friends," Walter replied, "I don't feel strong when I walk away from drags. I feel like nothing, like I always do. I only feel like I'm worth something when I use a little something [usually cocaine or heroin]. That's the only time I feel like a big shot." In order to com­ bat this drag-abuse-fostering attitude, the therapist focused Walter's attention on how he felt after he came down off his high, which was lower, weaker, more helpless, and less adequate than ever. This fact was reviewed time and time again, so as to deromanticize the false sense of pride that the use of drags seemed to induce temporarily. Another critical component of treatment in this instance was to help Walter to recognize things in his life in which he could take pride while in a drug-free and alcohol-free state of mind. These included his skill as a longshoreman, his physical strength, and the love and pride of his children. Still, Walter had difficulty focusing on these factors, especially when he felt frustrated in reaching his goals. At times of despair and anger, he would lapse into all-or-none thinking, believing that he had nothing worth living for. Walter liked to make a point of this by quoting Bob Dylan's immortal line, "When you ain't got nothing, you ain't got nothing to lose." This saying activated the belief that "I have nothing and I am nothing," and fos­ tered his tendency to turn to drags when he would suffer a disap­ pointment. It therefore became critical to (1) discuss the precious things he did indeed have, such as the love and respect of his chil­ dren, and (2) how he would lose them if he were to indulge in drug abuse. It should be noted that one of the most important opportuni­ ties for producing changes in basic beliefs occurs when those beliefs have been activated, as by a disappointment or frustration. On the other end of the socioeconomic spectram we see afflu­ ent patients who have become substance abusers. Some common rationales proffered by more financially successfully substance abus­ ers resemble those of lower socioeconomic status patients, revolving around core beliefs of being ineffective or socially undesirable. One patient stated that he began his cocaine use at a time (ten years ear­ lier) when it was still considered "cool" to get high, and when cocaine was viewed as a drug of particularly high status amongst those on the fast track. He noted that tiie drag accentuated his feelings of power, attractiveness, and invulnerability-feelings that were not only pleas­ ing to him but also congraent with his view of what a corporate vice- president should be. Although he started as a social user, his contin­ ued usage led to physiological and psychological tolerance, which necessitated that he increase the quantity and frequency of his usage. As a result of his financial affluence he was able to continue his habit Managing Life Problems 199 unabated. In this sense, his wealth served as a contributoty causa factor in the development of his full-blown cocaine dependence. Ultimately, his performance at work deteriorated and he was unceremoniously fired. Following this event, he was unable to attain another position at the same level. The patient hypothesized that he had been blackballed within the local corporate community. At this point, he found himself in financial crisis, as he had not diminished his lavish life-style or cocaine usage, although he had ceased to have an income. As a result, he fell into a state of considerable crisis, with a serious cocaine dependence, and faced the threat of bankruptcy. It was only at this advanced stage of life difficulties and cocaine abuse that he sought therapy. At this time, he has succeeded in remaining free of cocaine for a number of months, although he is still moder­ ately depressed about his significant loss of status, money, and friends. Therapy continues to focus on (1) problem-solving with regard to earning

a livelihood, (2) dealing with continuing cravings for cocaine (although they have diminished somewhat over time), and (3) rebuild­ ing his self-esteem, which deflated when he lost his high-profile lifestyle and had to look at himself without the masked feelings pro­ vided by cocaine. Another high socioeconomic status patient reasoned that he began using cocaine in order to increase his energy and confidence. He came to believe that he would be unable to perform his job without the "boost" that he received from using the drug. He also enjoyed the temporaty sense of a release of pressure, and therefore was convinced that cocaine was a stress reducer. Predictably, as his habit grew, his perceptions of the quality of his work became distorted and inflated way beyond reality. He became less productive, and more and more interpersonally aversive due to his advanced symptoms of paranoia and impulsive anger outbursts. In similar fashion to the patient men­ tioned previously, he lost his job and incurred financial debts. Addi­ tionally, he was shunned by his friends and lover after he repeatedly rejected their suggestions that he receive help for his problems. A factor that both high and low socioeconomic status drug-abusing patients have in common is a vulnerability to peer pressure. Lower socio­ economic status patients are frequently confronted by friends and associates who urge them to use, share, and sell drugs. Such patients may come to believe that they will be deprived of meaningful social contact if they avoid evety substance abuser they know. As we wit­ nessed in the case of Dee, the razzing that her brothers gave her for not joining them in their smoking binges provided a strong tempta­ tion for her to use again. In this case, it was vital to help her to seek and develop contacts and friendships with people who were abstinent from drags and alcohol. 200 COGNTTIVE THERAPY OF SUBSTANCE ABUSE Other lower socioeconomic status patients may feel pressured by their associates to prove that they are "one of the gang," or that they have the "guts" to use drags heavily. Walter told his therapist that he was challenged to shoot up a powerful combination of cocaine and heroin by a couple of guys he knew at a junkyard where he often hung out. When the therapist suggested some typical "middle-class liberal" (patient's description) assertive comments that he could make in response to their challenge, Walter replied, "You can't talk that kind of intellectual crap in m y neighborhood. Maybe it works in your neighborhood, but not in mine." With this, the therapist adjusted the intervention by putting the responsibility for generating "proud, drag- free comebacks" onto Walter's shoulders. This episode highlights the role that socioeconomic characteristics play in conceptualizing prob­ lems and designing interventions. The peer pressure may take on a different form among the upper socioeconomic status patients, but it exists nonetheless. For example, a patient at an exclusive party may become convinced that he must join in with the "recreational" cocaine users in order to be consid­ ered one of the "beautiful people." Fortunately, this distorted, over- romanticized view of the wealthy cocaine user has more recently fallen into disfavor, as the extent and ramifications of the nation's cocaine problems have come to light since the latter half of the 1980s. In another vein, the more affluent cocaine user may be motivated by a need to gain even more money, more power, and more success as a way to gain acceptance by the perceived upper echelon in his or her profession or community. In this drive to succeed, the patient may resort to a chemical stimulant such as cocaine in order to give him or her a sense of increased productivity and sociability. When peer pressure (in any form) is a factor that contributes to substance abuse, it becomes necessaty to address the patient's exces­ sive need for social acceptance. In this regard, self-esteem issues come to the fore. Furthermore, therapist and patient must work together to help the patient learn more adaptive social problem-solving skills, so that drags are no longer considered part of the "solution." DAILY STRESSORS Mundane problems or stressors can serve as triggers for patients' drug use. The daily hassles that people face, such as working at a job that entails pressure and deadlines, or one in which they feel trapped for financial reasons, or dealing with two or three unraly toddlers on a continual basis without a respite, can create feel- Managing Life Problems 201 ings that exceed a person's tolerance for discomfort. An accumula­ tion of such stressful events and feelings, experienced day after day, can serve as factors that encourage a person to resort to drags in order to "get through the day." "Dee" was a patient for w h o m the management of evetyday con­ cerns was central to her maintaining abstinence. An important thera­ peutic priority for Dee entailed helping her cope with taking care of her infant son. As a single parent. Dee was confronted with a great deal of change in her daily life when her son was born. She had been happy working long hours as an assistant in a nursing home; now she had to curtail her hours considerably. She had also been used to coming and going as she pleased, as her older children were adoles­ cents who could be trasted to take care of themselves in their mother's absence. Now, Dee's freedom of movement was restricted, and she had to plan her activities with her baby's care in mind. In addition. Dee was vety troubled and annoyed by her son's nocturnal ctying spells. In sum, the evetyday demands of being responsible for the well- being of an infant reinforced in Dee a sense of inadequacy (as a mother), helplessness (in taking care of all her responsibilities), and hopelessness (in ever enjoying life again). These core beliefs were sources of marked anger, frustration, and sadness, the likes of which touched off thoughts and cravings to smoke crack as her "only" source of pleasure and "freedom." It was clear that keeping Dee away from the lure of drugs would necessitate lessons in coping with her new life as the primaty caregiver for a helpless child. Among other therapeutic strategies, problem- solving and the planning of daily activities (working a few hours, arranging baby-sitting for her child, attending a support group meet­ ing, spending time with her baby, reading, etc.) were reviewed and employed. In addition, much time was spent in using Daily Thought Records to counteract Dee's thoughts that suggested to her that her life would now be intolerably restricted. Indeed, she was encouraged to consider rational responses that focused on various potential bene­ fits that she would now receive as a result of her newborn's arrival. For example, she would have the opportunity to disprove all the crit­ ics who told her that she would not be able to raise this child due to her drag habit. Naturally, this rational response was a powerful posi­ tive motivator to engage in the process of cognitive therapy; however, it also gave rise to another automatic thought. Specifically, Dee wor­ ried that a lapse into drag use would lead to evetyone's giving up on her, judging her negatively, telling her "I told you so," and taking her baby away from her. In order to safeguard against this possible eventuality. Dee and 202 COGNITIVE THERAPY OF SUBSTANCE ABUSE her therapist worked on strengthening her reactions to her areas o greatest vulnerability. Specifically, Dee was taught how to better deal with her baby's ctying fits (which perturbed her vety much). In ses­ sion. Dee was instracted to close her eyes and to imagine being at home at a given time when her baby was ctying without letting up. When Dee indicated that she could indeed "hear" her baby Ctying, she was encouraged to express her automatic thoughts. Given Dee's aptitude for vivid imaging, her automatic thoughts traly represented her "hot" (i.e., affectively charged) cognitions. These included "I can't stand it anymore. I have to get out of here," and "I'm going to go out and pick up again [use drugs again], and then m y aunt will have to come and take care of m y son. Then I'll be free again." After recit­ ing these thoughts. Dee was then instracted to imagine what her life would be like in the aftermath of these events. With this, she would become remorseful, and would begin to express rational responses to override the hot cognitions elicited above. As the exercise continued. Dee spontaneously generated more automatic thoughts that further explained her negative emotionality in response to her baby's ctying. She thought, "My baby doesn't love me. M y baby doesn't want to be with me. He'd rather be with the sitter. He'd rather be with m y aunt." These thoughts fed into Dee's sense of rejection. In response. Dee was encouraged to consider alternative meanings to her son's ctying, such as "He's ttying to tell me he's uncomfortable—that he's hungty, or cold, or needing to be changed," and "He's just being a baby. He's not telling me he doesn't love me. He doesn't even understand stuff like that yet. He's just being a baby." Still keeping her eyes closed. Dee was then told to imagine the self-satisfaction she would feel if she resisted the temptation to use drags, and instead proved to herself and others that she was capable of taking care of this child. In addition. Dee was assisted in imagin­ ing the distant future, when her then grown-up son would love her and think highly of her, because she was always there for him and because he saw her as a positive role model. These techniques, repeated in different variations over the course of a number of ses­ sions, were highly efficacious in helping Dee to see herself as being a person worthy of love, and effective in dealing with the stressors of life. As her core beliefs of helplessness and unlovability diminished. Dee became more consistent in remaining abstinent from drugs and alcohol. In turn, her mothering skills improved. Stressors of daily living can include any events that frigger a sense of frustration, anxiety, anger, fatigue, and loneliness. When such events occur frequently, or involve chronic conditions such as living Managing Life Problems 203 in a noxious home environment or in an undesirable neighborhood, then the therapist needs to help drug-abusing patients to anticipate and adaptively respond to their maladaptive reactions to these situa­ tions. In this way, patients may learn to become adept at coping with (and solving) the commonplace triggers that put them at daily risk for substance abuse. LEGAL PROBLEMS Where there is illicit substance abuse, there is the threat of criminal apprehension. This is so not only because of the patient's drag use itself, but because drag use often leads to additional illegal activities as well. For example. Dee was first arrested when she was caught ttying to cash forged checks. Walter's criminal convic­ tion resulted from his hijacking and fencing stolen goods from the docks where he worked. "Charleen" was apprehended when she sold crack to an undercover officer. When a patient enters treatment as a condition of parole or pro­ bation, the therapist needs to be aware not only of the difficulties that the patient will have in trying to stay away from drugs, but also of the difficulties that the patient will have in coping with the constraints and stigma associated with being in legal limbo. Charleen often expressed her consternation at having to be monitored by both the therapist and the parole office. For her, almost evety therapy session and urinalysis served as a reminder that her life was not entirely her own. As she dwelled on all the drawbacks of this situation (e.g., loss of freedom), her anger grew, and with it her defiant desires to renew her drug use as well. Figure 12.2 represents one of the Daily Thought Records that Charleen worked on in order to quell the anger and cravings associated with such negative thoughts about her parolee status. Until she used the Daily Thought Record, Charieen had never considered the advantages that her legal and psychological monitor­ ing provided for her-namely, added support and incentive

to remain drag free. Walter also experienced a significant degree of dysphoria in reaction to automatic thoughts about his legal statiis. In particular, he was frastrated and angered by the limitations set on his traveling by the terms of his parole. He believed that most of his problems had to do with his family members and his neighborhood acquaintances who abused drags. He rationalized (not to be conhised with rationally responding!) that he would stand a better chance of staying away from drags if he could move upstate to a more raral area. In fact, he applied Directions: W h e n you notice your mood getting worse, ask yourself, 'What's going through my mind right now?" and as soon as possible jot down the thought or mental image in the Automatic Thought Column. SITUATION AUTOMATIC THOUGHT(S) EMOTION(S) RATIONAL RESPONSE OUTCOME Describe: 1 Write automatic thought(s) 1. Specify sad. 1. Write rational response to automatic thought(s). 1. Re-rate belief 1. Actual event leading to that preceded emotion(s). anxious/ 2. Rate belief in rational response 0-100%. in automatic unpleasant emotion, or 2 Rate belief in automatic angry, etc. thought(s) 2. Stream of thoughts, day­ thought(s) 0-100%. 2. Rate degree 0-100% dreams or recollection, of emotion 2. Specify and leading to unpleasant 0-100%. rate emotion, or subsequent DATE/ 3. Distressing physical emotions TIME sensations 0-100%. I'm on the bus going 1 I hate this, I hate 1. Hate I have to be watched like this so I won't 1. Hate to my therapy session. this, I hate this! (100%) go back on drugs again and mess up my life (50%) I have to go. I don't 2 Why do I have to be 2. Anger again. (80%) 2. Anger have a choice or I'll watched like this? (100%) I may hate it, but it's better than being (50%) mess up my parole and a crackhead and going to jail. (100%) I'll have to go back 3 I hate being monitored. 3. Frustration 3. Frustration to jail. I hate ruining my day (100%) Being monitored makes me more careful to (80%) to go see my parole stay off drugs, so I guess it's helping 4. Calmed officer and my ther­ me. (60%) down 4 apist. It makes me mad. My parole officer and my therapist want me (50%) I should be allowed to to be OK. I guess they care about me, even 5 live my own life. if they annoy me sometimes. (100% I'm being treated like I guess these meetings could have a good a criminal and a baby. reason. (50% 6 I want my freedom back I'll get my freedom soon enough! (100% now! ! (100% for all thoughts) Questions to help formulate the rational response: (1) What is the evidence that the automatic thought is true? Not true? (2) is there an alternative explanation? (3) What's the worst that could happen? Could I live through it? What's the best that could happen? What's the mosf realistic outcome? (4) What should I do about it? (5) What's the effect of my believing the automatic thought? What could be the effect of changing my thinking? (6) If was in this situation and had this thought, what would I tell him/her? (friend's name) F I G U R E 12.2. Cfiarleen's Daily Tfiought Record. Managing Life Problems 205 to the parole office for a change of venue, but was told that he wo have to wait a considerable length of time before this could be approved. Walter reacted with a sense of hopelessness, along with cynicism and anger. Now, every time he came into contact with the "problem people" he alluded to above, Walter would think that he was now justified in using drags. His rationalization was, "Hey, I tried to get away from these people, but the cops wouldn't let me. If I fuck up, it's on their heads." Naturally, the therapist in this case spent con­ siderable time in getting Walter to modify his views on this matter, and to find alternative methods for steering clear of these "people and places." Another legal concern that drug-abusing patients sometimes face is loss of custody of their children, either to an ex-spouse, a relative, or to child protection services. Dee was one such patient. Her aunt in Baltimore had taken care of Dee's baby son for two months while Dee went through an inpatient drag abuse rehabilitation program. Dee knew that her aunt was waiting in the wings to assume custody of the child if Dee were to relapse. In fact. Dee's aunt had stated that she would seek to become the child's legal guardian if Dee were to demonstrate that she were an unfit mother. Rather than viewing her aunt as a safety net, she saw her as a threat, and often got herself agitated over the possibility that she would lose her baby. Such a perception produced added pressure on Dee, which exacerbated her urges to resume her use of crack cocaine. In this case. Dee was confronted with the vety real possibility of losing custody of her child. The therapist aimed to help Dee to view this as a motivator to stay straight, rather than as a sword of Damocles hanging over her head. Furthermore, rather than suspiciously treat­ ing her aunt as an evil adversaty. Dee was taught to cooperate and work with the aunt. It was noted that they both shared a common concern—namely, the welfare of the baby. It should also be noted that therapists who treat drag-abusing populations are confronted with more legal and ethical questions than are those who freat most other groups. For example, the association between drag use and domestic violence means that therapists have to be especially vigilant to situations that would require a duty to warn an intended victim, or a duty to contact agencies that monitor child or spouse abuse cases. Managing such situations, while also maintain­ ing a constructive therapeutic relationship with the patient, requires considerable skill indeed. Another legal-ethical situation that we have encountered occurs when a patient attends a therapy session in an intoxicated state. Whether or not the therapist chooses to continue with the session is 206 COGNITIVE THERAPY OF SUBSTANCE ABUSE subject to personal choice and/or case-by-case decision-making. Ho ever, if the therapist chooses to interact with the patient, it is impera­ tive that the therapist assess whether the patient has driven a vehicle to the therapist's office. If the patient has in fact driven to session, the therapist should make evety effort to keep the patient off the highway until such time as he or she is sober. In one case, we required a patient to remain in the waiting room for two hours before permit­ ting him to leave the Center. Other alternatives include calling a cab, using public transportation, or contacting a relative who could give the patient a lift. MEDICAL PROBLEMS Drug abuse is associated with a myriad of chronic medical conditions (O'Connor, Chang, & Shi, 1992). Even when the patient has given up drugs altogether, the medical consequences of the abuse may linger indefinitely, causing pain, worry, hopelessness, and renewed urges to "self-medicate" via the use of illicit drags. This phenomenon is well illustrated by a scene in the film Bird, the story of the legendary jazz saxophonist Charlie Parker. Parker, whose hard- drinking, hard-dragging, and hard-driving lifestyle led to the devel­ opment of excruciating gastric ulcers, used heroin to deaden the pain. This habit only served to hasten his death at the age of 34. In one scene, he tearfully tells his drug-free friend, Dizzy Gillespie, "Ain't it a bitch ... I go to the doctor and pay him $75, and it don't help m e . . . but 1 go to some 'cat' and pay him $10 for a bag of shit [heroin], and m y ulcers don't hurt, m y liver don't hurt, m y heart troubles is gone . . . and you tryin' to tell m e that this is the man I'm supposed to stay away from? Mr. Gillespie, m y comrade in arms, that is what I call a paradox." Therapists need to be aware of the medical issues involved in substance abuse, so as to be able to educate the patient about such matters, as well as to be able to detect physical signs that indicate that the patient may have relapsed. A patient who is actively abusing hard drags such as cocaine, heroin, or amphetamines, may be a malnourished and sleep-disturbed patient as well. Charieen had the misconception that the use of crack cocaine produced a beneficial weight loss. She lamented her weight gain during her recovery and treatment period, and talked about "the good old days" when she was using and was thin. The therapist taught her that her weight loss reflected a dysfunctional dependence on a drug that in no way substituted for nutrition. He emphasized that Managing Life Problems 207 weight loss needed to be achieved through exercise and sensible ea ing habits, not through drag-induced loss of appetite. It is common for patients who go on drag binges to be awake all night (while they are using), and to sleep most of the day. Another variation of sleep disturbance involves the patient's remaining awake for days at a time, and then crashing for a number of days. In either case, the body's natural sleep cycle has been badly disrupted, and normal vocational and social functioning ceases to be possible. It is noteworthy when a patient misses an afternoon therapy ses­ sion due to "oversleeping," or sounds groggy on the phone in the middle of the day. These are some of the telltale signs of substance abuse (Gawin & Kleber, 1988). In fact, a number of high-profile ath­ letes who were discovered to be cocaine abusers were first suspected as a result of oversleeping for important team practices, meetings, or travel obligations. Whenever practically feasible, we like to schedule our most dif­ ficult patients for the earliest appointments. A 9:00 a.m. appointment minimizes the likelihood that a patient will be able to mask the fact he or she used drugs the day before. At times when an early therapy session cannot be arranged, early-morning phone checks can be uti­ lized. For example, a patient may be instructed (as part of his home­ work assignment) to call the therapist each day at a specified midmoming time, just for a minute. When a patient complains that a session or a telephone call is too early to comply with, we respond by saying, "If you can't attend a therapy meeting at 9:00 a.m., will you be able to keep a job that requires you to be there at 9:00 a.m.?" In the same way that drag-abusing patients are often less than maximally compliant with therapy, they too are frequently noncom- pliant with medical advice. W h e n the patient has a steady physician, it is important to obtain the necessaty releases so as to have ongoing contact with the doctor. When the patient is on public assistance, and therefore may not see the same doctor with each medical visit, the situation is a bit more convoluted. Nevertheless, it is important to make evety effort to keep apprised of the patient's medical conditions and treatments. One of our most stubborn cases of medical noncompliance was "Ray," who suffered from a serious case of diabetes and was insulin- dependent. His maladaptive approach to his condition was a continual cause for concern in treatment. During one period in the course of therapy, he refused to inject himself with his required daily dosage of insulin, saying that he could control his blood-glucose level sim­ ply by watching his diet carefully. W h e n this was questioned, Ray added that the use of a needle (to inject insulin) would be a power- 208 COGNTTIVE THERAPY OF SUBSTANCE ABUSE ful inducement for him to shoot up "something a little harder than insulin." Needless to say, this posed quite a dilemma. To make matters worse, Ray refused to consult with his physi­ cian on this matter. The therapist attempted to appeal to the patient's sense of reason, and expressed a great deal of personal concern for Ray's life, all to no avail.

It was only after Ray began to experience significant physical malaise that he finally relented and contacted his doctor. Once there was open communication between patient, therapist, and physician, the next phase of treatment could begin. This entailed having Ray bring his insulin and needles to session, so that intensive rational responding could be practiced in the presence of the stimuli that elicited his most severe hot cognitions and cravings. Although this intervention involved a degree of risk for relapse, it was deemed by all parties to be less risky than the possibility of diabetic coma. Before long, Ray was successfully administering his own insulin shots, albeit sporadically. Ray's insulin compliance would need to be checked throughout the course of cognitive therapy. One of the most serious medical problems associated with the abuse of cocaine is the increase in prevalence of expectant mothers who are regular users. W e have witnessed a rise in the number of cases where babies are born prematurely, with life-threatening complica­ tions, and with physiological addictions to cocaine, as a result of their mother's habits during pregnancy (Closser, 1992; Grossman & Schottenfeld, 1992; Smart, 1991; Stimmel, 1991). (Additionally, there is recent preliminary evidence that traces of cocaine bind to sperm, thus also implicating fathers in transmitting the harmful effects of cocaine to their unborn children.) Again, education plays an important role in treatment. The thera­ pist can explain to pregnant patients the risks of drag use (and alco­ hol use, as well as cigarette smoking) to their unborn children. Such patients also need to be encouraged to receive regular prenatal medi­ cal attention. Imagety work can help these patients focus on mental pictures of healthy babies versus those who are seriously ill, so as to help sway the patients away from using drugs when cravings arise. One of our pregnant patients learned to recite a standard rational response to herself when she had an urge to go out and purchase crack cocaine. She would imagine a healthy, happy baby, and she would say to herself, "I'm going to start being a good mother right now. I'm going to start taking care of m y baby right now. I'm going to choose health for both of us. I'm not going to go out and use." Another medical problem related to substance abuse concerns communicable diseases, either by sexual contact or via the sharing Managing Life Problems 209 of needles. Substance abusers are notoriously reckless in their se behavior (e.g.. Goldsmith, 1988; Watkins, Metzger, Woody, & McLel­ lan, 1991). For example, they are prone to ignore simple measures of protection, such as the use of condoms. Remember (as stated in Chap­ ter 4, this volume) that Walt stated flippantly that it was against his religion to wear a condom. In extreme cases, people who are addicted to hard-core drugs will trade sex for drags. This translates to many people having indiscriminate sexual relations with many other people over time. The risks of contracting a wide range of venereal diseases, hepatitis, and the AIDS viras are substantially increased within this population (Chiasson et al., 1989; Fullilove, Fullilove, Bowser, & Gross, 1990; Stimmel, 1991). Given the fact that many of these people are intravenous drug users who use and share unsterile needles (Metzger et al., 1991), the risk of acquiring AIDS escalates even more. Once again, education is a crucial component of therapy. Three steps toward the goal of safety from transmittable diseases include (1) abstinence from intravenous drug use, (2) practicing safe sex, and (3) remaining monogamous with a trasted partner. W e have found that patients are fairly compliant with point (1), considerably less com­ pliant with point (2), and very rarely compliant with point (3), espe­ cially if they continue to abuse drags. It is extremely difficult to help a patient who is actively abusing substances to monitor his or her sexual behaviors. It is far more realistic to get patients to be sexually safer once they become abstinent from drug use. Nevertheless, every effort must be made to address this medical issue, whether or not the patient is using (see Appendix, page 329, for more information on managing patients who are HIV positive, as adapted from Fishman, 1992). Finally, a most basic and obvious area of medical complication associated with drug use involves the deleterious effects of the drugs themselves (Frances & Miller, 1991). Dangers include damage to the central nervous system, cardiac and respiratoty abnormalities, liver atrophy, and death by overdose. Yet another hazard is encountered when the drag is adulterated by foreign substances. In one notorious case in the Philadelphia area, three men were killed when they smoked the poison-laced crack that was smuggled into their prison cells. SUMMARY In this chapter we presented an overview of the life problems that typically plague substance-abusing populations. W e noted the vicious cycles that occur when such life problems trigger 210 COGNITIVE THERAPY OF SUBSTANCE ABUSE the abuse of substances, which in turn exacerbates the patients' ne tive life situations. C o m m o n problem areas include family and relationship dysfunc­ tion and discord, socioeconomic hardships, chronic, cumulative daily stressors at home and at work, difficulties associated with legal apprehension, and medical conditions and complications. W e dis­ cussed the roles that these problems play in the course of cognitive therapy, and provided case illustrations that highlight methods of addressing such life crises in the context of a comprehensive out­ patient treatment for substance abuse. W e would like to add that we have been strack by the way that some patients are able to make broad-sweeping changes in their lives as a result of their steadfast commitment to the treatment regimen. When positive life changes follow the patient's success in achieving and maintaining a drug-free existence, it behooves the therapist to make certain that the patient understands the nature of this positive feedback loop. In the same way that the therapist teaches the patient to recognize the dangers of the vicious cycles that drag abuse fosters, the therapist reminds the patient to take stock of the beneficial changes that take place as a result of abstinence. When patients fully come to realize how much they can improve their overall life situations, they gain even more motivation to make the most out of therapy and to work at preventing relapse. C H A P T E R 1 3 C r i s i s I n t e r v e n t i o n A crisis is an unplanned, sudden, and often un- desired change in a person's life that typically is associated with emotional distress. The Chinese symbol for "crisis" consists of the combined characters for "danger" and "opportunity." Similarly, in the cognitive therapy of substance abuse, a crisis is viewed as a stressor that simultaneously presents a danger of relapse and an opportunity for learning. To the extent that patients successfully cope with a cri­ sis without using drags, they increase self-confidence and furare cop­ ing skills. The individual who does not cope effectively with crises decreases self-confidence, resulting in weakened coping skills and increased potential for future relapse. Being available to patients in times of crisis is one of the therapist's most important responsibilities. For example, when a patient is acutely suicidal or otherwise in a potentially harmful situ­ ation, it is vital that the therapist be accessible for emergency con­ sultation. Furthermore, it is advisable that the therapist be highly skilled in managing such crises. Patients who abuse drags are espe­ cially prone to get into serious trouble, and a tĵ ical course of therapy with such patients often involves having to handle a number of criti­ cal incidents (Newman & Wright, in press). For example, it is not uncommon to see a patient who seems to be making therapeutic progress, only to have that patient suddenly stop showing up for therapy sessions, and fail to return the therapist's telephone calls. Sometime later, the patient calls the therapist in a highly agitated state, and it becomes clear that the patient has begun to abuse drags once again. Typical crises involve (but are not limited to) renewed drug use, breakups and/or violence in significant relationships, loss of employ- 211 212 COGNITIVE THERAPY OF SUBSTANCE ABUSE ment, depletion of finances, recurring legal difficulties, crimina involvement, medical emergencies, and overdose and suicidality. This chapter serves as a blueprint for the therapist in the han­ dling of the acute crises of drag-abusing patients. W e review and dis­ cuss the most common types of critical incidents, and we explicate some of the therapist's options in dealing with these difficult clinical situations. First, we examine some of the warning signs that should alert the therapist to the possibility of imminent or current crisis in the patient's life. W A R N I N G SIGNS Even when patients are not voluntarily forthcoming about serious difficulties that they are encountering, there are a num­ ber of common telltale signs that indicate that they may be in crisis. Such "crises" may or may not entail drag use. However, more often than not, a crisis that starts out as a non-drug-related event, left unmanaged, turns into an episode of drug use, thus compounding the cri­ sis (Kosten et al., 1986). One common sign is that the patient is habitually late for therapy sessions, or does not show up at all. When a patient misses a session, especially when he or she has not called to cancel officially, it often spells trouble. W e have witnessed many instances when patients not only missed sessions, but literally seemed to disappear for long stretches at a time. When the patient is incommunicado in this man­ ner, the chances are high that he or she is actively using drags, or is embroiled in other serious difficulties. After a patient has failed to show up for a session, the therapist should attempt to contact that patient as soon as possible. If the patient is easily reached by telephone, the therapist can ask the patient directly what is happening (e.g., "Do you know why I'm calling? Did you know that we were scheduled to meet a half hour ago? Can you tell me what's going on right now?"). A more problematic scenario occurs when the patient cannot be reached following a missed session. This scenario is typified by the patient's (1) telephone ringing at all hours of the day and night with­ out being picked up, (2) not returning the therapist's telephone mes­ sages, and (3) relatives or housemates sounding hostile, giving cryp­ tic answers in response to queries about the patient's whereabouts, or giving lip service to their willingness to pass on messages to a patient who ultimately does not return the call. Crisis Intervention 213 In a high percentage of cases in which the patient drops out of sight for more than a few days at a time, drug use has been involved. As such, missed sessions provide the therapist with a conspicuous red flag. Another warning sign that the patient may be in crisis is the patient's demonstrating a marked change in mood or behavior. Examples include (1) a patient who ordinarily speaks clearly and articulates his thoughts well sounds oddly incoherent or otherwise cognitively disorganized during a telephone contact; (2) a patient who almost routinely scores zero (i.e., no self-reported pathology) on such questionnaires as the Beck Depression Inventoty suddenly endorses the most extreme symptom items (e.g., "I would like to kill myself"); (3) a patient who has generally been cooperative and amicable evi­ dences hostility toward the therapist; (4) a patient evidences labile affect in session, such as shifts between agitation, ctying, and "silli­ ness"; and (5) a relative of the patient telephones to say that the patient is acting "out of control" and pleads for the therapist to intervene immediately. Although these examples are the most typical, they by no means represent an exhaustive list. Whenever possible, pronounced changes in patients' functioning should be addressed as soon as possible. A skilled mixture of accurate empathy and frank confrontation is called for in such instances. In this manner, the therapist may strengthen the therapeutic relation­ ship while also making strides to stabilize the patient and begin to deal with the sources of the

crisis in a constractive manner. Yet another red flag for the therapist to notice and address is the patient's sounding or looking abnormally groggy, on the telephone or in person, especially during the middle of the day. One of our patients used the excuse that she had worked "the late shift" the day before a session in which she appeared extremely fatigued and sub­ dued. The excuse seemed plausible at first but began to lose credibil­ ity when she began to miss sessions and then answered the therapist's telephone calls in similar states of retardation, regardless of the time of day that she was called. Furthermore, when asked to schedule her sessions in such a manner that she would not have to arrive after working "the late shift," this patient was unable to produce a work schedule of any sort. At this point, the therapist told her point blank that he believed that her drowsiness was drag or alcohol related, and that this was a vety serious problem indeed. Patients rarely are eager and enthusiastic to report on renewed drug episodes or their concomitant crises, but they may be more willing to discuss "close calls." W h e n a patient spontaneously reports 214 COGNITIVE THERAPY OF SUBSTANCE ABUSE that he "almost used drags this past week", it is a sure-fire bet he will have at least another close call this week. More than likely, he has actually already used drugs during the past week, and will probably use even more drags in the coming week. When this topic comes to light in session, the therapist must be alert to make it the number-one priority item for the session. The rest of the session must involve a concentrated effort to plan an emergency strategy to ward off drag use once the patient leaves the office. When the patient does not mention near-miss drag episodes on his or her own, the therapist may choose occasionally to ask, "Have there been any times this past week when you were tempted to use drugs? What's the closest that you came to using?" Such questioning may lead to the discussion of problems and potential crises that the patient may not have volunteered to discuss. When therapists are aware of the warning signs of renewed drug use and other crises, they stand a much better chance of keeping their drag-abusing patients in treatment. In doing so, they may also suc­ ceed in nipping major problems in the bud, and dealing with full­ blown emergencies before they lead to harm or incarceration (Newman & Wright, in press). CRISIS SITUATIONS In the sections that follow, we discuss some common crises that are often directly related to the use of alcohol and illicit drugs. These crises include overdose and suicidality, loss of domicile, disappearance, loss of employment, loss of close personal relation­ ships, medical emergencies, criminal involvement, and violent con­ frontation with the therapist. Overdose and Suicidality Whether by accident or in a deliberate attempt to harm one's self, an overdose of illicit drags (perhaps in combination with alcohol or legal controlled substances) represents a potentially life-threatening situation. In many instances, the therapist of a patient who overdoses will not know about the episode until after the damage is done. For example. Dee called her therapist from her hospital bed, saying that she would not be able to keep her therapy appointment due to her hospitalization for a severe asthma attack. Later, when the therapist consulted with the physician on her case, it was learned that her attack Crisis Intervention 2 1 5 was induced by cocaine intoxication, and that Dee had nearly asphyx ated. On rarer occasions, a patient will call the therapist to inform him or her of a drug overdose. W h e n the overdose is an accident, the patient will likely be in a state of panic and confusion. An attempt should be made to confirm the patient's location, and to explain that the therapist will be calling "911" in order to have an ambulance sent. If a friend or relative is on hand, that person may be asked to do the same. In cases in which the patient is alone and is too incoherent to provide an address, the therapist will need to have handy such basic information about the patient. A simple solution is for the therapist to keep copies of all patient telephone numbers and addresses both at work and at home. When the overdose is a suicidal gesture or attempt, the same need for an ambulance exists. (Needless to say, under such conditions the therapist does not have an obligation to protect the patient's anonym­ ity; therefore it is appropriate to break confidentiality in order to save the patient's life.) However, if possible, it is important to keep the patient on the line. This is particularly necessaty when the patient is unwilling to reveal his or her whereabouts. When Ray called his thera­ pist at home (and reversed the charges) on Thanksgiving Day, he claimed that he had taken enough heroin and cocaine "to kill an elephant." He refused to say where he was, save to note that he was at a telephone booth. Luckily, the therapist had two telephone lines at home and was able to write a message to his wife, signaling her to call an operator from the other line in order to trace Ray's call. Although the therapist kept the patient on the line for 50 minutes, the trace was unsuccessful. The therapist proceeded to call the police directly, giving a description of the patient and recommending places that he could most likely be located. Fortunately, Ray's claim was exaggerated, and the "suicide attempt" merely made him sleep for most of the next 24 hours. When the suicide call comes directly to the therapist's office, it may be possible to contact a colleague in the same manner that the therapist in the above example contacted his spouse. In sum, the therapist should (1) determine the patient's location, (2) keep the patient on the telephone, or in the office as the case may be, (3) assess the degree of severity of the suicidal behaviors, and (4) contact the police, an ambulance, or another person in the vicinity who can do this for the therapist while the therapist gives the patient undivided attention. In the case of a suicidal emergency where drugs have not already been ingested, but rather the patient is making threats that have not 216 COGNITIVE THERAPY OF SUBSTANCE ABUSE yet been carried through, the therapist can afford to respond a bi cautiously and methodically. If a patient is feeling hopeless as a result of renewed drag use, or has suffered a loss as a result of the abuse of substances, the suicidal wishes may be reduced by attacking the hope­ lessness directly. If successful, the therapist may not have to contact the police or an ambulance. Instead, a face-to-face therapy appoint­ ment should be arranged as soon as possible (e.g., same day or first thing the next morning). The situation becomes a bit more cloudy when the patient is suicidal while under the influence of drags or alcohol. Even when the patient has not taken an overdose the intoxication still may result in an irrational exacerbation of the patient's intentions to self-harm by other means (Marzuk et al., 1992). Under these conditions, the patient may not be capable of following the therapist's instractions, or of understanding the therapist's attempts to help. If the therapist senses that this is the case, it is necessary to take the safest course of action. This may entail contacting and instracting a relative of the patient to admit the patient to hospital or the therapist's calling for an emergency vehicle directly. Loss of Domicile Another drug-related crisis entails the patient's being expelled from his or her household, usually by a spouse, parent, housemate, or landlord who has gone beyond the limits of toleration for the substance abuser's extremely maladaptive behaviors. In some instances the patient has an alternative place to go, such as to another relative or a friend. Unfortunately, unless the patient changes his or her behavior, the same sequence of events may recur. Ultimately, such patients may find themselves with no place to go but out on the streets, in shelters for the homeless, in drug houses, or (if they are lucky) an inpatient facility. When a therapist becomes aware that a patient has been forced out of his or her domicile, it is advisable that the therapist elicit per­ mission from the patient to consult with the patient's parole officer or with a social worker. This allows the therapist to improve the patient's chances of gaining admission to an inpatient drug abuse rehabilitation program. Although the inpatient setting may provide only a short-term solution to a long-term problem (Cummings, 1993), it is far better for patients to be receiving medical supervision than to be languishing on the streets. Furtherniore, the patient's participa­ tion in an inpatient program may help him or her to regain favor Crisis Intervention 217 with family members, as well as to reestablish a program of outpati cognitive therapy after discharge. Disappearance At times, crises emerge not when patients are forced to leave their homes, but instead when they suddenly disappear from their residences. As noted earlier, this usually indicates that the patient is on a drag binge, and may be taking shelter with others who are also "on a ran." This type of crisis is one of a variety about which patients do not contact their therapist. Although patients who "dis­ appear" to places such as crack houses (for days at a time) may be losing money, jobs, family ties, physical health, and other important facets of life, they typically do not realize their own states of crisis until after the ran is over. At this point, the patients are forced to face the devastation that they have wrought. They have to face their lack of money; their now accentuated drug urges; their poor physical hygiene and condition; their irate relatives, employers, and landlords; and possible legal consequences, especially if their parole officers track them down. When patients leave their homes to go out on a drag run they may be almost impossible to locate. Still, it is important for thera­ pists to continue to try to reach them by telephone or by mail, and to be prepared to resume treatment with them once they return (or once they are apprehended). Therapists of patients who reappear following their binges may feel particularly frusttated, disappointed, and angry at the patients over thqir self-defeating and antitherapeutic behaviors. A common auto­ matic thought we have shared in response to this very type of situa­ tion might translate to the following: "After all the work I've done with this patient, he has some nerve leaving therapy to go out on a binge! How dare he rain all the painstaking work we've done together, just for a temporary high. N o w he expects me to pick up all the pieces for him again!" Such automatic thinking needs to be counterbalanced by rational responses that highlight the patient's genuine state of des­ peration, the therapist's opportunity to continue to offer earnest pro­ fessional help even under conditions of adversity (thus, hopefully strengthening the therapeutic relationship), and the realization that the therapist does not have to "pick up all the pieces." It is signifi­ cant just that the therapist demonstrates a willingness to continue to work with the patient. However, much work will need to be done to minimize the risk of a repetition of such an episode. This includes 218 COGNITIVE THERAPY OF SUBSTANCE ABUSE emphasizing the importance of the patient's contacting the therapi on a regular schedule, and whenever cravings for drags are elevated. Loss of Employment Persons who are actively abusing drags rarely make good employees. Often, drag use leads to the patient's arriving late for work, missing days of work, doing inefficient work, and at times losing the job altogether. A patient who loses a job may react with anger, anxiety, frastration, hopelessness, accentuated reductions in self-esteem, and a wide range of other negative emotions. These feel­ ings, along with their concomitant automatic thoughts and beliefs, represent significant threats to the

patient's abstinence. For example, a patient who derived most of his self-worth from his job will prob­ ably react quite adversely to being laid off. He may then believe that he has nothing to lose (including his pride, which has already been damaged) by going out and getting "stoned." In this scenario, the patient might think "1 worked so hard to stay off drags, and look what it got me—nothing but trouble! If that's the way it's going to be, 1 might as well just go out and get messed up!" When a therapist learns that a drug-abusing patient has lost a job, swift and concerted effort is called for to help the patient solve this life-disrupting problem. Obtaining new legal sources of income becomes a high priority agenda item. Such sources include unemploy­ ment insurance, public assistance, or another job. Newly fired patients must not be given tacit permission to give up, to abdicate all finan­ cial responsibilities to their families, to anesthetize themselves with more drags and alcohol, and to seek money through illegal means. Naturally, the substance abuse problem per se will continue to be addressed; otherwise the patient's resultant poor work habits will result in further losses of jobs. If the patient states that he or she no longer can afford to attend therapy, the therapist would do well to show sympathy for this posi­ tion. However, the therapist should impress upon the patient the importance of attending at least one session in order to address this crisis in a productive way. The therapist may then wish to consider making special arrangements for the patient to continue with treat­ ment, such as a reduced fee (when possible) and/or reduced frequency of sessions. In the case of parole-office-referred patients, or research cases, unemployment should have no bearing on the patient's con­ tinuing with therapy, as the expense typically is covered by a third party. The two focal points of the session should involve (1) assess­ ing and modifying the negative feelings, thoughts, and drag urges that Crisis Intervention 219 may have arisen as a result of the loss of the job, and (2) plannin and problem-solving with regard to making ends meet financially, while also beginning the process of looking for new employment. In addition, the therapist needs to engage the patient in an exploration of the possible results of attaining employment. This would involve discussions about such issues as (1) how the patient will "unwind" after work without resorting to drags or alcohol, (2) how the patient will spend and save money, (3) what kinds of thoughts and beliefs are likely to be triggered as a result of adjusting to a new job, and (4) how to respond rationally to drag urges that may crop up or escalate in reaction to the changes associated with the job. Loss of Close Interpersonal Relationships A rift with a significant other represents another cri­ sis that often develops in patients' lives as a result of their substance abuse. W e have seen patients become extremely dysphoric, angry, and hopeless when parents have "disowned" them, mates have broken up with them, contact with their children has been denied them (e.g., by an ex-spouse or the courts), and members of their family or circle of friends die. Needless to say, the therapeutic relationship takes on fremendous significance at these points, as patients may infer that they will lose the support of the therapist as well. Newly broken romantic and marital relationships in particular often put patients at risk for renewed drag use. Patients may seek solace in their drags of choice, hoping to anesthetize themselves from the pain of interpersonal loss. Even if such patients do not attempt to use drags as a form of self-medication, they may deliberately engage in self-destructive behavior out of anger, hopelessness, or a desire to manipulate the other party. In one case, a patient started an alcohol and cocaine binge in order to make his ex-girlfriend feel guilty and responsible for his "fall." He admitted later that he believed that she would take him back if he could convince her that the breakup "drove [him] to drink and drug again." In another case, a patient explained that when her boyfriend threw her out of the house she thought to herself that, "Nothing don't matter no more anyway, so I might as well get wasted." Consequently, she spent the night at a crack house and resumed her heavy use of the drag. A particularly disturbing form of interpersonal crisis related to substance abuse is domestic violence. W h e n the patient is the perpe- frator of physical or sexual abuse of a minor, the therapist will be legally obligated to inform child protection authorities. In order to 220 COGNITIVE THERAPY OF SUBSTANCE ABUSE keep the patient in treatment, and thereby help to control this se ous problem, we strongly advise therapists to encourage their patients to report themselves to the appropriate agencies while in the therapist's presence. The therapist who is willing and able to remain collaborative under these most trying of circumstances stands the best chance of helping abusive patients work toward change. When the patient is the victim of violence in the home, the thera­ pist will need to have ready access to telephone numbers of protec­ tive shelters and support groups. Therapists should pay particular attention to their female patients who are involved with substance- abusing males, as this type of relationship is significantly correlated with domestic violence, and with the victim's retreat into heavier drag and alcohol abuse as a "coping" mechanism (Amaro et al., 1990). When patients experience crises in their most important relation­ ships, it is helpful for therapists to provide support and to encourage their patients to discuss their sense of loss, anger, or guilt. Two criti­ cal points need to be emphasized to the patient: (1) that the therapist will not abandon the patient, even as others may have cut off emo­ tional ties, and (2) that the patient has some measure of control over these interpersonal losses, to wit, the patient is capable of doing things to precipitate loss, and the patient is also capable of changing behav­ iors in order to facilitate reconciliation or new relationships. In cases in which reconciliation is impossible, such as when a loved one dies, the therapist must react with great sensitivity to the patient's grief, yet still be willing to call attention to the patient's increased risk for drag use. As with any crisis, the patient's risk for suicide should be assessed. Medical Emergencies Patients who abuse drugs and alcohol incur substan­ tially greater risk for acute medical crises than does the general popu­ lation. Examples include the alcohol abuser who experiences bleed­ ing gastrointestinal ulcers, the pregnant crack abuser who goes into premature labor, the diabetic who neglects his insulin in favor of shooting heroin and then lapses into a coma, the asthmatic woman who begins to asphyxiate after smoking free-base cocaine, and others. An increasingly prevalent example is provided by "Roland," who discovered that he was seropositive for HIV. He called the therapist in an extreme state of agitation when he received the results of the test, saying over and over again that "I'm going to die, I'm going to die!" The therapist's first response was to offer a great deal of sympa- Crisis Intervention 221 thy and to let the patient vent. Next, the therapist assessed wheth there was any immediate risk for drag use or suicidal behavior. Finally, after spending 30 minutes on the telephone, the patient was suffi­ ciently calmed down to the point where he could be engaged in look­ ing for signs of hope. Specifically, it was noted that although the patient was seropositive, he was asymptomatic and might remain that way for many years to come. Therapist and patient agreed that in order to maximize this incubation period, Roland would have to live as healthy a lifestyle as possible, including abstention from drags. In this manner, Roland would increase his chances of surviving long enough to see the day when effective treatments or a cure could be devel­ oped. Later, when Roland arrived for a face-to-face therapy session, he was in a positive enough state of mind to address issues of sexual responsibility to his partners. In the meantime, the therapist contin­ ued to show empathy for Roland's medical condition, and disproved the patient's hypothesis that even the therapist would now treat him as if he were a social pariah. In many instances, therapists will be unaware of their patients' medical emergencies until after an acute crisis has passed, with treat­ ment already received. Here, the therapist's job is not so much to help solve the crisis as it is to reengage the patient in the work of therapy, with special emphasis on the ways that the patient's drag use and unhealthy lifestyle practices may have contributed to the emergency. The new goal is to take whatever steps are necessary in order to mini­ mize the risk of further medical complications. On those occasions when the patient informs the therapist that a serious medical problem is going untreated, it is imperative for the therapist to encourage the patient to consult a physician, or to go to the emergency room of the nearest (or most appropriate) hospital as soon as possible. W h e n the patient indicates that he or she is inca­ pacitated by the illness, injuty, or disorder, the therapist may be required to take the kinds of life-saving steps described earlier in the section on overdose and suicidal crises. At times, patients will strongly resist the therapist's pleas to seek medical help, either because the patients want to worsen the crisis (e.g., due to hopelessness and passive suicidality), because they resent the implied "weakness" or loss of control over their bodies, or because they fear that hospital tests will reveal their active drug abuse. In such cases, when the therapist's attempt at support and reason fall on deaf ears, the therapist may need to call for medical help without the patient's consent, and hope to repair the therapeutic alliance after the medical crisis remits. The therapist should emphasize that this does 222 COGNITIVE THERAPY OF SUBSTANCE ABUSE not spell the end of the therapeutic relationship, but rather a ne saty break while the patient receives medical treatment. Criminal Involvement When patients slip back into regular drag use they run the risk of getting involved in a wide range of criminal activities, the likes of which will certainly compound a state of crisis. For example, months after Walter precipitously and prematurely dropped out of therapy, it was learned from his parole officer that he had gone into hiding to escape apprehension and beatings at the hands of organized crime figures. It seems that as he resumed his use of cocaine, he began to borrow large sums of money from loan sharks. This was a stark example of his poor judgment and planning skills (especially when under the frequent influence of cocaine), as there was no way that he would ever be able to repay his exorbitant debts. In the end, Walter resorted to committing burglaries in order to attain enough money and goods to pacify the loan sharks. Eventually, he was caught by the police and reincarcerated for a number of months. When he was released, he resumed therapy. When therapists learn that a patient is actively breaking the law, they must first determine whether other parties are at risk of being harmed. If so, there is an obligation to alert the police, as well as the intended victim. Ideally, this can be avoided if the patient is either willing to work with the therapist to cease and desist from the crimi­ nal activity in question or willing to voluntarily come clean to his parole officer or voluntarily commit himself to hospital for inpatient or day hospital psychiatric and drag abuse treatment. If no others are at risk, the therapist and patient must work to find and implement problem-solving behaviors in place of the illegal behaviors. Cognitive techniques can help the patient to generate viable options, and to combat beliefs that "There's nothing I can do about this problem," or "There ain't no way out of this

situation except [to engage in the illegal activities]." Violent Confrontation with the Therapist Another related crisis—that we have been fortunate enough rarely to have encountered at the Center for Cognitive Therapy—is one in which the patient threatens the therapist with bodily harm. If the threat is purely verbal, and the patient has no weapon and does not make physically menacing gestures toward the therapist, the therapist may be able to defuse the situation simply by Crisis Intervention 223 showing a sjmipathetic interest in understanding the reasons for t patient's anger. W h e n the patient has calmed down a bit, the thera­ pist can explain that there is no need for the patient to make threats toward the therapist. For example, the therapist can say, "Mr. Smith, it's okay that you're angty with me. I want to know why, and I want to work things out between us so that we can continue to work together. However, it is extremely unhelpful if you say you're going to hurt me, because then I have to turn m y attention away from your needs and onto m y own need for safety and self-defense. If we are going to continue to work together, I must insist that you never make any threats or take any harmful actions toward m e again. Does this make sense to you?" When the threat is more serious, such as when a patient assaults the therapist or produces a weapon, the therapist has to choose quickly whether to flee from the situation (if possible) or to muster all his or her empathic skills in order to mollify the patient until the acute threat has passed. Under such conditions it is appropriate to enlist the assistance of colleagues and/or the police, and it is also within the therapist's ethical prerogatives to choose to discontinue seeing this patient (principle of self-preservation). If the therapist is out of dan­ ger and the patient is still present, the therapist may also choose to go on with the therapeutic contract after explaining the ground rales as noted previously. Fortunately, we have found the incidence rate of such crises to be especially low in an outpatient setting. SUMMARY: GENERAL PRINCIPLES IN MANAGING CRISES When patients respond to crises by using drags, it is highly advisable that they contact a helper in order to prevent the episode from progressing to a full-blown relapse. Therapists should stress to their patients that a lapse into drag use may serve as grist for the therapeutic mill, and therefore may be used advantageously in treatment. They can explain to their patients that the renewed use of drags does not mean that the patients are "failures" in treatment, and that the therapist still will be willing to help. Therapists can tell the patients that it is not necessaty—indeed, it is often harmful—to isolate themselves from helpful others after using drags. Once the therapist has succeeded in establishing contact with a patient during or following a crisis, there are at least four important principles that the therapist must follow. First, the therapist has to be aware that in such instances the patient will probably want to resort 224 COGNITIVE THERAPY OF SUBSTANCE ABUSE to the abuse of drags as a "coping" technique. Therefore, along wi offering support and accurate empathy, the therapist must assess the patient's intentions to use drags in this manner. A great deal of thera­ peutic work will be required in order to help the patient to remain drag-free under these conditions. Second, it is imperative that the therapist be alert to the patient's hopelessness and fatalism (e.g., "I'm always going to be jinxed. Why should I even tty to get m y life back together again? It won't do any good anyway."). Unless vigorously combatted, such an outlook will put the patient at risk for drag use, flight from treatment, and other self-defeating behaviors. Third, the therapist can help the patient to use the current crisis as an opportunity to practice coping skills without using drags as an escape or cratch. Here, the therapist assists the patient in viewing his or her predicament as a "test" that, if passed, may signal trae progress toward recovety. There is also the added benefit of the patient's gain­ ing a sense of self-efficacy with each such crisis that is handled suc­ cessfully without the use of psychoactive substances. Fourth, in order to prevent treatment from being reduced to focusing on one crisis after another, therapists can look for the com­ m o n dysfunctional beliefs and common problematic behaviors that underlie a seemingly disparate set of crises. For example, a patient's repeated interpersonal rifts (with spouse, family of origin, employer) may aU be linked by a common theme, such as the patient's resis­ tance to being told what to do and what not to do. In such a case, the therapist can help the patient to spell out the dysfunctional beliefs; for example, "If I listen to someone's advice, it means that I'm being controlled," and "If I'm being controlled, I'm not a real man." By focusing on these beliefs and their concomitant problematic behav­ iors, the therapist can maintain better structure and continuity in therapy. As therapists are human, they are not immune to being stymied by crises that patients present to them. At such times, we strongly recommend that therapists consult with other professionals, includ­ ing case workers, parole officers, and fellow clinicians. For example, when legal problems have arisen, we have held therapy sessions that included the patient's parole officer. O n occasion, this has provided an interesting "good cop-bad cop" scenario, with the parole officer reading the patient the riot act while the therapist intervened to reengage the patient in some serious and concerted therapeutic work in order to stay out of further trouble. Finally, we would like to reiterate a rather sobering as well as hopeful point. Substance abuse patients, as a group, will almost cer- Crisis Intervention 225 tainly experience and present with more crises than do most other types of patients. This means that the therapist will rarely be able to rest easily with such cases, no matter how well things may seem to be going in treatment. O n the other hand, therapists who are prepared to handle such crises, and who persevere in teaching patients to deal with them, have the opportunity to make significant positive impacts on their patients' lives. There is a great sense of intrinsic reward in seeing patients through their toughest times, and in witnessing them turn their lives around for the better. C H A P T E R 1 4 T h e r a p y o f D e p r e s s i o n i n A d d i c t e d I n d i v i d u a l s J. T X o o d disorders are a frequent concomitant of substance use disorders. The comorbidity ranges from 13.4% in alco­ holism to 2 6 % among general drag disorders (other than alcoholism) (Regier et al., 1990) to 30.5% specifically for cocaine abusers (Rounsaville et al., 1991). It is important to be aware of the presence of depression: first, because it may be a profound source of suffering for a given patient; second, because it reduces the prognosis for recovery from substance abuse; and third, because it is a clear-cut indication for interaction by a trained professional rather than a coun­ selor (Woody et al., 1983). It should be noted that it may not be possible in some cases to make an absolute diagnosis of depression for several months after a drag- addicted individual has completed a detoxification program since the use of drags may in itself produce a clinical picture similar to a mood disorder or some other syndromal disorder. However, a careful history and clinical evaluation based on an instrament such as the Stractured Clinical Interview for DSM-III-R Disorders (SCID: Spitzer, Williams, & Gibbon, 1987) and the Beck Depression Inventoty (BDI: Beck et al., 1961) may help to tease out a trae depression. In any event, the negative thinking and beliefs typical of depression ("depressotypic") should be addressed as part of the therapeutic regimen. APPLICATION OF THE COGNITIVE M O D E L OF DEPRESSION The approach to depressed drug abusers can be for­ mulated in ways similar to depression in general (Beck et al., 1979; 226 Depression 227 Carroll, 1992). It is helpful to inquire about the negative cogniti triad (Beck, 1967): patients' view of themselves, their immediate life situation, and their future. Much of this negative triad has a particu­ lar coloring relevant to drag use but, in most respects, it is the same for the drag-dependent depressive as for the nonabuser depressive. The depressogenic automatic thoughts and beliefs of each are phrased in similar or identical words and the thinking disorder and pervasive negativity are the same. Individuals who see themselves as trapped in a situation over which they have no control, believe that they are helpless or socially undesirable, and can see only a wall of difficulties and disappoint­ ments ahead are likely to (1) feel sad, (2) express pessimism about the future, (3) consider suicide as the only solution, (4) experience a subjective loss of energy, (5) lose motivation to attempt any construc­ tive activity ("because it is useless and I will only fail"), and (6) lose satisfaction from sex, eating, or other formerly pleasurable activities. In addition, such individuals are likely to become dependent and indecisive. Each of these symptoms can become a target for therapeutic intervention (see Beck et al., 1979). Because of the high risk of sui­ cide in depressed drug abusers and alcoholics (Mirin & Weiss, 1991), special attention must be directed toward suicidal wishes. BELIEFS ASSOCIATED W I T H DEPRESSION Certain negative beliefs are typical of depression but also are observed in some addicted individuals who are not depressed. As described in Chapters 2 and 3 (this volume), these dysfunctional beliefs have a powerful effect on the addicted individual's thinking, feeling, motivation, and behavior. The kinds of beliefs that are typi­ cal of depressed individuals who abuse drags or alcohol are listed below. NEGATIVE SELF-CONCEPT "I am helpless (because I can't control using)." "I am weak (because I can't resist craving)." "I am unlovable." "I am defective." "I am worthless/disgusting (because I have a 'dirty' habit)." "Everything I do is wrong." "I am a failure." "I am frapped." "I don't have the will power to stop using." 228 COGNITIVE THERAPY OF SUBSTANCE ABUSE NEGATIVE VIEW OF PAST "I have never done anything right." "Nothing has worked out for me." "I have always been unhappy." "I have messed up m y whole life." "My whole life is a big failure." NEGATIVE VIEW OF LIFE SITUATION "People despise m e for m y addiction." "My family has given up on me." "There are so many demands on me, I can't handle them." "My family is watching m e all the time." "My neighborhood is impossible." "My job is dull and depressing." NEGATIVE VIEW OF THE FUTURE "If I try something, it won't work out." "I will never get what I want." "My future is hopeless." "Things can only get worse." "I will never be able to stop using." "I don't have anything to look forward to." "I don't deserve anything better in life." These beliefs can be subjected to exploration by the therapist in a series of maneuvers such as (1) looking for evidence to counteract the dysfunctional belief, (2) examining the logical relation of these beliefs to actual experiences, and (3) testing the beliefs in planned experiments. The preceding checklists should be used to help focus on the specific negative beliefs and also to monitor the patient's progress. Therapists should note that it is particularly important to assess and modify patients' beliefs about their "complete inability" to overcome their drug abuse, especially in light of research that sug­ gests that self-efficacy beliefs profoundly affect treahnent outcome and maintenance (Buriing, Reilly, Moltzen, & Ziff, 1989). THERAPEUTIC APPROACH Timing of Intervention By preparing a comprehensive formulation of the patient's depression (see next section), the therapist can make tenta­ tive decisions as to the type and timing of interventions. In deciding

Depression 229 which strategies to use initially, the therapist should consider t following questions: 1. Is the depression so painful that the emphasis should be on symptom relief rather than on an immediate confrontation of the using or drinking problem? 2. Is it likely that relieving some of the symptoms will reduce the pressure to use? 3. Will providing the stracture (inherent in cognitive therapy) for controlling the craving and using itself reduce the depression? 4. Can the therapist use the drag control and antidepression pro­ grams concurrently? 5. Are the suicidal tendencies and hopelessness sufficiently sfrong to warrant robust antisuicidal intervention and precautions? Certainly, if the patient is acutely depressed and suicidal, the therapist's attention should be focused on this serious clinical prob­ lem. By focusing on improving the patient's mood, the therapist not only reduces the patient's risk of suicide, but also may help the patient to feel better equipped to manage the drag problem (Hall et al., 1991). Case Formulation The case formulation is as important in treating depression as it is in treating drag abuse in general. The following symptomatology should be covered: (1) cognitive, (2) affective, (3) physiological (e.g., sleep disturbance), (4) motivational (giving up, avoidance, lack of drive, ambivalence, suicidal wishes), and (5) behavioral ("retardation," inertia, agitation). The cognitive symptoms in particular should be identified—spe­ cifically, the automatic thoughts and the types of distortions. The therapist should also tty to link up the automatic thoughts with the consequent affective or motivational symptoms. For example, the thought "It's useless to do anything, I'll only feel worse" can be linked to the patient's loss of motivation and consequent avoidance or inactivity. The thought "Evetything is hopeless. I can never get what I want" can be linked to patient's suicidal wishes. The thought "I am all alone; nobody cares" can be tied to sadness. The thought "I have messed up m y life" can be tied into the patient's self-criticisms. As the patient's life histoty unfolds, the therapist can constract a diagram such as that shown in Figure 14.1. In this case, the patient attempted to break the rigid pattern of blind obedience to the group's demands, but by becoming isolated he experienced his childhood- based belief in his unlovability and became depressed. 230 COGNTTIVE THERAPY O F SUBSTANCE ABUSE Parents and siblings were hypercritical. Early history IVIother abandoned family at age 8. Nobody helped patient with his problem. i Core belief I am helpless, alone. Nobody cares for me. i Compensatory Join gang: Get acceptance through blind strategy obedience (e.g., using cocaine). Conditional belief If I want acceptance, I have to go along. T Specific stressor Rejected by peer group because won't engage in dealing, etc. I am alone. Sad, lonely Uses cocaine for seif-medication FIGURE 14.1. Relationship of developmental factors to basic beliefs and using. The therapist should prepare a conceptual diagram such as that s h o w n in Figure 14.1. At an appropriate time, the therapist should s h o w it to the patient and explain the sequence of external events and external reactions. SELECTING TARGET SYMPTOMS It is difficult to specify in advance which problems should be selected for intervention during an interview and at what level these problems can be approached most effectively. In general, for the moderately to severely depressed user, the focus of the thera­ peutic intervention is at the level of s y m p t o m s ("target symptom level"). The target s y m p t o m is defined as a c o m p o n e n t of the depres­ sive disorder that involves suffering or ftinctional disability. The spe­ cific target symptoms m a y be divided into five categories (see Beck, 1967, for a more complete description): Depression 231 1. Affective symptoms. Sadness, loss of gratification, apathy, loss feelings and affection toward others, loss of mirth response, anxiety. 2. Motivational. Wish to escape from life (usually via suicide), wish to avoid "problems" or even usual evetyday activities. 3. Cognitive. Cognitive errors such as dichotomous thinking, overgeneralization, selective abstraction, personalization; errors are frequently linked to harsh self-criticisms (e.g., "I am a worthless per­ son" or "I am no good and don't deserve to get any satisfactions"). 4. Behavioral Passivity, (e.g., lying in bed or sitting in a chair for hours on end); withdrawal from other people, retardation, and agitation; total reliance on drags or alcohol to escape or moderate dysphoria. 5. Physiological or vegetative. Sleep disturbance (either increased or diminished sleeping); appetite disturbance (either increased or decreased eating). The therapist (in collaboration with the patient) makes a deter­ mination as to which of the target symptoms should be addressed on the basis of many factors: 1. Which symptoms are the most distressing to the patient? 2. Which symptoms are most accessible to therapeutic interven­ tion? In general, the techniques may be classified as predominantly behavioral—engaging the patient in specific activities or projects, which, in themselves, help to ameliorate some of the suffering; and predominantly cognitive—in which the major focus is on the patient's thinking. When the patient's depression is less severe, the therapeutic focus is often on the kinds of problems that are related to the precipitation or aggravation of the depression. These problems include difficulties at home, school, or work. Many users become depressed after a loss, such as a disraption of a close personal relationship, particularly the death of a close person. Personal losses also occur through legal, financial, and medical problems associated with use. In severe depres­ sion, in which the patient feels totally out of control over drag using, the therapist might want to focus on ways to help the patient to con­ trol his craving and to structure his day in such a way as to distance him from his craving. Some depressed users engage in nondrag activities but derive little pleasure from them. This failure to derive gratification may result from (1) an attempt to engage in activities that were not satisfying even prior to the depressive episode, (2) the dominance of negative cogni- 232 COGNTTIVE THERAPY OF SUBSTANCE ABUSE tions that obscure any potential sense of pleasure, (3) selective tention to actual experiences of pleasure, or (4) a belief that none of these activities can ever replace the high that once was achieved by drug use. Other depressed drag users may not engage in vety many non-drag-related activities, due to lack of skills, inattention to the existence of such activities, a sense of apathy, or the sheer time- consuming nature of finding, using, and responding to drags. There­ fore, simple attention to the cessation of drag use is an incomplete therapeutic strategy. It is imperative that such patients develop sources of nondrug positive reinforcement, such as work, hobbies, and prosocial recreation (Stitzer et al., 1984). Eliciting Automatic Thoughts and Beliefs Some depressed patients find it difficult-or may be unwilling—to bring up topics for discussion in session; which may be the consequence of the depression itself. Among the depressive fac­ tors that may account for patients' lack of productivity are the fol­ lowing: difficulty remembering specific problems since the last ses­ sion, a general passivity and slowing down that interferes with their discussing their problems, avoidance of painful subjects, and inhibi­ tion due to concern of the therapist's possible disapproval. One way to elicit relevant material is to review specific responses to items on the Beck Depression Inventory or the Beck Hopelessness Scale. For example, the therapist's questioning of the patient regard­ ing the reasons for endorsing a particular alternative may start up a fiow of information: PT: I have nothing to talk about today. Nothing important has hap­ pened since last week. TH: {examining the Beck Hopelessness Scale) I see that you have checked the item "My future seems dark to me." When did you have that thought during the past week? PT: Well, last week m y girlfriend said she'd break up with me unless I stopped using. TH: Yes ... ? PT: Well, I don't think I can stop. TH: What did you feel when she said that? PT: I felt bad . . . hopeless. TH: What thought went through your mind when she said that? PT: If we break up, I have nothing. Depression 233 TH: And if you have nothing, what then? PT: There's no sense in going on. At this point, the therapist demonstrates the relationship between the sad feeling and the thought "I have nothing." The therapist then explores the basis for his beliefs "I don't think I can stop" and "If we break up, I have nothing." (Both beliefs, of course, contribute to the patient's hopelessness and suicidal tendencies.) The therapist could infer from these ideas that the patient has a set of core beliefs centering around "I am helpless (weak, defective)" and "I am unlovable." Without necessarily explicitly addressing the core beliefs at this time, the therapist could then utilize specific interventions to undercut the dysfunctional beliefs and, thus, to some degree defuse the depressive feelings and suicidal tendencies. The therapist, for example, might initiate a line of inquiry as follows: TH: You have shown a lot of control in the past—and, in fact, you haven't been using for several weeks now. What makes you think that you will not be able to continue [to abstain]? PT: I just feel I don't have any will power any more. TH: I respect your feeling. But remember you felt that way several weeks ago but you were able to apply the techniques and show control. Addressing the sense of loneliness and dependency on the girl­ friend might start with a question: TH: At the moment, your girlfriend seems to be devoted to you. But, if for some reason she did leave you, why does it follow that you are nothing? PT: But I feel like nothing. Depending on the interchange, the therapist might then proceed with a series of questions such as: TH: Did you feel like nothing before you met your girlfriend? PT: I felt O K then. TH: Did you have a girlfriend then? PT: No. TH: So it seems that you don't need to have a girlfriend in order to feel like something. PT: 1 suppose. ... But I still feel like nothing. 234 COGNITIVE THERAPY OF SUBSTANCE ABUSE TH: Just because you may feel like nothing, does it follow that you are nothing?" The therapist may then continue with the type of Socratic ques­ tioning described in Chapter 6 (this volume). The therapist should keep in mind that the ultimate goal of questioning patients' inter­ pretations of a particular event is not simply to disqualify an inter­ pretation but also to undermine the basic negative beliefs leading to that interpretation or conclusion. As noted in the above excerpts, many depressed patients are influenced largely by their unpleasant feelings. They "read" these feel­ ings to mean that they are all alone, that they are nothing, and that things are hopeless. The feelings are taken as a source of information to which the patients attach considerable credibility. The therapist needs to reduce the patients' tendency to accept their feelings as a factual representation and to help them turn to objective, verifiable evidence to form their conclusions and interpretations. One of the distressing aspects of the ideation in depressed drag abusers is its saturation with themes of self-deprecation relevant to the meanings attached to the addiction. For example, a patient may apply the following labels to him- or herself: weak character, bad, worthless, immoral, sick. These notions are woven into the patient's belief that he or she is the helpless victim of the addiction, that he or she is defective for having the craving and not exercising will power. The patients consequently may be severely self-critical for their presumed faults ("I'm worthless," "I'm disgusting," etc.). Such think­ ing is apt to perpetuate both depressive symptoms and drug use. It is therefore not surprising that there is an association between depres­ sion and rates of drag relapse (e.g., Hatsukami & Pickens, 1982). The therapist addresses depressotypic beliefs through explanations and reframing. The therapist, however, needs to use judgment to decide whether an explanation is indicated in a given case and how it should be presented to the patient. Table 14.1 illustrates this

pro­ cedure. Dealing with Suicidal Ideation The therapist needs to be alert to covert suicidal ten­ dencies and to deal with them frankly. A number of interventions are available (see Beck et al., 1979, Chapter 10). As in the preceding illustration, the therapist needs to address the underlying hopeless­ ness and negative self image. Another approach is to list the reasons for living as opposed to Depression 235 TABLE 14.1 Examples of Beliefs and Explanations and Reframing Belief Explanation/reframing Drug abuse is a pattern of behavior that becomes self- I am a bad person. defeating. The effects, thus, are bad but the addicted person is not necessarily bad or immoral. Your problem is a technical one. I am defective You need to understand more about because I can't the working of your addiction and control the habit. then learn special techniques to bring it under control. The painful empty feelings are due in part at least to your personal Life is painful, problems and in part to your empty, without dependence on drugs for relief. using. W e can work together to help solve the problems and develop other ways to get satisfaction. the reasons for dying. This strategy requires considerable skill because it presupposes that the reasons for living will outweigh the reasons for dying. Another strategy is to get the patient intrigued by the thera­ peutic process and the interesting questions that are raised—so that the patient will be motivated to return to the upcoming sessions in order to get some of the answers. Also, the patient can be instructed to call the therapist when the wish to escape via suicide is the stron­ gest. The therapist might say, for example, "You may feel better when you are here but you may feel worse later. That is the time when you can get the most out of therapy. If you call m e then-or at least write down your thoughts-we can help you the most." The therapist should also bear in mind the need to notify the patient's family and con­ sider hospitalization if the suicidal drive is not controllable. Above all, the therapist must appear confident, reasonably upbeat, and in control of the situation. Negative Reactions to Therapy A disruption of the therapeutic collaboration may occur in therapy if the patient lacks objectivity toward his or her negative thoughts. For example, patients often experience new disap­ pointments or frastrations due to traumatic environmental events. 236 COGNITIVE THERAPY OF SUBSTANCE ABUSE When this occurs, they may be fiooded with a stream of negative cognitions that they automatically regard as valid without subjecting them to further considerations. Consequently, they are likely to experience increasing depression and hopelessness. This symptomatic exacerbation may lead them to decide that cognitive therapy is inef­ fective and that their addiction is incurable. This is often a critical point for a relapse. Patients may also feel disillusioned with their therapist. Any of these reactions may lead the patient to stop cooperating, resist carrying out assignments, miss appointments, or drop out of therapy. Their reactions may be compounded if the therapist unques- tioningly accepts the patient's negative construction of his or her progress. If the patient begins to miss appointments, the therapist is advised to contact the patient and clarify the dysfunctional thinking that is disrupting the therapeutic collaboration. Also, the missed appointment may be due to a relapse, which needs to be investigated. Ironically, there is some evidence that patients diagnosed with antisocial personality disorder (ASPD) actually may be more likely to seek out and cooperate with therapists when they are most depressed (Alterman & Cacciola, 1991; Woody, McLellan, & O'Brien, 1990). In such cases, the patients' desire to feel better outweighs their charac­ teristic autonomous need to defy authority figures. By helping these ASPD/depressed drug abusers to improve their mood, therapists may be able to form an interpersonal alliance with patients who otherwise would not form a bond with a helper. Symptomatic recurrences of depression are common during treat­ ment (Rawson, Obert, McCann, Smith, & Ling, 1990; Ziedonis, 1992). The therapist should prepare the patient early in treatment to expect to have negative fluctuations. Such exacerbations provide a valuable opportunity to apply cognitive techniques and skills. Further, they provide "practice" to deal with the problems that inevitably occur after termination of treatment. The therapeutic focus for the depressed addicted person also is directed to external problems that are related to the precipitation or aggravation of the depression (cf. Kosten et al., 1986). These prob­ lems may include stresses or difficulties at home, work, or school. They frequently involve friction in close relationships, difficulties in reference to work, or financial or legal problems. Thus, for example, the therapist and patient may work on helping the patient to make important decisions regarding a problem contributing to or maintain­ ing his or her depression, discuss specific techniques to help to cope better with a difficult life situation, and consider ways of relieving stresses or external demands. Depression 237 This type of approach (concentration on external problems) is also used after the patient's acute or severe symptoms have been relieved. The therapist should bear in mind that situational problems, drag use, and depression may aggravate each other. This multiple reciprocal interaction may be modified to improve both the external stresses and the depressive symptomatology. APPLICATION O F B E H A V I O R A L TECHNIQUES Cognitive Change through Behavioral Change The cognitive therapy of the depressed abuser is based on the cognitive theory of depression (as well as the cognitive theoty of addiction). Working within the framework of the cognitive model, the therapist can vaty his or her therapeutic approach accord­ ing to the specific needs of a given patient at a particular time as long as the treatment is based on a cognitive formulation of the case (cf. Persons, 1989). The therapist is conducting cognitive therapy even though he or she is utilizing behavioral techniques. In the early stages of cognitive therapy with depressed users, it is often necessaty for the therapist to concentrate on restoring the patient's functioning to the predepressed level. Specifically, the thera­ pist attempts to induce the patient to counteract his or her withdrawal and to become involved in more constructive activities. Often the important people in their life have given up on the patient. These significant others may attribute the patient's low mood and impaired performance to drags and conclude that the patient is no longer capable of cartying out his or her expected functions as provider, homemaker, spouse, parent, or student. Furthermore, the patients can see no hope of gaining satisfaction from those activities (other than using) that had previously brought them pleasure. The depressed users are caught in a vicious cycle in which their reduced level of activity leads to labeling themselves as ineffectual or worthless. These negative self-evaluations often reflect the opinions of those around them—as well as society at large. These self-inflicted put-downs lead to further demoralization and ultimately to a drift into a state of immobility. In severe cases, it is difficult to carty out intel­ lectual functions (such as reasoning and planning) as well as perform­ ing complicated acts requiring specialized skill and training. Since these forms of behavior are generally instraments for achieving sat­ isfaction and maintaining one's self-esteem and the esteem of others. 238 COGNTTIVE THERAPY OF SUBSTANCE ABUSE the disruption of these functions as a result of diminished concen tration, fatigability, and low mood produces dissatisfaction and a reduction of self-esteem. The role of the therapist is clear. There is no easy way to "talk patients out" of their beliefs that they are weak, inept, or undesirable. Patients can observe that they simply are not doing those things that once were relatively easy and important to them. By helping patients change certain depressive behaviors (avoidance, passivity), the thera­ pist may demonstiate to them that their negative self-evaluations are biased views and can help to restore their morale. Once a patient begins to get involved in constructive activities, the therapist may show the patient that he or she has not, in fact, irretrievably lost the ability to function at his or her previous level. The therapist points out that the patient's discouragement, pessimism, and giving up make it difficult to mobilize resources to make the necessaty effort. The goal is to get the patient to recognize that one of the prime sources of the problem is a cognitive error: The patient thinks (absolutely) that he or she is weak, helpless, and worthless, and those beliefs seriously impair his or her motivation and behavior. The term "behavioral techniques" may suggest that the immedi­ ate therapeutic focus is exclusively on the patient's overt behavior; that is, the therapist simply prescribes some kind of goal-directed activity. In actual practice, the reporting of the patient's thoughts, feelings, and wishes remains a central factor for the successful appli­ cation of the behavioral techniques. The ultimate aim of these tech­ niques is to produce positive change in the dysfunctional negative attitudes so that the patient's performance will continue to improve. In a sense, the behavioral methods can be regarded as a series of small experiments designed to test the validity of the patients' nega­ tive hypotheses or ideas about themselves. As the negative ideas are contradicted by these "experiments," the patient gradually becomes less certain of their validity and is motivated to attempt more diffi­ cult assignments. Scheduling Activities Many depressed users report an overwhelming num­ ber of self-debasing and pessimistic thoughts when they are withdrawn and inactive. They criticize themselves for being "vegetables," for being addicted, and for having withdrawn from other people. At the same time, they may justify their withdrawal on the basis that activity and social interaction are meaningless or that they are a burden to oth­ ers. Thus, they sink into increasing passivity and social isolation. Depression 239 Depressed patients tend to interpret their inactivity and withdraw as evidence of their inadequacy and helplessness and they are caught up in a vicious cycle. Individuals with a drag problem tend to relapse readily if their life is not stractured. The prescription of special projects is based on the clinical observation that depressed patients find it difficult to undertake or complete jobs that they accomplished with relative ease prior to the depressive episode. W h e n depressed, they are prone to avoid complex tasks, or, if they do attempt such tasks, they are likely to have con­ siderable difficulty achieving their objective. Typically, the depressed patient avoids the project or stops ttying soon after he encounters some difficulty. Negative beliefs and attitudes appear to underlie the tendency to give up. Patients often report, "It's useless to tty," for they are con­ vinced they will fail. W h e n they engage in goal-directed activities, they tend to magnify the difficulties and minimize their ability to over­ come them and carty out the task. The use of activity schedules (see Chapter 9, this volume) serves to counteract the patient's loss of motivation, inactivity, and preoc­ cupation with depressive ideas. Scheduling the patient's time on an hour-by-hour basis is likely to maintain a certain momentum and prevent slipping back into immobility. By focusing on specific goal- oriented tasks the patient and therapist obtain concrete data on which to base realistic evaluations of the patient's functional capacity. As with other cognitive techniques, the therapist should present the patient with a rationale for scheduling activities. Often patients are aware that inactivity is associated with an increase in their pain­ ful feelings. At the vety least, the therapist can induce patients to engage in an "experiment" to determine whether activity diminishes their preoccupations and possibly improves their mood. The thera­ pist and patient determine specific activities and the patient agrees to monitor his or her thoughts and feelings while engaged in each task. If the patient is reluctant, the therapist may seriously question the patient, "What have you got to lose by ttying?" The therapist may choose to provide the patient with a schedule to plan his or her activities in advance and/or to record the actual activities during the day. A "graded task" hierarchy should be incor­ porated into the daily plan. Planning specific activities in collaboration with patients may be an important step in

demonstrating to them that they are capable of utilizing their time constractively. Severely depressed patients may report a sense of "going through the motions" with the notion that there is little purpose in their activities. By planning the day with the 240 COGNITIVE THERAPY OF SUBSTANCE ABUSE therapist, they are often able to set meaningful goals. Later, the patient's record of the actual activities (compared to what was planned for the day) can provide the therapist and patient with objective feed­ back about the patient's achievements. The record also provides a ref­ erence to self-ratings of mastety and satisfaction for successful attain­ ment. It m a y challenge the therapist's ingenuity to get the depressed user sufficiently involved in the idea of cartying out a program of activities or even filling his or her activity schedule retrospectively. Thus, the therapist (1) explains the rationale (e.g., that people gener­ ally function better when they have a schedule), (2) elicits the patient's objections or reservations, and (3) then proposes making a schedule as an interesting experiment. The therapist should emphasize to the patient that the immediate objective is to attempt to follow the sched­ ule rather than to seek symptomatic relief: Improved functioning fre­ quently comes before subjective relief. W h e n patients engage in various activities, it is useful to have them record the degree of mastery and pleasure associated with a prescribed activity. The term "mastery" refers to a sense of accom­ plishment when performing a specific task. "Pleasure" refers to pleas­ ant feelings associated with the activity. Mastery and pleasure can be rated on a 10-point scale with 0 representing no mastery/pleasure and 10 representing m a x i m u m mastery/pleasure. By using a rating scale, the patient can recognize practical successes and small degrees of plea­ sure. This technique tends to counteract the patient's all-or-nothing thinking. S U M M A R Y The approach to the depressed user or alcoholic is similar to the approach to depression in general with the added fea­ ture that the typical depressive negative bias against the self often revolves around the patient's reaction to being on drags or alcohol. Thus, patients m a y be filled with contempt for themselves, may con­ sider themselves lacking in character, and m a y perceive themselves as helpless, defective, and rejected by other people and by society in general. Given the negative bias against the self and the profound hopelessness that often accompanies it, individuals with the combi­ nation of depression and addiction constitute one of the highest-risk groups for suicide (Marzuk et al., 1992; Mirin & Weiss, 1991; Ziedonis, 1992). Consequently, the possibility of suicidal wishes must be Depression 241 addressed early in therapy. Aside from the attention to the suicidal wishes, the therapist can follow the usual guidelines: 1. Conceptualize the case. 2. Apply behavioral strategies if the patient is motorically regressed. 3. Utilize cognitive strategies to undercut the hopelessness and negative self image and suicidal tendencies. 4. Help the patient to acquire greater control over the cravings (see Chapter 10, this volume)-a strategy that in itself m a y help to stem the tide of the depression. The specific techniques include structured activity scheduling, greater task assignment, and improved time management as a way of counteracting patients' regressive tendencies. In addition, Socratic questioning and the use of daily thought records m a y help patients gain more distance from their dysfunctional thinking. As with the stan­ dard freatment of the addict, it is ultimately necessary to come to grips with patients' beliefs that are driving both their drug abuse and their depressive tendencies. C H A P T E R 1 5 A n g e r a n d A n x i e t y ^•nnecessary or exaggerated anger presents a major problem in human relations, whether intimate or casual. The fact is that people tend to overreact to disappointments, hurt, fancied slights, and imperfect behavior of others. This phenomenon is par­ ticularly apparent in substance abusers (Ellis et al., 1988; Walfish et al., 1990). Hostility takes its toll not only in its undesirable effects on other persons but also in terms of its effect on the person who is angered. Substance abusers are prone to act out hostile impulses when they are under the influence of drugs or alcohol. Although they are par­ ticularly prone to use or drink to dampen unpleasant feelings of anger, the substance paradoxically may increase the likelihood of the expres­ sion of anger via disinhibition. The angry reaction is greater if the "noxious action" by another person is perceived as avoidable, unnecessary, intentional, and attrib­ utable to a failing in the other person; the reaction is less (or not all) if the same event is viewed as unavoidable, necessary, unintentional, and not blameworthy. If individuals perceive that they (or somebody they are attached to) have been wronged, they may become angered to a degree that greatly exceeds the degree of damage or discomfort. Much of the anger is the result of the symbolic meaning that is attached to the event, the cognitive mechanism involved in magnify­ ing its impact, and the degree of responsibility attributed to the other person. L O W FRUSTRATION TOLERANCE One of the most common conditions prompting an addiction-prone individual to seek relief through substance use is a feeling of frastration. Patients with low frustration tolerance (LFT) are 242 Anger and Anxiety 243 hypersensitive to any thwarting, nonfulfillment, or interference w their goals, wishes, or actions. The typical patient with LFT goes through life judging situations in terms of the following: "Am I get­ ting what I want?" or "Are people getting in m y way?" Because of the "internal" pressure to attain the objectives of ftilfillment of wishes or completion of a particular task, patients overreact to situations that interfere with satisfying their wishes or reaching a goal. Thus, they tend to be chronically impatient, intolerant, and uneasy. Wish-oriented patients operate under the "now dimension": They experience continuing craving for immediate "reinforcement" (encouragement, praise, recognition) or help. Stuck in the receptive mode, they consider it imperative that their cravings and desires be satisfied without delay. Behind this pressure lurks a fear clothed in dichotomous thinking: "It's now or never." Patients react as follows: "If I can't get what I want right now, I never will." Consequently, any delay in satisfaction is particularly distressing. The fear that they will be prevented from getting what they want arouses anxiety. Any interference with or interraption of an enjoyable activity is perceived as a major deprivation and causes pain. Either pain or anxiety (or both) can lead to anger if the individual holds another person responsible for the deprivation or interference. "Lil," a young woman, told by her landlady to lower the volume of her stereo, felt a sharp pang of disappointment over being deprived of one of her pleasures (high-volume music). She generalized this disappointment to "Nobody ever lets m e do what I want." Conse­ quently, Lil became angry, stomped around the apartment, and started to drink. As Lil's actions demonstrate, rapid, fleeting experiences of disappointment or anxiety trigger and are overshadowed in the patient's awareness by anger toward the thwarting or disappointing individual. It should be noted that a rather subtle mechanism operates between the initial feeling of disappointment (or anxiety) and the experience of anger. This mechanism involves attributing responsi­ bility for the disappointment to the other person in a fashion akin to blame. Attribution of responsibility may be a focus for discussion in therapy. Patient and therapist can explore whether the attribution of responsibility is reasonable. Since these patients have not learned to modulate their wishes and urges, they tend to experience them as "needs" that demand prompt fulfillment. These demands on themselves and others are experienced as dire needs ("1 must") or imperatives ("people should"). Since the needs have a "do or die" quality, their nonfulfillment is experienced as a threat or painful deprivation. The intolerance for 244 COGNTTIVE THERAPY OF SUBSTANCE ABUSE frastration is paralleled by intolerance for the dysphoria produce disappointment. The distress is compounded by thoughts such as "This should not have happened. They have no right to freat me this way." The claims, expectations, and demands on these patients and oth­ ers are not only imperative but also rigid and unrealistic. Karen Homey (1950) refers to these phenomena as the "tyranny of the shoulds"; Albert Ellis (1962) applies the term "musturbation." These individu­ als impose the shoulds on themselves as well as on other people. Individuals driven to achieve success may experience a sense of being dominated by an internal "slave driver." Such individuals often experience stress symptoms (Beck, 1993) and may turn to drugs and/ or alcohol to relieve these symptoms. Action-oriented individuals with LFT operate according to the same rules as the wish- or receptive-oriented individual. They act on the assumption that they must attain a goal promptly. Any delay is per­ ceived as interminable; any interference as unconscionable. The prin­ ciple underlying sensitivity appears to be concerned with "the con­ servation of energy." Impediments to the forward progress of such individuals are experienced as an unacceptable drainage of power or energy, which leads to impatience and restlessness. Wish-oriented and action-oriented people are disposed to ascribe negative motives to other people. They operate on the following premise: "Anyone who does not help me or facilitate m y goals is self­ ish; others' noncompliance with m y wishes is the equivalent of op­ position." The LFT individual interprets lack of support or help as signs of indifference, negligence, or irresponsibility. Ironically, these individuals are not cognizant of the fact that their behavior is con­ trolled by their inner dictates (cravings, demands, imperatives) but perceive themselves as controlled and victimized by other people who are indifferent to their "legitimate needs." Since the patient's pattern of frustration is crade, inflexible, and indiscriminate, the thinking mechanisms are equally crude. Such patients tend to catastrophize and overgeneralize when their wishes are not met: "I will never be able to get the job done." "Others never cooperate with me." "People always get in m y way." In order to ward off the threats and prevent pain, patients with LFT attempt all the more strongly to control their environment and to impose regulations and expectations on other persons. However, the harder they try to compensate for their sense of inner vulnerabil­ ity, the more likely they are to be frustrated. As their demands and claims on others escalate, they are prone to feel let down, disappointed, or blocked. Thus, the strategy of hypercontrol of others is ultimately Anger and Anxiety 245 self-defeating. The social environment simply will not conform to these continual demands and expectations. Sooner or later, other people will fail to respond satisfactorily to their wishes or drives. At a deeper level, patients with LFT perceive themselves as pow­ eriess or helpless. Any obstacle to a goal that the patient encounters primes the sense of powerlessness and produces a transient feeling of weakness. The next step is attributing responsibility to the "frus­ trating" individual and wanting to punish that person for his or her fransgression. This sequence leads, of course, to the most notable char­ acteristic of LFT, namely, explosive rage over relatively trivial inci­ dents. An action-oriented person is enraged by a slow driver in front of him and aggressively and dangerously passes him—with his thumb pressed firmly on the horn. A husband rails over delays in meals or unsatisfactoty food; a wife is incensed at being kept waiting while her husband works late at the office. In each instance, the patient attaches an overgeneralized meaning to the "offense": "She doesn't care about me" or "He treats m e like a servant." These highly personalized mean­ ings attached to the event—not the event itself—lead to the inflamma- toty reaction: the sense of being wronged. This tendency to "person­ alize" situations, to interpret neutral behaviors as a perceived affront, is a hallmark of LFT. The generation and expression of anger and hostility serve sev­ eral related purposes. They constitute a robust attempt to establish control over other persons by "punishing" them for their action or inaction. When the expression of hostility is effective, there is no form of behavior that

exerts as powerful an influence on other people, particularly if the offended individual is in a position of strength. The patient assumes that punishment, whether in the form of a complaint, a reproach, or a tantrum, will help to shape the other person's future behavior properly. The implicit punishment will supposedly be a "leaming experience" for the offender. Further, the punishment con­ tained in the reproach in some vague way undoes or compensates for the damage to the patient's self-esteem: A person is not "a helpless, vulnerable wimp" if he or she can inflict pain on another person. Most important, the expression of anger gives a subjective sense of power (even though fleeting). By acting in a forceful, aggressive way, the patient is able to neutralize the sense of powerlessness activated by the delay in gratification. Since frustration or disappointment accentuates the perception of the self as ineffective, the expression of hostility shifts the self-concept from "I am helpless" to "I do have power." Of course, punishing other people as a consequence of one's own 246 COGNTTIVE THERAPY OF SUBSTANCE ABUSE "neediness" or sense of inadequacy is ultimately self-defeating. O people are pained and often angered when reproached and are prone to retaliate. Further, the frastrated individual is drawn into a power straggle with others over the issue of who will control whom. The individual gets tangled in a vicious cycle of increasingly futile attempts to control others, leading to increasing disappointment and rage. LFT individuals are prime candidates for addiction (Ellis et al., 1988). Using drags and/or drinking can serve several purposes. First, these substances satisfy the desire for instant gratification. Second, they reduce the anxiety and sadness engendered by frustrations. Finally, they can give a transient euphoria and sense of mastety to compensate for the feeling of helplessness and sense of inadequacy. It is important for the therapist to recognize that patients with LFT often are deficient in perceiving important social cues or recog­ nizing the rules that govern human behavior and allow people to interact with a minimum of friction. They are frequently unaware of and overstep the usual boundaries between people. Lil, for example, had no recognition of the fact that the loud noise from her stereo would bother the other boarders. Some patients addicted to drags have Attention Deficit Disorder (Gawin & Kleber, 1986; Glantz & Pickens, 1992; Weiss, 1992) and may rely on cocaine, for example, to sharpen their focus and increase their guarded awareness of other people. CHECKLIST FOR EVALUATING LFT 1. The patients' desires are viewed as imperative needs that requi prompt fulfillment. 2. Delays, interference, and blocks have idiosyncratic meanings such as "1 may not finish this job" or "I can't get what I want." 3. The frastration is overgeneralized to notions such as "I never get what I want" or "People always get in m y way." 4. Patients personalize such frastrations as though the frastrations are deliberately directed against them, and they regard the alleged agent of the frastration as culpable. Patients manifest the usual think­ ing errors associated with emotional distress: all-or-nothing thinking, selective abstraction, overgeneralization, catastrophizing, and person­ alizing. 5. Because of dichotomous thinking ("It's now or never"), patients build up arbitrary rules (shoulds and musts) to enforce their wishes, expressed as rights, entitiements, and claims: "People have no right to withhold what I want," "Others are wrong to get in m y way," and "1 should be able to work without interruption." Anger and Anxiety 247 6. The absolutistic expectations and demands are driven by catastrophizing: The consequences of a delay, obstacle, or difficulty are expected to be disastrous. 7. The patients' rales and demands also represent attempts to control others and prevent problems. W h e n controls break down (a rale is violated, a demand is unfulfilled), patients experience anxiety or hurt resulting from catastrophizing or exaggerated sense of loss. 8. The expression of anger is legitimized: "I have evety right to be angty." 9. Underlying the low threshold for frustration is a core belief such as "I am helpless" or "I am unlovable." Any delay, interference, or problem related to attainment of wishes or goals can evoke the sense of helplessness or unlovability, lead to catastrophizing, then distress, and, finally, to anger. 10. The accumulation of anxiety and anger and the mobilization to punish the offender lead to tension which patients attempt to relieve with drugs. In this context, drug-taking beliefs may take the following forms: "I can't stand the anxiety, sadness, or anger and must relieve it right away." "People want to stop m e from using or drinking because they consider me weak and want to control me." "They are wrong and bad for wanting to control me." "I will regain control by drinking or using when I want to." 11. People may be overly frustrated by their own mistakes, ineptitude, or deficiencies and will manifest extreme self-criticism. This mechanism is especially prominent in the addicted patient with low self-esteem (Tarter, Ott, & Mezzich, 1991). LFT A N D SUBSTANCE ABUSE: A CASE VIGNETTE "Charlotte," a 35-year-old single woman, was in out­ patient cognitive therapy following a 30-day inpatient stay for crack cocaine abuse at a nearby hospital. Charlotte and her cognitive thera­ pist identified many situational and cognitive triggers for her urges to use crack, including family disputes, worries about money, thoughts about deceased relatives, news about old boyfriends, and dissatisfac­ tion with her employment situation. These triggers tapped into a great sense of loss (e.g., "I don't have anything meaningful in m y life any- 248 COGNITIVE THERAPY OF SUBSTANCE ABUSE more, so I might as well drown my sorrows in alcohol and cocaine") and helplessness and hopelessness ("Evetything I do to try to help myself gets messed up in the end anyway, so there's no point in plan­ ning for a future"). Charlotte's dysfunctional beliefs (noted above) underlay her low tolerance for frastration. For example, when she tried to reach her new boyfriend by telephone one evening to talk about some upset­ ting things that happened on her job interview that day, she found that nobody was home. She tried to call numerous times within the following hour, but he still did not answer the telephone. Charlotte became quite agitated and she felt an increasingly strong urge to go out and find a crack dealer. In retrospect, her thoughts included: "I need him now. If I can't talk to him now, I'll go out of my mind." [Imperative: equating wish with need] "I can't stand being alone when I'm hurting. M y only alternative to feel better is to get high." [Myth of intolerability of pain] "Why isn't he home when I need him?" [Excessive demands] "I can't wait any longer!" [Intolerance for frastration] "Maybe he's out with another woman. Maybe he's cheating on me." [Catastrophizing] "If that's trae, then I don't have any reason to stay straight any­ more." [Justification] Charlotte noted that all of these thoughts transpired within only 90 minutes of trying to reach her boyfriend at home. Rather than attempting to solve the problem-calling him back at a later time, calling the therapist, or trying to utilize her own coping skills—she became vety upset that she could not find her boyfriend right at the time she "needed" him. This frustration led her to the premature and maladaptive conclusion that she had no choice but to go out and get high. She believed it was justifiable to do so; after all, if you are alone (as she believed), helpless, and have no future, why shouldn't you find a way to feel better now by any means possible, including drags? This vignette demonstrates how LFT is based on a thinking style that justifies impulsive and self-defeating behavior, and therefore feeds into urges to use drugs. The LFT patient is constantly in danger of exaggerating temporary inconveniences and upsets as indicative of (or equivalent to) unremitting denial of the most important goals in life- forever. Therefore, the LFT patient is likely to conclude that drug use (a short-term "solution") is necessary, given that long-term solutions are unseen or dismissed. Anger and Anxiety 249 DEALING V\^TH SPECIFIC ISSUES Imperativeness of "Needs" Patients require practice in delaying satisfaction of desires—whether in terms of resisting the desire to use or drink or of demands on others for their help, reassurance, and praise. 1. From a practical standpoint, patients should be coached to monitor their wishes: write down the specific desire, note its inten­ sity, and determine how long it takes for the desire to be reduced. This procedure enables patients to distance themselves from the desire and, thus, to control it better. 2. Distraction—for example, getting involved in an absorbing activity such as conversing with another person—often helps to reduce the imperativeness of the need. 3. By delaying their demand for instant gratification, patients can become aware of the exaggerated importance they attach to a particu­ lar desire. Their desires for affection and recognition are similar to the cravings for drugs and alcohol. W h e n an individual is frustrated in his or her yearning for affection (or expects to be frastrated), for example, he or she may channel this yearning into a craving for a drink or a "fix." The kind of pleasure from the chemical agent is perceived as a substitute for the pleasure desired through a loved one's affection. Meaning of the Event: Hidden Fears By delaying the expression of anger or impatience (acting out the frustration), the individual has an opportunity to become aware of (1) the immediate cognitive reaction to an event (e.g., a specific fear of disappointment) and (2) the immediate affective response such as anxiety or sadness. For example, the wife who railed at her husband for not coming home when she expected had an ini­ tial fear, "perhaps something happened to him," and felt consider­ able anxiety. W h e n he did appear, she was angty at him for "caus­ ing" her distress rather than for being late-although she scolded him for being late. In therapy, she recognized that the initial fear was related to her childlike concern of being abandoned. Thus, the therapy was directed toward her confronting her sense of vulnerability, which was aroused by his tardiness. In this case, her hostility served to pun­ ish him for making her feel vulnerable and, at the same time, gave her some sense of power. When an unpleasant event occurs, patients with LFT practically simultaneously assess the consequences and attach responsibility 250 COGNITFVE THERAPY OF SUBSTANCE ABUSE either to themselves or to others (or occasionally simply to bad lu The expectation of negative consequences is not a thought-out, reflective, rational procedure but is molded to a large degree by a ten­ dency to exaggerate. This tendency may take the form of foreseeing the possibility of catastrophic consequences or by exaggerating the current loss. The catastrophizing may be heralded by a "What if... ?" prelude; for example, "What if I hadn't caught the mistake in time.... It could have cost m e a large sum of money." The sense of loss places a great premium on the supposed diminution of some resource; for example, "They are costing m e valuable time by their inefficiency" or "I am forced to waste m y energy answering stupid questions." Issue of Responsibility The attribution of responsibility occurs almost at the same time as the estimate of loss or threat. The individual determines that some person is responsible for the loss or threat. The patient's reaction is to fix accountability on the other person as the cause of the frustration. This attribution of responsibility is evidenced in expressions (or thoughts) such as "You should have known better" or "You never pay attention." Words such as should, should not, never, and always are often expressions of attribution of responsibility and simultaneously of reprimand. A more subtle process is also discernible. Noxious events are likely to evoke a sense of helplessness in the LFT-prone patient. If such patients can fix responsibility on another and mobilize reproach, they can regain some of the sense of lost power. By blaming another per­ son, the patient in a sense is saying, "I'm not so weak and helpless. 1 am strong enough to punish you." Of course, if such

patients per­ ceive themselves as the cause of the problem, they criticize themselves and feel even more helpless. Issue of Control The problem of control is particularly pertinent to the treatment of drug abusers since the therapy involves, in essence, the therapist's attempting to control patients' behavior. This problem has been exacerbated in many instances in which patients have been sub­ jected to threats, coercion, and criticism by other well-intentioned persons trying to enforce abstinence or, at least, continence. These attempts by others can produce a chronic resentment in the substance abusers. Most significant from the therapeutic standpoint is the impact Anger and Anxiety 251 of such interventions on the individuals' self-esteem. They are li to respond covertly to the criticism with an increased sense of inad­ equacy, even helplessness, isolation, and unworthiness, although overfly they may devalue the other party. Many of these patients will become infuriated at their "torment- ers" and will attempt to regain a sense of power by counterattacking and thus alleviating the pain. Others will engage in elaborate decep­ tions to disguise their using or drinking and to protect their self-esteem from further blows. Thus, the confrontation by the therapist of the patients' deceptions is a delicate procedure since exposure of a pre­ tense may constitute a significant threat to the patients' self-esteem. Handling Anger Toward the Therapist Situations in which one person is attempting to influence the behavior of another person embody specific sets of prob­ lems. The drug-abusing patient's sensitivity to being controlled is practically a given in the therapeutic situation and is an issue that needs to be addressed by the therapist. As already indicated, the patient's strategy of covering up lapses in order to protect him- or herself from criticism may represent an important defensive maneu­ ver. Consequently, discussions of the patient's drinking and using or of his or her deceptions may arouse the patient's anger and shake the foundations of trast. O n the other hand, if the therapist appears to be "taken in" by the patient's deceptions, the therapist may appear to be a "pushover" and will not maintain the patient's confidence. Patients should be encouraged to discuss their angry feelings toward the therapist. However, while ventilation of the feelings may serve a constructive purpose—up to a point—the therapist must be prepared to set limits to the expression of anger as illustrated below. "I realize that you are very angry at me and I'm glad that you feel safe to express it here. But before we go any further, it is very important to find out all the factors that are fueling your anger. For example, what were you thinking just before you became angry?" Often the exploration of the automatic thoughts preceding the appearance of the angty response serves as a distraction, which allows the patient to gain some objectivity. Further, by eliciting the auto­ matic thoughts, the therapist can get a better grasp of the cognitive aspects of the patient's disposition to anger. The anger is often the consequence of a sequence of automatic thoughts relevant to secret fears or hidden doubts. For example: "He's ttying to control me" therefore "I won't be able to do anything on m y own" (secret fear) 252 COGNITFVE THERAPY OF SUBSTANCE ABUSE therefore "I can't let him get away with this" therefore anger; or must be a weakling to allow him to talk to m e that way" (self-doubt) therefore "He has no right to treat m e that way" therefore anger. However, this therapeutic strategy is not effective all the time, and many times the therapist has to be prepared to absorb verbal expressions of hostility until the patient's anger subsides. Abusive Behavior and Control of Anger Patients with an addictive personality profile are par­ ticularly likely to have difficulties in family relationships and may engage in abusive behavior (Amaro et al., 1990). The various steps for anger control are described in Love Is Never Enough (Beck, 1988). Briefly, these consist first of being able to recognize the earliest sub­ jective signs of anger, which may simply consist of somatic sensa­ tions such as a tightness in the chest or a stiffening of the muscles in the arms. When a patient experiences these premonitory signs, for example, he should then slow down his talking and examine whether he sounds angry. If he is indeed angry, he should stop talking until his anger subsides. If he is still angry, he should leave the room (assuming the hostile interaction is at home). He should then stay away from the targets of his hostility until he has cooled off. At times it may be necessary to leave the house for a while, although it is essential for patients to have a drug-free destination in mind (e.g., the public library). The therapist should explore with the patient various kinds of distraction. While in the heated situation, the patient might try to distract him- or herself by thinking of a pleasant experience. After leaving the room, the patient could get involved in some physical exercise or manual project to "work off" the anger. In order to pre­ pare the patient for using these self-control methods, the therapist should use the therapeutic session to get the patient to practice the techniques. One method is to have the patient recall in vivid detail a recent dysfunctional interaction and then, using imagery, go through the scenario but imagine using techniques of self-control. Induced imagery is also useful for ascertaining the individual's automatic thoughts (see Chapter 10, this volume). The patient is requested to relive a hostile encounter in imagination and then to pinpoint the automatic thoughts associated with the angry feeling. The material can then be dealt with using the standard way of fram­ ing reasonable responses to automatic thoughts. The therapist may also utilize role playing in order to model self- control techniques for the patient. Patient and therapist simulate or Anger and Anxiety 253 recreate a hostile encounter that will provide the patient with an opportunity to recognize his or her anger, detect the automatic thoughts, and rehearse rational responding. It should be kept in mind that an important facet of patients' anger proneness is their deep sense of helplessness and inadequacy in interpersonal conflict. Compounding this defective self-image is the lack of social problem-solving skills (Piatt & Hermalin, 1989) and frequently a lack of assertiveness. In his or her formulation of the patient's difficulty, the therapist needs to diagram the sequence: "I am helpless/inadequate" therefore "I am vulnerable to being insulted, controlled, etc." therefore "He has taken advantage of me" therefore "I have to protect myself" therefore anger and acting out. Patients capable of insight often are willing to accept this formu lation. The therapist's role then is to help the patient to reinforce a self-image of adequacy without having to prove it by expressing anger or abusing another person. Building up a positive core concept in­ cludes not only demonstrating to patients their potential capacity to handle conflicts with other people but also negating their dysfunc­ tional beliefs; for example, "If I don't succeed at something, it shows I'm inadequate" or "If somebody argues m e down, it means I'm infe­ rior." Further, the therapist has to communicate to the patient that "feeling helpless" is not an objective representation of reality; people tend to read their feelings as factual. But one can "feel helpless" and still function adequately. Assertive Training and Problem-Solving Some patients become unnecessarily provoked because they feel inhibited or lack social skills in presenting their own view­ point or expressing their own self-interest. Failure to assert themselves effectively makes them feel put upon and helpless and to view other people as controlling and domineering. These patients often have a particular problem with people in authority. "Steve," who has a histoty of dependence on alcohol and cocaine, worked as a contract mechanic in a automotive garage. W h e n he and the other mechanics were given jobs by the service manager, he noticed that the other mechanics were assigned the more expensive jobs. Although he realized that they were more assertive than he in asking for the good jobs, he concluded that he was being discrimi­ nated against. He was too inhibited, however, to complain to the service manager, w h o m he perceived as authoritarian and callous. One day he felt particularly frastrated because a big job on an 254 COGNITIVE THERAPY OF SUBSTANCE ABUSE expensive foreign car was given to one of the other mechanics. He wanted to say something to the manager but was too inhibited. As he left the garage, he became increasingly angty. He thought of going home but he had an image of his children squabbling with each other and that made him feel more frastrated. Although he had previously resolved not to drink or use, he decided to stop at a bar and have a drink to relieve his anger and frastration. After three or four beers he again thought of going home but this time he had visions of his wife scolding him for being late and drinking so when he left the bar he sought out a friend who gave him a joint of crack cocaine. When he got home eventually, he was filled with remorse for his lapse. The therapist pointed out to him iri their session that his lapse was not a full-blown relapse and that they could do something to overcome his repeated frastration and escalating anger. The therapist then engaged Steve in a series of exercises in assertive training (cf. Alberti & Emmons, 1974; Collner & Ross, 1978). He first modeled for the patient how to approach the manager and state his concern in a reasonable, straightforward manner. The therapist then role-played the manager and Steve rehearsed the approach. During the rehearsal, Steve had a number of automatic thoughts such as "He'll think I'm a troublemaker for complaining" or "He'll make m e look foolish." (These were his "secret fears" that he became aware of only as a result of the role play.) The therapist discussed these dysfunctional thoughts after the role play was concluded and tried to communicate that the important thing was being able to express a legitimate complaint irrespective of whether it was immediately effective. Also, since the other men seemed to be able to make requests or complaints with­ out being "put down," Steve certainly could try. Finally, Steve agreed that he would be no worse off than he was currently without com­ plaining. The result was that the manager denied that he was playing favorites and claimed (probably as an excuse) that he thought Steve preferred working on the American cars. Subsequently, the manager began to give him some jobs on foreign cars. Problem-Solving and Skills Training Addicted patients have difficulty in solving problems, particularly those dealing with interpersonal conflict (Piatt & Hermalin, 1989), for at least two reasons. First, a problematic situa­ tion can arouse such a degree of anger that they are driven to punish (scold, criticize) the other person so that the problem, far from being solved, becomes aggravated. Second, because of poor problem-solv- Anger and Anxiety 255 ing skills, the patient is rapidly frastrated in a difficult situat consequently becomes angry. It is interesting to note that addicted patients are not always de­ ficient in skills. Many know what to say in situations in which they do not have a strong personal involvement. Also, they may know how to advise others to handle difficult situations. However, when they themselves are involved in a difficult situation, they feel they cannot or do not want to tty to resolve the problem other than by berating the other person. Once the patients label others' behavior as an offense against them, they have slipped into the punishment mode. The remedy is for them to detach themselves from the personalized meanings and address the problem in a way leading to solution. "Bob," for example, became infuriated because a mail clerk denied having a package that Bob knew had been delivered to the mail room of his office. Instead of brainstorm­ ing about where the package could be, Bob started to yell at the clerk. The clerk became defensive and retaliated, "You can't speak to me that way." Bob then responded

sarcastically and walked away in a huff. He was so upset that he felt incUned to go to the bar next door and have a few drinks. Much later he learned that one of his colleagues had picked up the package and delivered it to Bob's secretary. In the therapy session, Bob realized that he had "personalized" the problem—as though the circumstances had directed this against him. He also had had a catastrophic thought, "If I don't locate the package, I may lose m y job." Once he took the situation personally, it was played out at a level of interpersonal fighting (conflict) instead of mutual problem-solving (cooperation). Thus, the following steps can be prescribed for situations in which one is angty about a difficulty: 1. Detach self from the personalized meanings (put down, dis­ criminated against, let down). 2. Tty to approach the problem cooperatively despite the pres­ ence of anger: a. Define the problem: the "lost" package and the clerk's dif­ ficulty in recalling where it might be. b. Experiment with explanations: Ask the clerk where it could have been mislaid. Inquire whether one of the other clerks may have seen it. Could somebody else have picked it up? Steve had a problem with his children-when they started to squabble, he became furious and wanted to yell at both of them. Applying the rale of "defining the problem," he discovered that the older child was continuously teasing the younger one. Steve then 256 COGNTTIVE THERAPY OF SUBSTANCE ABUSE could see the solution clearly: Issue a stern reprimand to the old child to stop teasing. This may seem to be a minor example, but it was built up into a major problem because of Steve's inner belief, "I am helpless when there is a conflict." This sense of helplessness made him reluctant to go home at the end of work because of his fear of friction in the fam­ ily. Hence, part of the program of anger management consisted of building up Steve's confidence in his parenting skills. Dealing with Catastrophizing The way that the underlying mechanisms of LFT can be increased is indicated in another interview with Steve. TH: Tell me more about why you decided to use [cocaine] last Thurs­ day. PT: Well, I came home and I asked m y wife whether she had taken the kids to the doctor—like she had promised. She said that they weren't really sick and besides she was too busy today. TH: What did you feel then? PT: 1 felt really mad but I controlled myself and left the room. Then 1 decided to have a line. TH: What went through your mind when she said she hadn't taken them to the doctor? PT: 1 thought, "Suppose they really are sick? Even getting there a day late could be very serious." TH: As you look at it now, do you still believe that it was a serious problem? PT: No, they really weren't sick, but at the time it seemed like they could get much worse. TH: So that's an example of what we know you are in the habit of doing—catastrophizing-a kind of exaggerated worrying. PT: Right. TH: Now, what did you feel in your body when you had that thought? PT: I felt weak all over. TH: Where? PT: Especially in m y muscles. TH: What else? PT: I felt a heavy feeling in m y stomach. Anger and Anxiety 257 TH: Was this weak feeling like a helpless feeling? PT: Well, I did feel helpless at the time. Because I told her some­ thing to do and she didn't do it. TH: So now, it seems that you have a tendency to expect a catastro­ phe when somebody doesn't follow instructions. You start to feel weak and helpless. Then you get mad at the person who you think made you feel that way and you want to punish her. Steve, of course, was more introspective than many patients and was adept at identitying his catastrophic thoughts and his feelings. But many, if not most, patients can be trained the way that Steve was, to recognize and evaluate these automatic thoughts. ANXIETY DISORDERS A N D SUBSTANCE ABUSE Background Anxiety disorders are among the most common of psychiatric disorders. Zung (1986), in a survey of 739 family practice patients, found that 2 0 % had clinically significant anxiety symptoms. In the Epidemiologic Catchment Area study, Regier and his colleagues (Regier et al., 1988; Regier et al., 1990) report that the lifetime preva­ lence of anxiety disorders in the general public is over 14%. In their study, this rate is exceeded only by substance use disorders, with a lifetime prevalence rate of over 16%. DSM-III-R distinguishes between seven Axis I anxiety disorders: Panic Disorder, Agoraphobia, Social Phobia, Simple Phobias, Obsessive- Compulsive Disorder, Post-Traumatic Sttess Disorder, and Generalized Anxiety Disorder. O n Axis II, there are at least two personality disor­ ders characterized by substantial anxiety: Obsessive-Compulsive and Avoidant Personality Disorders. Compared with those who do not have chronic anxiety, people with these disorders may be more vulnerable to abusing certain psychoactive substances (e.g., alcohol, benzodiaz­ epines, and nicotine). They may use drags in an attempt to cope with anxious feelings (i.e., as "compensatoty strategies"). In fact, numer­ ous studies and reviews suggest a significant positive relationship between anxiety disorders and substance abuse (e.g., Beeder & Mill- man, 1992; Helzer & Ptyzbeck, 1988; Hesselbrock, Meyer, & Kenner, 1985; Hudson, 1990; Kranzler & Liebowitz, 1988; Kushner, Sher, & Beitman, 1990; LaBounty et al., 1992; Linnoila, 1989; Mullaney & Trippett, 1979; Quitkin, Riflcin, Kaplan, & Klein, 1972; Regier et al.. 258 COGNTnVE THERAPY OF SUBSTANCE ABUSE 1990; Ross et al., 1988; Schuckit, 1985; Walfish et al., 1990; Wilson, 1988). According to Kushner et al. (1990), estimates of alcohol problems in anxious patients have ranged from 1 6 % to 25%. In alcohol treat­ ment programs, estimates of patients with coexisting anxiety prob­ lems have ranged from 22.6% to 68.7%. From a critical analysis of existing literature, these authors conclude that coexistence of alco­ hol problems and clinical anxiety is not a simple, unidirectional, causal relationship. Instead, Kushner et al. (1990) state: It appears more likely that alcohol has the potential to interact with clinical anxiety in a circular fashion, resulting in an upward spiral of both anxiety and problem drinking. For example, increasing alcohol consumption motivated by the short-term relief of anxi­ ety (or the belief that alcohol can relieve anxiety) may lead to increased anxiety related to autonomic nervous system hyper- excitability and anxiety-inducing environmental dismptions which, in turn, may lead to more alcohol consumption to relieve symp­ toms, (p. 692) T h e Cognitive Therapy of Anxiety A text on the cognitive therapy of anxiety disorders (Beck et al., 1985) emphasizes the importance of appraisal processes in anxiety. Anxiety is typically precipitated by a situation in which the individual regards himself as vulnerable to some threatening (i.e., unpleasant, dangerous, or harmful) event or situation. Prior to the onset of anxiety symptoms, a person makes a series of appraisals about a potentially threatening event or situation. The first appraisal, called a "primary appraisal," involves the initial assessment of risk. For example, in response to the social encounter, the socially anxious per­ son might make the primary appraisal "1 a m going to embarrass myself." In response to his rapid heartbeat, a person with a panic disorder might make the primary appraisal "I'm having a heart attack." After the primary appraisal, an individual makes a "secondary appraisal," followed by a series of "reappraisals." Secondaty apprais­ als involve the assessment of a person's resources for dealing with a potentially dangerous event or situation. The following are examples of positive secondary appraisals that enable an individual to cope effectively: "I have been able to handle sihiations like this before." "My heart is strong." "These are m y friends; I don't need to worty about them." Reappraisals involve "reality testing" and the constrac­ tion of risk:resources ratios. Anger and Anxiety 259 The central core of the anxiety process is the individual's chroni sense of vulnerability (i.e., uncertainty and insecurity). Anxious indi­ viduals tend to have core beliefs about their own helplessness that are activated in potentially threatening situations; that is, they char­ acteristically underestimate their ability to cope with the threat. As a natural consequence of this appraisal process, the patient with an anxiety disorder will experience an escalation of symptoms of anxiety: increased heart rate, sweating, shortness of breath (i.e., psychological symptoms); increased fear or terror (i.e., emotional symptoms); and increased ramination or obsessive thinking (i.e., cog­ nitive symptoms). The addicted person may respond to these symp­ toms by using psychoactive substances. For example, a person with a flying phobia may use alcohol as a sedative in order to cope with a necessary flight. A social phobic may use cocaine in order to feel more confident in a social siraation. A person with generalized anxiety dis­ order may smoke cigarettes in order to relax (in spite of the fact that cigarette smoking simultaneously creates autonomic stimulation and sedation). Treatment of the Anxious Substance Abuser Case Conceptualization An essential step in the cognitive therapy of the anx­ ious substance abuser is the case conceptualization. Similar to the freatment of other disorders, the therapist evaluates the patient's typi­ cal cognitive, behavioral, and emotional responses to relevant expe­ riences. The therapist traces these responses to earlier life experiences (e.g., famihal, social, educational) in order to understand their devel­ opment. An accurate case conceptualization ultimately facilitates the selection of appropriate techniques. For example, "Rick" is a 38-year-old unmarried accountant with a history of alcohol dependence and cocaine abuse, who entered treat­ ment as a result of a recent "driving while intoxicated" conviction. He explained that he had been drinking between one and two six- packs of beer per night for several years. He typically drank in the evenings until he would pass out while watching television. Rick admitted to blackouts, absenteeism, hangovers, and severe guilt related to his drinking. In an effort to develop an accurate and useful case conceptual­ ization. Rick's therapist inquired about his current functioning, his problematic situations, his developmental histoty, his basic beliefs and typical automatic thoughts, and his behavioral and emotional 260 COGNITIVE THERAPY OF SUBSTANCE ABUSE responses to these thoughts. The therapist related these to the de opment and maintenance of Rick's alcohol use. Rick described himself as a "loner." He stated: "I wish I had close friends, but I just get so nervous around people. I would like to get married some day, but I'm terribly uptight around women." Rick specifically described the following thoughts: "I have nothing intel­ ligent to say to people," "I will appear stupid," and "People, and especially women, will reject me." Rick occasionally dated women; however, he reported that he had to get intoxicated to tolerate the anxiety triggered by dating. Rick admitted to occasional panic attacks prior to social encoun­ ters. He viewed these panic attacks as supporting his beliefs about his extreme vulnerability. In fact. Rick dreaded all social encounters and he reported that his heavy drinking began when he was required to conduct financial audits for his company (requiring interpersonal contact with numerous business executives). He explained that his fear of criticism and failure, and his resulting anxiety, seemed to be intolerable until he began drinking. Although he was self-medicating with alcohol. Rick's symptoms of anxiety heightened as he developed a tolerance for alcohol. He viewed himself as increasingly fragile and vulnerable, until he was avoiding most social situations. As he reflected on his life, he felt increasingly depressed, typified by such global negative beliefs as "1 am totally worthless," "1 am a loser," and "1 am hopeless." Eventu­ ally, Rick began using cocaine to cope with his boredom, loneliness, and isolation. His cocaine abuse further supported his global nega­ tive beliefs (e.g., "Now I'm reaZ/y worthless!"). Upon inquiring about Rick's developmental history, his therapist learned that Rick's parents were "withdrawn and emotionally unre­ sponsive," only showing appreciation when he received perfect grades in school. Regarding his family history of alcohol use. Rick explained that his parents were "fundamentalists who never drank a drop." Their beliefs about alcohol were global and negative, including "Drinking is evil" and "No booze is good

booze." Such messages from his par­ ents taught Rick to have dichotomous beliefs about alcohol. For example, he believed, "If 1 am going to drink at all, 1 might as well get totally drunk." As a result of his extreme thoughts about alcohol, Rick had difficulty drinking in moderation. Rick's anxiety, loneliness, and isolation made him vulnerable to heavy alcohol and cocaine use. He responded to his symptoms with thoughts such as "1 need a drink or snort to handle this!" "What the hell!" and "Who cares if 1 use?" These thoughts resulted in urges and cravings to which Rick responded with permissive thoughts such as Anger and Anxiety 261 "I deserve this!" and "Besides, it doesn't really matter!" Following this permission, he focused o n the instrumental strategies necessaty for the acquisition of beer or cocaine. This often led h i m to drink until he passed out or to use cocaine until he had none left. The sequence for Rick's alcohol use is illustrated in Figure 15.1. A similar sequence occurred in his cocaine use. Educating the Patient about Anxiety a n d Substance Abuse Although most anxious individuals are somewhat aware that their fears are exaggerated, they continue to fear that some­ thing unpleasant will actually occur. In cognitive therapy the patient is educated in an alternative, more objective method for understand­ ing his or her problems. Specifically, the therapist teaches the cogni­ tive therapy models of anxiety and substance abuse to the patient as these models relate to the patient. Rick was taught that his anxiety resulted from his perception that he was weak or vulnerable. Further, the therapist taught Rick that his drag-related beliefs, automatic thoughts, and core beliefs fed into his Situation Basic Alcohol- Anxiety, Related Automatic Craving/ loneliness, and Beliefs Thoughts urges isolation Activated •What the "1 need a heil!" drinic to "Who cares?" handle this!" (... because I'm socially inept) Continued Focus on Facilitatiing Use or Instrumental Beliefs Relapse Strategies (permission) Drinl<s until (action) "I deserve he passes Goes to a this!" out store. "Besides it Buys some really doesn't beer. matter!" F I G U R E 15.1. Rick's alcohol use. 262 COGNITIVE THERAPY OF SUBSTANCE ABUSE drag use, more than did the environment or his circumstances. The therapist further taught Rick that there were multiple decision points at which he could control his substance abuse, including modifying the anxiety-arousing beliefs, rationally responding to permission-giving beliefs, and finding non-drag-related ways to cope with the anxiety. Techniques for Managing Anxiety There are numerous techniques used in the treatment of the anxious substance abuser. The particular technique chosen depends on the patient's presenting problem at each session, as well as the therapist's general conceptualization of the case. The efficacy of selected techniques depends on several factors: the quality of the therapeutic relationship, the accuracy of the case conceptualization, the therapist's appropriate use of the Socratic method, and the patient's socialization to the cognitive model. In the Socratic method, the therapist asks questions to guide the patient to more realistic conclusions about himself, his personal world, and his future. For example. Rick initially held the belief "I will be rejected if I allow others to see how anxious I am." At the same time, he held the belief "My mistakes are terrible." Using the Socratic method, the therapist asked questions to lead Rick to more adaptive conclusions: TH: What do you think about when you anticipate an upcoming social encounter? PT: I know that I'll make a fool of myself. TH: What do you mean by "a fool"? PT: I mean that I'll make an idiot of myself. TH: H o w do you picture yourself making "an idiot" of yourself? PT: I see myself having nothing to say. TH: Do you believe that all people who are quiet in public are "fools," and that they make "idiots" of themselves? PT: Well, I guess not. TH What else are other people likely to think about a quiet person? PT They might think that the person is just listening ... or that the person is just shy. TH: When you think of yourself as an "idiot" or a "fool," how do you feel? PT: Upset, nervous, uptight, depressed . . . Anger and Anxiety 263 TH: And when you view yourself as "just shy," rather than foolish and idiotic, how do you feel? PT: A little better. Prior to this dialogue, the patient equated "quietness" or "shy­ ness" with "foolishness" or "idiocy." Through the Socratic method, the therapist helped the patient to objectively examine this maladap­ tive assumption. There are five key questions that assist patients in responding rationally to the anxiogenic thoughts that may stimulate cravings and urges for drags (these questions are applicable to all patients across the diagnostic spectram): (1) What other perspectives can I take about this situation? (2) What concrete, factual evidence supports or refutes my automatic thoughts? (3) What is the worst that could happen in this situation and how would it ultimately affect m y life? (4) What constractive action can I take to manage this situation? and (5) What are the pros and cons of m y continuing to have negative automatic thoughts? W e summarize the rationale for each question here. 1. Other perspectives. Here, patients who might otherwise see thing in tunnel vision are asked to broaden their viewpoints in order to ascertain new possibilities. One way to facilitate this process is to ask the following questions: "Could there be a blessing in disguise here?" or "How might this situation not be as bad as it first seems?" This approach helps patients think more divergently and thus opens their minds to more constractive alternatives. 2. Evidence. Patients need to learn that their opinions are not necessarily synonymous with facts. Therefore, therapists teach them the importance of looking for information that can be objectively con­ firmed or disconfirmed. Therapists ask patients for evidence for and against their automatic thoughts because they want patients to be evenhanded in their evaluation—even if that means finding evidence that supports their negative impressions. One useful way to commu­ nicate the nature of "evidence" is to say, "If it wouldn't stand up in court, it's not factual enough to count as evidence." 3. What's the worst that could happen? This is not the rhetorical question that people sometimes ask sarcastically when they are mini­ mizing someone's problems. This is a factual question that asks patients to consider a realistic worst-case scenario and then to evalu­ ate the actual implications for their lives. In general, this question serves to decatasfrophize many patients' extreme worries, as they come to realize that the worst-case scenario is highly unlikely, and that they can cope with the outcome in any event. 264 COGNTTIVE THERAPY OF SUBSTANCE ABUSE 4. What action can I take? Even if patients find that their auto­ matic thoughts are borne out, this question reminds them that they can actively engage in problem solving. By asking this question, patients are more likely to apply higher cognitive processes such as planning and decision making, and less likely to lapse into lower-level catastrophizing. 5. Pros and cons of automatic thoughts. Some patients seem wed­ ded to their negative, dysfunctional ways of thinking. In such instances, it is useful to ask them what benefits (and drawbacks) they derive from such negativistic thinking. This tactic defuses the patients' resistance in that it respects their reasons for maintaining their points of view, however self-defeating they may be. For example, we have encountered the following "reasons" that patients are unwilling to relinquish their automatic thoughts: "I don't deserve to think positively." "Every time I get m y hopes up disaster strikes." "If I think the worst, I won't be disappointed when it actually happens." "Nobody will take me seriously unless I'm really upset." Understanding these beliefs helps therapists to conceptualize patients' core assumptions, helps establish rapport, and may provide a window of opportunity to help these patients find some advantages to improving the adaptive quality of their thinking. An example of Rick's completed Daily Thought Record (DTR) can be seen in Figure 15.2. Any and all of the five questions summarized above can be used effectively within the format of the DTR in order to generate rational responses. Not all of the questions need to be used—note that Rick's DTR seems to address the questions "What other perspectives can I take?" and "What constructive action can 1 take?" Behavioral Techniques Behavioral techniques are particularly useful in the treatment of anxiety disorders. Relaxation and assertiveness training (Alberti & Emmons, 1974; Collner & Ross, 1978), for example, pro­ vide the individual with behavioral skills for coping with potentially threatening situations. Relaxation skills involve the systematic, pro­ gressive relaxation of various muscle groups (Bernstein & Borkovec, 1973) and controlled breathing (Clark, Salkovskis, & Chalkley 1985), which allows the individual to be more physically comfortable dur­ ing times of physiologic distress. Assertiveness training involves learn- Directions: W h e n you notice your mood getting worse, asi< yourself, "What's going through m y mind right now?" and as soon as possible jot down the thought or mental image in the Automatic Thought Column. SITUATION AUTOMATIC THOUGHT(S) EMOTION(S) RATIONAL RESPONSE OUTCOI^E Describe: 1. Wrife automatic thought(s) 1. Specify sad, 1 Write rational response to automatic thought(s). 1. Re-rate belief 1. Actual event leading to that preceded emotion(s). anxious/ 2. Rate belief In rational response 0-100%. in automatic unpleasant emotion, or 2. Rate belief in automatic angry, etc. thought(s) 2. Stream of thoughts, day­ thought(s) 0-100%. 2. Rate degree 0-100%. dreams or recollection. of emotion 2. Specify and leading to unpleasant 0-100%. rate emotion, or subsequent DATE/ 3. Distressing physical emotions TII^E sensations. 0-100%. Friday Alone at home. "I'll mal^e a fool of Nervous "I'm generally quiet when I go out, 1. -Fool" myself." (75%) not foolish." (25%) Coworkers have invited (75%) (100%) 2. "Relieved" me to meet them at a (65%) restaurant. "I don't know how to Anxious "I know how to be polite toward 1. "Don't know act around people." (85%) others." how to act" (85%) (90%) (25%) 2.Relieved (65%) "111 be alone for the Sad "If I take risks, I'll meet someone 1. "Alone for rest of my life." (45%) eventually." life" (50%) (75%) (35%) 2. Hopeful (50%) Questions to help formulate the rational response: (1) What is the evidence that the automatic thougsh tt rue? Not true? (2) Is there an altemative explanation? (3) What's the worst that could happen? Could 1 live through it? What's the best that could happen? What's the mosf reaiistic outcome? (4) What should 1 do about it? (5) What's the effect of my believing the automatic thought? What could be the effect of changing my thinl<ing? (6)f was in this situation and had this thought, what would 1 tell him/ner!" (friend's name) F I G U R E 15.2. Rick's Daily T h o u g h t Record.. 266 COGNITIVE THERAPY OF SUBSTANCE ABUSE ing direct, adaptive expressive communication. As a result of incr skills, individuals develop a greater sense of competency. The learn­ ing of behavioral skills is facilitated by the use of covert rehearsal (imagety) and overt practice (role-play, behavioral experiments). Rick's therapist, for example, determined that Rick was particu­ larly vulnerable to feelings of inadequacy and rejection in social situ­ ations. Specifically, social situations tended to activate Rick's core beliefs about his vulnerability, which in turn activated his autonomic nervous system. His anxiety symptoms (perspiration, racing heart, etc.) then escalated into panic attacks, which he self-medicated with alco­ hol. To compensate for his feelings of vulnerability. Rick occasion­ ally used cocaine. Ultimately, Rick found that he had great difficulty controlling his cocaine use, which in turn exacerbated his feelings of vulnerability. In order to treat Rick's anxiety-related substance abuse problems effectively, his therapist began with progressive muscle relaxation and controlled breathing exercises. Such exercises prepared Rick for deal­ ing with the physiological arousal associated with panic attacks. After Rick had mastered relaxation, his therapist asked him to imagine a series of threatening social situations. O n becoming physiologically aroused at the thought of such situations. Rick practiced relaxing to decrease his autonomic arousal. Eventually, Rick began to modify his thinking from "I can't handle m y anxiety" to "I

can at least calm myself to avoid having a panic attack." Rick's therapist also engaged him in assertiveness training in order to provide him with essential communication skills. SUMMARY Unnecessary or exaggerated anger takes its toll not only on other persons but also on the person who is angered. Sub­ stance abusers are prone to act out their hostile impulses when they are under the influence of drugs or alcohol. Although they are par­ ticularly prone to use or drink to dampen the tension associated with anger, their substance use may, paradoxically, increase the likelihood of hostile behavior. Much of the generated anger is the result of the symbolic meaning attached to the provoking event, specifically, the notion "I have been wronged." Addicted individuals with low frustration tolerance (LFT) are hypersensitive to any blocking of their wishes or actions and are consequently prone to experience excessive and inappropriate anger. They fiequently have underlying beliefs such as "If I can't get [or do] Anger and Anxiety 267 what I want right now, I never will" or "Others are wrong to refuse me or get in m y way." Stemming from the other two beliefs is the belief "People are wrong and should be punished for blocking me." When activated, this belief leads to anger. Patients with LFT perceive themselves as powerless or helpless. The generation of anger helps restore a sense of control and power by punishing other people. Whether or not they act out their impulses to punish, however, they are more or less stuck with the anger, which can then lead to using or drinking. Additionally, the physiologic, cognitive, affective, and behavioral correlates of anxiety may place an individual at increased risk for substance abuse. W e presented some methods for conceptualizing LFT and anxious substance abusers, as well as some methods for inter­ vening with people who respond to LFT and/or anxiety with substance use. C H A P T E R 1 6 C o n c o m i t a n t P e r s o n a l i t y D i s o r d e r s T , he term "dual diagnosis" has been used widely to refer to the coJexLisht ence of substance abuse and other psychiatric disorders (Brown et al., 1989; Evans & Sullivan, 1990; O'Connell, 1990). In Chapters 14 and 15 (this volume) we discussed the dual diagnosis of substance abuse and major psychiatric syndromes (e.g., depression and anxiety). In this chapter we discuss the treatment of patients with dual diagnoses involving substance abuse and concomi­ tant personality disorders. Personality disorders (represented by Axis II of the DSM-llI-R clas­ sification system) (American Psychiatric Association, 1987), consist of longstanding affective-behavioral-cognitive patterns that are rigid, maladaptive, and resistant to modification. There have been numerous studies documenting the high preva­ lence of personality disorders among substance abusers. Regier et al. (1990), in their important comorbidity study, reported on data from over 20,000 subjects in the Epidemiologic Catchment Area study spon­ sored by the National Institute of Mental Health. These investigators found that of those with alcohol problems (abuse or dependence), 14% met criteria for antisocial personality disorder (ASPD). They found ASPD among those who abused marijuana (15%), cocaine (43%), opiates (37%), barbiturates (30%), amphetamines (25%), and halluci­ nogens (29%). In accord with these figures, many researchers acknowl­ edge a strong relationship between ASPD and substance abuse (e.g., Grande, Wolf, Schubert, Patterson, & Brocco, 1984; Helzer & Ptyzbek, 1988; Hesselbrock, Hesselbrock, & Stabenau, 1985; Hesselbrock, Meyer, & Kenner, 1985; Lewis, Robins, & Rice, 1985; Penick et al., 1984; Ross 268 Personality Disorders 269 et al., 1988; Schuckit, 1985; Stabenau, 1984; Wolf et al., 1988). Est mates of the lifetime prevalence of ASPD disorder in substance- dependent individuals vaty between 2 0 % and 50%. The high preva­ lence can be explained, to some extent, by the fact that substance abuse is one of the criteria for the diagnosis of ASPD. Several studies conclude that a wide range of personality disorders coexists with substance abuse. For example, Khantzian and Treece (1985) studied 133 opiate addicts and found that 6 5 % met criteria for at least one coexisting personality disorder. These investigators reported that "virtually the entire range of personality disorders [was] represented in [our] sample" (p. 1071). Drake and Vaillant (1985) evaluated 369 middle-age inner-city men, followed in a longitudinal study, for the presence of personality disorders and alcohol problems. Of those who had been alcohol dependent, 3 7 % had a concomitant personality disorder. Borderline personality disorder (BPD) is also commonly associ­ ated with substance abuse (Koenigsberg, Kaplan, Gilmore, & Cooper, 1985; Nace, Saxon, & Shore, 1983; Zanarini, Gunderson, & Franken- burg, 1989). These authors report consistently that among drug and alcohol abusers, BPD is second only to ASPD as a concomitant per­ sonality disorder diagnosis. Beck et al. (1990) have applied cognitive therapy to the treatment of personality disorders. The authors describe these patients as "often the most difficult in a clinician's caseload" (p. 5). Personality-disor­ dered patients typically share some of the following common features that make them especially challenging to treat: 1. Their most chronic symptoms are ego syntonic; that is, the patients do not perceive that there is anything substantially "wrong" with their personality. Although they may come into treatment in order to receive help for depression or anxiety problems, they rarely seek help for problems such as self-centeredness, avoidance of respon­ sibilities, lying, lack of empathy for others, lack of conscience, ten­ dency toward violence, interpersonal manipulativeness, defiance of authority, and other chronic aspects of their character that reflect common Axis II disorders. 2. Their behaviors and attitudes typically are noxious to others. Although these patients may feel that they are in emotional pain, it is common to find that they cause others a great deal of hardship as well. An example is a patient who bemoans the fact that his wife does not give him respect, but conveniently neglects to mention to the therapist that he steals her money, beats her, and cheats on her. 3. These patients are extremely resistant to change. Although they 270 COGNITIVE THERAPY OF SUBSTANCE ABUSE may enter therapy hoping to obtain some relief from their psycho­ logical suffering, they often do not wish to take an objective look at their own shortcomings, nor do they wish to alter their maladaptive behaviors or attitudes. An example is a patient who implicitly expects the therapist to offer him a "magic pill" to make him better, but balks at the idea of dealing directly with his problems at work and at home. 4. Personality-disordered patients have difficulty imagining being any other way. W h e n therapists tty to elicit their cooperation in making important changes in life, the patients often respond by say­ ing, "But this is who I am. I can't change," or "I've always been this way. H o w can I be any different?" or "If you tty to change me, I'll cease to exist." For these patients, their problems often are synony­ mous with their identities. Therefore, they resist change as staunchly as they would fend off personal annihilation. When a personality disorder contributes to drug use the pattern becomes more rigid and compulsive. Once the drug use has begun, the personality-disordered patient may be more likely to continue the pattern of drug use until it becomes a full-blown addiction. For example, the avoidant personality patient who cannot tolerate emo­ tional discomfort may receive temporary respite from his or her upsetting thoughts about their life situations via the use of crack cocaine. As a result, this patient may choose to continue using (and ultimately abusing) drugs as the preferred method to avoid facing up to problems. Furthermore, the patient who suffers from a concomitant person­ ality disorder will be at heightened risk for relapse following a period of successful abstinence. As a case in point, we have seen a border­ line patient relapse on drugs as a deliberate form of self-destructive- ness and interpersonal manipulation in response to arguments with a lover. Another patient, a woman who was diagnosed as having dependent personality disorder, believed that she would not be able to survive unless she had a man in her life. Consequently, she sought out male companionship at all costs, even if the men involved were active substance abusers. Each time she became involved with a new lover, she would subjugate her needs to his and would join him in his drug use. Thus, she knowingly and voluntarily lapsed back into drug use, even after having struggled so hard in therapy to achieve abstinence. Another ramification of patients' personality disorders is their capacity for cooperating and collaborating with the therapist (Carey, 1991). A substance abuser who does not suffer from a personality disorder may be an amiable and earnest patient with w h o m to deal Personality Disorders 271 in session after he or she has been free from drug use for some tim By contrast, we have seen patients who seem to be staying away from dmgs successfully but who have been extremely difficult to manage by virtue of their personality disorders. For example, "Ken" is a parole- office-referred patient who also has been diagnosed as having both paranoid and passive-aggressive personality disorders. In each therapy session, he adamantly insists that he has abstained from drugs in an ongoing fashion, and therefore concludes that his participation in treatment is "an idiotic waste of m y time." W h e n the therapist attempts to engage Ken in productive dialogues about the manage­ ment of general life stressors and maintenance of sobriety, he becomes reticent and projects an air of suspicion. At present, he remains drug- free, but he is not engaged in treatment to an optimal degree. His long-term prognosis, therefore, remains guarded. W e do not mean to imply that the treatment of drug abuse patients with concomitant Axis II disorders is an impossible task, nor do we mean to say that therapy with drug abuse patients without personality disorders is a breeze. The truth, obviously, is less dichoto­ mous and more complex. Nevertheless, it is vital that therapists per­ form a thorough diagnostic assessment of each substance abuse patient, so that the difficulties associated with the treatment of per­ sonality disorders may be anticipated. Treatment plans that address both the patients' drug abuse problems and characterological issues are better equipped to keep patients in treatment, and to prepare them for self-maintenance after regular sessions have been terminated (N.S. Miller, 1991). A S S E S S M E N T Personality disorders have been described in a num­ ber of ways. One simple formula that clinicians have followed over the years states that patients who have personality disorders cause as much or more grief for others in their lives as they suffer themselves (Cummings, 1993). Obvious examples of this are the narcissistic patient, who may be so blindly self-absorbed that he or she completely neglects the physical and emotional well-being of spouse and chil­ dren, and the ASPD patient, who may be having a "grand old time" while he or she lies, connives, and cheats in order to achieve his or her own ends. Another example is the borderline patient, whose affective lability, self-destructive impulsivity, and excessive interper­ sonal demands cause much consternation for significant others and for therapists. 272 COGNTTIVE THERAPY OF SUBSTANCE ABUSE If we use this rule alone, however, most substance-abusing patients will look as if they necessarily have personality disorders as well (Carey, 1991). When patients are in the throes of compulsive, addictive drug abuse, the following maladaptive behaviors are not uncommon: (1) stealing from family members, friends, or strangers in order to fulfill a desperate need to purchase and use drugs (symp­ tom of ASPD); (2) inveterate lying, in order to cover one's drug-abus­ ing tracks (symptom of ASPD); (3) almost delusional suspiciousness, a side effect of heavy usage of cocaine and/or amphetamines (symp­ tom of paranoid personality disorder); (4) angry outbursts and vio­ lence toward self and others due to overdose or withdrawal symptoms (symptoms of borderline, paranoid, and antisocial personality disor­ ders); (5) withdrawal from social activities into solitary drug use (symp­ tom of schizoid personality disorder); and (6) progressive self-absorp­ tion,

such that responsibilities toward others become ignored (symptom of narcissistic personality disorder), as well as numerous other behaviors. As we can see, the particular problems that accom­ pany the Axis I substance abuse disorder often look suspiciously like— and need to be distinguished from—full-blown Axis II disorders (cf. Gawin & Kleber, 1988). In order to clarify the assessment picture, we must look to the criteria of DSM-III-R, and to the diagnostic questionnaires that map onto the DSM-III-R, such as the Structured Clinical Interview for the DSM-IIl-R (SCID: Spitzer et al., 1987). Here, we are instructed to examine the enduring personality traits that exist and persist apart from the primary Axis I diagnoses. In this case the question becomes, "What personality traits were present during those times in the patients' lives when they had not yet begun to abuse drugs?" This question is espe­ cially useful when the substance abuse problem has an adult onset, and therefore there is an extensive premorbid histoty. Another useful source of assessment data comes from previous treatment experiences of patients. It is potentially enlightening to review notes and reports of patients' former therapists, as well as to ask patients themselves about their recollections of therapy contacts (and outcomes) in the past. Such information can shed light on the degree of chronicity of patients' drug problems, "changes" in person­ ality over the course of time as a result of drug use, and clues as to the kinds of interventions that were helpful or unhelpful. Another useful question is, "What personality traits are present during prolonged periods of abstinence, between drug use episodes?" Such an assessment question is salient when abstinence is the norm, punctuated by problematic lapses into drug bingeing. A simple interpretation of DSM-III-R suggests that a serious con- Personality Disorders 273 comitant Axis II disorder is less likely to be present when dysfun tional beliefs and behaviors did not predate the onset of the substance abuse disorder, and when they diminish or disappear during times of prolonged abstinence. The non-Axis-II-disordered patient is iden­ tifiable as one who seems to "change personality" dramatically when under the influence of psychoactive substances or cravings and urges associated with their abuse. This patient may seem quite compliant with the treatment regimen, and may relate in a friendly, coopera­ tive manner with the therapist when he or she is clean and sober. The same patient may break appointments, drop out of sight, and avoid or otherwise disregard the therapist when he or she resumes dmg use, and may engage in activities that seem to be completely "out of character." For example, one of our patients demonstrated herself to be a conscientious, hard-working, likable person when she had been free of dmgs for some time. Unfortunately, when she suffered a relapse, she resorted to thievery in order to obtain money for her habit. The crimes were not malicious but rather a matter of expediency—her values and priorities shifted dramatically when she used drugs (cf. Woody, Urschel, & Alterman, 1992). Family and friends (and the thera­ pist) were shocked when these facts came to light. Based on knowledge of her crimes of theft, and subsequent lies to cover up, this patient could be easily mistaken as having antisocial personality disorder. In actuahty, she did not meet DSM-III-R criteria for any of the Axis II diagnoses. W h e n she was abstinent from drugs, her most salient maladaptive beliefs were depressogenic, but not in­ dicative of longstanding characterological issues. As an example, she believed "I'll never be able to find a job in which I will be trusted to be in a position of responsibility again," and "Something always seems to set me back just when I think I'm getting on m y feet again." On the other hand, while she was actively using, this patient's beliefs became even more problematic (i.e., more similar to those espoused by Axis II patients; see Beck et al., 1990), such as "I have to do what I have to do, no matter what anybody tells me" (similar to beliefs held by ASPD and passive-aggressive patients), and "I have to use drugs because I can't bear to face life in a straight frame of mind" (similar to beliefs put forth by avoidant personality disor­ der patients). After successfully completing an inpatient detoxification program in a local hospital, this patient returned to cognitive therapy. She no longer gave credence to the beliefs that she had maintained during her relapse episode, and she once again became a vety compliant and actively involved outpatient. The depressive beliefs persisted, however. 274 COGNITIVE THERAPY OF SUBSTANCE ABUSE and became ongoing issues, along with a focus on drug relapse pre­ vention. In other cases, it is not necessary for patients to be in a phase of active drug use for them to evince dysfunctional beliefs and behav­ iors that earmark them as personality-disordered patients. For example, when "Lee" was going through his initial intake assessment, he ad­ mitted to abusing amphetamines, and highlighted a myriad of life problems. When asked how the drug use had led to the various life dilemmas, he quickly corrected the intake therapist, saying, "Speed doesn't cause all m y problems. I can get into all kinds of trouble, even if I'm not on the stuff, beheve me!" In this case, the substance abuse coexisted with a number of personality disorders, which would com­ plicate the drug abuse treatment and would need attention in their own right. To identify the presence of a personality disorder in a substance- abusing patient requires a careful evaluation of the patient's drug-free beliefs and behavior patterns, and a comparison of these with beliefs and behaviors that are activated by psychoactive substances. A gen­ eral rule of thumb holds that similarities between drug-free and dmg- using beliefs and actions suggest a high probability of an Axis II com­ ponent, while marked divergences give hope that the most noxious aspects of the patient's functioning will be extinguished if the dmg- taking is brought under control. For example, a patient who main­ tains the belief that he must never follow anyone's rules but his own— and espouses this view whether or not he is drug free—is likely to demonstrate antisocial personality traits that will complicate tteatment even if he is free from drugs. By contrast, a patient who typically believes that it is important to be an honest, responsible, cooperative person—that is, until she submits to her urges to use crack cocaine, whereupon she adopts an "I don't give a damn about anything" atti­ tude—is more likely to make significant therapeutic strides in all aspects of her life if she overcomes her addiction. Unfortunately, this assessment process is fraught with a number of difficulties. First, the optimal method of assessing drug-taking beliefs and actions is not readily apparent. W e have noted earlier in this volume (Chapter 12) that the patient's "hot" cognitions are most salient to assessment and treatment, and that such affect-laden thoughts are highly accessible under conditions that closely simulate the target problems in question. In the case of the patient who suf­ fers from a disorder such as panic, it is relatively safe and straightfor­ ward to contrive therapeutic situations that will elicit panic attacks in session, thus allowing the patient and therapist to have access to the hot cognitions. Personality Disorders 275 In the case of the drug-abuser, however, it is neither wise nor ethical to encourage the patient to engage in substance-taking as a means of gaining access to key cognitions. Instead, we must use more indirect methods. One method involves having patients rely on free recall to remember what they believed and what they did while in the active phase of drug use. This technique is simple and safe, but may be subject to inaccuracies, owing to patients' memoty distortions and deliberate confabulations. Another method involves gaining data from more objective sources, such as the verbal reports of family members and the writ­ ten reports of previous therapists and legal officers. Although these data help the therapist to ascertain the maladaptive actions of the drug- taking patient, they do not facilitate gaining an understanding of the subjective phenomenology of the patient who is under the influence of dmgs. W e have found that the two approaches explained above are most informative when combined with clinical data that are obtained through provocative imagety exercises, such as patients' closing their eyes and imagining the houses where they use crack. These exercises, described earlier in this volume (Chapter 10), provide patients with the kinds of covert stimuli that will likely produce a reasonable experience of their hot cognitions. These techniques, free recall, third- party behavioral reports, and in-session imagery exercises, combine to help therapists to assess patients' drug-using beliefs and behaviors. As these data are obtained, it becomes more likely that therapists will be able to separate those aspects of patients' functioning that are part of a consistent personality characteristic from those aspects that are dmg-related per se. Still, there are further diagnostic problems. For example, when a 35-year-old drug-abusing patient notes that he has used hard-core dmgs such as cocaine since the age of 15, the therapist is hard-pressed to view the patient's drug-related beliefs and actions as anything other than a major part of his personality characteristics. Here, there may not be a single person (the patient himself included) who could guess how the patient would function if he were not on drugs. In such cases, the line of distinction between Axis I and Axis II becomes quite clouded, and therapy may have to involve the teaching and nurtur­ ing of fundamental adaptive ways to view the self, the world, and the future practically from "scratch." Yet another problem in distinguishing personality issues from substance-abuse issues involves the medical complications of long- term drug abuse. Specifically, it is likely that prolonged, heavy use of certain psychoactive substances causes progressive damage to the 276 COGNITIVE THERAPY OF SUBSTANCE ABUSE central nervous system (Karan et al., 1991; Nace & Isbell, 1991; O'Connor et al., 1992). As time goes on, rather permanent changes in personality may result from the abuse of drugs, such that even complete cessation of drug use may still leave behind the residual characterological dysfunction. As an example, one of our patients ended a 20-year habit of daily marijuana use after his wife threatened to leave him. Although he succeeded in stopping cold turkey, the patient then was dismayed that he began to feel perpetually depersonalized, to the point that he was "unable to focus m y atten­ tion on anyone but myself and m y own feelings." In a sense, the after­ effects of the marijuana use seemed to be a narcissistic-schizoid per­ sonality style. This serious problem, along with his chronic anxiety, became the chief foci of therapy over the course of the next six months. Needless to say, there was a concomitant struggle to keep him from returning to regular marijuana use as well. In sum, it is difficult to separate substance abuse symptoms from long-standing personality issues and disorders. Our experience has taught us that the majority of all substance-use cases will necessitate some sort of therapeutic focus on Axis II disorders, either as a con­ tributory cause or as a consequence of the problematic taking of dmgs. The best barometer for the concomitant presence of personality dis­ orders is an assessment of the patient's beliefs—about drugs, and about the self, personal world, and future (the standard cognitive triad; cf. Beck, 1976), when the patient is free of psychoactive substances. BELIEFS THAT FACILITATE DRUG-TAKING ACROSS DIFFERENT PERSONALITY DISORDERS Beck et al. (1990) have outiined common beliefs that are held by patients who suffer from the various personality disorders as described by DSM-III-R. While it is beyond the scope of this text to describe fully each and every one of these disorders, along with their respective lists of beliefs, it is interesting to highlight some of the beliefs that feed into drug-taking. In each of the following examples, the treatment for drug addiction must attend to these con­ comitant maladaptive cognitive sets. Many of our drug-using patients are chronically anxious and fear­ ful, resorting to drugs as a way of feeling a greater degree of "safety" or "confidence." ̂ Ivoidant personality patients, who

experience chronic cross-situational, exaggerated feelings of threat to their physical and psychological well-being, fall into this category. They maintain beliefs that make the use of drugs seem quite enticing, certainly more Personality Disorders 277 attractive than facing up to (and solving) real problems. Such bel include "I cannot tolerate unpleasant feelings" and "If I feel or think something unpleasant, I should try to wipe it out with a drink or a dmg." Another type of anxious and fearful patient is the dependent per­ sonality. These patients do not believe that they can get by without "a little help from their friends" (e.g., drugs, and drug-using cohorts). Their drug-taking is facilitated by implicit assumptions such as "I can't cope as others can" and "I am needy and weak." This mind-set makes a patient easy prey for stimulants such as cocaine, crack, or amphet­ amines, which induce a temporary but powerful sense of confidence and control in the early stages of their abuse. In addition, peer pres­ sure to engage in drug use is magnified greatly when dependent patients think that they must win the approval of their acquaintan­ ces at all costs. Passive-aggressive patients are hesitant to act out their anger directly, instead choosing to ignore the rules set forth by authorities. They believe that it is safer simply to choose not to conform. For these patients, using drugs represents a way to rebel against disapproving members of their families and society, while maintaining erroneously that "It isn't hurting anybody, so what's the problem?" Meanwhile, these patients further neglect the obligations and responsibilities that they resent in the first place. Drug use is also facilitated by obsessive-compulsive beliefs such as "My way of doing things is generally the best way," and "I need to be in complete control." Here, obsessive-compulsive drug-users maladaptively convince themselves that they know what is best; there­ fore the use of drugs "must not be a problem." Additionally, the false sense of omnipotence that cocaine can induce reinforces their mis­ taken notion that they are in complete control of themselves and their life situations. Narcissistic patients believe that they are entitled to special treat­ ment and privileges, and that they do not have to abide by the rules that most others must follow. They believe that they are special and powerful, and that they deserve to have all of their desires fulfilled. Obviously, these implicit points of view contribute to narcissistic patients' feeling free to engage in any behaviors that give them a sense of satisfaction (including the use of illicit drugs) without feeling that they have done anything wrong or unwise. In the case of histrionic patients, drug use is facilitated by their beliefs that "I cannot tolerate boredom" and "If I feel like doing some­ thing, I should go ahead and do it." Furthermore, they want vety much to be the center of attention, something that is more likely to occur 278 COGNITFVE THERAPY OF SUBSTANCE ABUSE if they are acting outrageously as a result of chemical intoxica­ tion. Additionally, the beliefs of patients w h o suffer from paranoid, schizoid, and schizotypal personality disorders potentiate drug use in that these patients are looking for experiences that they can enjoy in solitude, and that help them to feel on their guard against intrusion or attack from others. The Axis II disorders described present significant challenges to the therapist w h e n patients also evidence substance abuse disorders. However, the most difficult and pernicious combinations involve, most notably, the borderline and antisocial personalities, which interact with chemical addictions to form a class of disorders that is substantially resistant to treatment of any kind. Therapy for such noxious composite disorders must involve skilled application of (1) structured sessions, (2) between-session assignments and monitoring, (3) therapeutic relationship-building behaviors, and (4) sophisticated and flexible conceptualizations of patients' problems. In order to approach these objectives, w e alert therapists to the following problems and complications that arise in the course of treatment with such patients. BPD A N D ASPD DRUG-ABUSING PATIENTS: CLINICAL MANIFESTATIONS A N D IMPLICATIONS FOR COURSE OF COGNITIVE THERAPY As separate entities, substance abuse disorders and per­ sonality disorders such as the borderline and antisocial syndromes are difficult to treat. In combination, the clinical picture becomes exfremely challenging indeed. For example, one useful method of eliciting the active cooperation of a dmg-taking patient is to help him or her to realize the long-term personal rewards of abstinence. For most patients, goals that appeal to self-interest are intrinsically atfractive. However, imagine that such a patient has the following schema: "I'm bad, and I deserve to have m y life fall apart." This schema, not u n c o m m o n for a borderiine patient, may motivate the patient to deliberately steer clear of any sfrat­ egy that might bring about personal success or recovery. As a result, the therapist's initial call to look at long-term benefits becomes meaning­ less to this patient. In such a case, the therapist must focus on modify­ ing the extiemely maladaptive beliefs that feed into self-hafred, fear of success, and compulsive self-sabotage. Personality Disorders 279 As another example of the immense difficulties involved in treat­ ing the severely personality-disordered substance-abusing patient, consider the case of a man with antisocial personality disorder. In contrast to the borderline patient described above, self-interest is markedly overdeveloped in the antisocial patient. In turn, the capacity for feeling guilt is profoundly underdeveloped. As a result, the therapist's focusing on the harm that the patient is causing others as a result of the substance abuse loses practically all its motivational leverage. Furthermore, the antisocial patient's thirst for freedom and control, and his or her antipathy for authority (e.g., therapists), often leads such a patient to work extremely hard to fight against the work of therapy altogether. The above examples offer but a glimpse at the dilemmas that face therapists who treat these most problematic of dual-diagnosis patients. W e would like to note that although we present them as separate diagnostic categories, the borderline and antisocial personalities some­ times overlap. In fact, the co-occurrence of multiple Axis II diagnoses in the same patient is a c o m m o n phenomenon (Gunderson & Zanarini, 1987). For the sake of clarity, however, we have chosen to discuss the special therapeutic issues of the borderline substance abuser and the antisocial substance abuser separately. Antisocial Antisocial patients rarely come into therapy of their own accord. They generally do not view themselves as having emo­ tional problems of any sort. Instead, they view others around them as being the source of any problems that they may have ("They're hassling me, ripping m e off, jerking m e around, on m y back all the time, setting m e up," etc.). As a result, their appearance in therapy is likely to have been imposed on them against their will. In the case of substance-abusing antisocial patients, the source of this mandate usually is a legal authority. It is only natural then for such patients initially to view the therapist as part of the "system" that is ttying to oppress them. As a result, patients try to assert control in any manner possible, usually by passive resistance, sometimes by more radical means such as intimidation. It is especially important at times such as these that cognitive therapists remember to be collaborative, rather than acting controlling in their own right. For example, when "Don" told one of us that he resented hav­ ing to see a "shrink," the therapist avoided a response with paternal- 280 COGNITIVE THERAPY OF SUBSTANCE ABUSE sounding imperatives such as "Too bad, but you have to be here, buddy!" Instead, he surprised Don by stating, "I know that if it were up to you, you wouldn't be here in this office. You weren't given much of a choice, and that pisses you off. But you know something? We've got something in common. / didn't ask to see you either. You were assigned to me, and I didn't have much of a choice, just like you. So ... since we seem to be stuck with each other, how are we going to make this a tolerable experience for both of us?" This vignette underscores another principle of successfully engaging antisocial patients^eeping them entertained. As they are insatiable stimulus seekers, antisocial patients easily become bored and inattentive in session unless the therapists are willing to be energetic, innovative, and just a little bit confrontational (Doren, 1987). To the extent that therapists can keep their antisocial patients stimulated, the patients will find the sessions engaging enough to want to return for fur­ ther appointments. These patients will be more likely to become actively involved in therapy if they feel that their relationships with their therapists pose interesting interpersonal challenges. As a caveat, thera­ pists need to watch out that they do not fall into the trap of being so entertained themselves by the antisocial patient's tales of daring and horror that they are led completely astray—away from the agenda of drug abuse and other patient maladaptive behaviors. Along these same lines, there is a great temptation on the part of therapists to fight to prevent their antisocial patients from gaining any degree of one-upmanship. These therapists believe that the only way to manage deceptive, manipulative patients is to browbeat them and to "show who's boss." This is a mistake. The therapist's job is not to impress the patient with his or her street smarts, nor to attempt to win evety argument. The goal is to keep the patient interested in a therapeutic process that has the best chance of influencing behavioral change (Cummings, 1993; Doren, 1987). This may mean that the therapist has to allow the patient to "win a few." By being reason­ ably flexible, unpretentious, and willing to own up to mistakes or shortcomings in knowledge, therapists stand a fair chance of disabus­ ing their antisocial patients of their commonly held beliefs that thera­ pists are simply into head games and power trips. To illustrate, one patient claimed that his barroom acquaintan­ ces, members of a local motorcycle gang, were responsible for his dmg relapses. He reasoned that they challenged his manhood by daring him to get high with them. W h e n the therapist suggested a few "middle-class, liberal" assertive comments for the patient to tty out on these hoods, the patient just laughed and said "That intellechial shit might work in your neighborhood, but not in mine!" Rather than Personality Disorders 281 defend his position, the therapist admitted he had goofed on the matter, and said, "Yeh, you're right. I'd probably get m y ass kicked if 1 said that to them." The therapist then proceeded to solicit sugges­ tions for snappy retorts from the patient himself, a strategy that was much better received. Another complicating factor in treating antisocial drug abusers is that their persistence in taking drugs throughout the course of therapy (and after treatment ends) seems to be worse than for the "average" substance abuser (Doren, 1987). While many addicted patients seem to decty their own drug use, and genuinely wish to kick the habit, antisocial patients often seem vety much to want to con­ tinue their drug use. Rather than using therapy as a way to stop, they attempt to use their treatment to learn how to conceal it better. This may account, in part, for the generally poor outcome that has been reported in the literature for the ASPD drug-abusing patient (Alterman & Cacciola, 1991; Woody et al., 1990). Similarly, ASPD patients' self- report data are more likely to be falsified than are the average drug patients'. This would include information on their drug history, present-time drug self-monitoring, attendance at support group meet­ ings, and adverse effects on their employment and relationships. For therapists to be best equipped to catch such patients at baldfaced lying, they must take extensive notes. While this is true in treating all patients, it is especially so in treating these most difficult of patients, who weave elaborate webs of falsehoods around unsus­ pecting, well-meaning therapists. By writing complete notes, not only do therapists protect themselves from unnecessary legal risks, they will also be able to spot the inevitable inconsistencies in patients' reports that

will serve as the basis for necessary therapeutic confron­ tations. For example, a patient who missed an appointment with his thera­ pist gave the alibi that, since his driver's license had been revoked, he was dependent on others to drive him to his sessions. He claimed that his "driver" failed to show up; therefore the missed session was not the patient's fault. The therapist documented this in great detail, including the "fact" that the license would not be restored for almost 6 months. One month later the same patient excused his lateness for a session by saying that he could not find a parking space. After determining that the patient had indeed driven himself to the therapist's office the therapist read aloud from his notes of the ear­ lier session when the patient had bemoaned his lack of a driver's license. The patient, dumbfounded and annoyed at having been caught in a lie, admitted that he was driving on a suspended license. W h e n pressed further, he was forced to admit that his lack of a license had 282 COGNITIVE THERAPY OF SUBSTANCE ABUSE been scarcely a deterrent to his driving, and that his absence fro the previous session was due to something else—taking a "detour" to a bar, where he got drunk. The remainder of the session was spent discussing the patient's drinking, his illegal driving, and his avoidance of therapy sessions. Without the detailed therapy notes, this patient's important area of deceit might have gone unnoticed at the time. When therapists challenge patients on the veracity of their state­ ments, they may do so in a respectful, supportive way, but they must not shy away from expressing disbelief where it is called for. The following dialogue serves as an example: TH: This is only the second time you've attended a therapy session in about two months, "Jackie." I'm wondering why it took so long for you to return m y phone calls. Have you been in some sort of trouble? PT: Oh, no! I've just been busy, you know. I'm working double shifts, you know, and I just didn't have time to get back to you, you know. [Therapist noticed that Jackie was overly animated, and won­ dered if she was anxiously withholding important information. Her overuse of "you know" statements was a giveaway that Jackie was pretty wound up.] TH: Jackie, I know that you work very hard at your job. You have always done so. But that never stopped you in the past from returning m y calls, except for those times when you were actively using drugs. PT: Oh, no, no, no! I haven't been using drugs. TH: OK. I want to believe that. But I'm puzzled that you didn't call me for four weeks, even though I left numerous messages. PT: 1 told you. I just didn't have time! TH: Jackie, it's hard for m e to accept what you're saying right now, because there have been many times in the past when you were very, very busy with your job, and with your family's problems, and you still managed to come to therapy sessions every week. You proved to me that you were capable of keeping your life organized and of not letting anything stand in the way of what you had to do. It's something I've always admired about you. PT: (starts to cry) TH: [Therapist wondered what Jackie was thinking as she started to cry, but did not want to get sidetracked just yet. He thought that the patient's tears might be real, but they might also be a Personality Disorders 283 manipulative ploy. Therefore, he decided to continue to press for the truth, albeit in a vety supportive manner.] I'm on your side Jackie. I'm totally on your side, and I want to help you. What kind of frouble have you been in lately with drugs? I really want to help, but you have to be honest with me. PT: {Crying) Everything's all messed up. [Jackie went on to say that she lost her job, and that she felt vety ashamed. She and the therapist spent the rest of the session assessing her degree of drug abuse, hopelessness, and suicidality, and they arranged to have daily phone contacts for the next three days until the next face-to-face session. None of this would have come out had the therapist accepted everything that Jackie had said at face value.] In this example, the therapist was very careful not to communi­ cate an air of self-satisfaction because he had successfully caught the patient in a deception. The goal was to elicit information that could lead to the appropriate assessment and intervention, not to demon­ strate that the therapist had superior savvy or to shame the patient. This underscores an important tenet of working with the antisocial dmg abuser, namely, always treat the patient with respect, even as you are "winning" the struggle for control of the therapeutic agenda. By leaving the patient with a sense of dignity, collaboration is main­ tained and facilitated. Antisocial patients are not altruists. They are not responsive to entreaties to "do the right thing"; they are turned off by the percep­ tion that others are moralizing, and they generally focus only on their own immediate wants and needs (e.g., a drug-based high). Therefore, the therapist can make headway in treatment only to the extent that the patient becomes convinced that it is in his or her best interest to change. Simply put, this involves helping patients to understand the relative costs and benefits of moditying the various aspects of their current functioning. Unfortunately, antisocial patients are notoriously inept assessors of the consequences of their actions. Likewise, they are deficient in ascertaining the consequences of their thoughts, beliefs, and emotions as well. Walter is an excellent case in point. Each time he got a new job, he went out and "celebrated" by getting drunk and then using cocaine. This invariably led to tardiness, absenteeism, and the loss of the job. Walter once again would become depressed, and adopt the attitude "I don't give a shit anymore. Fuck it!" This would in turn lead to more drug use, and the downward cycle would continue. A great deal of work (over a long time in therapy) was required in order for Walter to comprehend the amazingly simple fact that his 284 COGNITIVE THERAPY OF SUBSTANCE ABUSE own behaviors and thoughts were responsible for his downfall. A breakthrough of sorts was achieved when he was able to articulate the idea that his "celebrations" were at the root of his repeated fir­ ings. It took him considerably longer before he realized that the thoughts "I don't give a shit," and "Fuck it!" were also responsible for his continuing self-defeating behaviors. One of the most effective methods of teaching patients to moni­ tor the consequences of their responses is the advantages-disadvan­ tages analysis (Chapter 9, this volume). For example, when a patient experiences a strong craving for drugs, he or she is instructed to examine the advantages and disadvantages for using, as well as the advantages and disadvantages of not using. Similarly, this same tech­ nique may be used for other maladaptive aspects of functioning, such as the pros and cons of (1) acting on an urge to seek revenge, (2) quitting a job, (3) engaging in high-risk sexual behavior, (4) dwell­ ing on anger-producing thoughts, (5) discontinuing therapy, (6) con­ tacting drug-using associates, and (7) adopting an "I don't give a damn" attitude, as well as many other "popular" modes of operation of the drug-taking antisocial patient. As the patient begins to learn to use this technique with some degree of proficiency, he or she can be taught the next level on the hierarchy of difficulty, namely, distinguishing short-term pros and cons from long-term pros and cons. Typically, antisocial patients are driven by short-term rewards, and therefore often act out impulsively in a manner that is decidedly destructive in the long run. They seem to have a block in even thinking about or imagining long-term conse­ quences. It is a difficult task to help these patients to learn to envision the future and to delay gratification. One method is guided imagery exercises, where the therapist describes various outcomes for patients to picture in their mind's eye. Typically, the images depict aversive consequences for patients if they continue to use drugs, consequences such as loss of money, job, family, health, and freedom. It is impor­ tant to make these images as graphic as possible, such as waking up in one's own urine in a roach-infested house, facing an accusatory tirade from the spouse or children, or having handcuffs slapped on in front of the neighbors. As this exercise is repeated, patients are asked to provide more and more of the description themselves. Along with the negatives mentioned above, positive images of the outcomes asso­ ciated with abstinence from drugs should be generated in a detailed fashion as well. Therapists can teach their patients to practice these images every day, to help provide regular incentive to stay away from drugs. Personality Disorders 285 Another technique encourages patients to "test their ability" (an appealing challenge to antisocial patients) to withstand urges to engage in impulsive behavior, such as drug use, violence, unsafe sex, and so on. When they get one such urge, they are instructed to monitor how long they can "bravely" fight off the temptation. While they are doing this, the second part of the skill involves thinking (and writing) of ways that it might be advantageous to "lay low, size up the situation real carefully, stay cool, and bide your time." An important goal here is to appeal to patients' intelligence by suggesting that they exercise their ability to think before they act (out). As a warning, therapists should emphasize to their patients that this technique works best when it is applied to cravings that arise spontaneously. It is inappro­ priate to use this technique by contriving high-risk and craving situ­ ations under the guise of "testing" oneself. The methods described above are most effective when they are incorporated into regular therapeutic homework assignments. W e cannot overemphasize the importance of between-session practice of the techniques and principles that are reviewed in the therapist's office (Newman, 1993; Newman & Haaga, in press). Without such regular application of therapy techniques, much of what goes on in session will go "in one ear and out the other." W e advocate the taking of notes (by both the patient and the therapist) during the session. Fur­ ther, we suggest that patients keep a spiral notepad or folder where they can compile important therapeutic notes and homework assignments as they progress through treatment. In addition, we encourage patients to allow us to audiotape the goings-on in therapy, and for patients to take the tapes home to review the contents of the session again and again. Sometimes, we have increased compliance by asking patients to listen to the tapes in order to remind us where we as therapists "goofed up," and where the patients made their most astute observations in the previous ses­ sion. In addition to their benefits as teaching devices and general interventions, homework assignments serve a useful assessment func­ tion as well. Patients with the higher levels of motivation will be the ones who are most likely to make earnest efforts to complete the assignments. The antisocial patients who are simply "playing the game" of therapy will find dozens of elaborate excuses for fail­ ing to do the homework as assigned (Doren, 1987). In this sense, a patient's willingness to collaborate in the therapeutic process can be gauged via his or her responses to homework assignments. In any event, it is potentially useful to elicit the thoughts and beliefs that inhibited the patient from following through with the assignment. In the case of benign procrastination or lack of under- 286 COGNITIVE THERAPY OF SUBSTANCE ABUSE Standing, this is a rather straightforward task. On the other hand antisocial patient who is clandestinely ttying to subvert the therapy process will be more difficult to pin down. Here, it may be more useful to ask the patient to express his or her complaints about the home­ work as an entty point into the assessment of

thoughts that discour­ aged the patient from doing the assignment. Sometimes the responses are quite surprising, such as one patient who resisted reading a book that we had recommended, even to the point of purchasing it and then throwing it out when he got home. After much careful probing by the therapist, the patient was finally able to explain that he noticed that the author of the book had been employed by the Center for Cognitive Therapy. He then reasoned, "You all just want to make more money by asking people like me to buy your books. You don't really care about helping me\" The fact that he was receiving free treatment in the first place did not seem to fac­ tor into his thinking, nor did the fact that the book was a very inex­ pensive paperback. He was simply primed to assume that "everyone will always tty to get one over on me, if they get the chance." In his way of thinking, he certainly was not going to cooperate with any­ one who was trying to take advantage of him, even if it meant wast­ ing his money in the self-defeating act of throwing away a brand new book. This maladaptive mode of operation then became a focal point in therapy, along with his problematic substance abuse. Although not commonly associated with the treatment of sub­ stance abuse or antisocial personality disorder, assertiveness training (Alberti & Emmons, 1974; Collner & Ross, 1978) is a useful part of treatment with this population. Most of these patients engage in dichotomous, all-or-none social behavior, acting either passively or aggressively. Assertiveness is not typically part of their repertoire. This leads to many obvious problems, such as an inability to say no to peers who want them to use drugs, and inappropriate ways of deal­ ing with conflict situations with employers, significant others, parole officers, and therapists, among other people. Although one may argue that teaching antisocial patients the skills of assertiveness is tantamount to instructing them in ways to be better manipulators of others, we have found that this is generally not the case. These patients wreak far more havoc by being pas­ sive aggressive (e.g., "forgetting" to pay child support) and aggres­ sive (e.g, deliberate high-speed tailgating on the highway), rather than assertive (cf. Doren, 1987). Furthermore, the vicious cycles that are perpetiiated by dysftinctional thoughts that lead to dysftmctional interpersonal behaviors that then lead to interpersonal consequences that further refuel maladaptive thoughts are slowed down when Personality Disorders 287 assertive behaviors prevail. An additional potential benefit of assertive ness training is that it m a y help patients to identify and articulate thoughts and feelings that previously were ignored or deliberately suppressed by the patients. For example, a patient w h o believes that a therapist is trying to control and manipulate her m a y react in a hostile manner that results in greater attempts from the therapist to gain control over the patient's maladaptive behaviors, thus "confirming" the patient's suspicions. O n the other hand, if a patient assertively and directiy tells a therapist that she does not appreciate his heavy-handed therapeutic tactics and wishes that he would be a bit more lenient and understanding, the therapist is more likely to understand the patient's actions in a con- stmctive light. Thus, a potentially destructive power struggle m a y be averted. Borderline Similar to the antisocial drug abuser, the borderline patient tends to act out impulsively, to experience extremely low frus- ttation tolerance, to be deficient in learning from past mistakes, to use drugs in lieu of attempting to cope with stress, and to have diffi­ culty in establishing stable, trusting relationships with caregivers. In contrast to antisocial patients, borderline patients m a y use drugs more out of a sense of hopelessness than as thrill-seeking behavior. In addition, their behaviors are m u c h more deliberately self- destructive than are the behaviors of antisocial patients or other per­ sonality-disordered patients. Owing to the borderline patients' ten­ dency to maintain powerful negative beliefs about themselves, their hopelessness, and poor impulse control, the threat of suicide is far more prominent in the case of the borderline substance abuser. In contrast to antisocial patients, w h o often feel that they do not need to see a therapist, borderline patients more frequently are desperate for help. Nevertheless, they are apt to act in ways that sabotage the process of therapy, owing to their erratic emotionality, their difficulty in using thoughts (rather than raw emotions) to modulate their o w n behavior, and the extreme, conflicting demands that they make on their therapists. Borderline patients pose a peculiar challenge to therapists in that they so readily avoid acting in their o w n best interests. It is one thing to motivate a patient to give up the use of drugs by appealing to an improved quality of life in the future. It is quite another thing to attempt this with a patient w h o states "I don't give a d a m n about what happens to me, n o w or ever, so I might as well do whatever it takes 288 COGNITIVE THERAPY OF SUBSTANCE ABUSE to kill the pain," or who maintains the belief "I hate myself and 1 deserve to have a messed up life." O n the other hand, borderline patients may be more likely than antisocial patients to profess concern and love for others in their lives, in spite of their lack of caring about themselves. For example. Dee experienced a life filled with sexual, physical, and emotional abuse, leading her to conclude that she was a "bad" person who was "dirty and used up." At the start of treatment, she professed not to care for herself, and she took a cavalier attitude about the possibility that she might wind up killing herself with a drug overdose. It was clear that she was not intent on doing the work of therapy in order to obtain the benefits for herself. Instead, Dee and her therapist focused on her baby daughter, w h o m she loved dearly. Therapeutic goals were set on the basis of the premise that the baby would become the ultimate beneficiary. This served as the "hook" that got Dee involved in the active process of trying to curtail her drug taking. She knew that she would take proper care of her beloved baby if she practiced abstinence from drugs. O n the other hand, if she used drugs. Dee knew from experience that she would disregard everything in her life, even her own daughter. The prospect of this happening again frightened her and therefore served as an initial motivator for treatment. Dee's case typifies the difficulties involved in treating this popu­ lation. When she experienced a drug lapse, she would tend not to bring herself back into line by thinking about what was best for her daughter. Instead, she would look at her drug-using experience as "proof" that she was an evil person and a bad mother who deserved to die. She would think to herself that her baby would be better off if she had a different mother, and Dee would proceed to escalate her drug use into a full-blown relapse. As a result. Dee would get her "wish," in that her aunt would take custody of the baby. As she began the next cycle of recovery. Dee once again would yearn for her daugh­ ter, thus leading to more depressed affect, frustration, and rifts in the family. Such stressors precipitated significant drug-taking cravings and urges, which again would feed into Dee's extremely low self-esteem. In dealing with cases such as Dee's, the therapist must be on the lookout for a number of emotional and cognitive dysfunctions typi­ cal of the borderline patient that exacerbate the patient's drug-taking habits. First, as mentioned above, borderline patients are prone to self- sabotage, believing that they are undeserving and unable to change things for the better (Layden, Newman, Freeman, & Byers-Morse, 1993). Such a mindset is wholly unconducive to acting in one's own best inter­ est, in planning for the future, in collaborating with the therapist, and Personality Disorders 289 in resisting the temptation to abuse drugs. Therefore, the borderline patient's pathologically low self-image (i.e., beliefs about the self) must be assessed and addressed as part of the treatment for substance abuse. Another factor is the borderline patient's propensity for feeling profoundly lonely and empty. M a n y of our patients have reported that the use of cocaine, crack, or heroin temporarily alleviates these feel­ ings, causing them to feel no need for other people during the euphoric surge. It is therefore not surprising that borderline patients turn to drugs in response to interpersonal difficulties and the result­ ant dysphoria. Complicating matters is the fact that some of these patients view their drug-taking cohorts as their only source of "friends." The thought of losing these associates as a result of becom­ ing abstinent is simply too threatening to entertain, and the drug use continues. In order to deal with the problems associated with extreme lone­ liness, therapists must help their drug-using borderline patients to work at building new, drug-free relationships. In addition, it is often beneficial to assist these patients in trying to repair old relationships (e.g., with immediate family members) that were strained or severed due to their substance abuse. These are difficult tasks for the border­ line patient, but they are necessary in order to provide an attractive alternative to the "loneliness-killing" qualities of drugs. Yet another problem borderline patients pose is their hair-trig­ ger hopelessness. As w e discussed earlier in this volume (Chapter 12), patients are more likely to resort to the quick fix that drugs produce if they believe that the future holds no promise. The implication for therapy is that the patients' degree of hopelessness is a critical vari­ able with which to contend if w e are going to succeed in helping them become and remain drug-free. Borderline patients, w h o are notorious for their extreme black- and-white thinking (cf. Beck et al., 1990; Layden et al., 1993), can change their outlook on life from optimistic to completely hopeless with relatively little provocation. W e have seen such patients leave a productive therapy session in very high spirits, only to call the thera­ pist the same evening in a state of deep dysphoria due to an objec­ tively mild stressor. W h e n drug abuse is part of the clinical picture, these patients are more likely to resort to getting high than they are to call their therapists for assistance. The upshot is that the border­ line patient can flip vety quickly from hopeful and abstinent to hope­ less and using. Therefore, therapists need to remain vigilant in help­ ing these patients to learn, practice, and, it is hoped, master the skills of decatastrophizing and problem-solving in the face of setbacks. 290 COGNTnVE THERAPY OF SUBSTANCE ABUSE An offshoot of the borderline patient's problems with hopeless­ ness is a pronounced risk for suicide (Linehan, 1987). Ideally, the patient's wish to live or die should be assessed at every session, and a verbal antisuicide contract should be made and renewed frequently. For example. Dee was periodically suicidal, and the therapist repeat­ edly stressed that she was to call him at work or at home in the event that she wanted to harm herself. She agreed to do so, but she was agonizingly inept in keeping ready access to the therapist's telephone numbers. Therefore, as a matter of ritual, the therapist assigned Dee the task of memorizing his telephone numbers, and quizzed her sud­ denly at various points during every therapy session. As a population, substance-abusing borderline patients are more frequently in need of hospitalization than many other clinical patient groups. Even in an inpatient setting, such patients pose considerable management problems. If the patient is suicidal, the first order of business is to provide constant professional supervision. Although the acute risk for suicide may pass, this type of patient rarely is able to stay in the hospital long enough to deal adequately with the border­ line disorder. When these patients reenter outpatient treatment, the therapist's tasks are daunting. As with

the antisocial drug-abusing patient, it is vitally impor­ tant to carefully nurture and make clinical use of the therapeutic relationship (see Beck et al., 1990; Layden et al., 1993; Young, 1990; Chapter 4, this volume, for more detailed explications). The all-or- none thinking style of the borderline patient often shows itself in the way that the patient views the therapist, who may be seen as a savior one week (e.g., "Nobody else understands m e the way you do. You're cool and I can trust you.") and a villain the next week ("You're just like all the rest! You don't care about me! You don't know what you're talking about! You're not helping m e at all!"). Under such adverse interpersonal conditions, therapists must not allow themselves to buy into patients' extreme positive or negative feedback. Instead, they would do well simply to maintain their composure, to express an ongoing willingness to help, and to focus on the thoughts and beliefs that are behind the patients' polarized reactions. Likewise, therapists must make every effort to teach these patients to think carefully before acting. This involves painstaking training in problem-solving (D'Zurilla & Goldfried, 1971; Nezu et al., 1989) and rational responding (Beck et al., 1979; Newman & Beck, 1990), skills that are most difficult (yet most necessary) to employ at times of extreme distress. Also similar to the treatment of the antisocial patient is the focus on helping the borderiine patient acquire assertiveness skills, so as to reduce conflict in interpersonal situations. Personality Disorders 291 All these interventions require numerous repetitions with both antisocial and borderline patients, as their psychological learning skills and levels of motivation to change are notoriously deficient. SUMMARY The existence of severe Axis II disorders in substance- abusing patients significantly complicates the treatment picture with these patients. In particular, antisocial and borderline patients pose tremendous challenges to therapists w h o are working diligently to help them to overcome addictions to drugs. W e have suggested ways that therapists m a y detect concomitant personality disorders in substance abusers, taking into account the fact that patients w h o are actively on drugs often seem to "change personality," thus confusing the diagnostic picture. Nevertheless, w e posit that it is important to assess the presence or absence of person­ ality disorders in substance-abusing patients, as the more extreme personality disorders will require awareness of special issues in clini­ cal management, the likes of which w e have reviewed. Furthermore, assessing and understanding the substance abuser's characterological issues provides a window into the patient's beliefs or schemas about the self, world, and future, thus assisting the thera­ pist in reaching a sound conceptualization of the case (Beck et al., 1990; Persons, 1989). Finally, w e would like to add that there is no tried-and-true start­ ing point in the treatment of Axis II dual-diagnosis patients. The question whether to focus first on the substance abuse or on the personality disorder is a "chicken and egg" issue. Therapists must be prepared to attack these problems simultaneously, using their under­ standing of the patient's idiosyncratic dysfunctional beliefs to sym­ pathetically earn increments of trust and cooperation in cartying out the treatment for drug abuse. C H A P T E R 1 7 R e l a p s e P r e v e n t i o n i n t h e C o g n i t i v e T h e r a p y o f S u b s t a n c e A b u s e A f̂ter the patient has quit using drugs, an even more formidable challenge begins. It is relatively easy for many people to change undesired behaviors temporarily; however, maintaining behavioral changes is much more difficult. In fact, it was Mark Twain who said, "To cease smoking is the easiest thing 1 ever did. 1 ought to know because I've done it a thousand times" (Prochnow, 1969). Most people who quit using drugs have a lapse or a relapse (Saunders & Allsop, 1987; Vaillant, 1983), with the most hkely time being within 90 days of the initiation of abstinence (Mackay, Donovan, & Mariatt, 1991). A lapse, or a "slip," is defined as the ini­ tial use of a substance after an individual has made a commitment to abstain from that substance. A relapse, on the other hand, is a full return to the maladaptive behaviors originally associated with use of the substance. In the cognitive therapy of substance abuse, a major goal is to have patients learn from whatever setbacks in long-term abstinence goals they may have (Moorey, 1989). It is hoped that the lessons learned from such experiences will ultimately improve an addict's planning skills, resolve, and self-confidence. In this chapter we present cognitive therapy techniques for relapse prevention. Some of these techniques are an extension of methods used to help patients discontinue drug use in the first place (e.g., the development of a collaborative relationship, conceptualizing the patient's problems, guided discovery, structured sessions, advantages- 292 Relapse Prevention 2 9 3 disadvantages analysis, and homework assignments.) Other techniques (e.g., identification of high-risk stimuli) are uniquely designed to assist patients in the relapse prevention process. Prior to presenting these techniques, the cognitive model of drug use and relapse is reviewed. This model provides a framework within which these techniques may be understood and applied. THE COGNITIVE MODEL OF RELAPSE The cognitive model of relapse (nearly identical to the model of ongoing use) is presented in Figure 17-1. The cocaine addict is vulnerable to high-risk stimuli (HRS) (Marlatt & Gordon, 1985). These are internal and external "triggers" that stimulate the addict's appetite for drugs. Internal stimuli include emotional and physical factors such as depression, loneliness, bore­ dom, anger, frustration, and physical pain (Mackay et al., 1991). External stimuli include people, places, and things that are related in some way to drug use (Shulman, 1989). HRS vary greatly from per­ son to person. For example, possessing money may trigger some patients to use, while other patients may be more tempted to use when they feel upset about not having money. The difference lies in the personal meaning that the patient attaches to the stimulus. The importance of HRS is that they may activate basic drug-related beliefs that have been overlearned by the patient. Examples of these beliefs are the following: "Drugs are problems for some people, but they won't be for me." "Drugs enhance m y life by making it more fun." "People who are against drugs don't really understand them." High-risi< Basic stin'iuli drug-related Automatic Craving/ (internal or beliefs Thoughts urges external) activated Focus on Facilitating Lapse instrumental beliefs strategies (permission) (action) F I G U R E 17.1. Cognitive m o d e l of relapse. 294 COGNTnVE THERAPY OF SUBSTANCE ABUSE "As long as I am careful, drugs won't hurt me." "Without drugs, life would be boring." "When I have problems, drugs relieve the pain." From these basic beliefs, automatic thoughts may be stimulated (e.g., "I need a hit!" "Time to party!" or "I've got to get high!"). Automatic thoughts are associated with craving and urges, and like other feelings, craving and urges themselves become internal HRS. In response to craving and urges, the patient may engage in some facilitating beliefs. Facilitating beliefs are a subset of basic drug-related beliefs that give patients "permission" to use. Examples include the following: "I can use just one more time." "Nobody will find out." "It will be O K to use again. I'll keep it limited." Facilitating beliefs increase the likelihood that the addict will and eventually use drugs. Instrumental strategies are the actual behaviors and activities involved in seeking, acquiring, and using drugs. These sfrategies, like basic drug-related beliefs, may vary from person to person and from time to time. For some, the craving for cocaine is so intense that the addict becomes obsessed with drug seeking, to the exclusion of all other considerations. Like a heat-seeking missile, the craving addict may focus on a target and let nothing get in the way of drug acqui­ sition. As a result of this "drive" some addicts engage in violent or criminal behaviors to acquire cocaine. If a patient has proceeded through the steps above, he or she is likely to have a slip. This slip, or lapse, then becomes a high-risk stimu­ lus for the cycle to begin again. That is, on taking a hit, the addict may focus on his or her basic belief that "one hit means I'm out of control." This belief might be followed by a cascade of self-derogat­ ing automatic thoughts and negative feelings. The resulting urges may become stronger, and the use of drugs may escalate until the patient has a full-blown relapse. To illustrate this process, we review the case of Mike, a 35-year- old salesman who had been abstinent from cocaine for six months. After two years of heavy use Mike chose to quit while hospitalized for cardiac problems (tachycardia and arrhythmias) associated with his cocaine use. Mike agreed to inpatient addiction treatment and, on completing a 28-day program, he entered outpatient psychotherapy and Narcotics Anonymous to deal further with his addiction. Prior to Relapse Prevention 295 quitting, Mike had been having serious marital problems, as his wi "Judy" had recently given birth to their first child and had become increasingly frustrated with Mike's cocaine-related problems. In spite of Mike's extended inpatient and outpatient drug treat­ ment, he continued to crave cocaine. His internal (emotional) HRS included feelings of boredom, frustration, irritability, and anxiety. External high-risk stimuli included friends who used, parties, and arguments with Judy. After initial treatment Mike often "found him­ self" in high-risk situations. For example, he continued to spend time with his cocaine-using friends. Such behaviors made Judy quite angry and she eventually separated from him. In response to his separation, Mike felt depressed, anxious, bored, frustrated, and irritable; thus, he was at extremely high risk for relapse. Mike's cocaine-using friend "Ryan," seeing Mike in "such bad shape," offered him a line of coke to "boost his spirits" (introducing another high-risk situation as well as a facilitating belief). Mike chose to accept Ryan's offer and in doing so he perceived himself as being "out of control." His automatic thoughts included "I can't resist" and "I can't handle life without drugs." He began to imagine and antici­ pate the extreme positive feelings he had previously associated with getting high: "I'll feel great!" and "It will relieve all of m y pain." After vety brief deliberation he took his first "hit," which led to a second, a third, and so forth (this episode became his initial lapse). As the effects of the drugs wore off, Mike began to feel dysphoric, agitated, and confused. He perceived this experience as proof that his substance abuse was an irreversible condition and he thought "once an addict, always an addict!" and "Now that I've started again I can't possibly stop using." As a result of this dichotomous thinking, Mike set the stage for a full relapse. Relapses pose important challenges to both patient and therapist. They may trigger counterproductive thoughts in the therapist, such as "This patient's condition is hopeless!" and "I am wasting m y time." Counterproductive thoughts in the patient might include, "I have failed and I always will fail," "I can't ever tell m y therapist about m y using," and "My therapist would surely reject me, or even hate me, if he knew I've been using!" The disadvantage of such all-or-none think­ ing processes is that they may result in a sense of hopelessness or apathy in the patient or therapist. Ideally, the therapist helps the patient to construe lapses and relapses as opportunities to practice more adaptive ways of combat­ ting drug use. For example, Mike's therapist could help him to understand that there were more adaptive ways to deal with his dysphoria over his marital separation than to use drugs. Rather than 296 COGNITIVE THERAPY OF SUBSTANCE ABUSE believing that his situation was hopeless, and that his only relie be to accept his friend's offer of cocaine, he could focus on produc­ tive activities that would boost his mood and self-esteem. Further, even if he did begin to use cocaine, Mike did not

have to believe that he was in an irreversible free fall. Instead, he could practice reciting some control beliefs that might help him to "pull out of the nose dive" (Shiffman, 1992) if confronted with similar situations in the future. Such control beliefs might be "If I stop now, I can show myself that I am stronger than the dmg." "Walking away from drugs is the same as moving toward saving m y marriage." "I must seek drug-free friends when I feel this badly." When the patient has maintained abstinence for an extended period of time and the therapist and patient are confident in the patient's ability to maintain abstinence, formal therapy may be ter­ minated. At this time, "booster sessions" are arranged, which may include telephone calls, written correspondence, and face-to-face con­ tacts. These sessions serve several purposes. First, they focus the patient's attention on the need for vigilance in combatting the relapse process. Second, the therapist's continued interest in the patient pro­ vides social support that motivates further abstinence. Third, the thera­ pist can continue to provide expert guidance to a patient who may be at renewed risk for relapse. Substantial gains may result from extended contact with the patient. Each booster visit or telephone call might decrease the like­ lihood that the patient will relapse, or at least will remind the addict that the therapist is a potential resource for coping with HRS. If the addict does relapse after therapy has been terminated, it is recom­ mended that he or she be invited to return to therapy as soon as pos­ sible to work on improving coping skills. Again, careful analysis of each lapse and relapse provides the patient with an increased under­ standing and ultimately greater control over the relapse process. At the same time, we must note the following two caveats. First, although we as therapists endeavor to pro'vlde ongoing help to patients who continue to be in need, we agree with Gawin and Kleber (1992) that, "The goal of relapse prevention is gradually to decrease the external controls placed on the abuser, by family and therapist, dur­ ing initiation of abstinence, and gradually to facilitate development of the abuser's internal controls" (p. 47). Second, although it certainly is preferable for patients to learn from their lapses and relapses than merely to succumb to them, we must be careful not to convey the Relapse Prevention 297 wrong message that we encourage lapses and relapses as growth-enha ing experiences. Therapists must make it clear to patients that the best learning occurs via ongoing drug-free coping with life's evetyday and long-term demands. TECHNIQUES FOR THE PREDICTION A N D C O N T R O L OF RELAPSE Identification of High-Risk Stimuli People who suffer from drug and alcohol addictions inevitably encounter HRS. This is especially apparent when one con­ siders the fact that all human beings experience some sadness, anxi­ ety, nervousness, anger, or frustration (all internal HRS) at times. The identification of internal and external HRS is an extremely important component of the relapse prevention process, as it is not uncommon for the addict to lack an awareness of HRS. By increasing awareness of HRS, the addict may reduce the like­ lihood of exposure and/or reflexive reaction to them. In sessions, patients are encouraged to carefully review recent and remote memo­ ries of relapses in order to discover the full range of HRS that might lead to future relapse. Here, a therapist^atient dialogue is presented to illustrate this "trouble-shooting" process: TH: Mike, what types of situations are most likely to cause you to crave cocaine? PT: I don't know. TH: Well, let's tty to think of some, so that you might learn to pre­ pare and cope more effectively with them PT: Yeah, I guess that makes sense. N o w that you ask, when I am with certain friends I want to get high ... but 1 can't think of anything else. TH: OK, let's review your last binge. H o w did it begin? PT: I had been clean for almost a month when I suddenly got the urge one night. The next thing I knew, I was with Bob and Ryan, smoking crack. TH: Now think careftilly, what actually happened that night? I would like to hear the details. PT: Well, I was at home and m y wife was in one of those moods where nothing I did was right. TH: What happened next? 298 COGNTTIVE THERAPY OF SUBSTANCE ABUSE PT: I got pissed off at her and really let her have it. What I mea 1 really started to fight back. TH: What happened next? PT: She let me have it, even worse. She started screaming about how I'm a lazy, good for nothing so and so. TH: And then? PT: I stormed out of the house and went for a walk around the neighborhood. Sure enough, I saw Bob and Ryan sitting in front of the convenience store. TH: What happened next? PT: I walked over to them, and the next thing I knew we were all at Bob's house. The rest is history. TH: So, there were actually several high-risk stimuli involved, weren't there? They included your wife's criticisms, your anger and fms­ tration in response to her comments, your leaving home with the problem unresolved, walking around late at night, the con­ venience store where you guys typically meet. Bob and Ryan, and finally, being at Bob's house. PT: Yeah, now that you mention it, I guess the situation was kind of complicated. TH: H o w could you make this kind of situation "simpler" in the fu­ ture? PT: I don't know. TH: Well, this is something we have to work on. For starters, you could make things simpler if you know how to recognize when you're upset and looking for an escape. Does that ring tme for you? PT: Yeah. In this dialogue it became apparent that Mike had exposed him­ self to multiple HRS, thus dramatically increasing his likelihood of a lapse. Interestingly, Mike had difficulty seeing how he made decisions that led to a chain of events that led up to an ultimate lapse. Marlatt (1985) talks about such "accidental" exposures to HRS as resulting from "apparently irrelevant decisions." Carroll, Rounsaville, and Keller (1991) explain that such exposures may reflect underlying ambiva­ lence about changing addictive behaviors. Thus, it is important that the therapist, under these circumstances, evaluate the status of the patient's decision to become abstinent from drugs and alcohol. For example, Gorski and Miller (1982) highlight a number of telltale signs that signal a patient's starting along the road to relapse- Relapse Prevention 299 that is, being prone to exposure to high-risk situations, and ill- equipped to manage the thoughts, feelings, cravings, and actions that result. Such signs include, for example, (1) an increasingly noncha­ lant attitude about remaining actively involved in ongoing self-help activities (e.g., attending therapy or support-group meetings, con­ tinuing with homework derived from therapy), (2) regression into affective lability and hypersensitivity, (3) reduced willingness to talk about problems and concerns, (4) social withdrawal, (5) breakdown of healthy daily routines and structure, and (6) impulsive decision­ making. A concrete example is a patient who begins to cancel therapy sessions for spurious reasons, and who denies any symptoms on the Beck inventories (BDL BAI, BHS, etc.) in spite of obvious signs of stress. These early-warning indicators set the stage for the relapse process as depicted by the cognitive model flowchart, and must be given attention in their own right. It is apparent that HRS are highly idiosyncratic for individual drug abusers (Carroll, Rounsaville, & Keller, 1991). It is recommended that the cognitive therapist carefully evaluate HRS for each individual patient. A useful method for doing so is self-monitoring homework. Specifically, the patient keeps a journal of hourly changes in cocaine (use and) craving in relation to internal and external events. This journal is reviewed in session to assess "previously unseen patterns in cocaine use" (Carroll, Rounsaville, & Keller, 1991). At the end of each session, the patient is asked to anticipate specific HRS that might occur between the present and the next session. Furthermore, the patient is asked to plan strategies for coping with such HRS. Cognitive Strategies for Coping with HRS The significance of HRS is that they trigger basic drug- related beliefs that increase vulnerability to lapses and relapses. Thus, a fundamental cognitive strategy for relapse prevention is the devel­ opment of "control" beliefs that reduce vulnerability to lapses and relapses. The following are examples of control beliefs that reduce vulnerability to lapses and relapses: "I don't need drugs to have fun." "My life will improve without drugs." "I can cope with unpleasant emotions without using drugs." "I have control over m y own behaviors, including m y drug use." "Even if I slip I don't have to continue using drugs." "A lapse is not equivalent to failure." 300 COGNITIVE THERAPY OF SUBSTANCE ABUSE The following patient-therapist encounter illustrates the process of guided discovery, used here to modity drug-related beliefs and to build control beliefs. TH: I wonder how you make sense of your continued use of cocaine in spite of the destructive results? PT: Well, for one thing, I feel like I have no control over m y dmg use when I am emotionally upset. TH: What do you mean when you say "I have no control"? PT: I just don't feel like I have any control. TH: So you believe that you have "zero percent" control? PT: I guess. TH: Are you saying that each and every time you've ever had a crav­ ing for coke you've used? PT: Well, no. TH: H o w many times have you succeeded in not using in spite of having cravings and urges? PT: Well, lots of times, actually. I guess most of the time now that I think of it. TH: Does this mean that you've been able to demonstrate control? PT: It never seems that way to me, but I guess I do have some con­ trol. TH: And so you have just changed your belief from "I have no con­ trol" to "I have some control." As we have discussed, the first belief is an example of dichotomous, or "all-or-nothing" think­ ing, while the second thought is probably more objective, because it takes into account the "shades of grey." PT: Yes, I know. So when I think that I have no control I am more likely to give in to the urges and use cocaine. But when I look carefully and 1 see that 1 have greater control, I'm likely to tty to stay with m y program. TH: Yes, that's right. In this example the patient is helped to see the advantages of changing his drug-related belief "1 have no control" to "I have some control." Another cognitive strategy useful for dealing with HRS is distraction (Carroll, Rounsaville, & Keller, 1991). Specifically, the patient is encouraged to compile a list of distracting activities that might be used when HRS are encountered. Distracting activities may include any non-drug-related activity (e.g., exercise, singing, playing with children, and writing a letter). Although distraction techniques Relapse Prevention 301 are only a short-term coping device, they serve the all-important tion of providing a delay between the onset of cravings and the act of seeking and using drugs. Such a delay may provide patients with time to think of the full negative ramifications of using, as well as an opportunity to witness the diminishing of cravings if no drugs are taken (Carroll, Rounsaville, & Keller, 1991; Horvath, 1988). Behavioral Strategies for Coping with HRS After patients identify HRS and examine drug-related beliefs, they are helped to practice (i.e., rehearse) strategies for cop­ ing with certain HRS. For example, Mike might be assisted in devel­ oping and practicing methods for resolving conflicts with his wife. He might be encouraged to think about ways to give her and to ask her for more emotional support. He also might be encouraged to think about alternatives to going out alone, especially to areas where drug- using friends are located. Additionally, he would be helped to imag­ ine ways of saying "no thanks" to his friends when they invite him to use drugs,

in spite of his cravings and urges. As Mike's case illus- frates, relapse prevention necessitates that patients learn to cope with both general life stressors (e.g., marital discord) and discomfort that specifically is related to temptation to use drugs (Wills & Shiffman, 1985). A long-term goal for Mike, essential to most drug users' contin­ ued abstinence after termination from therapy, would be to establish and maintain meaningful relationships with people who are drug-free (Frances & Miller, 1991; Havassy et al., 1991). The phrase in 12-step groups that reflects this philosophy is "hugs, not drugs." Ultimately, the patient must learn to seek human contact, rather than drugs, in order to obtain gratification of dependency needs. As an important part of this process, Mike would be taught to examine the beliefs that led to his high-risk behaviors. For example, the therapist might ask, "What were you thinking when you yelled at your wife?" The patient's automatic thoughts might have included: "I'll put a stop to this!" or "She can't talk to m e that way." Underly­ ing these thoughts are Mike's beliefs about his personal inadequacy and powerlessness, beliefs that trigger both drug urges and the accompanying belief that giving in to the urges is the only way to feel powerful. Another behavioral relapse prevention strategy involves planned, gradual exposure to HRS (Mackay et al., 1991). Specifically, the patient and therapist construct a hierarchy of increasingly salient high-risk triggers and the patient is actually exposed to these in sessions {in 302 COGNITFVE THERAPY OF SUBSTANCE ABUSE vitro) and through homework {in vivo). The efficacy of this techniq may be attributable to the cognitive changes in patients as they per­ ceive themselves effectively dealing with these situations. In other words, patients gain an increased sense of self-efficacy as they suc­ cessfully cope with increasingly high-risk situations. Again, a word of caution is due. While we believe that exposure to high-risk situations is inevitable in life, and therefore gradually "inoculating" patients to these situations makes excellent clinical sense, this technique must be handled with care. Therapists must make sure that their patients have the requisite skills and motivation to man­ age their induced cravings in session before being asked to cope with cravings outside the office (Childress et al., 1990). Patients must be told that this type of assignment requires a detailed plan, as well as a safety valve contingency (e.g., calling the therapist) in case the cravings become hard to manage. Otherwise, patients may misuse this assignment by cavalierly "testing" themselves before they are ready to handle the resultant cravings (Carroll, Rounsaville, & Keller, 1991; Washton, 1988). An example is the patient who drives through the neighborhood where he used to buy heroin, "just to see if I can do it," without preparing a plan of action in advance, thus taking a huge gamble with his abstinence. By contrast, therapist-instructed graded- exposure assignments must be small, calculated risks that have excel­ lent chances for success and backup plans in case things go wrong. Keeping a Lapse from Becoming a Relapse As mentioned earlier, lapses provide the addict with opportunities to apply cognitive skills and promote further under­ standing of the mechanisms involved in relapse. Thus, a lapse is not necessarily perceived as "bad"; instead it is "grist for the mill." An important theme of relapse prevention is helping the patient keep lapses from becoming relapses (Mackay & Marlatt, 1991). There are many reasons why lapses occur. For example, addicts may choose to "slip" in order to test their ability to control their substance use. They might think, "I'll try' it just this once. It will prove that I am in control of m y addiction." As we've mentioned above, addicts may also "accidentally" or intentionally expose themselves to a high-risk stimulus without being prepared to respond cognitively or behaviorally to this stimulus. Another reason for a lapse may be that the addict once again believes that the advantages of using dmgs outweigh the disadvantages. Given the many reasons for lapses, an important component of relapse prevention involves the identification of decision points along Relapse Prevention 303 the cognitive model of relapse. For example, did the lapse occur because the addict failed to avoid an external HRS? Or, did the lapse occur because the addict lacked appropriate control beliefs for resist­ ing an inevitable HRS? Or, did the patient automatically engage in instrumental strategies for acquiring drugs in response to cravings, without waiting? A lapse usually becomes a relapse as a result of underlying all- or-none beliefs; for example, "A lapse means I don't have any con­ frol," "Since I could not stay abstinent, the therapy isn't working," and "A lapse is a failure." Marlatt and Gordon (1985) call this think­ ing process and the resulting relapse the "abstinence violation effect." An important strategy for relapse prevention, then, is to challenge such dichotomous thoughts about lapses so that they do not become relapses. When an addict has a lapse, imagety techniques are useful to reconstruct the sequence of stimuli, basic beliefs, automatic thoughts, and behaviors leading to the lapse. Additionally, it is important to use post-hoc rehearsal of techniques at each decision point to pre­ pare the addict for similar future circumstances. This sfrategy is illustrated in the following dialogue between Mike and his therapist. Prior to this interaction, the therapist learned that Mike had been walking near the areas where his drug-using friends, Ryan and Bob, get high together. In fact, Mike almost avoided his friends but he suddenly felt "obligated" to stop to talk to them. Bob asked Mike why he hadn't been around for awhile and he offered to drive Mike home. Mike accepted the offer and got into Bob's car. Once on the road, Bob explained that he had to make a quick stop at his house. While Bob was inside, Mike began to think that maybe he should join him in the house in order to be "sociable." Inside the house, Bob offered Mike "a quick hit." Although Mike felt that he did not really want any, he accepted the offer and took the hit. There were several choices and decision points leading to this lapse. First, Mike decided to walk through a "high-risk zone" instead of choosing a "safe" route. Upon seeing his friends, he had the option of choosing to keep walking. When he was in the car outside of Bob's house he had the option of staying outside, rather than entering the house. And finally, when Bob offered him the hit, he had the oppor­ tunity to accept or refuse. The following dialogue is the therapist's and patient's "postmortem" evaluation of the sequence of events lead­ ing to the lapse. Mike had his eyes closed and was imagining the scene as if it were happening now. TH: You were walking along and you happened to wander past the old hangout. What was the risk involved in taking that route? 304 COGNITIVE THERAPY OF SUBSTANCE ABUSE PT: I guess that I might have seen Bob and Ryan. I knew it could have been a problem in the long run, but I like spending time with them. W e really enjoy each other. TH: What happened when you saw Bob? PT: He called out to me and I thought "I ought to stop." TH: Is there any problem with that thought? PT: I couldn't turn m y back on m y friend. TH: What do you risk by following the dictates of that thought? Do you give up your freedom of choice? PT: I guess I could get into trouble that way. TH: What happened next? PT: I stopped and he offered to take m e home. TH: Did you have a choice? PT: I felt that 1 should take him up on the offer. TH: You often act instinctively but perhaps following those instincts may work against you . . . {pause) . . . Now, you're waiting in the car. What's going through your mind? PT: I'm thinking that maybe I should go up there and be sociable. TH: Are you feeling the urge to get high? PT: Maybe in the back of m y mind 1 may be thinking that it would feel good to take a hit. TH: What are you thinking when Bob offers you some? PT: I'm thinking that maybe I don't want to. But Bob offered it as a gesture of his friendship ... I felt I should accept. TH: Could you have asked yourself, "Am I being 'nice,' or am I just giving myself permission to use?" PT Yeh, 1 guess so. TH And what would the consequence have been of that thought? PT Maybe 1 would have realized how impulsive I was being. Maybe 1 wouldn't have taken the hit. TH: After you took the hit, what were the thoughts that ran through your mind? PT: 1 was thinking, "might as well finish the job I've started." TH And what was the meaning of that thought? PT; It meant "one hit and I'm over the edge." TH That sounds like "all-or-nothing thinking." PT: Yeh, you're right. Relapse Prevention 305 TH: And you proceeded to smoke crack at Bob's house all afternoon. PT: Yes. In this example, Mike was "guided" through the thoughts and images of his recent experiences so that he might review the deci­ sion points leading to his lapse. Upon arriving at his lapse, his thera­ pist helped him to see that his dichotomous thinking led him to con­ tinue using after his initial hit. In the section that follows, Mike's therapist encouraged Mike to rehearse alternative methods for cop­ ing at each decision point. TH: What could you have done instead of walking along the old route, past the old hangout? PT: I guess there were lots of walks I could have taken which would have avoided those guys. TH: And if you had still passed them, what could you have thought when they called out to you? PT: I could have thought "Those guys put m e at risk for relapse. Better keep walking." TH: And what would your resulting behaviors have been? PT: I would have waved in a friendly way and said "Hi guys! I'm in a hurry. Gotta run." TH: And what if Bob had pulled up in his car and offered you a ride anyway? PT: I could have continued to think "bad news. ..." TH: And the resulting behavior would have been what? PT: I would have turned down the ride. TH: And what if they convinced you to get in the car and you ended at Bob's house anyway? PT: 1 still could have thought "Don't get out of the car. The house is a danger zone." TH: And then what would have happened? PT: Bob probably would have come back to the car and taken m e home without using. TH: And if that would have happened, how would you have felt? PT: Probably relieved and proud of myself for exercising control over m y life. At this point, the therapist realized that Mike had other types of beliefs (relating to "loyalty" and "responsibility") that made him vul- 306 COGNTnVE THERAPY OF SUBSTANCE ABUSE nerable to lapses and relapse. In the next segment, the therapist explored these beliefs. TH: What were you thinking when you walked over to Bob's car? PT: I thought, "I have really neglected m y friends." TH: What do you mean by that? PT: I mean that I have to be friendly to these guys. After all, they are m y friends. TH: What do you mean when you use the phrase "have to"? PT: I like to think of myself as a loyal friend. TH: And what does "loyal" mean? PT: It means that I should go along with what they want. TH: Shoulds and musts. PT: Yes. TH: What do you know about shoulds and musts? PT: I know that they aren't the best. TH: And what are the results of shoulds and musts? PT: Well, they can get m e into trouble. Like in this case, they got me back into using. Advantages -Disadvantages Analysis Marlatt and Gordon (1985) explain that positive out­ come expectancies contribute to relapse. Specifically, individuals who have been addicted to a substance perceive substantial

positive advantages from using that substance. In fact, such people typically minimize or ignore the disadvantages of their drug use, especially when in high-risk situations. The advantages-disadvantages analysis is a technique commonly used in cognitive therapy (see Chapter 9, this volume). This technique is particularly useful in relapse prevention, where addicts selectively perceive the advantages of drug use. In the advantages-disadvantages analysis, the therapist constracts a four-cell matrix, with advantages-disadvantages on one axis, and use- nonuse on the other axis. Patients are asked to discuss the advantages and disadvantages of using and abstaining. The role of the therapist is to elicit objective data from patients, regarding such advantages and disadvantages. Additionally, patients are encouraged to understand that their exaggerated views about the advantages, along with their mini­ mized views of disadvantages, contribute to their addiction. Relapse Prevention 307 The following dialogue is a continuation of the session with Mike. It illustrates the advantages-disadvantages analysis: TH: Mike, when you were in the high-risk situation with your frien Bob and Ryan, what was going through your mind? PT: I guess I thought, "Soon I'll get some relief from this bullshit." TH: So you believed that an advantage of smoking crack was relief from your bad situation. PT: I'd forget about m y old lady for a few hours. TH: What other advantages could you see? PT: {pauses) Well, I can't think of any others. TH: What were the disadvantages of using? PT: Oh, the relief is only for a short time. I know that I am really making m y marriage and m y life worse. TH: What other disadvantages are there to using? PT: Well, I could lose m y kid if m y wife decides to split. TH: What else? PT: I can tell that m y life is changing for the worse from this stuff. When I am really honest with myself I know that m y health and self-confidence are going down the tubes. TH: OK, What are the advantages of abstaining? PT: Well, as I just said, I'm kinda wasting m y life. Maybe m y life would start to improve if I abstained. TH: Can you think of any other advantages of abstaining? PT: {pause, tears in his eyes) I guess I could start to be more of a father to m y son; I never really had a father myself. TH: Sounds like a vety big advantage! PT: Yeah, if I could only remember later how I feel right now! In fact, Carroll, Rounsaville, & Keller (1991) suggest a method that facilitates remembering the advantages-disadvantages analysis. They suggest that the addict list the advantages of using cocaine on one side of an index card. The therapist then helps the patient to see the ultimate negative consequences of all advantages. For example, the "terrific high" is following by the "dreaded crash." O n the flip side of the index card the patient lists the disadvantages of continued cocaine use. The card is then placed in the addict's wallet, near the money, so it will be accessible during periods of high vulnerability. 308 COGNITIVE THERAPY OF SUBSTANCE ABUSE Development of Social Support Netvkrorks Interpersonal conflict is a common high-risk stimu­ lus for many addicts (Mackay et al., 1991). In fact, Cummings, Gor­ don, and Marlatt (1980) found that 4 4 % of relapse experiences were linked to interpersonal conflicts. Loneliness is also a high-risk trig­ ger, for two reasons. First, loneliness is an uncomfortable emotion (i.e., internal trigger) that may be anesthetized temporarily by cocaine. Second, heroin and cocaine use often take place in a social environ­ ment, which temporarily provides the addict with a sense of having "company" to relieve the loneliness. From these assumptions, we assume that relapse prevention efforts will be enhanced by the patient's acquisition of a drug-free social support network (cf. Frances & Miller, 1991). It is important to understand that patients have numerous basic beliefs and automatic thoughts about relationships that influence their behaviors in relationships. For example, some addicts may believe "Only other users can understand me," "1 will never be accepted by those who have never used drugs," "Nonusers are boring," and so forth. Obviously, such beliefs result in social discomfort, or anxiety, and a certain degree of social avoidance. Patients can be helped by the therapist's understanding of this process, as well as by modifica­ tion of such beliefs. Friends and family who do not use drugs may be excellent sources of support to the patient. Many addicts have avoided their nonusing family members and friends for fear of criticism and rejection. Twelve- step programs (e.g.. Alcoholics Anonymous and Narcotics Anonymous) may also be sources of support for the addict. Therefore, patients should be encouraged to make use of these programs. By doing so they may experience multiple benefits, including social support, col­ laboration, an environment that is supportive of abstinence, and a way to spend time when suffering from boredom. Of course, cogni­ tive therapists strive to be part of the patients' support network as well, but it is vital that drug-abusing patients have drug-free friends in their everyday lives "for the long haul." SUMMARY In this chapter, relapse prevention was presented as part of the cognitive therapy framework. This model emphasizes the role of beliefs as well as high-risk situations in the relapse process. Techniques were presented for predicting and controlling relapse, Relapse Prevention 309 including identification of HRS, cognitive and behavioral strategi for coping with HRS, keeping a lapse from becoming a relapse, advantages-disadvantages analysis, and social support networks for relapse prevention. W e also alluded to the fact that long-term abstinence from drugs and alcohol entails broad-sweeping changes in attitudes and lifestyle. When therapists note that their patients are reverting back to former ways of viewing themselves, their world, and their future, it signals a need for the patients to reassess their self-efficacy and commitment to positive change. In sum, the conceptual and technical skills that patients learn during the actual course of cognitive therapy must be practiced and "lived" in the months and years after formal treatment has terminated. Toward that end, we emphasize to our patients that they must learn to become their own cognitive therapists, and that the work of therapy needs to become a way of life. POSTSCRIPT The following is our version of the "Serenity Prayer," adapted for cognitive therapy substance abuse patients: Cognitive Therapy Serenity Pledge I pledge that I will strive to gain the strength to stay away from those drug triggers that I can avoid, the serenity and know-how to cope with those drug trig­ gers that can't be avoided, and the wisdom to know the difference. APPENDIX 1 BELIEFS ABOUT SUBSTANCE USE* Name: Date: Listed below are some c o m m o n beliefs about d m g use. Please read each state­ ment and rate how much you agree or disagree with each one. Totally Disagree Disagree Neutral Agree Agree Totally Disagree Very Much Slightly Slightly Very Much Agree 1. life without using is boring. 2. Using is the only way to increase m y creativity and produc­ tivity. 3. I can't function without it. 4. This is the only way to cope with pain in m y life. 5. I'm not ready to stop using. 6. The cravings/urges make m e use. 7. M y life won't get any better, even if I stop using. 8. The only way to deal with m y anger is by using. 9. Life would be depressing if I stopped. 10. I don't deserve to recover from d m g use. 11. I'm not a strong enough person to stop. 12. I could not be social without using. 13. Substance use is not a problem for me. 14. The cravings/urges won't go away unless I use dmgs. 15. M y substance use is caused by someone else (e.g., spouse, boy­ friend/girlfriend, family member). 16. If someone has a problem with dmgs, it's all genetic. 17. I can't relax without dmgs. 18. Having this drug problem means I am fundamentally a bad person. 19. I can't control m y anxiety without using dmgs. 20. I can't make m y life fun unless I use. *This form was developed by Fred D. Wright, Ed.D. 311 312 Appendix 1 CRAVING BELIEFS QUESTIONNAIRE (CBQ)- Name: Date: Please read the statements below and rate how much you agree or disagree with each one. Totally Disagree Disagree Neutral Agree Agree Totally Disagree Very Much Slightly Slightly Very Much Agree 1. The craving is a physical reaction, therefore, I can't do anything about it. 2. If I don't stop the cravings they will get worse. 3. Craving can drive you crazy. 4. The craving makes m e use drugs. 5. I'll always have cravings for dmgs. 6. I don't have any control over the craving. 7. Once the craving starts I have no control over m y behavior. 8. I'll have cravings for dmgs the rest of m y life. 9. I can't stand the physical symptoms I have while craving dmgs. 10. The craving is m y punishment for using dmgs. 11. If you have never used dmgs then you have no idea what the craving is like (and you can't expect m e to resist). 12. The images/thoughts I have while craving dmgs are out of my control. 13. The craving makes m e so nervous 1 can't stand it. 14. I'll never be prepared to handle the craving. 15. Since I'll have the craving the rest of m y life I might as well go ahead and use drugs. 16. W h e n I'm really craving drugs I can't function. 17. Either I'm craving drugs or I'm not; there is nothing in between. 18. If the craving gets too intense, using drugs is the only way to cope with the feeling. 19. W h e n craving dmgs it's O K to use alcohol to cope. 20. The craving is stronger than m y will power. *This form was developed by Fred D. Wright, Ed.D. Appendix 1 313 RELAPSE PREDICTION SCALE* Name: Date: As you know, there are many situations that can trigger an urge to use cocaine or crack. This scale has two parts: (1) to determine how strong you think the urges will be in certain situations and (2) to determine the likelihood that you will use in these situations. Listed below are several situations that might trigger strong urges to use cocaine or crack. Read each item and imagine yourself in that situation. In the first column, "Strength of Urges," indicate h o w strong you think the urge will be. In the second column indicate the "Likelihood of Your Using" in these situations. 0 1 2 3 4_ None Weak Moderate Strong Very Strong PREDICTION Strength Likelihood of of Urges Using 1. I am in a place where I used cocaine or crack before. 2. I a m around people with w h o m I have previously used cocaine or crack. 3. I just got paid. 4. I see coworkers using. 5. I a m leaving work. 6. It's Friday night. 7. I a m at a party. 8. I a m thinking of the last time I used. 9. I start talking with someone about using. 10. I feel bored. 11. I feel great! 12. I see a lover/ex-lover. 13. I a m having a drink. 14. M y friend is offering m e some cocaine or crack. 15. I feel sad. (OVER) *This form was developed by Fred D. Wright, Ed.D. 314 Appendix 1 PREDICTION Strength Likelihood of of Urges Using 16. I see a prostitute. 17. I a m out looking for sex. 18. I feel sexy. 19. I remember how good the high feels. 20. I feel angry. 21. I feel stressed out. 22. I feel guilty. 23. I just used dmgs. 24. I just broke m y abstinence. 25. I a m getting ready for work. 26. I a m tired. 27. I a m fmstrated. 28. I see an anti-drug use poster. 29. I see a pipe. 30. I am out gambling. 31. I just had a "coke dream." 32. I a m watching sports. 33. I a m getting dressed up. 34. I a m

!2        P@.Q 4  ! 3B   ; 2 A 5     A 3  2     5   :    5      ; 3           5         :  3 5 ;       23  5              !    
 53 ;   5    2       A   5 2  2   :        2   2 23        2  3 3  3    ! 53        ;         2   3 2   3  2  5 3  ;       "  4 J"."@M    2  3    2      &  <5 G   "/J-M6@0/D@0% @
  45  G3 , J"./@M   2   53 2   ! 3    4 &  =060$/D0$% -
+15  5
4 J"./-M &    3 2  53 2  3    5       2   5 ; 32  &  <5 G  .#J-M6@/#D@== 0
+15  5
4 ;2 &   < J"./#M        2  ; 2A  1 
       4 &  #06#"-D#@0 /  G J".==M ! 53  5   3 2    53 4 &  "=0J=M6.-/D.0= = G F,
   J".#0M & 3  2     2  3     53   ! 2   J   M 5   + I &  ="6%0#D%0. #
  : J".=.M : 3 5!          2  *   0J/M6-@.D--@ %
 < (  I < 
 5  3 I  3 4 J".#%M ,;  !      2     33 2      2  !    I :    &  #=J-M6-%/D-%. . 
 K 12 ) ) O  &  4 J".%0M    2;! 2  3 5       6 4 A  3   5   A!      5 "/@J-M6=/.D==- "$ 
 C 2 ) 22 )
3 
 :    &  4 J".%@M : 3!  3    2    R
 ! AS    6 :            & 4 2 G .6@/@D@/# ""
  : +  3 F C 5 C W I J".%-M : 3  A !         2 6  3 5     5 "0#6@/.D@=$      
 "@  4 )  < 55 I     !K   F 43  B C J".%-M ,   3 2      6 4 ;        ;   5 "0#6@="D@=- "-  4 )  < 55 I4 &3; B   !K   F  43  B C J".%0M       2   5 "/$J"M60/D0. "0 )3 < &   &  
 C3 ; B , J".%0M     3     2  3 2     3    I ( &  "J"M6@@D @= "/   3  '   :3  & J".%%M 4 A  3   5 2  ;  !    3  2 ;          2    ; 4  3 5    I ( &  %J"M6-%D00 "= 3B I  &    &  :  G  FI J".%/M A     3!  2 6 4  3 5  5  5 "/=J"M6#-D## "# 3B I  &  :  G   &    I <   J".%=M A     3    2 6 4 2  5  5 & 5 .$J@M6"..D@$/ "% 3B I    < , < &  B 4   
 <  <4 +1 <I ) <C
5
 )33 I      4    F  
  F J".%%M *  3            5   6  3 5   2  3     5  5 "=%J-M6#@#D#-" ". 3B I  )  I   I    ) B I
  I< J"..@M &  !  2       ;  3B   6 A   23  3        *    "/6@."D@.# @$     , 3 3
 & 5 I  F12  :F   ) K1 <+ I 5 F: A ( F  3
  CI J"../M <  3!  3     5 2  3     53 I ( &  @"J0M6=//D==. @" W I '  +:  )
 4      +
 * C +   C & J"..@M <  ! W A   K!      3 2      53   *    "/6-".D-@# @@ & ; '   I  1  ( I22  C   ) J".%#M *               2   3     5 ,  I  -"=J"@M6#$"D#$= @- <L  F (;1
 + 2 C !
 4   * 3 +I J".%.M ! 1  *3     <2N   J&  M  +1!  G !  1  
 3 F  ""06"/"#D"/@- @0 (;1
 <L  F J"..@M &  3        6 :  5 2    A   ;  F     *    "/6@%/D@.$ @/ & 1  3 1  &   + F   K   F2 
  !  , J".%%M 1  3      +  2   <2N  
  F   ""-J"0M6/-%D/0@ @= ) 3
 & 1    I J"..@M :  & 1 &  6 *    !        23         *    "/6@.%D-$/ @# +1 4  1 <I & 
 5 F*  1  F I1 +: J".%=M :23 3 2   3  5      2  3     53 I &   0$6-$/D-$. @%   
 F  C  F & E 1 : );

   <   J".%.M   5    2  2    ! 536   5  5  5 "#$6"$--D"$-# @.  I  3B I  + 
J".."M :23   3  2 3 !   2  3     53 4 ( &  /60."D0..
                    !      4  2    ;                      2 2   2; * 5 ;         5    A  5  3  5    2 2      2    ;         ; )   !  2 A  3  5 22  ;       5 3     5  ;   5    3 5 *      2!A         ;    5   ! 3         3          2 A   !A       A  5           3    5 :  3        2 A  ;1 5 ; :  3      2       35   5  3     ;        23 2       P" @Q ;                5   *    5X P-Q  5  3                5 2!A     " 6     :  3        2  JY/M   JY/M F!    ;  3   3  2 "$    2 =$ 33 ; 5B J  "M :  F     2 2 3  2      J$"@ 33M ;    !                5   X +5 1  3   2    J

   "..=U &  "J-M60D#M 
!"#    7 7 $   #%  &  ' (   7 "7 $&     %) & "#   * (    +&       &       
,  
6   7 4A   ;         5 ;    $$@ 1 3      2 $" 1 2; 5 $$/ 1 3      2 $-/ 1 )   3  2     ; 3       5      )       3   5  3 !      3  J  @M 4  2     3  ;      5  53  2 J  -M :  3 3        

  ;     2          J  0M :     2 ;    *       !   ;   36 Y3
 5 
3  -  *   *   #    %   ' (  
 7 &7
#  & #%  * & &         &  . #       8 *   , 8 * '  
!"#   
6 7 F    3  2  3       23 2     :   3          23   !   5                     J  /M : 5     J"  @M     5       J
M

    7 4   ;     A  5   23  5 ; 23   3      " @ 
    23 6 A 5    ;1Y  5   ;1 FA
F  :  5  ;      ;   2   23!   ;     2    
3       7 :          5  !   ;     35   :     2  Y@
 
  4
@   :    A   2 2             !   @6   2Y  @

  
@  
  5       2  5    2    5   !   2 2    J  =M  Y  2
     5 @
 
 
   @         - 3       ;    3        A3   @$ 33   2      3   5  @ 33  3     4 2     2 @$ 33      3    3     2 -@ 33 :  !    !            ;U ;  3        3    3  :     ; 2     ;     0  A3!   A   2 " ,    2  @0 33   2      4    2 " ,  5 # 33    :       2      5  3 A        3   :           5   2   23  4   5  5     3 !    2     3        -  *   *   #    %   ' (  
 7 )7
#  & #%  * & &         &  ##  *  /   , * '  
!"#    7 *7 +& "#   * & (    .  & "#   *  #  0 & &  #. #      &  & +&% & '    12    12     #  .  1
2 7 +7 +&   #  *   ' (   & *   * & .  +& #     &      +&     &   &  ' 0 &   &   #   & '    3'  & '   % & & #  +    3    3   ;   ; !   3  J  =M :  3            3!   3  J"$ 33     5M           A3      ;  3   3  2 "$ 33 A       2 $$@      2 2      ; 2  -  *   *   #    %   ' (  
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
!"#    :   5            2      F!    35             A    3!    5    - 5      3   3   5            A    ;      ;   3A 35                    @$  0$V  ! 53      P%Q :     1     ;        '   F       1  *  33     5 ;    53  2  5    :     2  3 !   ;         2; 5 !   2        2 3    53  *      2   A        F   ; 5B 5 3 3  2         ! 3  4 5  5  2  A          2       ;     2 2  5       2        3  :      2  

   3 2 5  3  35       5 ;     3  5            5       :        2  3   B         2  3  3  B  2        " +5 I) ,;3 (&     F

  J"..=M 4 3   !         2      I ( *    #6-%"D-%/ @  1   &    C  < 4    &B
 C    < J"..0M &     5    5 2 A   3    6 *      ;  5  2   I ( *    /6#0/D#/$ -  5 +  5 :  & )C :    J"..=M &  A    ! 3  <2N  6 A  3  <         &;  &  &  "J-M60D# 0 3B I J"..-M *     6 :  !         !   43 I   "=$6="-D="% / &  B 4 3B I  :  G <   <  G   & J".%#M +!A!    5            #=J@M60/$D0/# = &   + & 5 F &  *C :  I< +     I   + C 23 ' &  22 I!  1 4 & ; ' J".."M    1 5  2       2    ;   5 F      %0J0M6"=-=D"=0- # <  +   <& J".%.M       3     2  !    3  5    56 :   3 2     3   @$J/M60##D0%/ % K 12 ) 4    & 13 <    &3; B C J"..@M  !    ;    3       2     2   ! 1   *    "/6@#@D@#% " #    $ %             +           3 2          4   31 3    2  A   2 2  ! 53   !  2     3  ;   P"D-Q * "..0    3      5    5    3B    5  3 2 !; :  ;        ! 3  :     ;     3     23 5   P" 0Q :  3  ; 3         5        23  3 5         :    3  2         5  2       1 2   1  2    ;    2   2     53  F  3B    ;   5 2      A   !     3        5 3           3B    X P/Q" 6    
# 7 & A 3B    ;   2 - 3     3!  2 -0 33   ; 1 2 $"/ 33 J@.0U I   I  *   &5 3 F ,IM ;  5  A : 5 ; 3    0 33   33     2      A     3 3   J  "M *3       5 2    ;    5 2  2   A  2 0 33  3!  ;  ; 1 2 $@ 33 3 2   5    &3!  2  A ; 3                     X +5 1  3   2    J

  "..=U @6.#D"$"M 
!"#    7 7 3'  * 4  )  % &   % *  .      " 7 "7 $    " ' 4  )  *  ". #   %   
4  7 & A    A     J  @M ; A  5    5  ;  3   3  2 =  "$ 33    2 0$ 33 J&  Z  [ 4< +L  &; BM   2    ;    ;   23    A  :   3  2     A   3  2     3      ; 3!      A 3        5         3  2 $/  J  -M 4 3  ;  33     3             3 3  J  0M :  A !   2   J+M ; 2  5    2    3  J
+M     3    3  J
+ 3M +

+
+3 5 " '  * 4  )   **   "#  
 7 &7 6&     * &  1 20 &      &  120  &  '  132 +& ' & &   &         0 #". , 0   , 0  7 )7 3 #       #        0  # , 0 #  , 0 ' ' :  A   2     ; 2  5    2     3  J
& 3  M 3          3    3  J
+ 3M &

& 3 
+ 3 :  ;  3   2            2;  A   5  2  22  3 
4   7 ;    A         ; ! 3 23        3  2  3    ; 3  J  - M &    ; 2  5    2   22!    3   ;  3     J
& 3  M 3    3   J
& 3M      3  2  3     J
& 3  M &

& 3  
& 3
& 3 
4    7 F 2      2      5 !  *   5B  2  2    2     2  3 23   ; 3    3  2  3             J  - M &  23  J
&M ; !       23   5     2    * 
!"#    ; 2  5    2   22 2    J
&  M 3    3      3  J
&  M          3  J
&  M
&

&   
& 
&   4    A   3  \&
 :  !   !  ;  2 3    ;          :   A  3   3       5   2    2 "%$ 33 JY$@- M  3     4  A        2 5  5  5  33  2  A A       ; 5   2  2    :     5   ;  3   2  5   ; 3         3      2  A J  /M    =  ;  A   2             2   2    ;   2  2    *  ;!    3  0     2  5  2      5  ;  3   2    :    2  !        2            ! 2   %$V ;       A  5 5 =$  =/V + ; 0  =     2         3   *   ;         A 3 5  3 2!          :   3  5 2      3!       A    ; 0  =  ; 3 5  3  5       2   F     2    !      2     3A 3 2  2      7 *7      *  "     7  *   %     &       " 1  2 5 " '  * 4  )   **   "#  

    3 2   2              3 2 1  !   
!"#       "
3  +)
 E :  &G   
F +1 ) )+  5 J"..%M :       F   I ( *    .60#"D0#% @ E 1 : F  C  F & );

 <   J".%%M &2!A !    2 6 4 A  3   5   43 I   "/"J0M6=#-D=#= - G 1 :   ) ( I E  I< G 1  &  C & B )4 !   )5 : E3B1 E J"..#M 4 3   3 5   6 4 3!    ;        2    I &  !  =.J@M6@//D@=# 0
3  +) :  G ) ^

 E J"..=M  5         ; 5  2 5  2  5!       I ( *    #J/M6=0"D=0. / G3   &  &) C   C 3 + <3  J"..=M !       2        2 6 22 2   3        2   & 5 "@$J"M6=$D#$ =  
 F
) G;    +    I   J"..=M          2  5    .0J.M6**!"%@D"%# # )3  4F +  I& A 4
   ' 5 , 1  J"...M :  !    2    !2          233 5           &  #J#M6#"$D#"0 % F  I<    ,I < 

H4   E  C B  IG J"..%M !   2 3      2  4 &  "--6@0=D@0. . & I!*   G E 1; I   1 : J"..%M 4 A  3   5 2      ;    2   22       !  3   *    @"60/D0. "$  B  5    I! 
; 4 <   J".%#M     5!     2         3    2  + 3   %6-D"" ""     ) 1        ,13     , J"..%M + 3  ! 5 2 5  !    6 :             
 0/6@$@D@$# "@ B  K  
  I: K  F C I J"../M 4  3  2  3 2  3     53      3    2  ;5          ."J.M6@0#$D@0## "-  5 +
 G  & )C &   I :    J"..#M
 1   <2N;  2 3 1 W 1 N   (4&4 @=6--D-% "0 3B I J"..%M  ; 2 53  2 3   2     !  I ( *    .J"M6#D"- "/ 4 E ;  FI (      I J@$$$M
 3 2  3 5!      A  22       I (  &  @$6@0"D @0. "= &  33  F   +  4 + !< F  & 3   J"...M 5 5     5     2     6 !  3 5       A  3      2    5 "#$6"=.D "#/ "# K 12 ) 4    & 13 <    &3; B C J"..@M  !    ;    3       2     2   ! 1   *    "/6@#@D@#% +& &    #    #%  &  ' (  
 "%   <  IF & 5 F  + 13   F1 )I  )  & )C &   + + CI ( ; 
J".."M 5    ;  ;     !A            %-J/M6"#%%D"#.% ". + C ( 3    I :1 : )2 &( ,;3 I I  ) !  
J"..=M   3   2   ;    3B    & 1 :  3     F         23    !    .-J"@M6@"="D@"=. @$ & ; B &   5 I< ; F
 3  4  (     3
J"..$M    2      5 4    2  3  !   5       %@J=M6@".$D@@$$ @"  + 13    <  FI  & 5 I +  +3 : &!  5 F J"..-M   5 2     2 3      5  2 6 G   23   5 A  2  -  -@$ 5 2    3    I 43   5 @"J"M60/D/0 @@ + C &  )   4 F &I J".#@M   2       36 *  3   4 &  "#/J"M6""%D"@# @- < 4  &  C 13 :  &   4 ; 4F J"..$M   2 5  2  5    2   2  5 2  5 I ( &  ""J=M6%-%D%0/ @0 ; 4F C 13 :   5 4 J".%=M   3 2    2     3   5 ;      2  3   3  3 3     :   43 I   "@-J@M6@@$D@-$ @/ &   4(    FI &  B
   J"..@M   2  !  5 5  2  5    2   2  5    2    !    & 5 """J=M6=##D=%@ @= & & + < <5 )  4  : J"..0M :   2          2     I ( &  ".J0M6/%/D/.- @# & ] F !
 5  , F B1 4  ; 4F &  ) J"..0M 2 2 2    B  2 3   
 2       2    2   I ( &  @$J0M6/0=D//= @%   ) ( I 53 I & B )4 +1 F &  F
 )5 : & ; B )   F  C   I J"../M  3  3 5       2 6  3 5      5  2   2        23  5      I ( &  @"J0M6/./D =$0 @. <  :I +5 <4 <  &) F  < E F 5 I J"../M    2   5 # 3  2   2   3   3 2  2  3!     53 I  &  @J0M6-0%D-// -$ ; 4F J".."M <2    B 6 :   2    3  3 I ( &  "-J/M6#-0D#-= )                           A   2    53      3   5U ;       53 3 5              !    5  5 A 3        3  :            53       !  5   3 2 2 A     1  2   2           G ;       3       3        53  2         2      3  ;   1 1  & 2   3   A  5  3   2 A  1      A 3  2  !        1 ;  2  2 A   3!    3     P"Q )  B  2     ! 3     22    2    1 2 2  2    3     22     P@Q +   2 A  5  !  2  A 3  2         !     3   15         
     !    2         ;          2 3     )   2     ;     1      A    P-D=Q  2    ;    2 1 2   ;    5    P/ =Q ;    ;  5          2 35      5   5  A    P#Q  ! 5   233 5   ;  2               2 A  A  3             ;  5   :   3        5   ; P%D"$Q G   5    2 5   1    5   233 5   P""Q :       D 2! 35         3  D   2       ;    

 ; 5 J  0M  !   35    3 3   ; 3  5     3    ;   3    2    3 !        3  2  :  ;   5    ! ;     5 2  5    *2    2 53 ! 5        2 2  2   ;     5  :    3     2  '    ; 3  ! 23    3 5 2 2    35   3 3     J  /M 3     ;     ;   5   ;    5 2 35          $    * "  * #%  ' #%  &  '   #   
 7 &7 9   &  *  # & 1 #2    &   *  '  % 1  .  2 ' & &  & * &  # &   +&    #   & 4!3   $ & %# *  # &  = %  # &    * &  %    +& &       * #    #  &  #" %  %  & '      5 ;   5   2 53 5  2    ;        B  ;   ;    ;  3      2 3    2      :   !    ;     3     ;     2   2     23    
!"#    7 )7      . 0 4!3 6#   *  .   %   * & & % ' #%  '  1    *  
2 5   .  &   & *   1    *  
2 4 3   2    3  ;     2   2          ;    5  2    :   1!  2    3  ; #0"\#- 3    2      * 315     2  5     5    J-@"\/% 3   -$"\/- 3   5 ,&M J  -M :  3      ;     5   5   5  ;  -  :;          3    5     ;  5  !   6 :  2 3   3 2      5        A   2  3    A  3!             ;    3!     3  3 !2   3 ;     5 3!  2 2     3 3       !  $    * "  * #%  ' #%  &  '   #   
 7 *7      . 0 4B3     * & # %0  .  #% . &    &  % * .   1    *  
2 *         1 ' 2   &  # 1    *  
2   !  5 P% "$ "" "-D"#Q :     3 1 ; -.$ 3     ;  5  P"/Q  0=$D0.$ 3   ;  !; :   P"$ "%Q :  2     !  ;     :     ; 3  5  ; 5!   P% "-D"/Q  : P"$ "# "%Q 5   ;     ; P"" "=Q   3    3   2  5 P. "" "0 "# "%Q    3      P% "/Q  ;      3         2         5 * 3 2     3     3   ;           5  3     ;                     ; :   3 !    3    P"" "0 "/Q        2  A    3       5      6    3  5         ; P%Q &       2        2    2  ;      ;    5     5   2   ; 2  ;  B 3  ;  5   ;       2     5       P% "- "0Q :  5  2   ;  5    ;  5      
!"#     2 3               ; 2  P".Q :         2   ;   !    5  2     3   2        P@$Q : 2  3 T2   H  2   !  ;  3     3   3   3          2             3 5  2    ;  5       5 3  233 5   ;              2   3    2 :,!  2;  3    P@" @@Q :    ;       3   ;   53 3    5 1  F      5       2 3   ;       ! 53 3 ) 233 5   2  2 3    !                2   ;    5 3   3      P@- @0Q     A3 !  2 3       ;      ; 2    2 3        3      ;   2   P@/Q :  '    3              ;1!  A   2    2  4 A        !       23    3     2            :  ;    5    !   P"" @=Q :  3   2 2  5 2   2      ; 3   A  3 5 5    B   !       3 5           2   ! 3   2 P@# @%Q :        3  22   5 23   ! 3   P@$ @.Q :  2   2 3 2    35      2! 2          5 2        ; ;   3  3   2        A  3!    3   2   2;  A    2     3   :  3      ;  5  !          P/ = @/Q :   2   !    2   ; 2  ;  5  B 3   !    A  5        5   3 23    ;       P= -$Q ;          ;  5             ; ;   ; 2  233 5      5   2 !  23   3     2   P#Q :      ;   2    '    :   5 5      A 3  A 2  233 5     A  2  2  35  ! 33 5   B 2   G 35        5   ;      5   5        5     3   23 2              ;    3 2 5        !    3      A 3     2 A    $    * "  * #%  ' #%  &  '   #   
 :   3          5 ;    A   2  '    :   2!A   3 !;           2A  5     

A   32   ! 3  :  3      ;         2   5 3    A   2 A  5     :  ;   2  3  ; 2       2    ! 3    3  2  4  5 22    5       ;    5      P"- "0 -"Q 4 A      2   5    2   2 !    5        2   A   35       ;      A   4   2     2  A 3 !  1 2   " 33  5  1; 2;    P-@Q       53    P--Q :     ; 3 !   5    A   2  2!A    3 !;       5           ;  A 3 3 2   3 !;      3   32   ! 3   3   2   1 A   *        2 A  ;   5  A             5     2  !    ;   3 5    3    !  ;        2 5   ;  :  2      3     3 2 2     !    3   3 2     ;    3 2   5             2 A  :   2!A   3 !;  3       3   2   1       "    )   + * C )     I + 1; I J"..%M !   444 A  6 F     ; 1       !2 3 ! _ I  &  /6-"$D-"# @ )3  4F F  3
I +  I& 1  J"...M A  3  3 2 A 3   !2 J* ( :M 2 A   3   2     I (  &  "#J"M6=$D=/ - 4 3 E& 12 ,   & 3 CC (  ,   < C B I!< I  J"..%M  2   3     53 2    2 ! 3   53  B    I ( &  @%J"M6"#%D"%- 0 : <+ C1  C B1 + '3   : J"..%M  2  3  !   53   5   5     2  I ( &  @%J"M6"%0D "%# / <      E
   E C  F   5  : 4 3   & A  ) , ) +  + 3  I!  I   ) J@$$$M   !    2 6 45  2 A        23  ' G&:4   5 I  :  #J@M6"$/D"@@ 
!"#    =    + 1; I * C IW I   I )   + J@$$$M !         2 2 A  2    !2   3  !      I (  &  ".J"M6/D"" # &  C   F C ;  I
 3 B I4 F  1 F  3 <   : J"...M   5    3  5 2 A   2    !    2 6   2 5       I  &  =J-M6@0=D@/$ % )3  4F +  I& A 4
   ' 5 , 1  J"...M :  !    2    !2          233 5           &  #J#M6#"$D#"0 . 3     E  5    < & 3  C  F
   E < !   J@$$$M      2    53  6    2   ! 5  2!A     !2  I  :  #J"M60#D=# "$
3  +) :  G ) ^

 E J"..=M  5         ; 5  2 5  2  5!       I ( *    #J/M6=0"D=0. "" B  K  
  I: K  F C I J"../M 4  3  2  3 2  3     53      3    2  ;5          ."J.M6@0#$D@0## "@  5 + ) 4I    4  & )C J@$$@M +   2 A  2 5!     5           3 I ( *    "-6=$"D=$# "- F  I<    ,I < 

H4   E  C B  IG J"..%M !   2 3      2  4 &  "--6@0=D@0. "0  
 F
) G;    +    I   J"..=M          2  5    .0J.M6**!"%@D"%# "/     C  1 <    
11  :
5  F  4 J"..%M F2 :    6 *       43 &  =0J"$M6"$$@D "$$= "= & I!*   G E 1; I   1 : J"..%M 4 A  3   5 2      ;    2   22       !  3   *    @"60/D0. "# G3   &  &) C   C 3 + <3  J"..=M !       2        2 6 22 2   3        2   & 5 "@$J"M6=$D#$ "% 3B I  :  G )  I  3   I   5 C
    J"../M ' 2     ; 5  2  5  A  3   3!     53 I ( *    =J=M6%#.D%%/ ". 3B I J"..%M  ; 2 53  2 3   2     !  I ( *    .J"M6#D"- @$ ; 4F J".."M <2    B 6 :   2    3  3 I ( &  "-J/M6#-0D#-= @"     43    + 3  I J"..#M    3 2 2 ! 3   53 5  &   5 36  3 5    2 #.  I ( &  @=J@M6"..D@$. @@ , ) &;   J"..#M +         3 2  3     53 I  &  06"=.D"#- @- 5  43    < 5 * 4 ;  F )   4  J"..0M *  A   ;     &5 3 " 2   3   ! 3 2       I  &  "6-"D0- @0 5B


    :  I &    J"..#M 4  233 5      ;  3  5  2  I  &  06-/0D-=$ $    * "  * #%  ' #%  &  '   #   
 @/ <  :I +5 <4 <  &) F  < E F 5 I J"../M    2   5 # 3  2   2   3   3 2  2  3!     53 I  &  @J0M6-0%D-// @=     ) 1        ,13     , J"..%M + 3  ! 5 2 5  !    6 :             
 "0/6@$@D@$# @# ; 4F C 13 :  ;  J".%=M   3 2    2 2   **       3 3    2     5  ! 2  5   I ( &  -J=M6%##D%%0 @% < 4  &  C 13 :  &   4 ; 4F J"..$M   2 5  2  5    2   2  5 2  5 I ( &  ""J=M6%-%D%0/ @. + C &  )   4 F &I J".#@M   2       36 *  3   4 &  "#/J"M6""%D"@# -$ F  4
5  (     F     :   C  1 < J"..=M
! 53             2   3                5  53 I  &  -6@=@D@=. -" K  C F  4 & ; 
 C 5 
   +  CI  5 :I J"...M 4 A   2           2  3    !  536             I ( &  @.6@.@D-$% -@   3  I&  12 ) J"..%M    A   2   1 2  !      2  3     53 I ( &  @%J-M60@@D0-$ -- *  C4 <  G   C   +  &
F I4 J"..#M   2  A 3    226 *3    2    53    I ( &  @=J-M60.@D/$" *                  53     ;5       

4 53 ;   2     3  @@ 33   3  ;    53 ;   3  2 -$  -/ 33 ;    5    3   5  ;    2    2 3   P@0D@.Q :     3  5 #V ;  5 ; ,  5!           2           !   3  5    5    2    ;    !  P@0Q *3     A     5 2  3   !      2;    3  ;   2  A   !     3   3   2   53  J  0M &  1 2   53        2  53   2;   3 ;       2   2      4 53   5    2    3   2  ;    ;  3   3  5  3  5 2 ""  --V P@0 -$D-@Q :  3   ;      5        7 "7
 %  % "  * # & 1=  #    # &2 & & *  # .   ## & +&  % *   & #        &  & 
!"#    7 &7
    #&% * * .     .# '  +&  %  &  % #  1* 2
# % # .  * &  % &  "  % #%  &  ' (   0 * '  1 2  #*  1 & 2 7 )7 +  '    & & & "   %    * & & 0  &  # &  1#%  &  ' (   2   # & ,  #  *      ,  "  * # & 3  2  3  5   3      ! 53    3 3 2        4 53  !      --V ;   2   ;1 P@0 -$ -@Q 4 ! 53 3 2 2   5 =$  .$V     2;!   P-$D-@Q : 3  ;    53 ;    5  53   1!     ;   2;   3       5    3   2      2         
    *  '     * ##  
        2 2         3 5      2               3 5 2     !3 3 ;   3   ;  22              3  5 :  3    ;     ;          3   P-- -0Q :              2     5 3U     ! 53  3 5 ".V    2;!  : 5   3!     3     3     4 "$V  3      3   ;   2;  A   P--Q I  3     3   ;         3 ;     1 2   P-/ -=Q %%V 2   53    ! 5 ;         5    ;  !    3  5 3         1 2      2   3   1 :      3    ;    2  3 ; 3     5  3!  5 ;    0/V            3! 3    53 A   4 53  ;  2 5 A      53      P-/Q
     7 :   53   5  2   !2      5   B    B   23  53    2 5  P@# -#D-.Q  5   P-%Q ;   5  3 2    53 3 5     2  :  2 P0$D0@Q ; 5  !2    4     3 2  2    5 2    2   2    53 ; A  2;     2  3        53 ;    5 ! !  3 :    2  3 ; ;    ; 3     3 4     2    1 1   3   2   53  A   2     3      3    :      22  !  ;   5    2  3  5 2 "@  =$V    ! 2  3      ;      2  3         !3  P@# -% 0"D0-Q 4  3    ;     2             2 2      53      ! 2    ;      4 53     5   5  !3 A2 2         ;1 2  3    P0-Q 4 2    53    3  2 @/  -/ 33  3       2;! 2 2     ;1 &    5 5     ; 3  23  ;!   53 A   1 1  2    2  ;    #/V    5 23  "..@     2 3   ;       @/V P-#Q *  3 !    !2  !    53 3 3 3 A       1     : 5 ;   5      4 ! 1 ;   5    2 ;  2   2 P0@Q   3 2  A 5  45     A 3 2 A    
!"#    P@.Q   1        53 P00Q :    ! 3  2      3    2    53   3!   3  3     B   5     ! 53 23   1 2;  A   4 5      ! 53  3   5  5     2     3!     53  3 2 P-# -. 0-Q

 0      7 :  3    5        3     53  3   ;    3  2 233 5       3    !  5 1   3  A!      5 ; 3!    P0/Q *  3   2     5  2         3  5             3      3 2 2  23  53 5   33 !3     2   P0=D0.Q
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

4 J".%@M : 3!  3    2    R
 ! AS    6 :            & 4 2 G .6@/@D@/# "@ 
 K 12 ) ) O  &  4 J".%0M    2;! 2  3 5       6 4 A  3   5   A!      5 "/@J-M6=/.D==- "-
  : +  3 F C 5 C W I J".%-M : 3  A !         2 6  3 5     5 "0#6@/.D@=$
    *  '     * ##  
 "0  4 )  < 55 I     !K   F 43  B C J".%-M ,   3 2      6 4 ;        ;   5 "0#6@="D@=- "/  4 )  < 55 I4 &3; B   !K   F  43  B C J".%0M       2   5 "/$J"M60/D0. "= 3B I  &    &  :  G  FI J".%/M A     3!  2 6 4  3 5  5  5 "/=J"M6#-D## "# 3B I    < , < &  B 4   
 <  <4 +1 <I ) <C
5
 )33 I      4    F  
  F J".%%M *  3            5   6  3 5   2  3     5  5 "=%J-M6#@#D#-" "% & ; '   I  1  ( I22  C   ) J".%#M *               2   3     5 ,  I  -"=J"@M6#$"D#$= ". + C &  )   4 F &I J".#@M   2       36 *  3   4 &  "#/J"M6""%D"@# @$   ) ( I 53 I & B )4 +1 F &  F
 )5 : & ; B )   F  C   I J"../M  3  3 5       2 6  3 5      5  2   2        23  5      I ( &  @"J0M6/./D =$0 @" <  :I +5 <4 <  &) F  < E F 5 I J"../M    2   5 # 3  2   2   3   3 2  2  3!     53 I  &  @J0M6-0%D-// @@    + 1; I * C IW I   I )   + J@$$$M !         2 2 A  2    !2   3  !      I (  &  ".J"M6/D"" @- &  C   F C ;  I
 3 B I4 F  1 F  3 <   : J"...M   5    3  5 2 A   2    !    2 6   2 5       I  &  =J-M6@0=D@/$ @0 (  
5  C  1 < &   F  4 J"../M 4        53 3 2   5 2  3  2  I *  &  %6-%"D -%% @/ +1 4  1 <I & 
 5 F*  1  F I1 +: J".%=M :23 3 2   3  5      2  3     53 I &   0$6-$/D-$. @= &5
 : 3   , 3 4 +  J"..0M      2  ! 3     536  3   2    A 3      + I &  %"J%M6""$#D"""$ @# < 
 4  I
H4   E  C &5 C&   FF ) ]  F     , B  IG J"..#M 4 ;     2 2 !   2  3     53 43 I &  "#-6"/.D"=0 @% 3     E  5    < & 3  C  F
   E <   J@$$$M      2    53  6    2   ! 5  2!A     !2  I  :  #J"M60#D=# @. &1  & &  &)  5 C:
 ) 3 + <3  J@$$@M      A  3   53 3 I ( &  -=6%"0D%"# -$ 3B I  :  G )  I  3   I   5 C
    J"../M ' 2     ; 5  2  5  A  3   3!     53 I ( *    =J=M6%#.D%%/ 
!"#    -" B  K  
  I: K  F C I J"../M 4  3  2  3 2  3     53      3    2  ;5          ."J.M6@0#$D@0## -@ + 4 3B I  :  G &   4    I  & J"...M 5  2  5!      !2 6 '  A  3   3     53 I ( *    "$6=$/D="@ --  
 F
 G;     +   G21 + +   +!  ;1  &  II J"..=M    2  3  2    A  3     53 3 I ( &  @-J/M6%".D%-" -0  
 F
) G;    +    I   J"..=M          2  5    .0J.M6**!"%@D"%# -/      F4 F5 I: )   I  :4 +   &  4   4 C5 +4 J"../M 4    53 3 2     2 !   A     I ( &  @@6-$=D-"/ -=   F4    +  4 C5 +4 J"..=M    2  53 !    ;   2       2  3     53     3 4 ( &  "$J@M6"==D"#- -# )  I   I  3  :  G + 
  I    3B I J"..@M *  3  5  2  3     536   5  5  5 "%06"%/D".$ -%  +  +! :  & )    3  1 J"..-M &2!A   3        2 23 A   2    53  6  3 5      *    "=6-..D-0@ -.      ! +  I! :      <   J"..0M : !  3!
 1   2    3 2    536 4 ! 3 5 A  3   5 I ( &  ".J0M6=.%D#$= 0$ & B )4   )   ) G 1 :    5 + & 
 G  & ( I J"..#M     53   3 3 6 4 A  3!  3  2     22  2      53 A !   I ( &  @=J@M6@@@D@-$ 0"   ) & B )4   )   )5 : G  & ( I J"..#M 4 A  3  3 2           2  ! 53   I  &  06@.$D@.# 0@  5 + & B )4 G 1 : F  4   ) 53 I   ) ( I J"..%M 1 2    2    2 A  3     ! 536
   3  B  ;    RS ;   53  _ I ( &  @#60/0D0=@ 0- F  : +    &  4    I + I
 <   J"..=M 4 ;  3 3 2  3     536 *        3 ! !;       I (  &  ""6.$D.# 00 &  1 <F 4  ,I + I  C )I &  B +1 I  J@$$$M A  3   5 2  2  2 1  B      !  53      2 3   + I &  %#6#"D#% 0/ ,;3 C I!   I )  35 F< :  
J"..0M 5 1       5      3     53    .$ P  @QJ/M6**!@@0D**!@@# 0= 4  & &B3 I!) <  
 )  3  I!    I!+ J"..$M  !   A  3   53       %@6.#-D.%" 0# 4  &
   +     < 3 < 2 < J"..@M   2 233 5 2    ;  3 2 A  3     53 I ( &  "=6"-.D"0# 0%   (I ,13 <+ <  * B5  &  I( 35 F< :  
J"..0M :    2   3 2 A  3     536 &5!
    *  '     * ##  
  5 2    2      ; 3  A      ! 233 5    I ( &  @$6/"D=$ 0. F  I( J"..0M 4 53 23     5      2 !    I ( &  @$J"M6"/-D"/0 /$  
 , 
 ; * <   )
   4 J".=@M    2 A  3   53 ;     &   3 %6@-#D@-. /" +; ) C   )  +4 F  * J".%%M +   536 : ! 3 ;      5     3     5  5

J/@VM ;   5 2      @#%    J0%VM 5  2  :    5 2!        2    *      ;   2  X +5 1  3   2    J

   "..=U @6@".D@@@M 
3    ##     2      3  3 5 J :M  ;    1 2 /  % 33 ;    ; 3  2      3  ; 1 2   3    2  A 3     1       ;  ;3   5     2   :     3  2   53   3   3  2  2   3 ;    !      3 3     ; 23 :   2    3 ;    5 3  5    3  2   ;    1 :   2  1 ;  3 2  3 ; 5   ;  3  3  2 "/ 33    5   ;   3   2   / 33 :  T  AH ;    5    2   53  3     5   53  P=Q  2   53 ; 
   A   3\&
 4       3\&
 3    ;   ; 3 5 3 ;5 4,G(4    5     5     $$/ ;     2      +     2 %.  53   2         5  2  53     J  "  :  "M :5  *  53    22   5  A 3       1        : 5    5 ""V :5  **  53     A 3 1 5 1   22   5     1 4 53  3   ! 7 7 7#&    *   *   %   &  %# 3 *   * *      %  &  #   '     %
 ' 7 +& ** %# *  % < +&  #&   &        *  #       A '      

4  H  K%L F F F "#")
4"  8 K L F F  
4"  5 K L F F  
8   5 K L F  
*     5 K L F F F 

 % 8 K L F F F "#""%
8 * #*    K L F F F 
6     "   F F F "#"* 
> #  K:L    "#"$& +   '  , ,  ,   &  #  ' ,  #  '   ,   #  ' ,  #  '  2   3    5  :  2         : 5    0@V :5  ***  53      2  5   3     2   3      2   : 5  2  23 :  3 5     5  22 23    5  *  53 :5  ***  53  3 2   0#V : 5    2  5    ;   , 2  5  *  53  5 ; 2  5  **  53 J/VM    3  53  3  2 /-\"- 33   ;!    1 2      2   2  23  53 2 5  ***          J".VM ; 5  ***  53 2 ;   3  53  3  ; #$\"% 33 -  4  2      5  2  53 ; 3    !     2     3     53 4 A! 3 1   3      2     3  !    2 A  2    * ;   /-V 2   53    5 *   3     /$V 2 2    53     2         2  A 3 1 P" #Q :       5 3 ;     3   2    
3    ##     5     2  1    5   B 2     23    :5  *  53     A 3     1    5 ""V   2       ;      P" #Q 4     2       53 3    2   2    5     ;     2 1   2 A    P%D"$Q :  3 2   2   2 ;  !      2  2      5 3       1 2   5    ;     2    2  ;   3    2    5     2     1 2    ;     A 3     1 *   !      2       1 ;   2 A   1!      P% "" "@Q : 2   53      ; 2      P% "-Q :5  ***  53  5  2  5 2         2 3 2    2   2 1!   A 3 1 :   53    3  3  2 #$ 33  ; 2  23 *   53  3    ;  53      53    P"-Q :    2  !  ; 3    ".V G            5 ;      3 2  23  53       2  5    1 2        53  2  5       P=Q * 2  ;           5  *  53 :    2   3     B      23 3     53 2        2   ! ;  2  23  53       2   :    53 ;  3 2   '2  5 5     5   3          ;      53 ;    5 2 3 23       :   5  3  3   :   2   3    !  5  3        2  :  ;           A  2      3   3A 3  3    15    3          53   ; 2     3  G5   :  ; 3        3   3 ! 3      2   3  P"0Q *         2   2 2  ! 53  2  5 2       ;       /$V    5  *  **  53     A 3 1 :5  *  ! 53 3 5    2    5    3   ;  3  3  1 2   :5  ***  53     !  B 5   2  A 3 1   3    1 2        5    5 3 *   * *      %  &  #   '     %
    " 43  E  2 &F ) 2 C F 1 &  < & F 3 F  ! 1  + J"..#M 4 3     5 2  3     53 2 !      + I &  %06"#%D"%$ @  :4 <  G   C
F I4 J"..-M *2    53 3 5 I ( &  "#J=M - 5 I F  < E  F F   &   &   I J"..=M    2 !  3  2  2  3     53      53 3 5 4 ( &  "$J-M6@/0D@=" 0  I  3B I  + 
J".."M :23   3  2 3 !   2  3     53 4 ( &  /60."D0.. /  5 +  & )C & W 1 F   ;5 I :    J"..=M
 !    3  4  53    3 <   1  ! 1 N <2N B &;  &  @6@".D@@@ = G   C < 
5  &   C   I
F I4 J"..@M 4    2 ;

1  2     25    2    ;  ;       3             5 3   A   2        5 2           ;  2  5 *('&
$      0  7 :     ;        2;!      22     ;  " 3 2  5   A  5 2   53        5  2  - =  "@ 3      2   ! 3         :  ;    2  ; 3    55 4   5 ; 23  5     2     1   53 A  
4  7 4       3 \&

22 ! ;   ;    ;  @         2  2    ;   $$/   
$    7 :  ;   2 @0 3    !   3    5  12  5    5  5 5  2      5            ;  6%    " '  *   '        *  '     % # 
 3 33 3   J:  "M ;!  23 #  0- 3  J-0\"= 3 M
@A47 4    ; 3 5 23 5     3      2  *('& 3       5  3 *('&   5      5 2    1     B 5        53        2      4   5 ;  5 :    22    ; !    53 ;    "-      2   2  !     2     4         ;    ;    "@    2 3      ;   ;  3  G   ;  A 3  *('&      2    5  2   5 ;   5        2    ;   2  2   *3!    53 A   ;         3        ! 53 ;     5 *('&     5   ;  A !   * 2   *('&    3    3 2      3          :  ;      22         2 *('& ; 3 3            5  ; 
6     7 -$!5     3  5 ;  @V J@8%%M   2       2     3   !   5         J/VM ; 5 ;    3   1  ;  3     3 :   2  ;    ;  3   3   1 4 ' "7 .& #   % 3  B7 4 
3  ' &  *   
>   **  % 

&% & 
3 # '  # &% 
4  #   $    

  '  
5   
>   %   
? * *  

&  )   '   1: * #   &     #2  12 %   =  &   &    
3    ##   ' &7 8 &0 '   %#       % '     '  3 
     #   
  +  +   /+0 &&1 -  M M      M M  '     CM M G C & *  %    &         &  *   * ' )7 +%#  *   *  # &  ' 
  G  4  @   '        

 .     
+     
+             ;  2   2 2    2  2 J 5  ***M 1 :   2  2      3 ;        53 ;       ,       22 3 3         ;    -$!5 3!  3  5 2 @@V J:  @M 4            2  /-    ;    5     J,& :  -M   3      2  3 3 ; 53   ;   2     2 =V )        53         %    J.VM     2;!  
            7 A 3 3    ! 3   23 0$  %- 33 J//\"@ 33M :   2  A ! 3   1 ; @@\- 33 ;   2 "%  -0 33    2 @"\"- 33 @$V J"%8%%M 2         1 ;    2 /\@ 33 2      3  4   ;  5 3 5 3         9=$
;   2 2      :5   2    2          :  0

 7 5 1 ;   -=V J-@8%%M    !  
A  5           "% A 3       3   1 J 5  *U @$VM /    !   J 5  **U =VM  0 2 2  J 5  ***U /VM )  ( 1  !+ :  2   ;   23 ; /0V A 3     1 J#8"-M 1 ;   2  5    ; 5 ! 6%    " '  *   '        *  '     % # 
 ' *7 /  * &      '        

  '    
6     
4 O 


  
> #    
!"  M  )    
5  '     ' ; & ""      2  2  /-    J:  /M    2  2    3   1       B ;     !  53  3  : 5 ;    5 5     2            2    2  A 3 1      53  3   3 3   ,       5          A 3     A   J  -M    3  B     1 ; 3 *  2 3  A 3   3   1    ! 7 &7 +  *     %  "   +& #  #*    B6
*  &  *   * &   *  0  "   #   # .   # % #   & *   
3    ##      A   G 3   3   1  4    53  B     3   2      ;     3    ;    2            2  @0 3  G   ;  A 3 1   5   2   53    3  2    
    : 5  ;         2 2     2  "@ 3    2 3 ;  53   1   3  B  2   3   2  3    5 ; 23   2  ; 3      *       3 1       ;      3  2   53
    G 2  2  5  *** 1 ;     3      ;      3       ;     2;!  :  5  *** 1 ;    2  3   5       !   3 
        5 2 ;        5     5 ;   3      ; 2;  !   3  

any observed changes in the experience of those af- fected by the law can be attributed to the law itself or whether changes in experience merely reflect the environment in which the law was passed. If the latter were true, this would not be to say that the ADA was unnecessary, just that it did not have the dramatic impact some opponents probably feared because the changes were already occurring. It is also possible that no change in experience of the disabled will be seen leading up to the passage or follow- ing implementation of the ADA. If this is the case and if the experience of disabled workers in the labor market remains inferior to that of the nondis- abled, then we clearly have to look beyond the ADA to improve that labor market situation. The analyses in this book show that while disabled workers are making progress in some dimensions of their labor market experience, the ADA does not seem to have had a striking impact in either a positive or negative direc- tion. Expanding or strengthening incentives to enter the labor force, providing training focused in high-growth and high-earnings occupations, and assistance in screening and matching workers with appropriate jobs are policies that would capitalize on the recent progress made by disabled workers and move them in the direction of greater labor market gains. xiv 1 Introduction DISABILITY LEGISLATION IN THE UNITED STATES The excitement, fear, and general controversy surrounding the pas- sage of the Americans with Disabilities Act (ADA) in 1990 might lead some to believe that this was the first time the United States had ever confronted the issue of potential discrimination against or differential treatment of people with disabilities. To the contrary, the nation has demonstrated some concern through legislation for individuals with disabilities since the 1920s. Not until the ADA, however, was there as sweeping a mandate, theoretically touching multiple dimensions of a disabled person’s life. The National Civilian Vocational Rehabilitation Act became law in 1920, was amended several times, then became the Vocational Rehabil- itation Act in 1954. Public Law 93-112 transformed it into the Rehabil- itation Act of 1973 (sections 503–504), which prohibits discrimination against the disabled by any program receiving federal assistance and requires federal agencies to take affirmative action to employ handi- capped individuals. In addition, the act dictates that companies having contracts of a certain size with the federal government ($10,000 or more, as of 1998) publicly state that the organization takes affirmative action to employ and accommodate workers with disabilities. Execu- tive Order 12086 in 1978 reassigned enforcement of the act to the U.S. Department of Labor. This strengthened the position of disabled and veteran workers by placing the regulation enforcement in line with protection from discrimination based on race, color, religion, sex, or national origin.1 So, while nondiscrimination against and employment of disabled workers have been of concern for firms doing business with the federal government for some time, it was not until the passage of the ADA that all firms in the United States (employing 15 or more people) would be held to the same standard regarding employment and accommodation of individuals with disabilities. In many ways individual states have taken the lead in providing workplace opportunities for the disabled. By 1990, all states had 1 2 Hotchkiss passed antidiscrimination legislation covering employment by state agencies and often employment by any firm doing business with the state.2 In addition, nearly all states by that time covered private em- ployers in some form or another, and many states covered all employ- ment, including that by very small firms (fewer than 15 workers). Common exclusions from the discrimination legislation included reli- gious organizations, social clubs, family members, American Indian tribes, and farm or domestic workers. Details of when each state passed legislation related to treatment of the disabled in the workplace and the exclusions of those laws are in Appendix D. The Rehabilitation Act of 1973, and the concern already expressed at the state level about the employment opportunities of disabled work- ers, culminated in the passage of the ADA in 1990. This act was unlike others before it, in that it provided for the civil rights of people with disabilities in the same way that all citizens are protected against dis- crimination based on race, color, sex, national origin, age, and religion. The ADA requires that employers treat workers, and potential workers, with disabilities identically to those without disabilities, with regard to hiring, compensation, and other aspects of employment. In addition, employers must make reasonable efforts to accommodate the nature of the worker’s disability in connection with the performance of the work- er’s job. Owners of places of public accommodation are required to provide facilities (e.g., entrances, elevators, bathrooms) fit for the dis- abled, and to provide services in such a way that people with disabili- ties are not restricted from receiving those services (e.g., requiring a driver’s license as the only way to provide proof of identification dis- criminates against the vision impaired). Public accommodation also includes equal access through telecommunication, such as access to the Internet. While admittedly only a small part of the entire legislation, the implications and impact of the labor market provisions (Title I) of the ADA provide the focus of this book. Any interpretation of the effect of the ADA or recommendations for enhancements must take into account policies already in place that may or may not influence a disabled worker’s labor market experience. The Social Security Administration manages two cash payment pro- grams for Americans with disabilities. Such programs are of great concern regarding labor market analyses for two main reasons. First, cash payment programs might crowd out labor market activity, and Introduction 3 second, they may be structured in such a way that labor market partici- pation is discouraged. The Social Security Disability Insurance Pro- gram (SSDI) provides benefits to workers who have been able to make enough contributions through the social security Federal Insurance Contributions Act (FICA) tax paid on their previous earnings. The Supplemental Security Income Program (SSI) is available to disabled Americans who have limited income and resources. The eligibility rules for payments from these two programs differ, but they both re- quire the applicant to be either not working or earning less than some specified amount. Both programs include incentives to get recipients back into the labor force. These incentives include a trial work period where some or all of the payments are retained for a certain period of time; continuation of Medicaid or Medicare even if cash payments have ended because of high earnings; reimbursement of impairment-related work expenses; exclusion of certain income from the earnings test if set aside for future self-sufficiency, such as education or starting a business; and referral and payment for vocational rehabilitation. Many of these incentives were only adopted recently as part of the Ticket to Work and Work Incentives Improvement Act of 1999 (Public Law 106- 170). The goals of the Ticket to Work initiative complement the goals of the ADA. Whereas these work incentive programs are designed to encourage the disabled to seek jobs and become self-sufficient, the ADA is intended to provide an environment in which these efforts are met with support and reasonable assistance. POLICY ISSUES The United States has a history of enacting legislation with strong social content, expressing society’s ethics and morals. Child labor laws and other civil rights legislation fall into this category. One could argue that such laws are grounded in economic concerns. For example, discrimination against workers with disabilities or against African Americans robs our economy of the efficient allocation and use of valuable resources. Also, with the prohibition of child labor, children really have no other option but to attend school, raising the human capital of our economy overall. While these arguments have merit in 4 Hotchkiss fact, it is also true that as a society we support these laws from an emotional and moral level. For example, the 1991 Harris poll on Public Attitudes toward People with Disabilities demonstrated that while most people were not even aware that the ADA had been passed (62 percent), they felt overwhelmingly (95 percent) that ‘‘Given how many difficul- ties disabled people face in their daily lives, the least society can do is make an extra effort to improve things for them.’’3 In addition, 81 percent of those surveyed thought that there should be an affirmative action program for people with disabilities. These are responses not entirely driven by economic concerns. When legislation is propelled by an evolution of ethical and moral concerns, we must face the question of whether it serves as a statement of where we (as a society) are rather than as a prediction of where we are going.4 For example, the establishment of a minimum age for employment (child labor laws) has been shown to have had little impact on the decline of child labor in the early part of the 20th century (Moehling 1999). The implication is that legislation of strong social content, rather than precipitating social change, is often actually a re- sponse to social change. In other words, the ADA might merely serve as a reflection of our moral and ethical beliefs rather than as a tool with which to improve the condition of a segment of society. Some argue that the ADA is ‘‘feel-good legislation that promises more than it deliv- ers’’ (Jay 1990, p. 23). A major criticism is that the ADA is absent of specifics necessary for effective compliance, particularly on the subject of what constitutes ‘‘undue hardship,’’ which serves as the measure of whether a firm must make the physical environment, service, or employment ‘‘accessible.’’ Some interpret the refusal of Congress to tackle the difficult issues that were sure to arise as indication that the ADA’s primary function was merely to be a statement of our morals. In addition, an amendment to the ADA that would have disallowed jury trials and punitive damages for disabled victims of discrimination (an amendment that would have been a clear sign that the ADA was not meant to have any teeth), was only narrowly defeated (Jay 1990). The implication, if the ADA serves merely as a statement of where we are, is that no impact of the law will be detected because, for the most part, we have already adopted the principles and practices laid out by the legislation. This outcome, then, begs the question of whether the ADA or child labor laws are necessary, or whether such Introduction 5 legislation is simply an expensive declaration of our morals. Even though one could argue that these laws might merely be statements of something we already knew, an even stronger argument might be made that public acknowledgment of a collective moral foundation serves an important purpose, one beyond quantification in economic terms.5 These laws strongly proclaim our social values and provide a legal mechanism with which to arrest the activities of those who have not yet adopted those ethics. In order to address whether the ADA merely serves as a statement of where we are rather than as a prediction of where we are headed, the analyses in this book will focus on two basic questions. First, how are disabled workers faring (relative to nondisabled workers) at any given point in time, and is their relative experience in the labor market improving? Second, did the ADA have any discernible impact on the relative experience of disabled workers? These questions will be asked in relation to as many dimensions of the labor market experience as possible. The questioning does not stop with the analyses, however. If it is discovered that the ADA has had or is having a positive impact on the labor market experience of disabled workers, then the ADA is accomp- lishing what it was designed to do. If the ADA has not had a measur- able effect on the relative labor market experience of disabled workers, and if their experience still falls short of that of nondisabled workers, then we may need to look toward additional or different legislation, specifically targeted at improving those dimensions identified as the most lacking. FOCUS AND STRATEGY OF ANALYSES This book is concerned with the labor market implications and impact of the ADA. In addition

to the multiple dimensions of the potential effect of the ADA on disabled workers, there are at least as many more ways in which the ADA influences the lives of all disabled Americans; these other outcomes are not the subject of the present discussion, but may in fact amount to a much greater overall impact than that felt by the disabled in the labor market. The strategy of analysis followed here for documenting the impact of the ADA on the 6 Hotchkiss labor market experience of disabled workers has been to assemble as much information on as many dimensions of that experience as possi- ble. The major contribution of the analyses that follow is the wide- ranging coverage and synthesis of a massive amount of information in such a way as to make recommendations for policy. The emphasis has not been on developing new ways to examine the labor market experi- ence of the disabled, but to broaden that examination. The focus is on labor demand issues, defining the environment that the disabled might face. As a result, the analyses of employment and wages, for example, will correspond to what a disabled person might encounter upon entering the labor market. The conclusions will not be conditional on the labor supply decisions of the disabled, but will take those decisions into account in presenting unconditional results that apply to the population of the disabled, instead of merely to the sample (of workers) on which the estimates are obtained. Other analyses, such as the incidence of voluntary part-time employment, job separation, or job search experience, will be generalizable only to that population for which the issues are relevant: the part-time employed, the employed only, or the unemployed only. These sample limitations are legitimate and logical given the population for which such questions are relevant. It is important to remember that the purpose of the labor market provisions of the ADA was to break down barriers to the disabled and to improve their experiences in the labor market. Although perhaps expected, the alteration of various voluntary behaviors (such as labor force participation) was not the goal of these provisions. A fair analy- sis of the ADA should only involve an evaluation of what it was de- signed to do. Regardless of its intent, however, any policy can have unintended consequences that should also be addressed. Outline of the Book Chapter 2 explores employment outcomes among the disabled. Both joint labor force and employment and unconditional employment probabilities are examined for the entire sample of disabled individu- als, controlling for selection into the labor force. The availability of firm size and the phased-in nature of the ADA are exploited in a differ- ences-in-differences analysis. Results by type of disability are also presented. The joint labor force participation and employment proba- Introduction 7 bility for disabled persons declined relative to this joint outcome among nondisabled individuals after the ADA was implemented. How- ever, the unconditional (i.e., controlling for selection into the labor market) employment probability did not change post-ADA, relative to the experience of the nondisabled. The source of the deteriorating joint outcome is explored in some depth. In addition, employment among the disabled was found to shift more toward larger firms than did em- ployment among nondisabled workers, suggesting that implementation of the ADA and the financial ability (of larger firms) to accommodate workers’ disabilities mattered in the employment experience of dis- abled workers. Chapter 3 looks at the wages earned by disabled and nondisabled workers. A pooled, cross-sectional analysis suggests that wages among disabled workers fell post-ADA, relative to wages among the nondis- abled. In addition, a standard decomposition of the wage differential observed between disabled and nondisabled workers is performed. The availability of benefits is also explored through a simple probit analy- sis. While the overall compensation experience of disabled workers is found to be deteriorating relative to nondisabled workers (in both wages and availability of employer-sponsored fringe benefits), the de- gree to which discrimination might be used to explain this differential is also declining. It is found, however, that wages of disabled workers explicitly covered by the ADA (based on the size of their employers) have not changed post-ADA, relative to their noncovered counterparts, suggesting the overall lower wages among the disabled are being driven by more than accommodation costs. A number of job quality issues are addressed in Chapter 4. First, hours of work and the incidence of part-time employment and type of part-time employment among disabled and nondisabled workers are explored. Second, the distribution of workers across occupations and industries is compared using a popular distributional index. Third, the representation of disabled workers in high-growth and high-wage jobs is evaluated. This chapter presents evidence that while the incidence of part-time employment is increasing for disabled workers, relative to nondisabled workers, the incidence of voluntary part-time employment is driving that increase, particularly among workers with mental disor- ders. The degree of dissimilarity and the growth in dissimilarity in 8 Hotchkiss occupation and industry distributions of disabled and nondisabled workers over the 1981–2000 period are striking. While showing some improvement since 1992, this is of concern since disabled workers also appear to be concentrated in low-growth, low-wage occupations. Job separation and unemployment experiences of the disabled are explored in Chapter 5. Results from a multinomial logit find that, among individuals who have separated from their job, disabled workers are more likely to have separated voluntarily and less likely to have separated involuntarily than nondisabled workers. A similar analysis then finds that, among the unemployed, disabled workers are more likely to be reentrants and new entrants into the labor market than nondisabled workers. A duration analysis shows that disabled job seekers are searching on average three weeks longer before finding a job than similar nondisabled persons, and that most of the difference in observed search length is explained by differences in individual characteristics. Taken together, these results suggest that while the endowments or characteristics of disabled and nondisabled workers ap- pear to be valued equally, employers may be going to greater lengths to discern the fit of a disabled worker’s skill set with a particular job, thus leading to longer searches, a better match, and less chance that a separation is for involuntary reasons. Chapter 6 explores the impact of state-level legislation on wages, employment, and hours of disabled workers in different states. The analyses in this chapter exploit the differential timing of protective legislation across a number of states. The results are consistent and support the findings from Chapters 2, 3, and 4 on these same issues. Namely, wages decline and overall employment probabilities are un- changed among disabled workers, post-legislation, relative to nondis- abled workers. In addition, part-time employment among disabled workers increases post-legislation. These results suggest that the wage and part-time employment effect of the ADA may have been much greater if the state legislation had not already absorbed some if its potential impact. Chapter 7 synthesizes the results of the previous chapters around policy implications and recommendations. It is suggested that three directions be followed to further enhance the labor market experience of disabled workers: 1) provide incentives to the disabled to enter the labor force and relief to employers for the cost of accommodating these Introduction 9 individuals; 2) expand the support of resources available for disabled workers to increase their general human capital and ability to move into high-paying occupations; and 3) provide mechanisms by which employers and disabled workers can find each other and determine the appropriateness of the employment match. Data Details and Estimation Issues The combined Current Population Survey (CPS) Annual Earnings files for the months of March, April, May, and June, for the years 1981 through 2000, were used to obtain demographic data, employment status, earnings, details related to the respondent’s job, and location information to control for local labor market conditions. These CPS Annual Earnings files were matched with the March CPS survey for each year to obtain data on disability status, other sources of income, and labor market information available for the previous year. This strategy resulted in a sample four times larger than any single month of current labor market statistics, yielding greater confidence in the precision of the results. Some have questioned whether self-reported disability status (as in the CPS) suffers from endogeneity (e.g., Parsons 1980; Haveman and Wolfe 1984). For example, it may be the case that someone less likely to enter the labor market or to be employed is also more likely to report the presence of a disability (i.e., the disability indicator and error term of the regression are not independent). Stern (1989) finds that ‘‘any bias due to potential endogeneity is small’’ (p. 363). Of course, endo- geneity may be more of a concern since the passage of the ADA. As will be addressed in Chapter 2, endogeneity among the population as a whole may be a greater problem than among only labor force partici- pants (also see Kreider 1999). Additional criticism has been lobbed at the use of the traditional ‘‘work disability’’ measure contained in the CPS for drawing conclusions about the overall experience of the dis- abled or the effectiveness of the ADA. Some argue that requiring a disability to be ‘‘work limiting’’ can be too narrow (Kruse and Schur 2002; McNeil 2000). Others contend that not appropriately defining what a work-limiting disability is results in too broad of an inclusion of respondents (Hale 2001 and Kirchner 1996). Yet, others provide evidence supporting the representative nature of the CPS for monitor- 10 Hotchkiss ing outcomes among the disabled (Burkhauser, Daly, and Houtenville 2001). It is because of this controversy that confirmatory evidence of the CPS results is sought from an additional data source. Regardless, the reader should be aware that this book makes use of ‘‘work-limiting disability’’ as the identifier of a disabled person. In addition, it is expected that when focusing on labor market outcomes, those who report a work-limiting disability are the most likely to feel the greatest impacts of the ADA, should they exist. Table 1.1 reports the potential sample sizes for each year obtained from the CPS. Actual sample sizes for each analysis may differ be- cause of missing data or the use of specific subsamples (e.g., the unem- ployed only).6 While the sample sizes vary somewhat from year to year, the proportion of disabled to nondisabled remains fairly constant, and most analyses benefit from roughly 1,500 disabled workers and 50,000 nondisabled workers. Table 1.1 Sample Sizes for Merged CPS Data Files Disabled Nondisabled Labor force Labor force Year Total All participants Employed All participants Employed 1981 100,291 9,818 2,022 1,744 90,473 60,873 56,656 1982 94,351 9,617 1,962 1,661 84,734 57,006 52,015 1983 93,720 9,119 1,788 1,490 84,601 56,606 51,114 1984 94,683 9,654 1,922 1,661 85,029 57,591 53,507 1985 95,075 9,832 1,931 1,648 85,243 58,111 54,192 1986 90,341 8,931 1,848 1,581 81,410 55,604 51,935 1987 88,507 8,513 1,805 1,560 79,994 54,829 51,591 1988 85,371 7,811 1,697 1,493 77,560 52,258 49,625 1989 85,224 7,913 1,713 1,533 77,311 53,364 50,789 1990 93,625 8,745 1,919 1,692 84,880 58,896 56,005 1991 92,958 8,681 1,833 1,598 84,277 58,172 54,558 1992 90,520 8,547 1,913 1,614 81,973 56,617 52,779 1993 90,056 8,842 1,950 1,684 81,214 55,926 52,316 1994 88,674 9,709 1,810 1,602 78,965 55,341 52,246 1995 77,674 8,654 1,507 1,336 69,020 48,217 45,775 1996 77,188 8,396 1,535 1,379 68,792 48,356 45,892 1997 78,322 8,418 1,609 1,456 69,904 49,437 47,112 1998 77,583 7,796 1,468 1,332 69,787 49,403 47,370 1999 77,487 7,625 1,392 1,266 69,862 49,406 47,542 2000 79,242 7,917 1,488 1,358 71,325 50,825 48,999 Introduction 11 Since the ADA (and similar legislation) was designed to improve the labor market conditions of a group of workers, the analyses pre- sented here will be almost purely cross-sectional. The result is a com- prehensive comparison of the labor market experiences of one group of workers (the disabled) with that of another group of workers (the nondisabled). When making comparisons across groups of people, there will surely always be exceptions to the norm. It is important to recognize, however, that policy is rarely designed around exceptions. The use of individual data in the analyses does allow for control of identifiable individual characteristics (other than the group-defining characteristic of being disabled) in the determination of workers’ expe- riences. The

premise, of course, is that identical disabled and nondis- abled workers should have the same labor market experience. This presupposition, which holds in making any comparisons across groups of workers (i.e., men versus women, or blacks versus whites), is more problematic in making comparisons across disability status; there are likely more unobservable characteristics across disability status than, for example, across gender. In addition, since most of the analyses consider the experience of those in the labor market, or at least control for selection into the labor market, no restriction is imposed on age.7 For each of the analyses, it is important to distinguish any changes in outcomes that might have resulted from the enactment of the ADA from any long-term trend. In other words, changes in the labor market experiences of workers with disabilities may reflect an evolving social awareness that culminated in the passage of the ADA, rather than the other way around. Consequently, this book documents labor market outcomes from 1981 through 2000.8 In addition, since a major over- haul of the CPS questionnaire was undertaken in 1994, care is taken to differentiate any ADA impact from a potential statistical artifact (see Polivka 1996). Also, due to the complicated matching across one to four months of the CPS, all analyses have been performed unweighted. According to Wooldridge (1999), ‘‘stratification based on exogenous variables does not cause any problems: estimators that ignore the stratification are consistent and asymptotically normal, and the usual variance matrix estimators are consistent’’ (p. 1386). Since stratification in the CPS sampling design is based on exogenous variables (geographic and de- mographic), and the attrition that results from the matching procedure 12 Hotchkiss is likely unsystematic, weights should be unnecessary (for further evi- dence on this point, see DuMouchel and Duncan 1983; Manski and McFadden 1981). In addition, any effect of stratification on the estima- tion can be accounted for by including indicator variables that corre- spond to the strata (Ginther and Hayes 2001), so demographic variables (such as disability status) should control for any observable effect sam- pling based on those characteristics might have (either initially or through attrition of matching). Any systematic attrition or sample loss due to unobservables will not be accounted for, but also cannot be corrected using weights. DISABLED AMERICANS As a first look at the data used for the analyses in the following chapters, Figure 1.1 depicts the percentage of the sample in each year and the percentage of workers in each year indicating a work-limiting disability.9 The vertical lines correspond to the phase-in years of the ADA. It is of interest to know whether there is any noticeable change Figure 1.1 Percentage of Sample and of Workers in the CPS Data Set Indicating a Work-Limiting Disability, 1981–2000 12 10 % of 8 sample 6 4 2 % of workers 0 1981 1983 1985 1987 1989 1991 1993 1995 1997 1999 (a) (b) (c) (a) ADA Enactment (b) ADA Phase I (c) ADA Phase II Introduction 13 in the reporting of having a work-limiting disability, particularly on the part of workers. Over the time period from 1981 to 2000, an average of 10 percent of the entire sample indicated having a work-limiting disability.10 There is a significant 1 percentage point difference between the aver- ages prior to and including 1991, and 1994 and later.11 Kreider (1999) finds evidence of substantial overreporting of limitations by nonwork- ers, a behavior which may be enhanced in the presence of protective legislation. It has also been found that the passage of the Personal Responsibility and Work Opportunity Reconciliation Act (PRWORA) in 1996, which essentially put a time limit on welfare payments, re- sulted in a movement of qualified recipients from the welfare rolls to SSI (Lewin Group 1999). The Lewin Group found ‘‘a very substantial flow of program participants from AFDC to SSI during the pre-reform period’’ (p. ES-3). The ‘‘pre-reform’’ period would coincide with the rise in the percentage reporting being disabled between 1993 and 1995 in Figure 1.1. However, the percentage has been declining fairly stead- ily since 1995. This issue of increased reporting of a work-limiting disability among the entire population is taken up in greater detail in Chapter 2, and again points to the potential endogeneity problems in- herent in using a self-reported disability classification. The proportion of workers indicating a work-limiting disability has remained fairly constant at about 3 percent across the entire time span; there is no significant difference in the 1981–1991 and 1994–2000 periods. So, while heightened awareness of the ADA and other pro- gram changes may have increased the reporting of work-limiting dis- abilities among the population, the primary individuals of focus for this study, i.e., workers, do not seem to have changed their reporting behavior in a way that might be expected to bias the analysis. In addi- tion, given that the reporting percentage of the population has begun to decline again, and that the share of workers seems unaffected, it is safe to say that CPS survey design changes that occurred in 1994 do not seem to have impacted the reporting of those with work-limiting dis- abilities. Comparing raw averages of disabled and nondisabled workers across the time period, one can see that there are some significant de- mographic differences among these categories of workers. Table 1.2 reports averages across time for a variety of demographics for disabled 14 Hotchkiss Table 1.2 Means of Select Demographic Variables for Disabled and Nondisabled Workers over Entire Time Period, CPS, 1981–2000 Variable Disabled workers Nondisabled workers Hours of work 33.97 38.01 Female  1 0.47 0.48 Single  1 0.48 0.40 Nonwhite  1 0.13 0.13 College degree  1a 0.10 0.18 Midwest  1 0.26 0.25 South  1 0.29 0.30 West  1 0.24 0.22 Norteast  1 0.21 0.23 Age 43.43 37.49 a Coding of education changed substantially in 1992; these averages reflect the average across years 1992–2000. and nondisabled workers. The distribution of workers across occupa- tions and industries is of interest, as well, but that will be explored in great detail in Chapter 4. While females and nonwhites seem to be equally represented among disabled and nondisabled workers, and each group of workers appears to be equally distributed geographically, there are some notable differences in demographics. Disabled workers, on average, work fewer hours, are less likely to have a college degree, are older, and are more likely to be single. While means across time give us some idea of the relative differences between worker catego- ries, they tell us nothing about trends. One trend of particular interest is the change in average hours per week over time among workers. Figure 1.2 depicts the average hours of disabled and nondisabled workers for each year between 1981 and 2000. While the average weekly hours of nondisabled workers rise fairly steadily over this time period from 37.5 in 1981 to 38.7 in 2000, the hours of disabled workers fall from an average of 34.7 in 1981 to 33.8 in 2000. Since a dramatic part of this decline occurred after 1992, one might suggest that the ADA was a factor. Full-time jobs may be less available to disabled workers; the ADA may have induced employ- ers to be more flexible regarding hours of work in accommodating a worker’s disability; or workers with more serious disabilities, unable to work full-time, may have begun to enter the labor market (Kaye Introduction 15 Figure 1.2 Average Weekly Hours of Disabled and Nondisabled Workers, CPS, 1981–2000 40 39 38 Nondisabled workers 37 36 35 34 33 32 31 Disabled workers 30 1981 1983 1985 1987 1989 1991 1993 1995 1997 1999 (a) ADA Enactment (a) (b) (c) (b) ADA Phase I (c) ADA Phase II 2002). Issues related to differences in part-time employment across disability status will be evaluated in greater detail in Chapter 4. SURVEY OF INCOME AND PROGRAM PARTICIPATION (SIPP) Data from the SIPP were used to construct a sample to supplement the analyses from the CPS. The goal in employing the SIPP is twofold. First, it provides validation of the results obtained using the CPS. Sec- ond, given that the SIPP allows identification of the nature of a respon- dent’s disability, some questions regarding the importance of the type of disability can be addressed. The samples from the SIPP have been constructed to match those from the CPS as closely as possible (e.g., regarding variable definitions, etc.).12 While providing more detail re- lated to the respondent’s disability, the SIPP does not provide as long or as large a data set with which to study labor market experience. Table 1.3 provides sample size details for the SIPP samples constructed 16 Hotchkiss Table 1.3 Sample Sizes for SIPP Data Files Disabled Nondisabled Labor force Labor force Year Total All participants Employed All participants Employed 1986 18,290 2,102 759 650 16,188 12,036 11,191 1987 33,884 3,939 1,470 1,297 29,945 22,278 20,932 1988 34,284 3,995 1,476 1,324 30,289 22,623 21,579 1989 16,274 1,826 651 579 14,448 10,949 10,505 1990 34,010 3,788 1,404 1,233 30,222 22,771 21,629 1991 51,140 5,596 1,998 1,755 45,544 34,392 32,328 1992 76,496 8,231 2,936 2,570 68,265 52,105 48,582 1993 73,831 8,112 2,839 2,442 65,719 49,861 46,694 1994 50,384 5,495 1,881 1,680 44,889 34,338 32,529 1995 23,753 2,610 918 829 21,143 16,313 15,550 1996 57,625 5,865 2,049 1,871 51,760 41,158 39,357 1997 46,914 4,706 1,584 1,470 42,208 33,718 32,558 for each year. Due to the sampling structure of the SIPP, the sample sizes varied from just over 16,000 to over 76,000. However, as Figure 1.3 illustrates, the representation of the disabled within the whole sam- ple and within the working subsample has remained consistent, al- though slightly declining over the period.13 In addition, there does not seem to be any shift in the trends during the ADA phase-in period. The percentages of the sample and of workers indicating a work- limiting disability are slightly higher in the SIPP than in the CPS. This occurs for two reasons. SIPP respondents are given two opportunities to answer a disability question positively. In addition, since the sample came from Wave 2 (the second survey within a panel), the respondent is reminded if he or she indicated a disability in Wave 1 (the first sur- vey), increasing the chances for a positive response (also see Kruse and Schur 2002). The percentages reflected in Figure 1.3 are consistent with those found by other researchers using the SIPP (e.g., DeLeire 2000; Kruse and Schur 2002). The nature of a person’s disability is placed into one of 30 different categories (including ‘‘other’’). In order to be able to include controls for type of disability, these categories were combined to correspond to the groupings used by the Social Security Administration.14 Aggrega- tion was necessary due to category size limitations; the four groups included as controls were: 1) musculoskeletal systems and special Introduction 17 Figure 1.3 Percentage of Sample and of Workers in the SIPP Data Set Indicating a Work-Limiting Disability, 1986–1997 14 12 10 8 % of sample 6 4 2 % of workers 0 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 (a) ADA Enactment (a) (b) (c) (b) ADA Phase I (c) ADA Phase II senses; 2) internal systems; 3) neurological systems and mental disor- ders; and 4) other. Figure 1.4 presents the distribution of all disabled individuals and disabled workers across these categorizations. The largest group by type of disability contains those with musculoskeletal and special senses disabilities; the internal systems category is gener- ally the next largest, followed by neurological and mental disorders (typically), and other. One can observe a slight upward trend in the Figure 1.4 Distribution of Disabled Individuals across Disability Type, SIPP, 1986–1997 50 45 40 Musculoskeletal and special 35 senses (%) 30 Internal systems 25 20 Neurological and mental disorders 15 10 Other 5 0 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 18 Hotchkiss neurological and mental disorders classification, while the proportion for internal systems has declined slightly, and that for musculoskeletal and special senses has remained fairly consistent over the time period. These four classifications will be used to determine whether the labor market experience varies across type of disability, an important consid- eration when trying to mold policy to impact those most affected. Notes 1.

Further details of the history and provision of the Rehabilitation Act of 1973 can be found in Ellner and Bender (1980). 2. See Advisory Commission on Intergovernmental Relations (1989). 3. Data provided by the Roper Center for Public Opinion Research, University of Connecticut, Storrs, Connecticut. By 1999, 67 percent of those surveyed by the same polling group had heard of the ADA. 4. This issue has often been raised by historians. For example, see Landes and Solmon (1972). Donohue and Heckman (1991) also empirically address the sub- ject with regard to civil rights legislation. They conclude that federal civil rights legislation did play a major role in the progress of blacks beginning in 1965. 5. Some have even suggested that our analyses of the labor market impact of the ADA are misguided, and that attempts to quantify an impact in the labor market are merely arrogant efforts to justify our assumptions about how the labor market should operate (see Schwochau and Blanck 2000). 6. Appendix A contains additional information pertaining to the matching and merg- ing of the CPS files across months and concerning other details learned in the process. 7. The exceptions are analyses of employment where age is restricted to 15–65 years. 8. Prior to 1981, identification of a disability in the CPS was made only in the context of why a respondent was not working. 9. See Table C.1, in Appendix C, for percentages used to generate Figure 1.1. 10. The percentages of the entire sample that are disabled are slightly higher than those reported by Burkhauser, Daly, and Houtenville (2000, 2001). This is likely the result of the matching technique employed here, allowing for a much larger sample, and thus greater opportunity to be classified as disabled. 11. The Z statistic corresponding to the hypothesis of equal means over these time periods is 3.18, leading to a rejection of the null hypothesis of equality at the 99 percent confidence level. 12. Details of the construction of the SIPP samples are contained in Appendix B. 13. There are two check variables in the topical module used to identify a work- limiting disability for the 1986–1993 panels. The 1996 panel has only one check variable, which may explain the slightly lower incidence of a work-limiting dis- ability in the 1996 and 1997 SIPP samples. Kruse and Schur (2002) make use of Introduction 19 the functional limitations module (rather than the work disability module used here) and note a similar decline in disability percentages in later years due to question placement in that module. These nuances in the survey design among panels are clearly important and raise, once again, the concerns associated with using a self-reported disability indicator. 14. The Social Security Administration’s listing of impairments for disability status purposes can be found on the Internet at http://www.ssa.gov/OP_Home/cfr20/ 404/404-ap09.htm. The mental disorder category does include those classified as mentally retarded. The mentally retarded group is not broken out into a sepa- rate category in order to be consistent with the classification used by the Social Security Administration, to correspond with the groupings used by others (e.g., DeLeire 2000; Kruse and Schur 2002), and to preserve reasonable sample sizes within the categories. The employment and wage analyses were reestimated with mental retardation as a separate category, and none of the results or conclusions changed. This Page Intentionally Left Blank 2 Employment (Co-authored with Ludmila Rovba) Employment levels of the disabled are affected by both labor sup- ply and labor demand issues. Individuals suffering from a functional disability will also experience a larger cost to entering the labor market as, holding all else constant, greater effort or sacrifices must be made relative to nondisabled workers. The net result is that the reservation wage (the wage at which a person is willing to enter the labor market) for disabled individuals will be higher than for the nondisabled, and fewer disabled people will choose to enter the labor market, ceteris paribus. In addition, a person’s functional disability will be more likely to render him or her less productive than an otherwise identical, nondisabled person. Consequently, the disabled worker will be less likely to qualify for a given job and therefore less likely to be hired. Merely a perception of lower productivity or a greater difficulty of predicting a disabled worker’s productivity will reduce the likelihood of the individual being hired. So, for both supply and demand reasons, the employment levels of disabled workers would be expected to be lower than those of nondisabled workers.1 Figure 2.1 presents evidence from the CPS consistent with this prediction.2 The proportion of dis- abled individuals employed in any given year is at least 44 percentage points lower than the share of nondisabled individuals employed in that year. Other observations are worth mentioning in comparing employ- ment percentages. The recession dips of the early 1980s and early 1990s are obvious for the nondisabled, but not nearly as severe (in percentage terms) for the disabled. In addition, the employment per- centage for the nondisabled has made a fairly steady climb over the entire period compared with the relatively stagnant, then declining, em- ployment percentage of the disabled. Legislation that potentially affects the costs of either labor force participation or of hiring a group of workers can be expected to impact the employment levels of that group. The ADA, through its required 21 22 Hotchkiss and Rovba Figure 2.1 Percentage Employed of Disabled and Nondisabled Individuals, CPS, 1981–2000 80 70 Nondisabled 60 50 40 30 20 10 Disabled 0 1981 1983 1985 1987 1989 1991 1993 1995 1997 1999 (a) (b) (c) (a) ADA Enactment (b) ADA Phase I (c) ADA Phase II NOTE: Individuals in this figure refer to all people, regardless of labor force participa- tion status. accommodations, can be anticipated to reduce the cost to a disabled individual of entering the labor force, thus promoting labor force par- ticipation.3 It might also be argued that greater accommodation of a disabled worker’s limitations will result in enhanced productivity of that disabled worker, thus increasing the likelihood of employment. (This will be the case, however, only if employers are able to accurately predict the cost and productivity gains of such accommodations.) If those required accommodations, however, are ‘‘binding’’ in the sense that the employer would not undertake them in the absence of the ADA, it must be the case that increased productivity of the disabled worker does not offset the cost of implementing those accommodations.4 In other words, the value of the productivity gains is not as great as the cost of accommodation. This may result in decreased employment probabilities of disabled workers, since the cost of hiring a disabled worker has increased. Referring back to Figure 2.1, there does not seem to be any noticeable, or permanent, change in the employment Employment 23 percentages for disabled workers around the time of passage of the ADA, although there is a slight drop around the second phase-in period. This chapter explores more fully the employment probabilities of dis- abled and nondisabled workers between 1981 and 2000, controlling for observable individual characteristics and labor force participation. The issue of joint versus unconditional employment differences is explored through estimating a bivariate probit model, controlling for selection into the labor market. This is followed by an examination of how employment has changed across size of firm. The analyses indi- cate that, at worst, employment probabilities of disabled workers have not deteriorated relative to nondisabled workers, and that employment of disabled workers has shifted from medium and small firms to large firms. UNCONDITIONAL AND JOINT PROBABILITIES The labor market provisions of the ADA were motivated by a de- sire to eliminate barriers to disabled individuals that might exist in the labor market. An appropriate assessment of the success of the ADA in this endeavor would involve evaluation of unconditional employment outcomes. In other words, the question to be answered is whether there has been any progress in employment outcomes for the disabled person drawn from random, controlling for the likelihood that he or she is a labor force participant. The resulting probability of interest is an unconditional probability of employment. An alternative question, which has been the source of recent condemnation of the employment impacts of the ADA, is whether there has been any progress in employ- ment among all disabled people. This second question involves evalu- ation of a joint outcome: what is the probability of entering the labor force and being employed? While the impact of the ADA on labor force participation may be of interest from a social, resource, and de- mographic perspective, the unconditional probability will tell us more about the barriers disabled workers face, which is the focus of the employment provisions of the ADA. Consideration of the joint out- come (or, employment among all disabled people) confounds conclu- sions regarding the employment impact of the ADA with labor supply decisions. 24 Hotchkiss and Rovba When considering the unconditional probability, one must control for unobservable characteristics that might both affect the labor force participation decision and the employment outcome. Without control- ling for this potential self-selection, any differences measured in the employment probabilities may actually be confounded by variations between characteristics that affect the labor supply decision of disabled and nondisabled persons. If these characteristics change in a system- atic way over time, the problem is magnified. A bivariate probit model with selection will be estimated in order to obtain information on un- conditional employment outcomes and to control for selection into the labor market at the same time. The bivariate specification allows for the two outcomes (labor force participation and employment) to be impacted by the same unobservable factors (e.g., motivation). The selectivity part of the model is merely a recognition that we do not get to see the employment outcome unless the person is in the labor market to begin with, and that those we observe in the labor market may have systematically different employment outcomes than those not in the labor market. Correcting for selectivity allows us to make inferences for anyone from the population, not just those found in the labor mar- ket; this is what makes the probability unconditional.5 The following model defines the relationship assumed between labor force participation of person i (lfpi), employment (empi), and individual characteristics that are believed to affect the labor force par- ticipation decision (X1i) and the employment outcome (X2i): (2.1) lfpi 
1  1
X1i  1disablei  1i 
1 if person i is in the labor force 0 otherwise (2.2) empi 
2  2
X2i  2disablei  2i 
1 if person i is employed 0 otherwise disablei is equal to 1 if person i is disabled, 0 otherwise, and 1i and 2i are distributed as a bivariate normal with means equal to 0, variances equal to 1, and correlation equal to . In addition, of course, empi is Employment 25 only observed if lfpi  1.6 X1i and X2i both include age; age squared; female, nonwhite, education, and regional dummies; and state unem- ployment rate. The labor force participation equation regressors (X1i) also include nonlabor income, marital status, and a worked-last-year indicator. The employment equation regressors (X2i) also include num- ber of weeks worked last year. The impact of having a work-limiting disability on employment, then, is determined by calculating the proba- bility of interest for each person (using the estimated parameter coeffi- cients,
1,
2, 1, 2, 1, and 2), varying the disability index between 0 and 1, then averaging the difference across the sample.7 The model is estimated for each year separately, and the marginal impact of having a work-limiting disability is calculated.8 The significance of having a work-limiting disability is determined from the significance of the esti- mated coefficient. Figure 2.2 reflects the marginal effect of having a work-limiting disability on the predicted joint probability of labor force participation and employment in each year.9 The impact of having a work-limiting disability on the joint labor force and employment probability intensifies (becomes more negative), rather dramatically, in 1994, corresponding to the second phase of the ADA. The marginal effect increases from an average of 15 per- centage points prior to 1994 to an average of 19 percentage points Figure 2.2 Impact of Disability on Joint Labor Force Participation and Employment Probabilities,

CPS, 1981–2000 Difference in 0 probability -0.03 -0.06 -0.09 -0.12 -0.15 -0.18 -0.21 1981 1983 1985 1987 1989 1991 1993 1995 1997 1999 (a) ADA Enactment (a) (b) (c) (b) ADA Phase I (c) ADA Phase II 26 Hotchkiss and Rovba between 1994 and 2000. In other words, having a work-limiting dis- ability decreased an individual’s joint probability of being in the labor force and employed by 4 percentage points more after 1994 than it did prior to 1994. This dramatic relative decline in the joint probability for the disabled is the result on which DeLeire (2000) and Acemoglu and Angrist (2001) base their warnings regarding the ADA. Breaking the joint probability into its employment and labor force participation components, it becomes clear that this salient change in 1994 is driven by decreases in labor force participation among the disabled. Figure 2.3 plots the predicted (unconditional) employment and labor force participation probabilities for the disabled alone using the same parameter estimates that generated Figure 2.2. After increas- ing fairly steadily, the predicted labor force participation rate declines in 1994 and stays below 1986 levels. At the same time, and with the exception of the recession years of 1991–1993, the predicted uncondi- tional employment probability among the disabled has increased fairly steadily. Figure 2.3 Separate Predictions of Employment and Labor Force Participation Probabilities for the Disabled, CPS, 1987–2000 Predicted probability 0.90 of employment 0.85 0.80 0.75 0.70 0.65 Predicted probability of labor force participation 0.60 0.55 0.50 1981 1983 1985 1987 1989 1991 1993 1995 1997 1999 (a) ADA Enactment (a) (b) (c) (b) ADA Phase I (c) ADA Phase II Employment 27 Again, in evaluating the barriers facing disabled workers, change in the unconditional employment probability is a more appropriate measure than the change in the joint labor force and employment out- come. Consequently, the condemnation of the employment impact of the ADA by DeLeire (2000) and Acemoglu and Angrist (2001) is misplaced, since both of these analyses confound their evaluation of employment changes with changes in labor supply decisions. The re- sults in Figure 2.3 show that the decline in employment probabilities among all disabled people is labor-supply driven and does not reflect an increase in employment barriers for individuals with disabilities. One may argue that the disabled have decreased their labor supply behavior in response to a real or perceived change in employment prob- abilities (demonstrating a potential ‘‘feedback effect’’), but the pre- dicted unconditional employment probabilities are not consistent with this view.10 It may be the case, however, that the condemnation of the ADA by recent studies should be aimed at its apparent impact on labor force participation. For example, if the ADA resulted in lower wages for the disabled (their employment has now become more expensive through required accommodations), it is possible that the wage would fall below the reservation wage of a significant number of disabled labor force participants, causing them to drop out of the labor market. It may also be the case that the severity of disabilities has been growing over time, resulting in declining labor force participation rates (Kaye 2002). The next section explores this drop in labor force participation rates among the disabled in greater detail. EXPLAINING THE DECLINE IN LABOR FORCE PARTICIPATION RATES Even if the ADA has not had a negative employment impact but has inadvertently discouraged the disabled from seeking employment, there would be a concern worthy of further policy consideration. The disabled and nondisabled populations can be represented in the follow- ing chart: 28 Hotchkiss and Rovba In labor Not in force labor force Nondisabled A B Disabled C D Cells A through D contain a given number of people at any given time period. A decrease in the disabled labor force participation rate (lfprd) corresponds to a decline in the ratio C/(CD). This ratio can decline if C decreases and/or if D increases.11 If C falls, these people must go somewhere; it is most likely that they either go to A (stay in the labor force but change their identification to nondisabled), or go to D (keep their identification as disabled, but leave the labor force). It is this latter possibility that is of potential concern. Although the ADA was not designed as a policy to necessarily increase the labor force partici- pation rate among the disabled, a precipitous drop in such participation, even remotely attributable to the ADA, is considered by many as unde- sirable. The lfprd may also fall, however, as a result of an increase in D. Again, the increase in D must come from somewhere; the most likely candidates are C (disabled leaving the labor force) or B (nondis- abled, nonparticipants in the labor force changing their identity to dis- abled). The latter (movement from B to D) is what might result, for example, from a shift of (nonparticipant) welfare recipients away from welfare programs and into disability programs; this movement follow- ing the reform of welfare has been documented (the Lewin Group 1999; also see Davies, Iams, and Rupp 2000). Greater effort to be classified (and identified) as disabled might also result from increased generosity of the disability programs themselves (see Autor and Dug- gan 2001; Bound and Waidmann 2002).12 This could be consistent with the finding by Kreider (1999) that nonworkers substantially over- report a work limitation. So, the question is, which is it? Are the disabled moving out of the labor force (from C to D) or are more nonparticipants identifying themselves as disabled (from B to D)? One way to get an indication of the movement across these cells is to evaluate the trends in the percentages represented in each cell. These results are depicted in the Employment 29 following chart. The percentage in each cell represents the growth, or decline, experienced within that cell. The cells exhaust the population, so the changes sum to zero.13 In labor Not in force labor force 0.1504** 0.2309** Nondisabled (A) (B) 0.0225** 0.1030* Disabled (C) (D) **  significant at the 95 percent confidence level. *  significant at the 90 percent confidence level. The largest net change in the cells was movement of the nondis- abled out of the nonparticipant category (cell B). The coefficient indi- cates that the nondisabled, nonparticipant percentage declined an average of about 0.23 percentage points per year between 1990 and 2000. Even if the entire increase in cell A (nondisabled labor force participants) came from cell B, that still means that the overwhelming bulk of the increase in cell D (disabled nonparticipants) came from cell B as well, not cell C. In fact, the smallest net cell change was among the disabled labor force participants. This result provides strong evi- dence that the observed decline in the lfprd was not the result of the disabled fleeing the labor force but was most likely due to the reidenti- fication of some nonparticipants from nondisabled to disabled (move- ment from B to D). While providing an explanation for the decline in lfprd, this movement from cell B to cell D is a reminder of the criti- cism of using self-reported disability status in statistical analyses. This also suggests that endogeneity will be less of a concern for analyses that focus exclusively on labor force participants in evaluating the labor market experience of the disabled (using a self-reported measure of disability). In other words, there is less movement across disability status among labor market participants than among nonparticipants. In addition, these results indicate that the observed decline in the lfprd should not be considered as casting a shadow on the measured impacts of the ADA on employment. 30 Hotchkiss and Rovba POOLED, CROSS-SECTIONAL ANALYSIS Along with cross-sectional analyses, an analysis across time is per- formed to help quantify any difference in predicted probabilities of employment between disabled and nondisabled individuals after the ADA relative to before the ADA. The strategy used to accomplish this is to estimate a cross-sectional, time-series bivariate probit model with dummy variables representing whether the observation shows up in the data pre-ADA or post-ADA and whether the observation is a disabled or nondisabled person. These dummy variables are also interacted to determine whether being disabled had any greater impact on employ- ment after the ADA than before the ADA, relative to the experience of a nondisabled person.14 While this type of pooled, cross-sectional anal- ysis has been applied by many researchers (for example, Card 1992; Gruber 1994 and 1996; Zveglich and Rodgers 1996; and Hamermesh and Trejo 2000), the technique also has its critics (such as Heckman 1996). The primary criticism of this approach is that it is impossible to control for unobserved changes in the environment that occurred at the same time as the event of interest. For example, the second phase of the ADA occurred in 1994. This was also when the CPS underwent a major overhaul, and there is no way to disentangle these two events. In addition, the economy began its longest-running expansion in recent history at the same time that the ADA was being phased in, which could potentially confound any measurable impact of the ADA through this estimation strategy. One advantage of the analysis here is that the CPS survey changes should not have a differential impact on the dis- abled and nondisabled (as the changes did not affect measurement or classification by disability),15 and general business cycle activity should essentially impact the disabled and nondisabled in relatively the same proportions.16 Nonetheless, the state unemployment rate is included as a regressor in order to capture any general business cycle influences. The empirical model looks just like the bivariate probit with selec- tion estimated in one year, except with the additional time-period dummy variables: (2.3) lfpi 
1  1
X1i  1disablei  1posti  1disablei  posti  1i Employment 31 (2.4) empi 
2  2
X2i  2disablei  2posti  2disablei  posti  2i. Again, lfpi  1 if person i is in the labor force, 0 otherwise, and empi is not observed unless lfpi  1. disablei is equal to 1 if person i is disabled, 0 otherwise; posti is equal to 1 if person i is observed in 1992 or later; X1i and X2i include individual demographic characteristics; and 1i and 2i are distributed as a bivariate normal with means equal to 0, variances equal to 1, and correlation equal to . In this framework, the affected group (the disabled) is controlled for by a dummy variable indicating whether the individual has a work- limiting disability (disable), and the time period is controlled for by a dummy variable indicating whether the ADA had been implemented yet or not (post); 1 and 2 are the estimated parameter coefficients for the time-period dummies. Given the nonlinearity of the bivariate probit estimation procedure, a single parameter coefficient does not tell us the additional impact the ADA had on the difference in employment probabilities between the disabled and nondisabled. The difference in the impact of having a work-limiting disability on employment across the two time periods can be calculated by evaluating the probabilities of interest for each person, varying the disable and post dummy vari- ables, then taking the difference between these probabilities, and aver- aging this difference across the sample.17 The significance of the coefficient on the interacted disable  post (1, and 2) will, however, yield significance levels of the calculated marginal effects. The decision of when one would expect the ADA to have its stron- gest impact (i.e., how to define post) is debatable. One might expect some impact when the ADA was enacted (1990). However, employers were not required to respond until 1992 (for employers with 25 or more employees) and 1994 (for employers with 15 or more employees). The year 1992 was chosen for defining post since that is the first year of enforcement of the law. Table 2.1 details the regression results. The coefficients on disable  post presented in Table 2.1 are consistent with the conclusions drawn from Figure 2.2. Namely, labor force participation among the disabled declined significantly after im- plementation of the ADA, relative to labor force participation among the nondisabled. In addition, while all other regressors contribute sig- nificantly to explaining employment (all at the 99 percent confidence 32 Hotchkiss and Rovba Table 2.1 Labor

Force Participation and Employment Bivariate Probit with Selection Results, CPS, Combined Years 1981–2000 Labor force participation Employment Regressor equation equation Intercept 2.9988*** 0.6421*** (0.0152) (0.0215) Age (00) 13.3928*** 1.3265*** (0.0742) (0.1176) Age2 (0000) 16.6844*** 2.2159*** (0.0905) (0.1544) Female1 0.4651*** 0.1811*** (0.0032) (0.0047) Nonwhite1 0.0176*** 0.2393*** (0.0045) (0.0058) High school grad1 0.2094*** 0.0762*** (0.0041) (0.0061) Some college1 0.1127*** 0.2076*** (0.0046) (0.0067) College grad1 0.2754*** 0.3730*** (0.0053) (0.0083) Advanced degree1 0.3378*** 0.3479*** (0.0082) (0.0134) Central city1 0.0398*** 0.0321*** (0.0051) (0.0070) Midwest1 0.0530*** 0.0337*** (0.0045) (0.0065) South1 0.0196*** 0.0813*** (0.0042) (0.0064) West1 0.0233 0.0218*** (0.0046) (0.0067) Single household1 0.2148*** — (0.0036) Nonlabor income (000000) 16.6473*** — (0.2922) Worked last year1 2.0763*** — (0.0035) Weeks worked last year (00) — 3.2187*** (0.0151) State unemployment rate (0) 0.0901*** 0.6357*** (0.0076) (0.0103) disable1 0.7624*** 0.2012*** (0.0080) (0.0143) Employment 33 Table 2.1 (continued) Labor force participation Employment Regressor equation equation post (year1992 or later)1 0.0677*** 0.0139*** (0.0035) (0.0051) disablepost1 0.1706*** 0.0298 (0.0120) (0.0225) Rho 0.0371*** (0.0065) Log-likelihood 596,816 Number of observations 1,359,885 NOTE: Standard errors are in parentheses. ***  significant at the 99 percent confidence level. Notation of, for example, (00) indicates regressor has been scaled by dividing by 100. level), being disabled after ADA implementation is not one of them; the disabled are no more or less likely to be employed than the nondis- abled, post-ADA relative to pre-ADA. In other words, the ADA has not changed the relative employment probability between disabled and nondisabled workers.18 Table 2.2 translates the parameter coefficients in Table 2.1 into marginal effects. These marginal effects indicate that the employment probability of disabled labor force participants, relative to nondisabled labor force participants, declines at most 0.6 of a percentage point post- Table 2.2 Change in Marginal Effect of Disability on Labor Force Participation and Employment Probabilities Probability Probability Probability (lfp1) (emp1) (emp1, lfp1) Before After Before After Before After ADA ADA ADA ADA ADA ADA Nondisabled 0.7284 0.7397 0.8592 0.8569 0.6718 0.6798 Disabled 0.5693 0.5431 0.8233 0.8148 0.5202 0.4942 Marginal effect 0.1591 0.1966 0.0359 0.0421 0.1516 0.1856 Change in 0.0375 0.0062 0.0340 marginal effect NOTE: Probabilities calculated using parameter coefficients from Table 2.1. 34 Hotchkiss and Rovba ADA (see column 2, last row). However, this effect is not significantly different from zero; in a model where all other coefficients are signifi- cantly different from zero, this is notable. On the other hand, the labor force participation rate declined significantly, by nearly 4 percentage points more for the disabled than for the nondisabled, post-ADA.19 EVIDENCE FROM THE SIPP The analysis detailed in Equations 2.3 and 2.4 is reestimated using the sample obtained from the SIPP for the years 1986 through 1997. Table 2.3 reports the coefficients of interest from estimating the bivari- ate probit model with selection using the SIPP data. The results re- ported in Table 2.3 mirror those in Table 2.1, with one difference: employment among the disabled increased more post- versus pre-ADA than did the employment of the nondisabled. This positive 0.0768 coefficient on disable  post translates into a 0.8 of a percentage point higher employment probability for the disabled relative to the Table 2.3 Labor Force Participation and Employment Bivariate Probit with Selection Results, SIPP Combined Years 1986–1997 Labor force participation Employment Regressor equation equation disable1 0.9404*** 0.2435*** (0.0105) (0.0211) post (year1992 or later)1 0.0293*** 0.0348*** (0.0048) (0.0077) disablepost1 0.1360*** 0.0768*** (0.0129) (0.0250) Rho 0.4811*** (0.0204) Log-likelihood 292,341 Number of observations 500,560 NOTE: Additional regressors included age; age squared; state unemployment rate; female, nonwhite, education, regional dummy variables; an indicator for SMSA resi- dence (employment); and non-labor income and marital status (labor force participa- tion). Standard errors are in parentheses. ***  significant at the 99 percent confidence level. Employment 35 nondisabled. In addition, the relative decline in labor force participa- tion among the disabled found in the CPS data is also seen using the SIPP data as well. Along with the reestimation of Equations 2.3 and 2.4, a specifica- tion is estimated in which the impact of having a disability post-ADA is allowed to vary by type of impairment:20 (2.5) lfpi 
1  1
X1i  S 1 musculoskeletali  I 1 internali  M 1 mentali O 1 otheri  1posti  S 1 musculoskeletali  posti  I 1 internali  posti  M 1 mentali  posti  O 1 otheri  posti  1i (2.6) empi 
2  2
X2i  S 2 musculoskeletali  I 2 internali  M 2 mentali O 2 otheri  2posti  S 2 musculoskeletali  posti  I 2 internali  posti  M 2 mentali  posti  O 2 otheri  posti  2i where lfpi is equal to 1 if person i is in the labor force, 0 otherwise, empi is equal to 1 if person i is employed, 0 otherwise, Xi is a set of covariates for each person (individual demographic characteristics), musculoskeletali is equal to 1 if person i has a musculoskel- etal disability,21 internali is equal to 1 if person i has a disability involving the internal systems, mentali is equal to 1 if person i has a mental disability, otheri is equal to 1 if person i has a disability classified as ‘‘other,’’ and posti is equal to 1 if person i is observed in 1992 or later. Again, these equations are estimated via maximum likelihood as a bi- variate probit with selection, where empi is only observed if lfpi  1. In this framework, the type of disability is controlled for by dummy variables indicating whether the individual has a musculoskeletal, in- ternal systems, mental, or other disability; and the time period is con- trolled for by a dummy variable indicating whether the ADA had been implemented yet or not. The coefficients of particular interest (j 1 and 36 Hotchkiss and Rovba j 2, j  S,I,M,O), therefore, allow us to calculate the labor force partici- pation and employment changes among disabled workers post- versus pre-ADA relative to the changes for nondisabled workers. Table 2.4 provides selected estimated coefficients and regression details. The estimation results presented in Table 2.4 from the SIPP data set are also consistent with the conclusions drawn using the CPS data: labor force participation declined more for all classifications of disabil- ity, relative to nondisability, post- versus pre-ADA. However, employ- Table 2.4 Labor Force Participation and Employment Bivariate Probit with Selection Results by Type of Disability, SIPP Combined Years, 1986–1997 Labor force participation Employment Regressor equation equation musculoskeletal1 0.8253*** 0.2798*** (0.0150) (0.0281) internal1 0.9597*** 0.1660*** (0.0192) (0.0409) mental  1 1.2722*** 0.1599*** (0.0237) (0.0487) other1 0.8396*** 0.3429*** (0.0280) (0.0508) post (year1992 or later)1 0.0305*** 0.0348*** (0.0048) (0.0077) musculoskeletalpost1 0.1416*** 0.0542 (0.0187) (0.0348) internalpost1 0.1137*** 0.1693 (0.0247) (0.0532) mentalpost1 0.0850*** 0.1187** (0.0282) (0.0573) otherpost1 0.0829** 0.1484** (0.0360) (0.0675) Rho 0.4799*** (0.0205) Log-likelihood 292,164 Number of observations 500,560 NOTE: See notes to Table 2.3 regarding additional regressors. Standard errors are in parentheses. ***  significant at the 99 percent confidence level. **  significant at the 95 percent confidence level. Employment 37 ment probabilities (controlling for labor force participation) increased significantly more for the disability classifications of mental and other than for the nondisabled, post- versus pre-ADA. Relative em- ployment probabilities did not change significantly for those with mus- culoskeletal or internal disabilities. While it is difficult to interpret the employment impact for those with disabilities classified as other, the major role that those with mental disorders play in explaining the overall relative employment improvement is not surprising, given the attention paid to and policies developed for those with mental disabilities in recent years.22 In addi- tion, if we expect costs of accommodation to influence employment outcomes of the disabled, these results might suggest that accommo- dating workers with mental disabilities (such as through flexible work scheduling) has been relatively less expensive for employers than ac- commodating workers with musculoskeletal or internal disabilities (for example, through physical modification of the work environment). EMPLOYMENT PROBABILITY AND FIRM SIZE The phased-in nature of the ADA yields an additional dimension across which to examine its impact on employment.23 After enactment in 1990, the ADA covered employers with 25 or more employees starting in 1992, and employers with 15 or more employees starting in 1994. One might expect a differential employment impact of the ADA based on whether a particular firm is covered by the legislation. In addition, because of the potential costs of accommodating workers’ disabilities, there is reason to expect that disabled workers might migrate toward covered employers (based on size), toward employers who are more able to absorb the cost of accommodation (larger firms may have more re- sources to devote to such investments), and toward employers who can spread the fixed costs of accommodation across more workers (again, this would be true of larger firms). While most estimates indicate that per-worker costs of accommodations only range between $100 and $1,000, this expenditure is clearly easier to absorb for larger, more afflu- ent firms (LaPlante 1992; Kujala 1996).24 The federal government recog- nizes this burden to small business by making a targeted tax credit available for up to half of an accommodation expenditure that exceeds $250 but is less than $10,250 (Dykxhoorn and Sinning 1993; Hays 1999). 38 Hotchkiss and Rovba Figure 2.4 plots the distributions of disabled and nondisabled workers across firm sizes. Here, a small firm is one that employs fewer than 25 workers, a medium firm employs at least 25 but fewer than 100, and a large firm employs at least 100 workers.25 The CPS began asking about the size (number of employees) of a worker’s firm in 1988. This question refers to a person’s main job during the previous year and is therefore available for the years 1987–1999. Large firms employ by far the greatest percentage of both disabled and nondisabled workers. While the average (over time) percentage in medium-sized firms is practically identical across disability status (14 percent), a greater proportion of nondisabled workers (62 percent versus 58 per- cent) is employed in large firms, and a greater proportion of disabled workers (28 percent versus 24 percent) is employed in small firms. As far as trends are concerned, nothing obvious is apparent from Figure 2.4. Trend regression indicates that there have been statistically sig- nificant declines in medium-firm employment among both disabled and nondisabled workers. While most of this decline among the nondis- abled shifted toward small firms, the shift among the disabled was toward large firms. The analysis that follows will allow quantification of these movements and a direct comparison across disability status. A multinomial logit analysis was undertaken to determine how the relative employment of disabled and nondisabled workers in different- sized firms has shifted over the entire time period for which firm size is available.26 This approach allows us to specify multiple possible outcomes (e.g., employment in a small, medium, or large firm) as a function of a variety of observed characteristics and unobservable fac- tors, recognizing that as one’s probability of being in one firm size increases, the probability of being in another firm size necessarily de- creases. A person’s employment outcome is divided into three catego- ries (where n refers to the number of employees at the worker’s firm): 1) employed by a small (n  25) firm, 2) employed by a medium (25  n  100) firm, and 3) employed by a large (n 100) firm.27 It is assumed that the individual selects the firm size (ceteris pari- bus) that maximizes the utility gained from that choice. The employer plays a role in that decision by making different job packages available, such as wages and other characteristics. The probability of person i being employed in firm size 1 is defined as (where u refers to utility): Employment 39 Figure 2.4 Distribution of Disabled and Nondisabled Workers across Firm Size, CPS, 1987–1999 (A) Disabled workers 70 Large 60 (Firm size ² 100) 50 40 Small (Firm size < 25) 30 20 Medium (25 ¢ firm size < 100) 10 0 1987 1989 1991 1993 1995 1997 1999 (B) Nondisabled workers 70 Large (Firm size ² 100) 60 50 40 (%) 30 Small (Firm size < 25) 20 10 Medium 0 (25 ¢ firm size < 100) 1987 1989 1991 1993 1995 1997 1999 40 Hotchkiss and Rovba (2.7) P1  P(ni 

1)  P(ui1  uij) for j  2,3. Let Pj (2.8)  F(
X) for j  1,2, P j j  P3 where F () is the cumulative distribution function, X are individual characteristics, and  are parameter coefficients. This means that Pj F(
jX) (2.9)   G(
X) for j  1,2. P j 3 1  F(
jX) Because of the rules of summation, (2.10) P3  1/1  2 G(
jX) G(
jX) and Pj  j1 1  2 . G(
jX) j1 If we let (2.11) G(
jX)  exp(
jX) and Yij if person i falls in firm size category j 
1 0 otherwise the log likelihood function (ln L) can be written as (2.12) ln L  3 3 YijlnPij, i1 j1 exp(X
ij) 1 where Pij  2 and Pi3  2 1  exp(X
ik) 1   . exp(X
ik) k1 k1 The multinomial logit results in three sets of parameter estimates, each set describing the probability of one of the firm size outcomes. Employment 41 Every person has a probability of being employed by each size firm, and those three probabilities sum to one (since the analysis is restricted to employed individuals). Figure 2.5 summarizes the predicted proba- bilities of disabled workers, relative to the predicted probabilities of nondisabled workers, being employed by each size firm for the years from 1987 to 1999.28 The probability of employment of disabled workers relative to non- disabled workers in both small- and medium-size firms declined over this time period, whereas the relative probability of employment of disabled workers in large firms increased.29 This means that relative to nondisabled workers, disabled workers were increasingly likely to be employed in large firms between 1987 and 1999. This result is consis- tent with Kaye (2002), who finds growing employment rates among the disabled in ‘‘big-business’’ industries (500 or more employees). The increased probability of employment among larger firms may sug- gest that they have been able to accommodate (i.e., afford, spread costs over greater numbers of workers, find appropriate job matches) work- ers’ disabilities more than small- or even medium-sized firms, and that disabled workers have found it fruitful to seek out jobs at the largest firms. In fact, large companies have typically been at the forefront of implementing costly accommodations, either because of public rela- tions initiatives or because of other considerations not faced by smaller Figure 2.5 Ratio of Predicted Employment Probabilities for Disabled versus Nondisabled Workers by Firm Size, CPS, 1987–1999 1.25 Small firms 1.20 1.15 1.10 Medium firms 1.05 1.00 0.95 0.90 Large firms 0.85 0.80 1987 1989 1991 1993 1995 1997 1999 42 Hotchkiss and Rovba businesses (Johnson 1997). These visible efforts make larger compa- nies more attractive for disabled job seekers. In addition, one study has found that large firms are significantly more likely to comply with the ADA and to have specific policies in place guiding the hiring of workers with disabilities (Scheid 1998). It is important to point out, however, that since the ADA has no affirmative action component, the relative shift in employment of disabled workers toward larger firms is not likely the result of active recruitment efforts. Since employers with 25 or more workers were covered by the ADA beginning in 1992, one additional computation can help to quan- tify any adjustment that may have occurred at that time in the relative employment probabilities. Table 2.5 presents a form of differences-in- differences-in-differences (DDD) calculation for the average predicted probabilities of employment across firm size, time, and disability status. These DDD results are not derived from an estimation proce- dure, but are merely the differences in predicted probabilities across coverage, firm size, and time. The predicted probabilities from the Table 2.5 DDD Calculation for Average Predicted Probability of Employment by Firm Size, Disability Status, and across Time, CPS, 1987–1999 Time difference for Firm size / year 1987–91 1992–98 a given firm size A. Disabled workers n 25 0.7434 0.7280 0.0154 n25 0.2566 0.2720 0.0154 Firm size difference at a point in time 0.4868 0.4560 Differences-in-differences 0.0308 B. Nondisabled workers n 25 0.7798 0.7571 0.0227 n25 0.2202 0.2429 0.0227 Firm size difference at a point in time 0.5596 0.5142 Differences-in-differences 0.0454 DDD: 0.0146 NOTE: n refers to the number of employees in the firm. Predicted probabilities for each firm size from the multinomial logit results presented in Figure 2.3 are averaged across the years indicated and disability status to obtain the average predicted proba- bilities. Employment 43 multinomial logit estimation are used for differencing, and the proba- bilities for the medium and large firm sizes are combined to correspond to the coverage of the ADA beginning in 1992. This analysis is not as precise as we might like, since employers with 15 or more workers were covered by the ADA beginning in 1994, and since there are no standard errors available to determine significance of the results. Con- sequently, the results in Table 2.5 should be viewed only as suggestive. The DDD analysis suggests that covered disabled workers have, at most, a 1.5 percentage point greater probability of being employed, relative to noncovered disabled workers, post-ADA, relative to pre- ADA, relative to the employment probability differences among non- disabled workers. CONCLUSIONS The purpose of this chapter was to evaluate the relative employ- ment experiences of disabled and nondisabled workers. A pooled, cross-sectional analysis determined that the joint labor force and em- ployment probability of the disabled decreased significantly after im- plementation of the ADA relative to a nondisabled person’s employment probability. This joint probability is influenced by both supply and demand factors and therefore confounds the employment experience of disabled workers with labor supply influences. In order to get a picture of the expected employment outcome, the unconditional em- ployment probability was calculated. It was found that the uncondi- tional employment probability among the disabled did not change after implementation of the ADA relative to the employment probability among the nondisabled. In other words, although improvements in relative employment outcomes have not been realized, there has not been the deterioration of the employment position of the disabled as claimed by others. This suggests that adjustments in the labor supply of disabled workers are not likely the result of feedback effects or fear of negative outcomes, since the employment outcomes for disabled workers relative to those of nondisabled workers did not deteriorate post-ADA. It was also shown that the decline in labor force participa- tion was likely the result of the reclassification of nondisabled, nonpar- 44 Hotchkiss and Rovba ticipants as disabled, post-welfare reform and potentially in response to the growing generosity of disability benefit policies. Analysis of the SIPP data produced a confirmation of the CPS results and allowed a closer evaluation of employment probabilities by type of disability. In fact, the SIPP results suggest that the relative employment position actually improved, with greater unconditional employment probabilities among the disabled post-ADA, compared to the nondisabled. It was found that those with mental disorders and those with disabilities classified as other experienced the greatest pos- itive employment impact of the ADA. Workers with musculoskeletal and internal system disabilities did not experience any different em- ployment probability growth from those without disabilities. Evidence that the cost of accommodation is not irrelevant in the labor market’s adjustment to the ADA was found in a DDD analysis, which accounted for the size of a worker’s employer, allowing for identification of disabled workers who were covered by the legislation and those who were not. Disabled workers employed by large or me- dium firms (covered employers) have, at most, a 1.5 percentage point greater probability of being employed, relative to disabled workers in small firms, post-ADA, relative to pre-ADA, relative to nondisabled workers. In addition, employment of disabled workers was shown to shift more towards large firms post-ADA than did employment of non- disabled workers. Since the fixed cost of disability accommodations can be spread over a greater number of workers in large firms, this result suggests that larger firms were better poised and able to absorb the costs of accommodations dictated by the ADA and/or better able to match disabled workers’ job skills with recruitment needs. Notes 1. Also see DeLeire (1997, Section 3). 2. See Table C.2 in Appendix C for the percentages used to generate Figure 2.1. 3. See Stern (1996). 4. It has also been suggested that persons with disabilities entering the labor force after the ADA will have more severe disabilities than those employed prior to the ADA, making the potential for ‘‘binding’’ accommodation requirements that much more likely and expensive (Chirikos 1991). 5. This model specification is similar to that familiar to most labor economists: con- trolling for selection into the labor market (or employment) when estimating a Employment 45 wage equation. In that problem, we are interested in the (unconditional) expected wage for anyone in the population. By controlling for selection into the labor market (since we can only estimate the wage equation on those for whom we observe wages), we are able to make unconditional predictions that correspond to the population. If selection is not controlled for, the only prediction of wages that can be made is that conditional on labor force participation. 6. The bivariate probit model with selection gives rise to the following likelihood function: ln L  LFP1, EMP1 ln 2
1X1i,
2X2i,   LFP1,EMP0 ln 2
1X1i,  
2X2i,   LFP0 ln 
1X1i, where 2 is the bivariate normal cumulative distribution function and is the univariate normal cumulative distribution function. 7. This method of calculating the marginal effect of a change in a dummy variable is referred to as a measure of discrete change and is described in greater detail by Long (1997, pp. 135–138). Specifically, the average marginal impact of having a disability on the joint labor force and employment outcome is calculated as 1 N N Pilfp1, emp1 Xi, disable1Pilfp1, emp1 Xi, disable0 , i1 and the average marginal impact of having a disability on the unconditional prob- ability of employment is calculated as 1 P N N iemp  1 Xi, disable  1  Piemp 1 Xi, disable  0 . Both of i1 these are calculated, of course, using the parameter estimates obtained from the bivariate probit model with selection detailed in endnote 6. 8. This model specification allows a comparison to results with earlier studies, as well (through calculation of the joint probability). 9. See Table C.3 in Appendix C for the numbers used to generate Figure 2.2 (num- bers in column 3 minus numbers in column 1). 10. In addition, Stern (1996) presents empirical evidence that labor supply decisions of disabled people are driven more by labor supply factors than by labor demand factors. Also see Averett, et al. (1999) for further evidence on this point. 11. One can easily show that, for C  0 for lfprd  C/(C  D), lfprd/D  0 and lfprd/C  0. 12. Acemoglu and Angrist (2001) dismiss this theory by showing that controlling for receipt of disability benefits only marginally impacts their results. They fail to point out, however, that the receipt of benefits will reflect only a fraction of the desire to receive benefits (see Kubik 1999). Consequently, the actual impact of increasing program generosity on the disability status change for nonparticipants could be much larger than that measured by growing recipiency. 13. These trend coefficient estimates for each cell were obtained from simple linear regressions of the percentage of people represented in that cell as a function of a time trend corresponding to the period 1990–2000, in order to focus on post- ADA changes. 46 Hotchkiss and Rovba 14. The strategy described here can be likened to the popular differences-in-differ- ences (DD) methodology, but it is applied to a nonlinear statistical model. 15. Acemoglu and Angrist (2001, Appendix A) show that results are fairly consistent across a variety of sample restrictions based on differences between the 1993 and 1994 samples (crossing the survey modification time period). Consequently, it is not expected that the results reported here are significantly biased by changes in the CPS survey design. 16. The cyclicality of disabled and nondisabled employment is explored by Burk- hauser, Daly, and Houtenville (2000), although, like Acemoglu

and Angrist (2001) and DeLeire (2000), their analysis confounds employment outcomes with labor supply effects. 17. See endnote 7. 18. The remaining parameter estimates are consistent with labor/leisure choice the- ory. For example, higher nonlabor income and being female lead to lower labor force participation, and the age/participation profile is concave. They also con- form to standard human capital theory with more education and greater labor market experience (measured through number of weeks worked last year) leading to a greater probability of employment. 19. The marginal effect on the joint probability outcome was 3 percentage points (column 3, last row). While not directly comparable, DeLeire (2000) estimates a 7.2 percentage point drop in employment among all disabled men, and Acemoglu and Angrist (2001) estimate a 10–15 percentage point drop in the number of weeks worked by the disabled. Again, these results are analogous to the joint probability calculated here, although not surprisingly of slightly varying magni- tude given the differences in data used (DeLeire) and in estimation procedure and model specification (DeLeire and Acemoglu and Angrist). 20. See Appendix B for the source of classification of disability. 21. The musculoskeletal grouping includes disabilities involving the special senses (e.g., hearing, sight). 22. The President’s Committee on Employment of People with Disabilities had placed an emphasis in the late 1990s on individuals with mental impairments. This committee has more recently been replaced by the Presidential Task Force on Employment of Adults with Disabilities. Information about the activities of this task force can be found on the U.S. Department of Labor web site, http:// www.dol.gov/ sec/programs/ptfead/. 23. Chay (1996) and Carrington, McCue, and Pierce (2000) represent other research exploiting the natural phase-in periods across firm size or geographic differences in dates of implementation in order to measure the labor market impact of social policy legislation. 24. In addition to the direct costs of accommodation, efficiency costs not directly absorbed by the employer, but felt by the labor market as a whole, are identified by Rosen (1991). 25. This definition of small, medium, and large firms follows that of Acemoglu and Angrist (2001); the definition will change in the next chapter. Employment 47 26. Multinomial logits have come under frequent criticism because of the assumption of independence of irrelevant alternatives (IIA) that is implied by the logit speci- fication. Alternative specifications that retain the desired probability structure (i.e., multinomial probit) are riddled with their own problems and not considered here to add value greater than the cost imposed. It has been pointed out that under the framework of what is called a ‘‘universal’’ logit, the estimation procedure can be applied, but the utility interpretation of the structural estimates is lost. In addition, the more regressors included to describe the multiple outcomes, the less bothersome is the assumption of IIA. For these reasons, the logit structure is retained. For further discussion on these points, see Ben-Akiva and Lerman (1985, section 5.2) and Moffitt (1999, pp. 1382–1387). 27. The results of this analysis are relevant for workers only and not generalizable to the entire population. 28. The predicted probabilities are found in Table C.4 in Appendix C. 29. A firm-size analysis was also undertaken by Acemoglu and Angrist (2001). Since they did not restrict their analysis to workers, they found that relative employment declined across all firm sizes (compared with not working), and that there was no change in relative employment of disabled workers in large firms, as compared with the nondisabled. This Page Intentionally Left Blank 3 Compensation: Wages and Benefits The issue of compensation has generated numerous contributions to the demand side of the disability literature. For example, Haveman and Wolfe (1990) evaluate the economic well-being (in which a major factor is earnings, or compensation) of the disabled over an extended period of time (1962–1984). A large part of their measure of well- being, however, is accounted for by transfer income (a nonlabor market source of income).1 In addition, while Salkever and Domino (1997), Johnson and Lambrinos (1985), and Baldwin and Johnson (2000) have examined the issue of wage discrimination against the disabled, evi- dence on how these measures of discrimination have changed over time is sparse (see DeLeire 2001). Depending on the nature of the impairment, one would expect a disabled worker to be less productive than an otherwise identical non- disabled worker; thus, lower wages would be seen for disabled workers. The implementation of a policy that is expected to raise productivity, however, would increase those individuals’ wages. The ADA, through its accommodation requirements, should unambiguously increase the productivity of disabled workers. The impact of this process on work- ers’ earnings, however, is uncertain. If productivity is increased by more than the cost of accommodations, wages of disabled workers should rise. If, on the other hand, the cost of accommodation exceeds the gains in productivity, disabled workers are likely to bear some of the increased costs through lower wages. In addition, since accommo- dation should not impact the productivity of workers not in need of those accommodations (i.e., nondisabled workers), we should not ob- serve a substantial wage change for nondisabled workers post-ADA. The CPS contains data on wages paid and hours employed for all workers. Information on the availability of health insurance and pen- sion plans through one’s employer is also available. This chapter com- pares how relative earnings for workers with disabilities have changed over time and if there was any significant alteration coinciding with the implementation of the ADA. These comparisons are also made across types of disability with the help of the SIPP data set. Earnings of 49 50 Hotchkiss disabled workers are compared with those of workers without disabili- ties to determine whether there has been any improvement in the com- pensation disparity over time and how much of that disparity is left unexplained by differences in productivity (i.e., potential discrimina- tion), particularly around the period that the ADA became law and was fully implemented. The methodology employed will also allow an examination of how much of the earnings disparity is accounted for by different representations of disabled workers across occupations and/ or industries. Nonwage compensation is growing in importance for all employees and may be of particular importance to disabled workers. As such, the probability of being covered by health insurance and a pension plan is evaluated as a function of disability status, also across time. These probabilities are compared, again, to see if any change occurred when the ADA was implemented. WAGE LEVELS Figure 3.1 depicts the average real (1982–1984  100) hourly wages for disabled and non-disabled workers for each year from 1981 Figure 3.1 Average Real Hourly Wages, CPS, 1981–2000 10 9 Nondisabled workers 8 ($) 7 6 Disabled workers 5 1981 1983 1985 1987 1989 1991 1993 1995 1997 1999 (a) (b) (c) (a) ADA Enactment (b) ADA Phase I (c) ADA Phase II Compensation: Wages and Benefits 51 to 2000. As theorized, wages for disabled workers lie below those of nondisabled workers for every year, although these raw figures do not control for differences in human capital or other demographic and job characteristics. Real wages of nondisabled workers exhibit a clear up- ward trend (a significant raw trend average of about 0.03 of a percent- age point per year), and real wages of disabled workers exhibit a downward trend (a significant average of about –0.02 of a percentage point per year, in spite of the recent upward swing). The net result is a growing differential between wages of nondisabled and disabled workers. These relative trends will be examined to determine whether there was a significant difference pre- and post-ADA after controlling for individual and job characteristics. In addition, the wage differential will be decomposed to determine what factors are the greatest contribu- tors to its level and its growth. For example, it could be the case that wages are falling among disabled workers because the nature of disabilities is becoming more severe in the population (Kaye 2002), making disabled workers as a whole less productive. Alternatively, the human capital of disabled workers may be deteriorating for some rea- son, disabled workers are shifting to lower-paying jobs, or employers may be passing along the costs of accommodations through lower wages for disabled workers. Splitting the time series, the negative trend in wages for disabled workers observed in Figure 3.1 is clearly driven by the early years (a significant average decline of 1 percentage point per year from 1981 to 1991), while the years between 1992 and 2000 exhibit no trend at all, statistically. One could interpret this as a positive outcome of the ADA if the policy actually halted, or mitigated, a long-running downward trend in wages of disabled workers. However, this observation tells us nothing about the relative wage trend or components of the differential. DIFFERENCES IN WAGES OVER TIME Pooled, Cross-Sectional Analysis The first analysis of wages in this chapter is a simple, pooled, cross-sectional one.2 A linear relationship is specified in which the log of real wages is a function of demographic and job characteristics, as 52 Hotchkiss well as indicators for disability status, time period, and the interaction between disability status and time. The following specification is esti- mated via OLS for the time period 1984–2000:3 (3.1) lnrwagei 
 
Xi  1disablei  2posti  3disablei  posti  i where lnrwagei is the natural log hourly real (1982–1984  100) wage of worker i, Xi is a set of covariates for each person (demographic and job characteristics), disablei is equal to 1 if person i has a work-limiting disability, and posti is equal to one if person i is observed in 1992 or later. The affected group (the disabled) is controlled for by a dummy variable indicating whether the individual has a work-limiting disability, and the time period is controlled for by a dummy variable indicating whether the ADA had been implemented yet or not. The coefficient of interest (3) measures the change in real wages of disabled workers, relative to nondisabled workers, after implementation of the ADA, rela- tive to before implementation. In other words, 3 tells us how wages changed for disabled workers versus nondisabled workers. Xi includes individual demographic and job characteristics detailed in Table 3.1, which contains the estimation results;  are additional parameter coef- ficients to be estimated, and i is the random error term. Since wages are observed for workers only, and since the charac- teristics of workers may be changing over time in unobservable ways, it is important to control for any potential unobserved self-selection into the labor market. Consequently, Equation 3.1 is modified by sim- ply adding the standard inverse-Mills ratio obtained from a first-stage probit estimation of a labor force participation/employment equation. This standard Heckman (1979) two-step procedure for controlling for self-selection is presented in greater detail in the section on wage de- compositions. Briefly, including the selection term in the regression allows 3 to be interpreted as relevant for the entire disabled and non- disabled population, even though the sample for the regression in- cluded workers only. The parameters of the model are identified Compensation: Wages and Benefits 53 Table 3.1 OLS Selectivity-Corrected Regression Results for Log Real Wages, across Disability Status and ADA Implementation, CPS, 1984–2000 Regressor Parameter estimates Intercept 1.1277*** (0.0121) Age 0.0498*** (0.0059) Age squared 0.0005 (0.0023) Female  1 0.2249*** (0.0016) Nonwhite  1 0.0434*** (0.0018) Single household  1 0.0524*** (0.0015) High school grad  1 0.0834*** (0.0020) Some college  1 0.1614*** (0.0022) College grad  1 0.2806*** (0.0025) Advanced degree  1 0.3072*** (0.0013) Hours 0.0024 (0.0055) Union  1 0.1976*** (0.0005) disable  1 0.1719*** (0.0018) post (year1992 or later)1 0.0033*** (0.0011) disablepost  1 0.0286*** (0.0064) ̂ (selection term) 0.0327*** (0.0043) Number of observations 766,060 F statistic 18,486*** Adjusted R2 0.4279 NOTE: Other regressors included in the estimation, but not reported here, include seven industry and five occupation dummy variables, region, and government em- ployer dummy variables. ***  significant at the 99 percent confidence level. Asymptotically consistent stan- dard errors are in parentheses. 54 Hotchkiss through inclusion of some regressors in the first-stage probit estimation that are not in the wage regression; these regressors include nonlabor income and an indicator of whether the person worked last year or not. Since the purpose of

this two-stage estimation approach is merely to obtain unbiased estimates of the coefficients in the wage equation, in- terpretation of those coefficients is not changed by controlling for selection. All of the parameter estimates in Table 3.1 are of the magnitude and direction one might expect from standard human capital and other labor market theories. For example, women and nonwhites earn lower wages, union workers earn higher wages, and increased wages accrue to those with greater levels of education. The positive coefficient on the selection term ( ̂) indicates positive selection into the labor market; the more likely someone is to enter the labor market, the more likely he or she will earn a wage above the population average. The coefficient on the interaction term disable  post is 0.0286 (and significantly different from zero), indicating that dis- abled workers experienced about a 3 percent decline in wages, relative to nondisabled workers, post-ADA implementation, relative to pre-im- plementation. In other words, wages of the disabled fell by 3 percent more post-ADA than the wages of the nondisabled. This finding is not consistent with the result of DeLeire (2000), who found no significant change in the wages of disabled workers relative to those of nondis- abled workers, post-ADA.4 However, the result does suggest that the cost of accommodating disabled workers, overall, potentially exceeded their gains in productivity.5 It is important to note that this result was obtained by controlling for job characteristic differences, such as hours of work, occupation, and industry. The potential implication of shifts in occupation and industry distributions on these results is explored in Chapter 4. Evidence from the SIPP The pooled, cross-sectional analysis specified in Equation 3.1 was reestimated using the SIPP data set for the years 1986–1997 (and, again, controlling for selection of workers into the labor market through a two-step estimation strategy). Selected coefficient estimates from the reestimation are presented in Table 3.2. While not statistically Compensation: Wages and Benefits 55 Table 3.2 OLS Selectivity-Corrected DD Regression Results for Log Real Wages, across Disability Status and Type of Disability Status, SIPP 1986–1997 Parameter estimates Disability indicator Type of disability Regressor only indicated disable  1 0.1535*** — (0.0027) post (year  1992 or later)  1 0.0515*** 0.0305*** (0.0017) (0.0017) disable  post  1 0.0070 — (0.0072) musculoskeletal  1 — 0.1230*** (0.0084) internal  1 — 0.1537*** (0.0115) mental  1 — 0.3427*** (0.0157) other  1 — 0.1124*** (0.0153) musculoskeletal  post  1 — 0.0424 *** (0.0098) internal  post  1 — 0.0125 (0.0148) mental  post  1 — 0.0220 (0.0183) other  post  1 — 0.0130 (0.0193) ̂ (selection term) 0.0046 0.0116* (0.0067) (0.0065) Number of observations 353,651 287,343 F statistic 8,855 6,502 Adjusted R2 0.4289 0.4489 NOTE: Other regressors included in the estimation, but not reported here, include seven industry and five occupation dummy variables, hours of work, age, age squared, and race, education, union, gender, marital status, region, and government employer dummy variables. Asymptotically consistent standard errors are in parentheses. ***  significant at the 99 percent confidence level. *  significant at the 90 percent confidence level. 56 Hotchkiss significant, the negative coefficient on the disable  post regressor is at least consistent (in sign) with the results obtained from the CPS data. Table 3.2 presents an additional specification, which identifies type of disability. In the following equation, the impact of a worker’s dis- ability on the real wage is allowed to vary by type of impairment:6 (3.2) lnrwagei 
 
Xi  S 1 musculoskeletali  I 1 internali  M 1 mentali  O 1 otheri  2posti  S 3 musculoskeletali  posti  I 3 internali  posti  M 3 mentali  posti  O 3 other  posti  i where lnrwagei is the natural log hourly real (1982–1984  100) wage of worker i, Xi is a set of covariates for each person (individual demographic characteristics), musculoskeletali is equal to 1 if person i has a musculoskel- etal disability,7 internali is equal to 1 if person i has a disability involving the internal systems, mentali is equal to 1 if person i has a mental disability, otheri is equal to 1 if person i has a disability classified as ‘‘other,’’ and posti is equal to 1 if person i is observed in 1992 or later. In this framework, the type of disability is controlled for by dummy variables indicating whether the individual has a musculoskeletal, in- ternal systems, mental, or other limitation, and the time period is con- trolled for by a dummy variable indicating whether the ADA had been implemented yet or not. The coefficients of interest (j 3, j  S,I,M,O), therefore, measure the change in log real wages of workers with each type of disability, relative to nondisabled workers, after implementa- tion of the ADA, relative to before implementation. Xi includes indi- vidual demographic and job characteristics, detailed in Table 3.2. Again, selection into the labor market has been controlled for.8 As it turns out, the type of disability that appears to be driving the observed overall decline in real wages of disabled workers relative to nondisabled workers, post-ADA, is musculoskeletal. The real wages Compensation: Wages and Benefits 57 of workers with musculoskeletal disabilities declined 4 percent more than for workers without disabilities post-ADA, relative to pre-ADA (the coefficient on musculoskeletal  post is 0.0424). This is of interest because it lends support to the theory that wages of disabled workers are sensitive to the degree of accommodation required of the employer. Whereas accommodation of a worker with a mental disor- der, such as depression, may simply be a flexible work schedule, indi- viduals with musculoskeletal disabilities might require more investment in infrastructure, such as specially constructed office furni- ture.9 In light of evidence that the typical per-worker cost of accommo- dation is fairly modest (on the order of $100–$1,000), according to Kujala (1996), employers may be setting wages on some perceived higher cost. Firm Size Analysis The CPS contains a question about how large (i.e., number of em- ployees) a worker’s main employer was in the previous year. Given that the ADA covers employers only of certain size, this information can be exploited to perform an additional analysis across covered and noncovered disabled workers. Covered disabled workers would be those employed by a firm with 25 or more employees in 1992 or later or employed by a firm with 15 or more employees in 1994 or later. Unfortunately, classifications of firms with fewer than 25 employees were not made until the 1992 survey year, which limits the amount of pre-ADA data available for the analysis. The post-ADA years were restricted to balance this survey-imposed limitation. The following model will be estimated twice: once for a large versus not-large firm classification, and a second time for a medium versus small firm classi- fication. Selection into the labor market will be controlled for in both estimations, using the standard Heckman (1979) two-step procedure detailed later in this chapter. The impact on wages across firm size (ADA coverage) is obtained from the following linear specification: (3.3) lnrwagei 
 
Xi  1disablei  2posti  3coveredi  4disablei  posti  5disablei  coveredi  6posti  coveredi  7disablei  posti  coveredi  i 58 Hotchkiss where lnrwagei is the natural log hourly real (1982–1984  100) wage of worker i, Xi is a set of covariates for each person (demographic and job characteristics), disablei is equal to 1 if person i has a work-limiting disability, posti is equal to 1 if person i is observed in 1992–1993 for the large firm analysis and equal to 1 if person i is observed in 1994–1996 for the medium firm analysis,10 and coveredi is equal to 1 if person i is employed by a firm covered by ADA legislation. The dummy variables (disable, post, and covered) control for the time-invariant characteristics of the affected group, disabled workers (1); the time-series changes in wages (2); and the time-invariant char- acteristics of the covered firm size, large or medium (3). The second- level interactions control for changes over time for disabled workers (4), time-invariant characteristics of disabled workers in the covered firm size (5), and changes over time within a covered firm size (6). The third-level interaction (7) captures all variation in wages specific to the disabled workers (relative to nondisabled workers) in the covered firm size (relative to uncovered firms) in the years after the firm was covered by ADA (relative to before ADA). The uncovered firms for the large firm analysis contain both small- and medium-sized organiza- tions (n  25). Uncovered firms for the medium firm analysis contain small entities only (n  10); large firms are not included in the medium firm analysis.11 The results of the medium firm analysis are somewhat contaminated by the fact that the ADA covers firms with 15 or more employees, so the indicator for medium firms contains some employers not technically covered by the ADA (those who employ more than 10 but fewer than 15 workers). The covariates included in the regression are detailed in Table 3.3, which presents the estimation results. The results in Table 3.3 indicate the following. Workers in large and medium (covered) firms earn higher wages than workers in small firms (see the coefficient on covered); disabled workers in large and medium firms earn higher wages, holding everything else constant, than nondisabled workers in those firms (see the coefficient on disable  covered); and wages of disabled workers covered by the ADA did not change post-ADA relative to disabled workers not covered (see the Compensation: Wages and Benefits 59 Table 3.3 OLS Selectivity-Corrected Regression Results for Log Real Wages, across Disability Status, Covered Firm Size, and ADA Implementation, CPS Medium firms Large firms (n 25) (10n25) as Regressor as covered group covered group Intercept 4.4252*** 3.9121*** (0.0368) (0.0820) Age 0.0838*** 0.0966*** (0.0169) (0.0266) Age squared 0.0009 0.0011 (0.0061) (0.0131) Female  1 0.3505*** 0.4264*** (0.0237) (0.0357) Nonwhite  1 0.0079* 0.0041 (0.0044) (0.0098) Single household  1 0.0905*** 0.0683*** (0.0013) (0.0023) High school grad  1 0.1374*** 0.2670*** (0.0032) (0.0099) Some college  1 0.2023*** 0.3086*** (0.0055) (0.0112) College grad  1 0.3093*** 0.3965*** (0.0055) (0.0102) Advanced degree  1 0.4246*** 0.5948*** (0.0064) (0.0124) disable  1 0.2780*** 0.4490*** (0.0047) (0.0210) post  1 0.0273*** 0.1071** (0.0057) (0.0081) covered  1 0.0982*** 0.1174*** (0.0047) (0.0086) disable  post  1 0.0570* 0.0317 (0.0322) (0.0431) disable  covered  1 0.0572** 0.1074** (0.0264) (0.0492) postcovered  1 0.0245*** 0.0146 (0.0066) (0.0123) disablepostcovered1 0.0517 0.0879 (0.0382) (0.0707) 60 Hotchkiss Table 3.3 (continued) Medium firms Large firms (n 25) (10n25) as Regressor as covered group covered group (selection term) 0.2767*** 0.0897 (0.0279) (0.0735) Number of observations 182,318 55,459 F statistic 5,253 1,048 Adjusted R2 0.51 0.40 NOTE: Reference group for large firm analysis is small and medium firms (n25); reference group for medium firm analysis is small firms only (n10). Other regress- ors included in the estimation, but not reported here, include seven industry and five occupation dummy variables, and region, central city, benefit receipt, and government employer dummy variables. For the large firm comparison, post0 for 1990–91 and post1 for 1992–93. For the medium firm comparison, post0 for 1991–93 and post1 for 1994–96. Asymptotically consistent standard errors are in parentheses. ***  significant at the 99 percent confidence level. **  significant at the 95 percent confidence level. *  significant at the 90 percent confidence level. coefficient on disable  post  covered, which is not significant). The implication is that the decline in wages of disabled workers relative to those of nondisabled workers found in Tables 3.1 and 3.2 (and in Table 3.3 by the coefficient on disable  post) is attributable to something other than ADA coverage. It is tempting to attribute the wage decline among disabled workers relative to nondisabled workers found earlier as indication that firms are passing accommodation costs on to disabled workers through lower pay. However, given that the wage difference between covered and noncovered disabled workers does not change post-ADA, there must be some other explanation than direct accommodation costs for the decline in wages relative to those of nondisabled workers. In other words,

if the lower wages among disabled workers were the result of accommodation costs directly, then the wages of covered workers should fall relative to those of noncovered workers (whose employers are not required to incur the cost of accommodation). This is not what we see from the firm-size analysis. It appears that all disabled workers (covered or not) are suffering some ramifications of the ADA not di- rectly attributable to the costs of accommodating their disabilities. The Compensation: Wages and Benefits 61 ADA may have created an environment in which firms view all dis- abled workers as a hiring risk (perhaps through fear of litigation upon termination), and are passing that perceived greater risk on through lower wages. WAGE DECOMPOSITION This section decomposes the wage differentials observed in Figure 3.1 to determine which factors have the greatest influence over their levels and growth. Standard log wage equations are estimated sepa- rately for disabled and nondisabled workers. The following specifica- tion, presented for person i, is estimated separately for each year. In these equations, ‘‘nd’’ denotes nondisabled and ‘‘d’’ denotes disabled: (3.4) lnWind  Xindnd  ind lnWid  Xidd  id where lnWi is the natural log hourly wage of workers, Xi are explanatory variables,  are coefficients to be estimated, and i is the random error term. As was seen in Chapter 2, there may exist significant self-selection into the labor market, particularly among the disabled population. In order to obtain an estimate of  representative of the population, this selec- tion is controlled for using the standard Heckman (1979) two-step pro- cedure.12 The first stage of this procedure involves estimating a binary choice model of the following form: (3.5) Ỹi  Z
i  ui , ui  N(0,1) where Zi are explanatory variables,  are parameters to be estimated, ui is the normally distributed random error, and individual i enters the labor force if Ỹi  0. Since Ỹi is unobserved, a binary variable, Yi, is defined as 62 Hotchkiss  (3.6) Yi 
1 as Ỹ 0 i 0.  The parameters, , are estimated via maximum likelihood probit, and the inverse-Mill’s ratio is constructed for inclusion in the wage equa- tions, which are then estimated via OLS. The modified wage equations are (3.7) lnWind  X
ind nd  nd ̂ind  vind, lnWid  X
id d  d ̂id  vid, (Zijj) where all variables are as previously defined, ̂ij  , j  d, nd, (Zijj) and vi is the newly defined random error. The parameter values that result from OLS estimation of the rela- tionships in Equation 7 can be used to decompose the wage differential between disabled and nondisabled workers as follows:13 (3.8) lnWnd  lnWd  ̂knd(Xknd  Xkd)  Xkd(̂knd  ̂d) k k  ( ̂nd nd  ̂d d). The first term on the right-hand side reflects the role of differences in characteristics (endowments) that disabled and nondisabled workers bring to the labor market; it is referred to as the ‘‘endowment effect.’’ The second term represents the differences among groups of workers in how their characteristics are valued in the workplace. This second term is often referred to as the ‘‘coefficient effect’’ or the ‘‘unexplained portion’’ and is cautiously attributed to discriminatory behavior on the part of the employer. The third term reflects the role of selection into the labor market (across disability status). The selectivity-corrected wage differential is calculated by subtracting the third (selectivity ef- fect) term from the observed wage differential. Figure 3.2 presents the results from this empirical analysis. The solid and dashed lines that move together toward the bottom of the figure represent the observed wage differential and the wage differen- Compensation: Wages and Benefits 63 Figure 3.2 Observed and Selectivity-Corrected Wage Differentials and the Coefficient Effect as Percentage of Corrected Wage Differential, CPS, 1981–2000 1 120 0.9 100 0.8 Coefficient effect as 0.7 percent of corrected wage differential 80 Coefficient Log wage 0.6 effect as a differential nondisabled 60 percent of corrected vs. disabled 0.5 wage differential 0.4 40 Wage differential 0.3 corrected for selection 20 0.2 Observed wage differential 0.1 0 1981 1983 1985 1987 1989 1991 1993 1995 1997 1999 (a) ADA Enactment (a) (b) (c) (b) ADA Phase I (c) ADA Phase II tial corrected for selection, respectively. The dotted line represents the coefficient effect as a percentage of the corrected wage differential.14 Observed and Selectivity-Corrected Wage Differentials Since 1981 (the earliest year of data), there is a clear and persistent increase in both the observed and selectivity-corrected wage differen- tials between disabled and nondisabled workers. Both differentials show that over this whole time period, nondisabled workers earned, on average, wages that were 23 percent higher than those of disabled workers. In addition, the corrected wage differential increased from 13 percent to 30 percent, indicating a deterioration of earnings of disabled workers relative to nondisabled workers over this time period. The decline in relative earnings is consistent with a downward trend identi- fied by Haveman and Wolfe (1990) beginning in 1974. However, the growth in the wage differential appears to have been mitigated since 1992; the selectivity-corrected wage differential grew from 13 to 29 percent between 1981 and 1992, and has hovered around a mean of 28 percent since 1992, which was the first year of implementation of the ADA. 64 Hotchkiss In addition, and particularly since 1996, individual selection into the labor market, or differences in selection between disabled and non- disabled workers, does not seem to be biasing the observed wage dif- ferentials between the two groups. The one exception to this might be the period 1992–1995. During this four-year span, the observed wage differential underrepresented the wage differential corrected for selec- tion. The implication of this is that disabled workers were positively selecting into the labor market to a greater extent than nondisabled workers, driving the observed wage of disabled workers as a whole upward (and the wage differential downward). This is consistent with the labor force participation decline observed in Chapter 2, if the dis- abled labor force nonparticipants beginning around 1992 had systemat- ically lower earnings potential than the disabled persons who stayed in the labor market. This would likely be the case as a result of the flow of Aid to Families with Dependent Children (AFDC) recipients to SSI (Lewin Group 1999). This result is not consistent, however, with the conjecture that the disabled workers entering the labor force post-ADA were those with the most limiting disabilities (Chirikos 1991). What- ever might have been making the observed and selectivity-corrected wage differentials diverge in the mid 1990s seems to have disappeared, since the two series have basically followed identical paths since 1996. The implication of this is that since 1996, selection into the labor mar- ket has had essentially the same impact on wages for disabled and nondisabled workers; self-selection explains none of the remaining wage differentials since that time. Potential Wage Discrimination against Disabled Workers The dotted line in the top part of Figure 3.2 reflects the coefficient effect as a percent of the corrected wage differential. Over the entire time period, the coefficient effect averages 77 percent of the corrected wage differential, or clearly a majority of the difference in wages be- tween disabled and nondisabled workers. While there is quite a bit of variation over the years, the coefficient effect dominates the endow- ment effect in each year. The regressors in each year explain the usual 30–40 percent of the variation in wages of disabled workers and about 45 percent of the variation in wages of nondisabled workers (as indi- cated by the adjusted R2 of the regressions). Consequently, interpreting Compensation: Wages and Benefits 65 the entire coefficient effect as an indication of discrimination would not be prudent. However, given the relative magnitude of the coeffi- cient effect, the expected success in explaining wage variation, and the number of observable characteristics included in the regression, it is also unlikely that the coefficient effect can be completely dismissed as the result of unmeasured characteristics of either the disabled or nondisabled. Using data from the SIPP, Baldwin and Johnson (1995, 2000) also find that the coefficient effect is larger than the endowment effect as a percentage of the selectivity-corrected differential in 1984 and in 1990. Using a similar methodology and data from 1972, John- son and Lambrinos (1985) show that only 34–40 percent of the wage differential between disabled and nondisabled workers was left unex- plained by differences in endowments. Examining SIPP data from 1984 and 1993, DeLeire (2001) finds that only between 5 and 8 percent of the earnings gap is attributable to the coefficient effect. Examining the endowment and coefficient effects in greater detail, it is of interest to see which set of regressors makes the largest contri- butions to these components. Table 3.4 presents the median contribu- tion (across years) of the groups of regressors that control for occupation, industry, and education. While the contributions vary across the years, these median values represent the typical scenario (i.e., there is no obvious trend in any of these contributing factors), and it is usually the case that occupation and education were the largest Table 3.4 Contribution of Regressors to Log Wage Differentials, Median across Years, CPS, 1981–2000 Contribution Contribution to the to the endowment coefficient effect effect Occupation 0.0452 0.0424 Industry 0.0085 0.0817 Education 0.0363 0.0419 Median total effects 0.0768 0.1738 NOTE: The contributions of occupation, industry, and education do not add up to the total effect because these numbers represent the median across all years and also do not represent all regressors in the wage regression. 66 Hotchkiss contributors to the endowment effect and that industry was the largest contributor to the coefficient effect. It is also of interest to note that in 18 of the 20 years, the return to education acted to decrease the wage differential between disabled and nondisabled workers (the contribu- tion of the education regressors to the coefficient effect was negative). In other words, disabled workers typically received a greater return to their educational investment than nondisabled workers.15 Regarding endowments, it is clear that nondisabled workers bring greater educa- tional attainment to the labor market and are more likely to locate in the higher-paying occupations and industries (a phenomenon that will be explored more fully in Chapter 4); these observations are evidenced by positive contributions made by the occupation, industry, and educa- tion regressors to the endowment effect. An additional feature provided by the analysis is that the relative importance of the coefficient effect over time can be evaluated. While perhaps not very obvious in Figure 3.2, there is actually a (slightly) significant negative trend in the coefficient effect as a percentage of the corrected wage differential. On average, the contribution of the coefficient effect to the overall wage differential declines an average of 1.7 percentage points per year from 1981 to 2000.16 In addition, the endowment effect as a percentage of the corrected wage differential increased an average of 0.6 of a percentage point per year over the same time period.17 Consequently, another silver lining to the rising wage differential between disabled and nondisabled workers, and to the large portion of that differential not explained by differences in endowments, is that any potential discrimination against disabled workers, as measured by the coefficient effect, is declining. Addition- ally, this result suggests that one way to combat the rising wage differ- ential is to improve disabled workers’ endowments (e.g., greater investments in human capital, or placement in higher-paying occupa- tions or industries). It is also important to note, however, that these improvements appear to be a continuation of a trend rather than any dramatic post-ADA shift. BENEFIT ANALYSIS As the percentage of fringe benefits in total compensation contin- ues to increase, benefits become an increasingly important contributor Compensation: Wages and Benefits 67 to workers’ labor market experience. The CPS allows identification of a worker receiving two fringe benefits from his or her employer: health insurance and a pension plan. The data used for this analysis were obtained from the CPS March supplemental questionnaire and there- fore refer to benefit coverage in the years 1980–1999. The probabili- ties of being included in an employer’s pension plan or of receiving health insurance through an employer were fairly stable

across the years; however, as Figure 3.3 shows, the proportion of nondisabled workers relative to disabled workers included in a pension plan has grown over the time period. In 1980, nearly 11 percent more of nondisabled workers were in- cluded in a pension plan than disabled workers were. This difference grew to 17 percentage points by 1999. The greater proportion of non- disabled workers receiving either benefit could be closely related to the types of jobs disabled versus nondisabled workers hold. The increase in the difference in proportions could also be related to disabled work- ers moving into jobs less likely to offer these benefits. For example, Chapter 4 will detail the growth in part-time employment among dis- abled workers. Neither phenomenon, however, appears to have been impacted by events surrounding the passage and implementation of the ADA. The goal of the analysis of this section is to determine whether Figure 3.3 Difference in Proportion of Nondisabled and Disabled Workers Receiving Benefits, CPS, 1980–1999 Percentage 19 points 18 17 Health 16 insurance 15 14 13 12 Included in pension plan 11 10 1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 (a) ADA Enactment (a) (b) (c) (b) ADA Phase I (c) ADA Phase II 68 Hotchkiss disabled workers are more or less likely than nondisabled workers to be included in their employers’ pension plan or to have their employers pay for health insurance, while controlling for all other individual and job characteristics. Since the observed model is an indicator of coverage/inclusion or not, the empirical model is specified as a probit: (3.9) B̃ij  Q
ij  vi , i  N(0,1) where individual i receives benefit j ( j  health insurance, pension plan inclusion) if B̃ij  0. Since B̃ij is unobserved, a binary variable, B̃ij, is defined as  (3.10) Bij 
1 as B̃ij 0. 0  The parameters, , are estimated via maximum likelihood probit. Qi comprises various individual and job characteristics for worker i, in- cluding a dummy variable indicating whether the worker is disabled or not. The model is estimated on a sample of workers only; thus, the results are generalizable solely to workers. The marginal effect of being disabled on receiving a benefit is calculated as the partial deriva- tive for each worker, then averaged over the entire sample.18 The esti- mation results are depicted in Figure 3.4.19 A reliable measure of annual earnings is available only since 1987; therefore, the marginal effects only cover the period 1987–1999.20 The marginal impact of being disabled on fringe benefit receipt follows the same path for both health insurance and pension plan inclu- sion: an increasingly negative impact of being disabled on the probabil- ity of receiving benefits. Since these marginal effects are calculated from specifications which included a control for earnings, the increase cannot be attributed to disabled workers merely being employed in jobs that are lower paying (thus, less likely to offer fringe benefits). The specification also included controls for hours of work (thus the poten- tial impact of part-time employment), occupation, industry, and indi- vidual human capital characteristics. While there does seem to be a fairly significant intensification in the negative impact of being dis- abled in 1995, it would be hard to attribute that to anything other than Compensation: Wages and Benefits 69 Figure 3.4 Marginal Effect of Disability on the Probability of Fringe Benefit Receipt, CPS, 1987–1999 Impact on 0 probability -0.02 -0.04 -0.06 Health insurance -0.08 -0.10 Pension plan -0.12 1987 1989 1991 1993 1995 1997 1999 (a) ADA Enactment (a) (b) (c) (b) ADA Phase I (c) ADA Phase II the continuation of a trend, since the negative impact has diminished somewhat in recent years. However, the negative trend cannot be ig- nored. There is some evidence that a weakening in 1989 of antidis- crimination laws governing provision of health insurance resulted in lower rates of benefit incidence among ‘‘peripheral’’ workers, such as recent hires or those working part-time (Farber and Levy 2000). If disabled workers fall into the category of ‘‘peripheral,’’ this might ex- plain some of the deterioration in health insurance receipt among dis- abled relative to nondisabled workers. An additional consideration that might explain the deterioration of relative health insurance coverage is that the Employment Opportuni- ties for Disabled Americans Act and the Omnibus Budget Reconcilia- tion Act of 1990 amended Title XVI of the Social Security Act to allow SSI recipients to continue participating in Medicaid (under specific circumstances) even if their earnings exceeded the SSI qualifying level (59 FR 41403, 12 August 1994). This amendment became effective in August 1994. The change would not have typically affected disabled workers already employed but would have encouraged disabled indi- viduals who previously had not sought employment because of a lack of health insurance to do so. While this law says nothing about the 70 Hotchkiss provision of pension plans, fringe benefits are highly correlated with one another, so it is not a surprise that as the proportion of disabled workers without employer-provided health insurance increased, the proportion without a pension plan also increased. CONCLUSIONS The purpose of this chapter was to evaluate the relative compensa- tion experience of disabled and nondisabled workers over time, and to determine whether any change in that experience is evident in relation to the implementation of the ADA. A pooled, cross-sectional analysis using the CPS data indicated that overall, disabled workers experienced a 3 percent decline in real wages, relative to nondisabled workers, post- ADA, relative to pre-ADA. Results from the SIPP are consistent with the CPS results and show that the wage experience of those with mus- culoskeletal disabilities is driving that observed wage decline. The real wages of workers with musculoskeletal disabilities declined 4 percent more than for workers without disabilities post-ADA, relative to pre- ADA; workers with other types of disabilities did not experience any different wage change than did nondisabled workers. These results together indicate that overall, the cost of accommodating workers’ dis- abilities exceeds the gain in productivity of those workers. In addition, if accommodating musculoskeletal disabilities results in more costly structural investments, the SIPP results lend support to the theory that wages of disabled workers are sensitive to the degree of accommoda- tion (in terms of cost) required of the employer. This result is consis- tent with the finding by Gunderson and Hyatt (1996) that firms pass on (through lower wages) the cost of workplace modifications to the hired disabled worker. In light of this, the tax credits in place to assist certain employers in absorbing these costs (Hays 1999) either have not gone far enough or are not being widely used. In contrast with these overall (disabled versus nondisabled) results, it was found that disabled workers employed by firms covered by the ADA did not experience any wage deterioration relative to disabled workers not covered by the ADA. This combination of results suggests that the wage decline experience by disabled workers relative to non- Compensation: Wages and Benefits 71 disabled workers is not the result of explicit accommodation costs being passed on to disabled workers through lower wages (in which case, the wages of covered disabled workers should have deteriorated the most). Rather, all disabled workers seem to be bearing the burden of a perceived additional hiring risk associated with them that may exist post-ADA. Consistent with the overall relative wage decline identified through the pooled, cross-sectional analysis, further study illustrates that the positive wage differential between disabled and nondisabled workers has risen considerably since 1981. However, the trend does seem to have flattened out since the early 1990s. The large portion of that differential not explained by differences in endowments has also been declining over time, indicating that any potential wage discrimination against disabled workers is at least falling. Additionally, the fact that worker endowments (occupation, education, etc.) contribute positively to the measured wage differential suggests that one way to fight the rising wage differential is to enhance disabled workers’ endowments (e.g., greater investments in human capital, or placement in higher- paying occupations or industries). In addition to the rising wage differentials, the negative impact of being disabled on the probability of receiving benefits has also been rising. One potential contributor to this situation is the allowance (as of August 1994) for some disabled workers to continue receiving Medicaid even when their earnings surpass SSI cut-off levels. The combination of the rising wage differential and decreasing probability of disabled workers receiving employer-sponsored health and pension benefits leads to the conclusion that the relative position of disabled workers regarding compensation is deteriorating overall. The ADA provides a clear mandate ‘‘for the elimination of dis- crimination against individuals with disabilities’’ (section 2 of the ADA). The evidence provided in this chapter suggests that while po- tential wage discrimination against the disabled appears to be declin- ing, the improvement is taking place at a very slow rate, and not necessarily as a result of the ADA, as the trend goes back at least to 1981. In addition, even if discrimination is decreasing, the overall compensation experience (including benefit provision) for disabled workers, relative to nondisabled workers, is declining, as well. 72 Hotchkiss Notes 1. See also Burkhauser, Haveman, and Wolfe (1993) for an analysis of the well- being of the disabled. 2. Examples of other applications of this method of analysis can be found in Card (1992), Gruber (1994, 1996), Zveglich and Rodgers (1996), and Hamermesh and Trejo (2000). The reader is reminded of the caveats detailed by Heckman (1996). Also see further discussion in Chapter 2. 3. The first year in the analysis was 1984 due to the poor representation of disabled workers in some occupations in 1983, and since union status was not indicated in 1981 or 1982. 4. While insignificantly different from zero, the coefficient in DeLeire’s pooled, cross-sectional analysis was also negative. The results reported in Table 3.1 are robust to defining post as years past 1990 and to defining post as 1994 and later. 5. It is important to note that since there is no measure of labor market experience available in the CPS survey, and since disabled workers likely have less labor market experience than nondisabled workers of the same age, the coefficient on the disable dummy variable may be capturing some of the impact of labor market experience on the wage. The inability to control for labor market experience should have a smaller impact, however, on the coefficient for the interaction term of disable  post, since the consequence of the absence of an experience vari- able should be similar across time. 6. See Appendix B for disability types grouped for these classifications. 7. The musculoskeletal grouping includes disabilities of the special senses (e.g., hearing, seeing). 8. See the wage decomposition section in this chapter for details of the procedure used to control for self-selection. 9. Whereas the wages of those with mental disorders did not deteriorate post-ADA, it is of interest to note that the largest (negative) coefficient among types of dis- ability (indicating overall relative wage performance) is that on mental (0.3427). This is consistent with the findings of Baldwin (1999). 10. The pre and post periods were chosen to achieve balance in the number of years; the results did not change if the number of years were extended. 11. These definitions of ‘‘large’’ and ‘‘small’’ differ from those used in Chapter 2. 12. This specification could also be modified to account for the likely joint determina- tion of wages and hours worked; Moffitt (1984), Lundberg (1985), Altonji and Paxson (1988), and Tummers and Woittiez (1991) all demonstrate the importance of the simultaneous determination of wages and hours. This joint model is not estimated here for simplification, but hours are included as a covariate in the wage equation. 13. Since only 3 percent of the working sample is disabled, nearly all of the coeffi- cient effect is attributed to the disadvantage experienced by the disabled, since the linear combination of the two worlds yields estimates very close to those experienced in the nondisabled world (see Cotton 1988). Other renditions of this Compensation: Wages and Benefits 73 decomposition, such as Oaxaca and Ransom (1994) would result in a similar outcome, since the disabled make up such a small portion of the whole workforce.

This same strategy is followed by Baldwin and Johnson (2000) in their analysis of labor market discrimination against disabled workers. 14. The data generating this figure are found in Table C.5 in Appendix C. 15. This finding is consistent with the results reported by Hollenbeck and Kimmel (2001) that people with poor health or disability earn a positive return to educa- tion and training, although they find that return to be equal to that of nondisabled individuals. 16. The coefficient of 0.017 on a linear trend estimation had a standard error of 0.0067, making it significantly different from zero at the 95 percent confidence level. 17. The coefficient of 0.006 on a linear trend estimation had a standard error of 0.0027, making it significantly different from zero at the 95 percent confidence level. 18. This is preferable to calculating the marginal benefit for the average person, since we are most likely to be interested in the marginal effect for a worker drawn at random, rather than the marginal effect for the average person in the sample. 19. Table C.6 contains the marginal effects used to generate Figure 3.4. 20. An identical analysis was performed for the entire 1980–1999 time period, ex- cluding the control for earnings. The marginal effect of being disabled was sig- nificantly larger, as would be expected, but exhibited the same trend going back to 1980 as that depicted in Figure 3.4. This Page Intentionally Left Blank 4 Hours of Work, Distribution, and Representation In addition to the wage, there are a number of other characteristics that can be used to quantify the quality of a worker’s job. One feature is whether a job is full-time or part-time. While the availability of part-time employment may be important to disabled workers (and per- haps more so than to nondisabled workers), part-time jobs are often accompanied by lower pay, fewer benefits, and less stability.1 The first part of this chapter compares and evaluates the incidence of part-time employment and type of part-time employment (voluntary versus in- voluntary) across disability status and across time. If disabled workers are considered marginal workers, then they would be more likely to be employed part-time. If, however, disabled part-time workers are more likely to be voluntarily, versus involuntarily, employed part-time, then their part-time status may indicate a greater flexibility that might be needed to accommodate the worker’s situation. The chapter then ex- plores the full-time wage premium earned by disabled and nondisabled workers. Disabled workers may not earn as great a premium for com- mitting to a full-time schedule as nondisabled workers. Given the po- tentially higher fixed cost of accommodating the worker’s disability, the individual may have to commit to a greater number of hours before seeing the premium; this could show up in a lower premium at any given definition of part-time employment.2 A major characteristic of one’s job is its occupation or industry. A popular indicator of the quality of employment of a disadvantaged group is how well that group is represented in desirable occupations relative to some comparison group, and how the disadvantaged group’s distribution across occupations compares with that of the comparison group. The occupation that a worker holds, or the industry in which someone works, can play an important role in that person’s satisfaction and potential advancement in the labor market. Dual labor market theory suggests that some workers are relegated to undesirable (e.g., low-paying, dead-end) jobs from which they have virtually no escape.3 75 76 Hotchkiss The second part of this chapter will explore the distribution of disabled workers across occupations and industries, relative to the distribution of nondisabled workers, as well as examine the representation of dis- abled workers in ‘‘desirable’’ jobs. The emphasis will be on how the relative distribution and representation have changed over time and whether the ADA seems to have played a role in their current determi- nation. HOURS OF WORK Figure 1.2 in Chapter 1 highlighted a growing disparity in average hours worked per week between disabled and nondisabled workers. This section looks more closely at the role part-time employment plays in that observed decline in hours and determines whether it reflects voluntary or involuntary behavior on the part of disabled workers. Part-time employment among the disabled may not be a sign of mar- ginalization or discrimination because of these individuals’ unique physical or mental capabilities and potential income sources. Such employment may be sought by disabled workers (and employers for their disabled workers) as a way to accommodate health limitations. In addition, part-time employment may provide additional earnings that do not jeopardize disability benefits based on income levels. Incidence of Part-time Employment Figure 4.1 depicts the percentage of both disabled and nondisabled workers that are employed part-time for each year, 1981 to 2000. Al- though there is some discrepancy as to the appropriate definition of part-time employment (see Hotchkiss 1991), the CPS definition of ‘‘less than 35 hours per week’’ is retained here. The use of respondent- supplied reasons (later in the chapter) for working less than 35 hours per week makes this the practical choice. As has been suggested in Chapters 1 and 3, part-time employment has grown among disabled workers between 1981 and 2000 and has declined somewhat among nondisabled workers. By itself, this obser- vation is consistent with the contention that disabled workers are being Hours of Work, Distribution, and Representation 77 Figure 4.1 Percentage of Disabled and Nondisabled Workers That Are Employed Part-Time, CPS, 1981–2000 45 Disabled 40 35 30 25 Nondisabled 20 15 10 5 0 1981 1983 1985 1987 1989 1991 1993 1995 1997 1999 (a) ADA Enactment (a) (b) (c) (b) ADA Phase I (c) ADA Phase II pushed to the fringe and becoming more marginalized. However, these raw numbers do not control for other job or individual characteristics. In order to appropriately model the impact of having a work- limiting disability on the incidence of part-time employment among workers, a bivariate probit with selection model, as in Chapter 2, is specified. This model estimates the probability of being employed part-time while controlling for unobservable determinants of being both employed and employed part-time. The bivariate specification allows for the two outcomes (employment and part-time employment) to be impacted by the same unobservable factors (e.g., motivation). The selectivity part of the model is merely a recognition that we do not get to see the part-time employment outcome unless the person is employed to begin with, and that those we observe as employed may have systematically different part-time options or make different hours choices than those not employed. Correcting for selectivity allows us to make inferences for anyone from the population, not just those we observe as employed; this is what makes the probability unconditional. The following model defines the relationship assumed between the employment of person i (empi), the probability of being employed part- 78 Hotchkiss time (pti), and individual characteristics that are believed to affect the employment outcome (X1i) and the incidence of part-time employment (X2i): (4.1) empi 
1  
1X1i  1 disablei  1i 
1 if person i is employed 0 otherwise (4.2) pti 
2  
2X2i2 disablei  2i 
1 if person i is employed part-time 0 otherwise. disablei is equal to 1 if person i is disabled, 0 otherwise; 1i and 2i are distributed as a bivariate normal with means equal to 0, variances equal to 1, and correlation equal to ; and
j, j, and j ( j1, 2) are parame- ter coefficients to be estimated. In addition, of course, pti is only observed if empi1.4 X1i and X2i include individual demographic char- acteristics detailed in the notes to Table 4.1. The impact of having a work-limiting disability on part-time employment, then, is determined by calculating the unconditional probability of being employed part- time for each individual, varying the disability index between 0 and 1, then averaging across the sample.5 Separate specifications are estimated for each year, and the marginal impact of having a work-limiting dis- ability is calculated. The estimation results are depicted in Figure 4.2.6 The line in Figure 4.2 reflects an increase in the impact of being disabled on the determinant of the unconditional probability of being employed part-time. A work-limiting disability increased the probabil- ity of a worker being observed as employed part-time by 12 percentage points in 2000. This is double the 6 percentage points impact of a disability on being employed part-time estimated for 1981. What is also apparent from the graph is that this effect has experienced a rather consistent upward trend during the entire time period, with the largest adjustment occuring during the ADA phase-in period. In order to quantify the apparent growth in selectivity-corrected part-time employment among disabled workers, relative to nondisabled workers, the pooled, cross-sectional analysis introduced in Chapter 2 is applied here. The idea behind the analysis is to estimate a cross- sectional, time-series bivariate probit model with dummy variables rep- Hours of Work, Distribution, and Representation 79 Figure 4.2 Impact of Having a Disability on Being Employed Part-Time, CPS, 1981–2000 Percentage-point 18 impact on 16 probability of 14 being disabled 12 10 8 6 4 2 0 1981 1983 1985 1987 1989 1991 1993 1995 1997 1999 (a) ADA Enactment (a) (b) (c) (b) ADA Phase I (c) ADA Phase II resenting whether the observation shows up in the data pre-ADA or post-ADA and whether the observation is a disabled or nondisabled person. These dummy variables are also interacted to determine whether being disabled had any greater impact on employment after the ADA than before the ADA, relative to the experience of a nondis- abled person.7 The pooled, cross-sectional analysis looks just like the bivariate probit with selection estimated in one year, except with the additional time-period dummy variables: (4.3) empi 
1  
1X1i  1 disablei  1 posti  1 disablei  posti  1i (4.4) pti 
2  
2X2i  2 disablei  2 posti  2 disablei  posti  2i. Again, empi  1 if person i is in the labor force, 0 otherwise, and pti is not observed unless empi  1. disablei is equal to 1 if person i is disabled, 0 otherwise; posti is equal to 1 if person i is observed in 1992 or later (as in Chapter 2); X1i and X2i include individual demographic characteristics; 1i and 2i are distributed as a bivariate normal with means equal to 0, variances equal to 1, and correlation equal to ; and j and j ( j  1, 2) are additional coefficients to be estimated. 80 Hotchkiss In this framework, the affected group (the disabled) is controlled for by a dummy variable indicating whether the individual has a work- limiting disability, and the time period is controlled for by a dummy variable indicating whether the ADA had been implemented yet or not. Because of the model’s nonlinearity, a single parameter coefficient does not tell us the additional impact the ADA had on the difference in employment probabilities between the disabled and nondisabled. Table 4.1 details the regression results. Using the parameter estimates, the difference in the impact of hav- ing a work-limiting disability on part-time employment across the two time periods can be calculated by evaluating the probabilities of inter- est for each person, varying the disable and post dummy variables, then taking the difference between these probabilities and averaging those differences across the sample. This calculation translates the Table 4.1 Employment and Part-Time Employment Bivariate Probit with Selection Results, CPS Combined Years 1981–2000 Part-time Employment employment Regressor equation equation disable  1 0.1118*** 0.3474*** (0.0146) (0.0130) post (year  1992 or later)  1 0.0112** 0.0017 (0.0051) (0.0037) disable  post  1 0.0298 0.1775*** (0.0224) (0.0190) Rho 0.7983*** (0.0054) Log-likelihood 477,354 Number of observations 906,646 NOTE: Regressors included both in the employment and part-time employment equa- tions (but not reported here) include age, education, region, race, gender, marital status, and a central city residence indicator. Regressors unique to the employment equation include the state unemployment rate and the number of weeks worked last year. Regressors unique to the part-time employment equation include occupation and industry dummies, nonlabor income, and a government employer indicator. Stan- dard errors are in parentheses. ***  significant at the 99

percent confidence level. **  significant at the 95 percent confidence level. Hours of Work, Distribution, and Representation 81 estimated coefficients from the bivariate probit into a 5 percentage point greater probability of disabled workers being employed part-time than nondisabled workers, post-ADA relative to pre-ADA. In addition, the probability of nondisabled workers being employed part-time changed by less than one-hundredth of a percentage point post- versus pre-ADA. Type of Part-Time Employment Given the conclusion that disabled workers are more likely than the nondisabled to be employed part-time and that the disparity is growing, an important consideration is what is the nature of the part- time jobs? Are disabled workers more likely to be employed other than full-time by choice? In order to answer this question, a univariate probit analysis is performed. The purpose of the probit analysis is to determine, among part-time workers, whether the probability of being voluntarily (versus involuntarily) employed part-time has increased or decreased for disabled workers, relative to nondisabled part-time work- ers, holding constant other factors that may determine the classifica- tion.8 The results of this probit estimation can be found in Figure 4.3.9 The graph depicts the marginal effect of being disabled on the probability that a part-time worker’s status is voluntary. The results are generalizable to part-time workers only. The observation of inter- est from Figure 4.3 is that prior to 1992, being disabled decreased a part-time worker’s probability of being voluntarily (versus involun- tarily) employed part-time; however after 1992, disabled part-time workers became more likely to be voluntarily employed part-time than nondisabled part-time workers. The implication is that much of the growth in part-time employment has been voluntary (for a given set of individual characteristics) and may actually be in response to the better accommodation of a worker’s disability.10 In order to quantify what may be obvious from Figure 4.3, a pooled, cross-sectional approach is taken to determine the extent to which the disabled are more likely than the nondisabled to be voluntar- ily employed part-time post-ADA versus pre-ADA. Since this analysis is concerned only with part-time workers as a group, a linear probabil- ity model is estimated so that the parameter coefficient is directly inter- pretable as a marginal effect; again, the results are generalizable only 82 Hotchkiss Figure 4.3 Impact of Having a Disability on Being Voluntarily Employed Part-Time, CPS, 1981–2000 Percentage- 8 point impact on 6 probability of 4 being disabled 2 0 -2 -4 -6 -8 -10 1981 1983 1985 1987 1989 1991 1993 1995 1997 1999 (a) ADA Enactment (a) (b) (c) (b) ADA Phase I (c) ADA Phase II to the part-time employed population. The linear probability model takes the following form: (4.5) vpti 
3  
3X3i  3 disablei  3 posti  3 disablei  posti  3i where vpti is equal to 1 if person i is voluntarily employed part-time, 0 if involuntarily part-time, Xi is a set of covariates for each person (individual demographic characteristics), disablei is equal to 1 if person i has a work-limiting disability, and posti is equal to 1 if person i is observed in 1992 or later. In this framework, the affected group (the disabled) is controlled for by a dummy variable indicating whether the individual has a work- limiting disability, and the time period is controlled for by a dummy variable indicating whether the ADA had been implemented yet or not. The coefficient of interest (3), therefore, measures the change in em- ployment probability of disabled workers, relative to nondisabled workers, after implementation of the ADA, relative to before imple- mentation (the other parameter coefficients are analagous to their coun- Hours of Work, Distribution, and Representation 83 terparts in Equations 4.3 and 4.4). X3i includes individual demographic characteristics. Table 4.2 details the regressors included in the estima- tion and the regression results. The coefficient on disable  post confirms that the probability of being voluntarily (versus involuntarily) employed part-time increased 4 percentage points more for disabled part-time workers than for non- disabled part-time workers, post-ADA. This result, taken with the overall growth in part-time employment, suggests that part-time em- ployment and flexible hours may be a mechanism by which employers are able and willing to accommodate workers’ disabilities. Evidence from the SIPP The pooled, cross-sectional methodologies are appealed to again in order to determine the impact of different types of disabilities on the incidence of part-time employment and on the kind of part-time Table 4.2 Linear Probability, Voluntary Part-Time Employment Results, CPS Combined Years 1981–2000 Regressor Coefficient disable  1 0.0158** (0.0072) post (year  1992 or later)  1 0.0549*** (0.0023) disable  post  1 0.0401*** (0.0102) Adjusted R2 0.08 F statistic 481.91 Number of observations 170,870 NOTE: Observations from 1983 were not included because of the unreliable represen- tation across occupational categories. Regressors included both in the employment and part-time employment equations (but not reported here) include age, education, region, race, gender, marital status, and a central city residence indicator. Regressors unique to the employment equation include the state unemployment rate and the num- ber of weeks worked last year. Regressors unique to the part-time employment equa- tion include occupation and industry dummies, nonlabor income, and a government employer indicator. Standard errors are in parentheses. ***  significant at the 99 percent confidence level. **  significant at the 95 percent confidence level. 84 Hotchkiss employment across time using the SIPP data set. The bivariate probit model with selection is estimated with two different specifications. The first includes only a single dummy variable indicator for having a work-limiting disability; the second includes multiple dummies indi- cating the type of disability a person might have. The general structure of the estimation strategies looks like this: (4.6) empi 
1  
1X1i n k 1 distypek i  1 posti k(S,I,M,O)  n k 1 distypek i  posti  1i k(S,I,M,O) (4.7) pti 
2  
2X2i n k 2 distypek i  2 posti k(S,I,M,O)  n k 2 distypek i  posti  2i k(S,I,M,O) where distypei  disablei in specification 1 musculoskeletali(S); internali(I); mentali(M); and otheri(O) in specification 2. Again, empi  1 if person i is in the labor force and employed, 0 otherwise, and pti is not observed unless empi  1. X1i and X2i include individual demographic characteristics; posti is equal to 1 if person i is observed in 1992 or later; disablei is equal to 1 if person i has a work-limiting disability; musculoskeletali is equal to 1 if person i has a musculoskeletal disability;11 internali is equal to 1 if person i has a disability involving the internal systems; mentali is equal to 1 if person i has a mental disability; otheri is equal to 1 if person i has a disability classified as ‘‘other’’; and 1i and 2i are distributed as a bivar- iate normal with means equal to 0, variances equal to 1; and correlation equal to ; and
, , , , and  are all parameters to be estimated. In this framework, the affected group (the disabled) is controlled for by a dummy variable (or set of dummy variables) indicating whether the individual has a work-limiting disability (or type of dis- ability), and the time period is controlled for by a dummy variable indicating whether the ADA had been implemented yet or not. The Hours of Work, Distribution, and Representation 85 difference in the impact of having a work-limiting disability on em- ployment across the two time periods can be calculated by evaluating the probabilities of interest for each person, varying the distype and post dummy variables, then taking the difference between these proba- bilities, and averaging across the sample. An additional estimation is performed on a subsample of part-time workers only to determine whether the type of disability impacts the incidence of voluntary part-time employment. The following equation is estimated via OLS: (4.8) vpti 
3  
1X3i k 3 distypek i  3 posti k(S,I,M,O)  n k 3 distypek i  posti  3i. k(S,I,M,O) Table 4.3 contains the results from estimating both specifications of Equation 4.7 and both specifications of Equation 4.8. The two speci- fications for each equation correspond to how distype is defined. Con- sistent with the findings from the CPS, the results in Table 4.3 indicate that the probability that a disabled worker is employed part-time in- creased more than the probability for a nondisabled worker, post-ADA relative to pre-ADA. The coefficient of 0.0925 in column 1 of Table 4.3 translates into a relative 2 percentage point greater probability for part-time employment for the disabled post-ADA. In addition, the strongest impact was experienced by those with musculoskeletal (0.1084) and mental (0.1730) disabilities. Regarding voluntary part- time employment, the results suggest that while the impact of having a disability on the probability of being voluntarily employed part-time increased post-ADA, that rise was not significantly different from zero for disabled workers as a group. However, those with mental disabili- ties seem to have experienced a greater increase in the probability of being voluntarily employed part-time than nondisabled part-time work- ers, post-ADA. This makes sense if mental disorders are the type of disability most effectively accommodated by a flexible or reduced- hours work schedule.12 Full-Time Wage Premium Even beyond whether a part-time job is voluntary or involuntary is the wage penalty experienced by part-time workers. It is well docu- 86 Hotchkiss Table 4.3 Employment and Part-Time Employment Bivariate Probit with Selection and Linear Probability Model for Voluntary Part-Time Employment, SIPP 1986–1997 Probability of part-time employmenta Probability of voluntary part-timeb Disability Type of Disability Type of Regressor indicator only disability indicated indicator only disability indicated disable  1 0.2024*** — 0.0151 — (0.0188) (0.0102) post (year  1992 or later)  1 0.0090 0.0088 0.0380*** 0.0683*** (0.0059) (0.0059) (0.0035) (0.0038) disable  post  1 0.0925*** — 0.0095 — (0.0238) (0.0127) Musculoskeletal disability  1 — 0.0821*** — 0.0221 (0.0262) (0.0156) Internal Systems disability  1 — 0.2743*** — 0.0275 (0.0356) (0.0194) Mental disorder disability  1 — 0.4177*** — 0.0398* (0.0464) (0.0238) Other disability  1 — 0.2587*** — 0.0414 (0.0481) (0.0261) Hours of Work, Distribution, and Representation 87 musculoskeletal  post  1 — 0.1084*** — 0.0238 (0.0331) (0.0195) internal  post  1 — 0.0013 — 0.0068 (0.0476) (0.0258) mental  post  1 — 0.1730*** — 0.0907*** (0.0550) (0.0281) other  post 1 — 0.0302 — 0.0266 (0.0627) (0.0344) Adjusted R2 — — 0.06 0.07 Log-likelihood 199,110 199,021 Number of observations 360,036 360,036 72,890 59,059 NOTE: Standard errors are in parentheses. ***  significant at the 99 percent confidence level. **  significant at the 95 percent confidence level. *  significant at the 90 percent confidence level. aThese results are from estimation of a bivariate probit with selection. Other regressors included in the part-time employment equation include age; age squared; nonlabor income; and gender, education, marital status, race, education, region, urban, government, industry, and occupational dummy variables. The selection equation (not reported here) is an employment equation. bThese results are from estimation of a linear probability model via OLS and are generalizable to the part-time population only. Other regressors included in the voluntary part-time employment equation include age; age squared; and gender, education, marital status, race, education, region, government, industry, and occupational dummy variables 88 Hotchkiss mented that part-time workers earn considerably less per hour for not making a full-time commitment to his/her employer. This penalty can range from 30 to 60 percent lower wages depending on gender and race groups (Averett and Hotchkiss 1996), and it is a main reason that part- time jobs are considered undesirable (Blank 1990). The reason typi- cally given as to why part-time workers earn a lower wage is the pres- ence of fixed costs associated with hiring personnel. Employers are able to spread these fixed costs over more hours for full-time workers, allowing them to pay higher wages to such individuals. One concern might be that the fixed costs of hiring disabled workers are even greater than for nondisabled workers so that the wage differential between full- time and part-time disabled workers is larger than the differential be- tween full-time and part-time nondisabled workers. This section presents the full-time/part-time wage differentials ex- perienced by disabled and nondisabled workers, controlling for their

selection into the labor market. Standard log wage equations are esti- mated separately for disabled and nondisabled workers. In the equa- tions for person i, ‘‘ft’’ denotes full-time, and ‘‘pt’’ denotes part-time: (4.9) ln Wift  Xiftft  ift ln Wipt  Xiptpt  ipt where lnWi is the natural log hourly wage of workers, Xi is the explanatory variable,  is the set of coefficients to be estimated, and i is the random error term. As seen in Chapter 2, there may exist significant self-selection into the labor market, particularly among the disabled population. In order to obtain an estimate of  representative of the population, this selection is controlled for using the standard Heckman (1979) two-step proce- dure. The first stage of this procedure involves estimating a binary choice model of the following form: (4.10) Ỹi  Z
i  ui , ui  N(0,1) where Zi are regressors expected to affect the labor supply decision,  are parameter coefficients, ui is the normally distributed random Hours of Work, Distribution, and Representation 89 error, and individual i enters the labor force if Ỹi  0. Since Ỹi is unobserved, a binary variable, Yi, is defined as  (4.11) Yi 
1 as Ỹ 0 i 0.  The parameters, , are estimated via maximum likelihood probit, and the inverse Mill’s ratio is constructed for inclusion in the wage equa- tions, which are then estimated via OLS. The modified wage equations are (4.12) ln Wift  X
ift ft  ft ̂ift  ift ln Wipt  X
ipt pt  pt ̂ipt  ipt where Xi and  are defined in Equation 4.9, is the coefficient on the selectivity term, i is the modified random error term, and all variables (Zijj) are as just defined and ̂ij  , j  ft, pt. (Zijj) The parameter estimates that result from OLS estimation of the pair of equations in (4.12) can be used to decompose the wage differen- tial between disabled and nondisabled workers as follows: (4.13) ln Wft  ln Wpt  X
ft̂ft  X
pt̂pt  ( ̂ft ̂ft  ̂pt ̂pt). The third term on the right-hand side in the parentheses reflects the role differences in selection into the labor market (across part-time status) play in observing differential wages. In order to obtain the selectivity-corrected wage differential, this selectivity term (or differ- ence in selection) is subtracted from the observed wage differential between full-time and part-time workers. This estimation procedure and calculation are performed for each year in the data set to see how different full-time and part-time wages are across disability status and whether that differential has changed over time. If the ADA has forced firms to make environmental changes that also enhance or facilitate a disabled worker’s employment, this full-time/part-time wage differen- tial might fall post-ADA. This would be because what used to be an 90 Hotchkiss extra fixed cost to hiring a disabled worker has been shifted to general access requirements mandated by other provisions of the ADA. Figure 4.4 plots these selectivity-corrected full-time/part-time wage differentials for disabled and nondisabled workers across time. Over the entire period, the part-time wage penalty is declining for both disabled and nondisabled workers. After full implementation of the ADA, however, the part-time penalty for disabled workers is less than the part-time wage penalty for nondisabled workers for all but one year. There are two potential explanations for this phenomenon. First, it may be the case that disabled workers are able to negotiate part-time hours in occupations or jobs that would not typically accommodate part-time work arrangements. Second, the situation may reflect a change in social attitude about what is ‘‘acceptable’’ behavior of a com- mitted worker. Either way, a smaller (and shrinking) part-time wage penalty (although still at roughly 40 percent) is good news for disabled workers who may require a shorter workweek to accommodate their impairments.13 DISTRIBUTION OF WORKERS An indication of how mobile workers with disabilities are com- pared to workers without disabilities (and how this mobility has Figure 4.4 Full-Time/Part-Time Wage Differentials for Disabled and Nondisabled Workers, CPS, 1981–2000 75 70 65 60 55 (%) 50 Nondisabled 45 40 Disabled 35 30 1981 1983 1985 1987 1989 1991 1993 1995 1997 1999 Hours of Work, Distribution, and Representation 91 changed over time) is their distribution over different occupations and industries. Figure 4.5 presents the distribution of disabled and nondis- abled workers among occupations and industries in 2000. Approxi- mately the same proportions of disabled and nondisabled workers are found in the technical support area and in the farming, fishing, and forestry occupational category. Disabled workers, however, are more heavily concentrated in service and laborer occupations, with nondis- abled workers more concentrated in managerial and craft occupations. As we will see later (and as was seen in Chapter 3), this concentration is split along earnings levels, with the occupations in which disabled workers are concentrated being the lower-paying ones. There does not seem to be as wide a disparity in the distribution of workers across industries. However, disabled workers are slightly more concentrated than nondisabled workers in the trade and service industries. Again, these are the lower-paying fields, on average. Figure 4.5 Distribution of Disabled and Nondisabled Workers across Occupations and Industries, CPS, 2000 (A) Occupational category Proportion 35 of workers Nondisabled 30 (%) Disabled 25 20 15 10 5 0 Professional Technical Service Farm, Craft Laborers & managerial support fishing & forestry (B) Industry category Proportion 45 of workers 40 Nondisabled (%) 35 Disabled 30 25 20 15 10 5 0 Agriculture Mining & Manu. Trans., Trade Fin., ins. Service Public constr. comm. & & real admin. utilities estate 92 Hotchkiss The Duncan Index The Duncan Index is useful for comparing the distributions of dif- ferent workers over various occupational and/or industry groups.14 It can be applied to analyze the distribution of disabled workers in rela- tion to that of nondisabled workers across occupations and industries. The Duncan Index (I) is calculated as follows (Duncan and Duncan 1955): 1 2  K (4.14) I  ndj  dj , j1 where K is the total number of occupations or industries and ndj and dj are the proportions of all nondisabled and disabled workers, respec- tively, in occupation or industry j. The index is equal to one-half the sum of the absolute differences between the proportion of nondisabled and disabled persons in each occupation or industry. An index equal to zero means that these groups of workers have identical employment distributions across occupations or industries. An index equal to one corresponds to the extreme situation of complete segregation (no dis- abled and nondisabled workers in the same occupation or industry). Another way to interpret I is as the percentage of disabled (or nondis- abled) workers that would have to change occupations (industries) in order to eliminate the difference in occupational (industry) distribu- tions. Figure 4.6 presents the Duncan Index calculated for each year and plotted along with the Duncan Index for white and nonwhite groups of workers as a frame of reference.15 The first noticeable characteristic of these graphs is the growing disparity between the distributions of dis- abled and nondisabled workers in both occupations and industries. This is in comparison to the declining trend in disparity between non- white and white workers. The contrast is particularly interesting since the average occupational index for both disabled versus nondisabled and nonwhite versus white is 0.13, but with very different end points. In other words, an average of 13 percent of disabled or of nondisabled workers would have to change occupations to equalize their distribu- tions across occupations. Hours of Work, Distribution, and Representation 93 Figure 4.6 Duncan Indices of Dissimilarity across Occupations and Industries, Disabled versus Nondisabled and Nonwhite versus White, CPS, 1981–2000 (A) Occupational dissimilarity 0.19 Nonwhite 0.17 vs. white 0.15 0.13 0.11 Disabled vs. nondisabled 0.09 1981 1983 1985 1987 1989 1991 1993 1995 1997 1999 (a) ADA Enactment (a) (b) (c) (b) ADA Phase I (c) ADA Phase II (B) Industry dissimilarity 0.12 0.10 0.08 Nonwhite vs. white 0.06 0.04 Disabled vs. nondisabled 0.02 1981 1983 1985 1987 1989 1991 1993 1995 1997 1999 (a) ADA Enactment (a) (b) (c) (b) ADA Phase I (c) ADA Phase II A second observation is that it appears as though the ADA may have helped to halt the early growth in occupational disparity, since that series (panel A) seems to have begun a new downward trend since 1992 (the first year of implementation of the ADA). This suggests that disabled workers may be able to take advantage of opportunities not available to them or that they merely may not have utilized pre-ADA. To the extent that matching the distribution of nondisabled workers is an objective, this is a positive outcome for disabled workers. 94 Hotchkiss It is of interest to note that other literature concerned with the distribution of workers may present different goals than equalization of distributions. For example, a dissimilarity in occupational distributions between native and immigrant workers can be considered a positive indicator for an economy, as immigrants fill in occupational gaps left by native workers (see Green 1999). In addition, it may not be clear that the equality of occupational distributions across disability status is desirable. The unique skills and abilities of typical disabled workers may make them fundamentally better suited for occupations not held by the typical nondisabled worker. The growth in disparity among industries does not seem to have followed the pattern of improvement seen among occupations; the distribution of disabled and nondisabled workers has become increasingly disparate, particularly in recent years. Evidence from the SIPP Analysis of the SIPP data reveals a similar pattern of growth in the dissimilarity in distributions across occupations between disabled and nondisabled workers. In addition, the SIPP allows for an evaluation of which type of disabilities results in the least similar distribution. Fig- ure 4.7 provides the Duncan Index calculated to compare the distribu- Figure 4.7 Duncan Indices of Dissimilarity across Occupations and Industries, by Type of Disability, SIPP, 1997 Duncan 0.45 Index of dissimilarity 0.40 0.35 Occupation 0.30 Industry 0.25 0.20 0.15 0.10 0.05 0 Musculoskeletal Internal systems Mental disorder Other Hours of Work, Distribution, and Representation 95 tion of workers with each type of disability with the distribution of nondisabled workers across both occupations and industries. Workers with mental disorders are distributed most differently across both in- dustries and occupations from nondisabled workers: 38 percent of ei- ther workers with mental disabilities or nondisabled workers would have to change occupations to equalize the distributions, compared with only 11 percent of workers with a musculoskeletal disability. Workers with mental disorders are much less likely than nondisabled workers to be employed in professional, technical, and craft occupa- tions, whereas they are much more likely to be employed in service and as laborers. Thirteen percent of workers with mental disorders (or of nondis- abled workers) would have to change industries to equalize the distri- bution, whereas only 6 percent of workers with internal system disabilities would have to switch industries. Workers with mental dis- orders are less likely to be employed in the transportation, communica- tion, and utility industry, and more likely to be employed in the service industry. REPRESENTATION OF WORKERS The equality in the distribution of disabled and nondisabled work- ers is a goal only if the disparity in distribution reflects characteristics of jobs or industries that are desirable. In other words, if nondisabled workers are systematically more concentrated in jobs that are more attractive than jobs in which disabled workers are concentrated, there is a call to make the distributions more equal. A ‘‘share of workers’’ measure can be used to determine whether disabled workers are more or less concentrated in occupations and industries with appealing quali- ties, such as higher wages or growth. The desirability of higher wages is obvious, but growth of an occupation could also be considered at- tractive since it may be indicative of stronger demand for workers, perhaps leading to greater wage growth and occupational advancement. Representation and Growth A high-growth indicator for each occupation (and industry) in the sample is constructed as follows: 96 Hotchkiss 1 if emp t1 k  emp t k / emp t k  (4.15) hgk  K

(1 / K ) emp t1 t j  emp j / emp t j j1 0 otherwise where occupation (or industry) k is defined as high-growth if the per- centage change in employment between year t and t  1 for that occu- pation exceeds the average percentage change in employment in all occupations represented in the sample (K corresponds to the total num- ber of occupations or industries). The probability of being employed in a high-growth occupation (and/or industry) is determined as a func- tion of individual characteristics, including disability status. The sample was split into three time periods, and the employment growth of each occupation and industry represented in the sample was determined by a source external to the data file.16 Table 4.4 contains the growth rates of each occupation and industry represented in the sample. For example, for the 1983–1989 period, service occupations, managerial and professional specialty, and technical, sales, and admin- istrative support are considered ‘‘high growth,’’ since their growth ex- ceeds the average for all occupations. Similarly, for the same period, mining and construction; transportation and public utilities; retail trade; finance, insurance, and real estate; services; and public adminis- tration are all considered high-growth industries. Simple probit models were estimated to determine whether dis- abled workers are more or less likely to be employed in growing occu- pations or industries: (4.16) Ỹi  Xij  i , i  N(0,1) where individual i’s job is in a growing occupation/industry if Ỹi  0. Since Ỹi is unobserved, a binary variable, Yij, is defined as  (4.17) Yij 
1 as Ŷij 0. 0  The set of parameters, , were estimated via maximum likelihood probit. Xi comprises various individual and job characteristics for Hours of Work, Distribution, and Representation 97 Table 4.4 Employment Growth Rates for Industry and Occupational Classifications Growth rate (%) 1983–89 1987–93 1991–98 Occupation Managerial and professional specialty 28.8 16.2 25.9 Technical, sales, and administrative support 15.6 5.6 6.1 Service 12.3 11.7 9.7 Precision production, craft, and repair 12.1 1.0 8.8 Operators, fabricators, and laborers 12.0 0.8 4.6 Farming, forestry, and fishing 7.5 3.6 0.1 Average Growth Rate 12.2 4.7 9.2 Industry Agriculture 9.7 2.9 3.3 Mining and construction 18.8 3.9 16.1 Manufacturing 8.6 5.8 0.7 Transportation and public utilities 15.8 8.2 13.0 Wholesale and retail trade 14.6 7.5 11.4 Finance, insurance, and real estate 22.7 2.7 10.2 Services 23.6 17.7 18.4 Public administration 17.9 10.2 4.1 Average growth rate 14.0 4.2 9.7 SOURCE: Author’s calculations from Jacobs (1998). worker i, including a dummy variable indicating whether the worker is disabled or not. The model was estimated on a sample of workers only; therefore, the results are generalizable to workers only. The mar- ginal effect of disability on being employed in a high-growth occupa- tion or industry was calculated as the marginal benefit for each worker, then averaged over the entire sample. Table 4.5 presents the estimated marginal effects of a work-limiting disability on having a job in a high- growth occupation or industry. In each of the three years analyzed, the probability of a disabled worker being employed in a high-growth occupation was from 2 to 5 percentage points less than the probability of a nondisabled worker being employed in a high-growth occupation. In addition, the marginal (negative) effect was the highest during the post-ADA years, suggest- ing that the ADA has not improved the opportunity of disabled workers to move into high-growth occupations. On the other hand, disabled 98 Hotchkiss Table 4.5 Marginal Effect of Disability on the Probability of Employment in a ‘‘High Growth’’ Occupation or Industry, CPS prob (high growth prob (high growth Year occupation)/disable industry)/disable 1989 0.0299 0.0123 (0.0092) (0.0108) 1993 0.0194 0.0469 (0.0085) (0.0121) 1998 0.0453 0.0261 (0.0140) (0.0129) All three years 0.0381 0.0268 (0.0068) (0.0070) NOTE: Probit estimations included the following regressors, in addition to a disability dummy variable: age; age squared; regional, education, marital, female, and nonwhite dummy variables; and occupation (for the industry probit) and industry (for the occu- pation probit) dummy variables. Standard errors are in parentheses. workers have been more likely to be employed in high-growth indus- tries. Unfortunately, a worker’s industry does not reflect as much on an individual’s job opportunities as one’s occupation does. For exam- ple, for someone with skills suited to a secretarial job, a decline in manufacturing as an industry is not as devastating to the person’s op- portunities as a decline in the administrative support occupation. Of course, occupational representation within an industry, such as the pro- portion of those in the precision production occupation in manufactur- ing industry, could be an important consideration. Representation and High Wages A simple correlation between wages in an industry or occupation and the percentage of workers in that industry or occupation that are disabled was performed to determine whether disabled workers are concentrated in low-paying occupations. The wage decomposition re- sults in Chapter 3 suggest that disabled workers are concentrated in the lower-paying occupations and industries, since occupation and industry regressors contribute positively and significantly to the observed wage differential between disabled and nondisabled workers. Figure 4.8 plots this correlation coefficient for each year for both occupations and industries. First, over the entire time period, the corre- Hours of Work, Distribution, and Representation 99 Figure 4.8 Correlation Coefficients, Concentration of Disabled Workers and Industry/Occupation Wage, CPS, 1981–2000 0 -0.2 Industry -0.4 -0.6 Occupation -0.8 -1.0 1981 1983 1985 1987 1989 1991 1993 1995 1997 1999 (a) (b) (c) (a) ADA Enactment (b) ADA Phase I (c) ADA Phase II NOTE: Correlation coefficients plotted include projections for industry years 1984 and 1999 and occupation year 1983, as these coefficients appeared to be extreme outliers. lation coefficient between concentration of disabled workers in an oc- cupation and the average hourly wage in that occupation is 0.82; the average across industries is 0.55. Disabled workers are more concentrated in the low-wage occupations and industries. Conse- quently, concern about the disparity in distributions across occupations and industries is warranted. While not particularly dramatic, there ap- pears to be some modest improvement (becoming less negative) in these correlation coefficients since 1992 (the first year of ADA imple- mentation). If the ADA did enhance the mobility of disabled workers across occupations and industries, it appears as though these individu- als are moving slightly toward the more desirable jobs, in terms of wage levels. CONCLUSIONS This chapter explored some dimensions of employment not cap- tured merely by employment probabilities or overall compensation. Initially, the incidence of and type of part-time employment were com- 100 Hotchkiss pared across disability status. Then, the distribution of disabled work- ers relative to nondisabled workers across occupations and industries was analyzed. Finally, the representation of disabled workers in what might be considered ‘‘desirable’’ occupations was evaluated. Having a work-limiting disability increased the probability of a worker being observed as employed part-time by 12 percentage points in 2000; this is double the 6-percentage-point impact of a disability on being employed part-time estimated for 1981. The upward trend in the relative occurrence of part-time employment among disabled workers does seem to have gotten a boost during the phase-in period of the ADA. There also seems to have been an ADA-related shift in the type of part-time employment experienced by disabled workers. After 1992, disabled part-time workers went from being less likely voluntar- ily employed part-time to more likely relative to nondisabled part-time workers. The implication is that much of the growth in part-time em- ployment has been voluntary and may actually reflect the required ac- commodation of a worker’s disability. Analysis with the SIPP data indicates that the growth in part-time employment has occurred pri- marily among those with musculoskeletal and mental disabilities, but that only part-time workers with mental disabilities have experienced any significant relative increase in voluntary part-time employment. This may reflect the fact that part-time or flexible work hours may be the best and least costly way to accommodate a worker with a mental disorder. An additional indication that part-time jobs are becoming an ac- ceptable alternative for employing disabled workers is the greater de- cline between 1981 and 2000 in the full-time/part-time wage differential among disabled workers than among nondisabled workers. In 2000, full-time nondisabled workers earned a wage (ceteris paribus) 45 per- cent higher than part-time nondisabled workers, while the full-time/ part-time wage differential between disabled workers was only 39 per- cent. It may also be the case that required accommodation enables disabled workers to negotiate part-time hours on a job that would not normally allow for flexible hours. Either of these reasons translates into an improved labor market experience for disabled workers in the dimension of hours of work. The movement of nondisabled workers away from part-time em- ployment and the movement of disabled workers towards part-time Hours of Work, Distribution, and Representation 101 employment may be one explanation why we observe a growing dis- parity in the distribution of workers across occupations (and indus- tries). It may be the case that some occupations, such as service occupations, are better suited to accommodate workers with disabili- ties. Workers with mental disorders are the most dissimilar in their distribution across both occupations and industries as compared with workers without disabilities. This might be expected considering that workers with mental disorders are the most likely to be voluntarily moving into part-time employment, which may mean that they are re- stricted in the types of occupations or industries open to this type of accommodation. The overall growth in occupational distributional dis- parity, however, is mitigated somewhat post-1992. This may mean that, post-ADA, disabled workers have been able to profit from oppor- tunities not previously available to them or that they merely may not have taken advantage of pre-ADA. To the extent that occupations and industries in which nondisabled workers are concentrated are desirable, this mitigation of dissimilarity since 1992 is a positive outcome for disabled workers. Further analysis found that in each of the three years analyzed, nondisabled workers were more likely to be employed in high-growth occupations and in the highest-paying occupations and industries. Consequently, concern about the disparity in distributions across occupations and industries is warranted, since nondisabled workers seem to be more concentrated in what might be considered desirable occupations and industries than are disabled workers. On the upside, there does appear to be modest movement of disabled workers into more high-paying occupations and industries since 1992. Notes 1. See Averett and Hotchkiss (1995, 1996, and 1997), Hotchkiss (1991), and Farber and Levy (2000). 2. This phenomenon, identified by Averett and Hotchkiss (1996), has been the expe- rience of women in the labor market. 3. For a discussion of labor market segmentation and dual labor market theory and references to this literature, see Kaufman and Hotchkiss (2000, Chapter 6). 4. The bivariate probit model with selection gives rise to the following likelihood function: ln L  EMP1, PT1log 2 
1X1i,
2X2i,   EMP1, PT0ln 2 
1X1i,  
2X2i,   EMP0ln 
1X1i, 102 Hotchkiss where 2 is the bivariate normal cumulative distribution function and is the univariate normal cumulative distribution function. 5. This method of calculating the marginal effect of a change in a dummy variable is referred to as a measure of discrete change and is described in greater detail by Long (1997, pp. 135–138). 6. Table C.7 (column 1) in Appendix C contains the marginal effects used to gener- ate Figure 4.2. 7. This type of pooled, cross-sectional analysis has been applied by many research- ers (for example, Card 1992; Gruber 1994 and 1996; Zveglich and Rodgers 1996; and Hamermesh and Trejo 2000). The technique, however, also has its critics (such as Heckman 1996). The primary criticism of this pooled, cross-sectional approach is that it is impossible to control for unobserved changes in the environ- ment that occurred at the same time as the event of interest. Issues of potential concern in this regard are explored in Chapter 2. 8. Voluntary part-time is defined as (1994–2000) working less than 35 hours per week and did not want to work full-time, and (1981–1993) reason for working less than 35 hours per week coded as 07–15 (see Stratton 1994 for justification). Category reasons 07–15 are holiday, labor dispute, bad weather, own illness, on vacation, too busy with school

or house, did not want to work full-time, full-time work week is less than 35 hours, or other. 9. Table C.7 (column 2) in Appendix C contains the marginal effects used to gener- ate Figure 4.3. 10. These results are consistent with those of Schur (2002), who finds that part-time and contingent work grows among the disabled during tight labor markets (where demand is strong relative to supply), which would be expected if these arrange- ments are voluntary. 11. The musculoskeletal category includes disabilities of the special senses (e.g., hearing, seeing). 12. See Magill (1997) for a detailed discussion about how flexible and part-time work schedules can often serve as low- or no-cost solutions to the accommodating problem. However, indirect costs, such as having to hire additional workers to cover lost hours of a disabled part-time worker, should not be ignored. 13. The SIPP data showed the same differential decline for both disabled and nondis- abled workers as seen in the CPS data, but essentially inconsequential differences across disability status in any given year. 14. An additional distribution index, the IP Index (see Watts 1992), was also evalu- ated, and the conclusions were the same. While there is a direct mathematical relationship between these two indices, the IP index reflects the percentage of workers in the labor market that would have to switch occupation or industry in order to equalize the distribution of disabled and nondisabled workers across occupations or industries, while maintaining the original occupational structure. Since the conclusions were the same regardless of the index employed, the more familiar Duncan Index is detailed here. 15. The actual indices plotted in Figure 4.6 are contained in Table C.8 in Appendix C. The indices calculated for the distribution across industries are consistent with Hours of Work, Distribution, and Representation 103 what Yelin and Cisternas (1996) calculated using the National Health Interview Survey; their data indicate that the relative distributions of disabled and nondis- abled workers have been fairly steady as far back as 1970. The occupational categories in their data were not comparable to those in the CPS, however. 16. Employment growth was calculated as the percentage change in the number of workers in an occupation from one time period to the next. This Page Intentionally Left Blank 5 Separation, Unemployment, and Job Search Separation from one’s job is an important dimension of the experi- ence of a worker. If separations are dominated by involuntary actions, such as a layoff or being fired, the worker’s experience is obviously diminished. Voluntary separation, however, may or may not be an indicator of a positive situation. On the one hand, voluntary separation (quitting) may indicate that workers are able to respond to better job opportunities through labor market mobility. On the other hand, exces- sive voluntary separations may reflect instability among that group of workers. This may be of particular importance for disabled workers who may need to voluntarily quit jobs for health reasons. The first analysis in this chapter considers a group of labor force participants who have experienced a recent job separation and evaluates the deter- minants, including disability status, of the type of separation.1 Unemployment is another important dimension of the labor market experience. In a given month in 2000, an average of 3.3 million people flowed into unemployment.2 Between the ages of 18 and 27, individu- als average 4.4 unemployment spells and spend an average of 31 weeks unemployed (Veum and Weiss 1993). In addition, Figure 5.1 shows that, in 1999, workers in the CPS data used here spent from roughly one to two weeks on average looking for work.3 Figure 5.1 also shows that while the disabled clearly spend more of their time in a given year looking for work, the movement of weeks spent in this activity essentially mirrors the trend for the nondisabled. The next sections will look more closely at this time spent looking for work. Job separation is only one reason why a worker might be unem- ployed. Workers entering the labor market for the first time and rejoin- ing the labor market after an absence are also considered unemployed until they find a job. Examining the composition of the unemployed over time can tell us something about flows into and out of the labor market. The second analysis of this chapter will explore the probabili- ties of unemployment categories as a function of disability status. 105 106 Hotchkiss Figure 5.1 Weeks Spent Looking for Work, by Disability Status, CPS, 1980–1999 Number 5 of weeks Disabled looking 4 3 2 Nondisabled 1 0 1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 (a) ADA Enactment (a) (b) (c) (b) ADA Phase I (c) ADA Phase II The worker’s situation while unemployed, namely the job search experience, will also be explored in this chapter. For the same reasons argued in Chapter 2, one would expect that disabled workers would have a more difficult time finding a job that suits their skills, and would thus have longer spells of search duration, ceteris paribus. In addition, if disabled workers are subjected to hiring discrimination, search dura- tion would also increase. The third analysis in this chapter will evalu- ate the search spells of disabled and nondisabled workers and determine whether differential search strategies are more successful for one group or the other. SEPARATION If disabled workers have a more difficult time finding employment or employers that will accommodate their disabilities, they may experi- ence greater voluntary turnover as they continue to search for the job that will best match their skills. On the other hand, the fear of losing one’s health benefits (‘‘job-lock’’) may be more severe for disabled workers, leading to fewer voluntary separations relative to nondisabled workers (see Kapur 1998; Buchmueller and Valletta 1996). The impact of the ADA on voluntary separations among the disabled is ambiguous. Voluntary separations may increase as more opportunities become Separation, Unemployment, and Job Search 107 available to disabled workers, but may also decrease as employers make disability accommodations, which have been shown to increase tenure and reduce voluntary turnover among disabled workers (Burk- hauser, Butler, and Kim 1995). If disabled workers suffer from discrimination, or overall have less labor market experience or tenure with their employers, they might suffer more frequent involuntary separations through layoffs. Based purely on anecdotal evidence, Yelin (1991) concludes that ‘‘persons with disabilities, like those from minority races, constitute a contingent labor force, suffering displacement first [as an industry declines]’’ (p. 135). However, if disabled workers are a more selected group in the sense that they are less substitutable with other inputs (perhaps the accommodations employers have made for them enhance their produc- tivity beyond that of nondisabled workers, ceteris paribus), they will be less likely to be laid off as marginal workers. In other words, the discrimination or marginalization might be taking place at the hiring stage, rather than at the separation stage of the relationship. The pas- sage of the ADA might decrease involuntary separations among dis- abled workers if employers are fearful of accusations of discrimination. Among workers who have separated from their jobs, the CPS con- tains information about why a separation occurred. Figure 5.2 presents the percentage of disabled and nondisabled individuals who have expe- rienced a recent job separation and the reason for that separation. The series are split into pre- and post-1994, since the universe for the job separation question changed at that time.4 Overall, disabled workers have a greater incidence of voluntary separations, and nondisabled workers have a greater incidence of involuntary separations. The break in 1994 makes it difficult to draw any conclusions regarding trends. In order to look more closely at any possible trends, and to control for individual characteristics, a multinomial logit estimation was per- formed.5 A multinomial logit specification allows us to specify multi- ple possible outcomes (e.g., separation types) as a function of a variety of observed characteristics and unobservable factors, recognizing that as one’s probability of having separated voluntarily increases, the prob- ability of having separated involuntarily necessarily decreases. Job separations have been divided into three categories: 1) volun- tary separation, 2) involuntary separation, and 3) ‘‘other’’ separation.6 108 Hotchkiss Figure 5.2 Percentage of Job Separations by Type and Disability Status, CPS, 1981–2000 (A) Voluntary separations 85 80 Disabled 75 70 65 Nondisabled 60 55 50 1981 1983 1985 1987 1989 1991 1993 1995 1997 1999 (B) Involuntary separations 40 35 30 25 Nondisabled 20 15 10 Disabled 5 0 1981 1983 1985 1987 1989 1991 1993 1995 1997 1999 Separation, Unemployment, and Job Search 109 The separation is modeled as a multinomial logit, where the probability of observing a job separation (S) of type j for person i is equal to (5.1) Pj  P(Si  j), j  1,2,3. Pj (5.2)  F(
jX) for j  1,2, Pj  P3 where F() is a cumulative distribution function, X corresponds to char- acteristics expected to influence the type of separations, and j dictates how those characteristics affect separation j. This means P
j F(jX) (5.3)   G(
j  1,2. P jX) for 3 1  F(
jX) Because of the rules of summation: G(
jX) (5.4) P3  1/ 2 1  G(
jX) and Pj  2 . j1 1  G(
jX) j1 If we let (5.5) G(
jX)  exp (j
X) and Yij gory j 
1 if person i falls in separation cate , 0 otherwise the log likelihood function (ln L) can be written as (5.6) In L  3 3 Yij ln Pij, i1 j1 exp(X
i j) 1 where Pij  and Pi3  2 . 1  2 exp(X
i k) 1  exp(X
i k) k1 k1 The multinomial logit gives three sets of parameter estimates, each set describing the probability of one of the separation types. Each per- 110 Hotchkiss son has a probability of every type of separation, and those three proba- bilities sum to one (since the sample contains only those who experienced a separation). Figure 5.3 summarizes the marginal effects of having a work-limiting disability on the probability that the separa- tion was involuntary or voluntary for years 1981–2000.7 Over the period, the probability of a separation being voluntary is, on average, 12 percentage points higher for workers with disabilities, relative to workers without disabilities. This result lends support for the theory that disabled workers may have more difficulty finding the right ‘‘match,’’ and are therefore more likely to quit in search of a better accommodation for their disability. It could also point to the higher frequency of health-related quits among the disabled. In addi- tion, with the exception of the most recent two years, the impact of disability on voluntary turnover seems to be declining since the phase- in of the ADA. This suggests that mandated accommodations relieve the disabled worker of the burden of changing jobs in order to search for a better fit; the worker’s current job (or employer) does the chang- ing to better suit the disabled worker. It could also mean that employ- ers are accommodating potential health cycles that in the past would Figure 5.3 Marginal Effect of Being Disabled on the Probability of Separation Type, CPS, 1981–2000 0.20 0.15 0.10 0.05 Impact on probability of voluntary separation 0 -0.05 -0.10 Impact on probability -0.15 of involuntary separation -0.20 1981 1983 1985 1987 1989 1991 1993 1995 1997 1999 (a) ADA Enactment (a) (b) (c) (b) ADA Phase I (c) ADA Phase II Separation, Unemployment, and Job Search 111 have necessitated a job separation. This result is marginally consistent with the research of Baldwin and Schumacher (1999), who find, using 1990 data from the SIPP, that disabled workers are more likely (but insignificantly so) to voluntarily separate from their employers, relative to nondisabled workers. The probability of a separation being involuntary is, on average over the time period, 11 percentage points lower for workers with dis- abilities. This does not support the notion that disabled workers are considered ‘‘marginal.’’ Since this negative impact of being disabled on involuntary separations is in evidence well before the ADA, it may also help alleviate employers’ fears that the ADA makes it more ‘‘dif- ficult’’ to dismiss disabled workers, on average. These results are not consistent with these

of Baldwin and Schumacher (1999), who find that disabled workers are slightly more likely to experience an involuntary separation than nondisabled workers. Baldwin and Schumacher (1999) explore overall job turnover, where nonseparation plays a large role in the outcomes of workers analyzed; their results are primarily driven by the fact that disabled workers experience more separations overall. The analysis here compares only types of separation and does not consider the nonseparation outcome. This approach is more relevant when con- sidering the separation experience of workers, rather than the question of turnover. In other words, the question answered here is, ‘‘Among those who have separated, what is the most common reason?’’ Further evidence that disabled workers are not marginalized is the experience during the recession of the early 1990s. Separation during the early nineties for disabled workers was even less likely to be the result of involuntary action, relative to nondisabled workers, compared with the years prior to and since the recession. UNEMPLOYMENT Examining the composition of the unemployed over time can tell us something about flows into and out of the labor market. For exam- ple, if most of the unemployed are new entrants or reentrants, these individuals would constitute an inflow of workers into the labor market. In 2000, job losers (fired or laid-off) made up the largest category of all the unemployed (44 percent), and re-entrants represented the next 112 Hotchkiss largest category (36 percent of the unemployed).8 As far as disabled workers are concerned, a policy such as the ADA is expected to de- crease the cost of entering the labor market (with improved accommo- dation and fewer barriers to employment), thus potentially increasing the incidence of new entrants and reentrants to the labor market, rela- tive to these categories for nondisabled workers. Another multinomial logit was estimated to evaluate the determi- nants of unemployment categories, with the type of unemployment di- vided into four classifications: 1) job loser (fired or laid-off), 2) job leaver (quit), 3) reentrant to the labor force, and 4) new entrant to the labor force. The first two categories were considered in detail in the previous section. The focus of the analysis in this section is on the relative probabilities that unemployment spells for disabled workers are of the new entrant or reentrant variety, and on whether the composi- tion of the unemployed was altered by the ADA. The structure of the multinomial logit estimated for this analysis is the same as described by Equations 5.1 through 5.5, except that there are four categories in- stead of just three. The multinomial logit provides four sets of parameter estimates. Each set describes the probability of one of the unemployment types; every person has a probability of each type of unemployment, and those four probabilities sum to one (since the sample contains only those who are unemployed). The CPS started categorizing types of unemployment in 1989, so that is the first year of analysis. The four panels in Figure 5.4 summarize the predicted probabilities of the job- less experiencing each of the categories of unemployment.9 The actual probabilities have been smoothed using a third-order polynomial.10 The smoothing process amounts to taking a plotted series and drawing a smooth line through the middle of the plotted points. The effect is to highlight any trends that are more difficult to decipher from the raw data. The higher the order of the polynomial (e.g., third- versus first- order), the more changes in direction will be captured (the higher the order, the less restrictive is the smoothing). The probabilities of being a job loser (fired or laid-off) and a job leaver (quit) are consistent with the results reported for separations in the previous section (see Figures 5.2 and 5.3). The disabled are less likely to be job losers than the nondisabled (panel A) and (typically) more likely to be job leavers (panel B). The disabled unemployed are Separation, Unemployment, and Job Search 113 Figure 5.4 Probability of Different Types of Unemployment by Disability Status, CPS, 1989–2000 (A) Unemployed is a job loser 0.75 Nondisabled 0.60 0.45 Disabled 0.30 0.15 0 1989 1991 1993 1995 1997 1999 (B) Unemployed is a job leaver 0.24 0.20 Disabled 0.16 0.12 Nondisabled 0.08 0.04 0 1989 1991 1993 1995 1997 1999 also predicted to be more likely to be new entrants and reentrants into the labor market (panels C and D) than the nondisabled. This indicates that the disabled move into and out of the labor market more than do nondisabled workers. This is not good news for the disabled, as shifts between jobs (even with intervening unemployment) typically result in better subsequent outcomes than movements out of and back into the labor market (Horvath and Shack 1986). 114 Hotchkiss Figure 5.4 (continued) Probability of Different Types of Unemployment by Disability Status, CPS, 1989–2000 (C) Unemployed is a new entrant 0.025 0.020 0.015 Disabled 0.010 0.005 Nondisabled 0 1989 1991 1993 1995 1997 1999 (D) Unemployed is a reentrant 0.7 0.6 0.5 Disabled 0.4 0.3 0.2 Nondisabled 0.1 0 1989 1991 1993 1995 1997 1999 Separation, Unemployment, and Job Search 115 There does not seem to be any noticeable effect of the passage and phase-in of the ADA on the probability of unemployed disabled work- ers being in one category of unemployment or another. There is an upswing in the probability of being a reentrant beginning about 1992, but that change is shared by the nondisabled, which likely means it was a general labor market response (by all workers) of recovery from the early 1990s recession; workers discouraged by the recession began to reenter the labor market as the economy recovered. The fairly steady decline in the probability of being a job loser among disabled workers since 1993 deviates somewhat from that observed for nondisabled workers, perhaps signaling that additional accommodations have made disabled workers even that much less marginalized than before. It could also be signaling employers’ fears of being accused of inappro- priately dismissing disabled workers. Perhaps, because of these fears, employers are even more scrutinizing when hiring a disabled worker, improving the chances of a good fit. JOB SEARCH The theory proposed so far as to why the disabled have a lower probability (among those who separate) of being a job loser, or experi- encing an involuntary separation, is that employers are more careful in their hiring of disabled workers. Employers may feel that hiring a disabled worker is more risky, or they may fear the consequences if they would have to dismiss the worker. The greater ‘‘care’’ in hiring a disabled worker should show up in longer search spells. Of course, the observation of longer search spells is also consistent with a theory of discrimination against disabled workers, but it is necessary to provide support for the preceding theory. Figure 5.5 presents the average dif- ference between the expected search spells of disabled workers and of nondisabled workers.11 While exhibiting some degree of variation from year to year, the average length differential ranges from a high of 24 weeks in 1987 to a low of 0.66 weeks in 1998. The median spell length differential (not controlling for any individual characteristics) for the entire time period is approximately 14 weeks.12 The expected impact of the ADA on search spell length of disabled workers is ambiguous. The legislation may reduce search spells by 116 Hotchkiss Figure 5.5 Average Search Spell Length Differential between Unemployed Disabled and Nondisabled Searchers, CPS, 1981–2000 Number 25 of weeks 20 15 10 5 0 1981 1983 1985 1987 1989 1991 1993 1995 1997 1999 (a) (b) (c) (a) ADA Enactment (b) ADA Phase I (c) ADA Phase II making overt hiring discrimination more difficult. It may also shorten spells merely by raising the awareness of employers to the capabilities of workers with disabilities. On the other hand, it may lengthen search spells if the fear of dismissing disabled workers is so great that employ- ers increase their scrutiny of such individuals before hiring them. It appears from Figure 5.5 that the ADA did not have an impact on the average search spell length differential. Fitting a trend line through the data points in Figure 5.5 results in a zero slope; while there is a wide variation in the average from year to year, and while the difference does not exhibit any trend, it is positive throughout the time period. Search Duration Estimation The difficulty of estimating job search spells using CPS data is notorious (Kiefer, Lundberg, and Neumann 1985). Individuals who are currently searching for a job are asked how long they have been searching, so that everyone in the sample is in the middle of a censored spell. Akerlof and Main’s (1981) approach to using CPS data is to double the observed censored search spells and then treat them as com- pleted. This results in an accurate representation of completed search spells under the assumption of a steady state.13 Under this assumption, Separation, Unemployment, and Job Search 117 the impact of various demographic characteristics on the length of a search spell can be determined. The demographic of particular inter- est, of course, is whether someone has a work-limiting disability. The accelerated failure time model that will be described allows for the estimation of these spell lengths, taking into account how long some- one has already been searching. This is of interest if, for example, the longer someone has been searching the harder it is for him or her to escape the search (by finding a job). If a person has a completed search spell length, t, the contribution to likelihood is f(t), where f(.) is the probability density function of the random search duration, T. In order to describe the variation in T conditional on a set of explanatory variables, X, a specific distribution is specified for T as a function of a set of parameters, . If T0 is a random time duration sampled from the baseline distribution for an individual whose covariates are all zero, then for nonzero covariates, X, the event time will be T(X)  exp(X
)T0 (see Kalbfleisch and Pren- tice 1980; Kiefer 1988). This model specification allows writing the log duration as a linear function of the covariates, lnt  X
  . Assuming T is distributed as a Weibull, the following likelihood func- tion results: 1 (5.7) L(ti,Xi)   g((lnti  X
i)/), i  where g(.) is the probability density function of the transformed search duration; in this case, g(.) takes on the form of an extreme value distri- bution. Regressors for the duration analysis include age; age squared; nonlabor income; female, single, nonwhite, education, and regional dummies; dummy variables for availability for employment, whether searcher wanted a full-time job, and whether searcher worked last year; and dummy variables for search methods, disability status, and search methods interacted with disability status. The specification of a duration model, as opposed to merely esti- mating the relationship by OLS, for example, allows for the likelihood that the chance of a search spell ending in time t is related to how large t is (i.e., the probability that someone finds a jobs and stops searching in t depends on how long the individual has already been searching). 118 Hotchkiss This relationship between the chances of finding a job and how long a worker has been searching is referred to as duration dependence. Figure 5.6 plots the predicted expected search duration for the total sample, by disability status, for each year.14 Over this entire time pe- riod, on average, disabled individuals could expect to be looking for a job 3.2 weeks longer than nondisabled workers with identical demo- graphic characteristics and search strategies. The only trends that ap- pear in this graph correspond to the general conditions of the labor market; predicted expected duration for both disabled and nondisabled searchers moves together. There is an obvious rise in search length beginning in the early nineties (corresponding to the recession of the period), and duration begins to fall again in the mid 1990s. Controlling for demographic and search strategy reduces the aver- age marginal effect of being disabled on expected duration over this time period from approximately 14 weeks to approximately 3 weeks. The implication is that the majority of the observed search duration differential is explained by demographic

and search strategy differ- ences. This suggests that discrimination in hiring may not be of great concern, but it does not rule out that employers are more ‘‘careful’’ in screening disabled workers. In other words, it is possible that the characteristics of disabled and nondisabled workers are being valued equally, but that employers go to greater length to discern the fit of a Figure 5.6 Predicted Expected Search Duration by Disability Status, CPS, 1981–2000 Predicted 60 duration (weeks) 50 40 Disabled 30 Nondisabled 20 10 0 1981 1983 1985 1987 1989 1991 1993 1995 1997 1999 Separation, Unemployment, and Job Search 119 disabled worker with a particular job. This behavior would be consis- tent with a risk-averse employer who has more difficulty determining the productivity of disabled workers than of nondisabled workers from observable traits. This could be due to lack of experience on the part of the employer or because there is greater variance in the productivity among disabled workers for any given set of observable characteristics. Effectiveness of Search Methods One important quality that an individual brings to the job search is the type of strategies used. It is of interest to determine whether dis- abled job seekers consistently use different search strategies and whether those methods are consistently more or less effective for dis- abled searchers than for the nondisabled. Several researchers have found that certain search methods are more effective in finding em- ployment than others (Bortnick and Ports 1992; Thomas 1991), so if disabled individuals are systematically relying on less effective meth- ods, or systematically have less effective methods available, this leads to an obvious remedy. Figure 5.7 details the average percentage of disabled and nondisabled job seekers using each of six search strategies (‘‘did nothing’’ is excluded as a strategy from the illustration). Figure 5.7 Percentage of Disabled and Nondisabled Searchers Using Each Search Strategy, Averaged over the Period 1981–2000, CPS 80 70 Disabled 60 Nonisabled 50 40 30 20 10 0 Public Private Checked Other Checked with Checked agency agency with friends employer ads 120 Hotchkiss Disabled searchers are marginally more likely to check with friends, public agencies, and advertisements. None of the differences across strategies, however, is significantly different from zero. The implication, then, is that disabled and nondisabled job seekers rely, in the same percentages, on the same search strategies. In addition, there is no trend over time in the proportion of people using each method across disabled or nondisabled searchers. Furthermore, out of the 100 possible coefficients (across years and across search strategies) included to control for potential differential impacts of search strategies on length of search for disabled workers, only 17 were significantly different from zero and not in any consistent way. In other words, each strategy never helped or hindered disabled searchers, relative to nondisabled searchers, in any systematic way. The implication is that personal demographics were the primary deter- minants of search outcome, and that merely changing strategies will not improve those outcomes for disabled workers. CONCLUSIONS The ‘‘not working’’ side of the labor market can be important in shaping the overall experience of a worker. This chapter compares the separation, unemployment, and job search experiences of disabled and nondisabled workers over time. The lower probability for disabled workers that a separation is involuntary is clearly good news for dis- abled workers. It implies that disabled individuals are not ‘‘marginal’’ workers, as some have theorized. The higher probability for this group that a separation is voluntary indicates, at a minimum, that disabled workers do enjoy some labor market mobility with which they can seek out the most accommodating employment setting. It may also show that disabled workers have a more difficult time finding a good employ- ment match, which is not entirely unexpected, or that health conditions necessitate more frequent voluntary movements in and out of jobs. There is no indication from the analysis performed here that the ADA has had much impact on the probability that a disabled worker’s sepa- ration is either voluntary or involuntary. Disabled workers are more likely to be reentrants and new entrants into unemployment than nondisabled workers. This is unfortunate be- Separation, Unemployment, and Job Search 121 cause leaving and reentering the labor market typically results in less advantageous outcomes than movements between jobs.15 There are not a lot of policy options, sadly, that might alter this pattern, since much of the movement into and out of the labor market may be dictated by the nature of the disabled worker’s mental or physical condition. On average over the time period, an unemployed disabled worker searches three weeks longer than a similar nondisabled worker before locating a job. This result, in combination with the finding that dis- abled workers are less likely to separate from his/her job for involun- tary reasons, suggests that employers are being more ‘‘careful’’ in their hiring of disabled workers. There is a fine line between discrimination and being discriminating. However, it was also found that differences in individual characteristics between disabled and nondisabled workers explain most of the difference in their search spell lengths. Therefore, it is possible that the characteristics of disabled and nondisabled work- ers are being valued equally (the typical measure for discrimination), but that employers go to greater lengths to discern the fit of a disabled worker’s endowments with a particular job. This behavior on the part of the employer would result both in longer search spells for the dis- abled and in a lower probability that a separation is involuntary. These results, taken together, suggest that policies that assist em- ployers and disabled workers in finding each other would go a long way to improving search outcomes. Active job placement and objective screening of skills might be useful and serve the needs of the ‘‘discern- ing’’ employer. Notes 1. While it may seem natural to first focus on the incidence of separation, the em- ployment analysis contained in Chapter 2 essentially contains the inverse of an analysis of incidence. The CPS does not ask separation questions of those who are currently working, so that an evaluation of separation incidence would amount to an analysis of unemployment incidence (the inverse of what was evaluated in Chapter 2). The results from Chapter 2 indicate that the disabled are more likely at any given time to have separated from their job, and that the difference in probability across disability status is unchanged over time. 2. Unpublished data from the U.S. Bureau of Labor Statistics; obtained November 2001. 3. The sample for whom these statistics were plotted consisted of those who spent at least some (no matter how little) time working in the previous year. This gives 122 Hotchkiss us a better picture of the looking-for-work activity of those with at least some marginal attachment to the labor market. 4. The data used to generate Figure 5.2 and details of the questionnaire change are found in Table C.9 in Appendix C. 5. Multinomial logits have come under frequent criticism because of the assumption of independence of irrelevant alternatives (IIA) that is implied by the logit speci- fication. Alternative specifications that retain the desired probability structure (i.e., multinomial probit) are riddled with their own problems and not considered here to add value greater than the cost that would be imposed. It has been pointed out that under the framework of what is called a ‘‘universal’’ logit, the estimation procedure can be applied, but the utility interpretation of the structural estimates is lost. In addition, the more regressors included to describe the multiple out- comes, the less bothersome is the assumption of IIA. For these reasons, the logit structure is retained. For further discussion on these points, see Ben-Akiva and Lerman (1985, section 5.2) and Moffitt (1999, pp. 1382–1387). 6. Voluntary separations included the following reasons: personal, family, school; personal/family (includes pregnancy); return to school; health; retirement/old age; and unsatisfactory work arrangements (hours/pay/etc.). Involuntary separation included the following reasons: seasonal job completed, temporary seasonal or intermittent job completed, slack work/business, and temporary nonseasonal job completed. The analysis in this chapter is based only on those who experienced a separation either in the past five years (1981–1993) or in the previous year (1994–1999); the results, therefore, are not generalizable to the entire labor force. 7. The predicted probabilities are found in Table C.10 in Appendix C. Whereas the impact of the questionnaire change was obvious in the raw data (see Figure 5.2), the predicted probabilities are comparable across years in the cross section, with the primary impact being larger sample sizes prior to 1994. 8. U.S. Department of Labor, Employment and Earnings (January 2001, p. 203). To be counted as unemployed, the respondent must have undertaken some action within the past four weeks to find a job. 9. The predicted probabilities from the multinomial logit analysis used to generate these figures are found in Table C.11 in Appendix C. The figures plot a ‘‘smoothed’’ version of the series of numbers reported in Table C.11. 10. This simply involves calculating a least squares third-order polynomial fit through the data points. 11. Following the suggestion of Akerlof and Main (1981) to compensate for the defi- ciencies of the job search data collected by the CPS, expected search duration is calculated as twice the observed duration at a given point in time. This represen- tation of expected duration relies on the assumption of a steady state, which may be difficult to support for a statistic across time (i.e., business cycles; Sider 1985); however, this may be less critical when making cross-sectional comparisons, as is done here. It is also assumed that this assumption has the same implication across disability status. 12. The numbers used to prepare Figure 5.5 can be found in Table C.12 (column 1) in Appendix C. Separation, Unemployment, and Job Search 123 13. See note 11. 14. Predicted expected duration was calculated using the coefficient estimates re- sulting from the accelerated failure time model. Where search duration is as- sumed to be distributed as a Weibull, expected search duration is calculated as Etiti  0;Xi,,  exp(X
i )(1  ), where  is the gamma distribution function. The characteristics (other than disability) for which the predicted ex- pected durations were calculated (X) are the means corresponding to the entire sample. See Long (1997). The numbers used to prepare Figure 5.6 can be found in Table C.12 (columns 2 and 3) in Appendix C. 15. There is even some evidence that disabled workers experience more discrimina- tion between employers (from switching jobs) than they experience with any given employer (on their current job). See O’Hara (2000). The implication is that any job separation may worsen a disabled person’s labor market outcome. This Page Intentionally Left Blank 6 State versus Federal Legislation Whenever major federal legislation to regulate the functioning of a market is enacted, a key question raised is whether that law is redun- dant or whether it has the potential of actually altering the functioning of the market (i.e., is ‘‘binding’’). By the time the ADA was passed, all states had some form of legislation addressing discrimination against the disabled (see Table 6.1).1 Thus, the environment in which the ADA was approved was arguably already a post-ADA one. One might suggest that the ADA was superfluous; the states were already addressing the problem of discrimination against the disabled and there was no need for federal legislation. This situation may be an explana- tion for finding no or very little labor market impact attributable to the ADA in previous chapters. In other words, it may be the case that the state legislation ‘‘crowded out’’ any potential impact of the ADA. On the other hand, legislation at the state level may have served as a state- ment of ethical beliefs already integrated into the economy. To determine whether state-level protective legislation ‘‘crowded out’’ or had a differential impact on the experience of disabled workers than the federal ADA, employment, wage, and hours analyses mirror- ing those contained in Chapters 2, 3, and 4 are repeated here, but only for a subsample of states that enacted disability legislation between 1981 (the beginning of available data) and 1991 (the last year before implementation of the federal legislation). The employment impact is determined by estimation of a pooled, cross-sectional bivariate probit

 posts  is where lnrwageis is the natural log hourly real (1982–19841) wage of worker i in state s, Xi is a set of covariates for each person (demographic and job characteristics), disablei is equal to 1 if person i has a work-limiting disability, and posts is equal to 1 if person is observed in state s post-legislation for that state. 134 Hotchkiss The affected group (the disabled) is controlled for by a dummy variable indicating whether the individual has a work-limiting disability, and the post-legislation time period in each state is controlled for by a dummy variable indicating whether disability legislation was in place yet or not. The coefficient of interest (3) measures the change in real wages, relative to nondisabled workers, after passage of disability legislation, relative to before legislation was in place. Xi includes indi- vidual demographic and job characteristics. Wages are observed for workers only, and because the characteris- tics of workers may be changing over time in unobservable ways, it is important to control for any potential unobserved self-selection into the labor market. Consequently, Equation 6.2 is modified by simply adding the standard inverse-Mills ratio obtained from the first-stage probit estimation of a labor force participation/employment equation. This standard Heckman (1979) two-step procedure for controlling for self-selection is presented in greater detail in Chapter 3. The parame- ters of the model are identified through some regressors in the first- stage probit estimation that are not in the wage regression; these re- gressors include nonlabor income and an indicator of whether the per- son worked last year or not. Since the purpose of this two-stage approach is merely to obtain unbiased estimates of the coefficients in the wage equation, interpretation of those coefficients is not changed by controlling for selection. Table 6.3 contains the estimation results from the log wage equation estimation. As with the employment analysis, the results in Table 6.3 mirror those at the national level, as reported in Table 3.1 in Chapter 3; most coefficients on the control variables are of the same sign, and they are roughly of the same magnitude. Two exceptions are the coefficient on the advanced degree dummy variable, which is about one third the size of that estimated at the national level, and the coefficient on the selec- tion term. In Chapter 3, there was evidence of positive self-selection, meaning that those entering the labor market could expect to earn higher wages than the population as a whole. In the state-level analy- sis, the coefficient on the selection term is negative, indicating that those entering the labor market typically will earn less than the popula- tion as a whole. Since the purpose of controlling for selection is to obtain consistent estimates of the other coefficients in the wage equa- tion, this difference in sign between the national and state-level analy- State versus Federal Legislation 135 Table 6.3 Log Real Wage OLS Estimation with Selection, CPS, 1981–1991 Labor force Regressor participation equation Intercept 2.5415*** (0.0509) Age (00) 0.0415*** (0.0051) Age Squared (0000) 0.0004 (0.0140) Female  1 0.2266*** (0.0048) Nonwhite  1 0.0416*** (0.0042) High school grad  1 0.1230*** (0.0046) Some college  1 0.2072*** (0.0055) College grad  1 0.3863*** (0.0067) Advanced degree  1 0.1071*** (0.0032) Midwest  1 0.3072*** (0.0010) South  1 0.1441*** (0.0000) West  1 0.1125*** (0.0001) Hours worked per week 0.0017 (0.0142) Union  1 0.1496*** (0.0008) Single household  1 0.0659*** (0.0033) State unemployment rate (0) 0.0109*** (0.0003) Real gross state product (000000) 0.0081 (0.0052) Log population 0.0953*** (0.0047) disable  1 0.0802*** (0.0033) 136 Hotchkiss Table 6.3 (continued) Labor force participation Regressor equation post legislation  1 0.0307*** (0.0039) disable  post  1 0.0452** (0.0176) (selection term) 0.0811*** (0.0057) R2 0.46 F statistic 2,354 Number of observations 95,604 NOTE: States included in the analysis are Alaska, Arizona, Delaware, Idaho, Massa- chusetts, North Carolina, North Dakota, South Dakota, Texas, and Wyoming. The wage regression also included seven industry and five occupational dummy variables, and a government dummy variable not reported here. The first-stage probit estimation included age; age squared; gender, race, education, and disability dummies; and the state unemployment rate. Nonlabor income and an indicator of working last year were included as identifying regressors. All estimated coefficients in the first-stage estimation were significantly different from zero at the 99 percent confidence level. Standard errors are in parentheses. ***  significant at the 99 percent confidence level. **  significant at the 95 percent confidence level. ses is not a concern here but may be worth exploring elsewhere from a behavioral perspective. The coefficient on the interaction term (disable  post) is 0.0452, indicating that disabled workers experienced a 4.5 percent decline in wages, relative to nondisabled workers, post-disability legis- lation, relative to pre-disability legislation. This is roughly of the same magnitude as the 3 percent decline in wages experienced post-ADA implementation (see Table 3.1). The implication is that we see the same relative decline in wages of disabled versus nondisabled persons at the state and national levels. This may mean that the measured impact of the federal ADA on relative wages is muted, given that some adjustment to disability legislation had already taken place as a result of state laws. In addition, this lower wage of disabled workers relative to that of nondisabled workers may not be the consequence of adjust- ment cost, but, rather, reflect the overall negative impact on the dis- State versus Federal Legislation 137 abled, not just those covered by the legislation; this was the result discovered in Chapter 3. IMPACT ON HOURS The increase in part-time employment post-ADA found in Chapter 4 is worth exploring at the state level, as well. It is of interest since flexibility in hours may serve as an important mechanism through which employers can accommodate many types of disabilities. If this is the case, then requirements to accommodate workers’ disabilities at the state level should result in similar adjustments as seen post-ADA. Figure 6.3 plots the average proportion across states of disabled and nondisabled workers that are employed part-time. The reference verti- cal line corresponds to the time when legislation was in place in each state. There appears to be an increase in the proportion of disabled workers that are employed part-time, as well as a modest divergence in the two series. The pooled, cross-sectional analysis of Chapter 4 is repeated here in order to determine whether there is any significant growth in part- time employment among disabled workers, post-legislation, relative to Figure 6.3 Proportion of Disabled and Nondisabled Workers That Are Employed Part-Time, CPS, 1981–1991 45 40 35 Disabled 30 (%) 25 20 15 Nondisabled 10 5 0 -9 -8 -7 -6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6 7 8 Event period 138 Hotchkiss nondisabled workers. The model estimated is the bivariate probit with selection: (6.3) empis 
1  
1X1i  1 disablei  1 posts  1 disablei  posts  1is ptis 
2  
2X2i  2 disablei  2 posts  2 disablei  posts  2is. empis  1 if person i in state s is employed, 0 otherwise, and ptis  1 if person i in state s is employed part-time and is not observed unless empis  1. disablei is equal to 1 if person i is disabled, 0 otherwise; posts is equal to 1 if the person is observed after passage of the state legislation; X1i and X2i include individual demographic characteristics; 1is and 2is are distributed as a bivariate normal with means equal to 0, variances equal to 1, and correlation equal to . Again, the coefficient in the part-time equation on the disable  post regressor is what tells us whether there is any change in the proba- bility of part-time employment among disabled workers, post-legisla- tion, relative to nondisabled workers. Table 6.4 details the regression results. Using the parameter estimates, the difference in the impact of having a work-limiting disability on part-time employment across the two time periods can be calculated by evaluating the probabilities of interest for each person, varying the disable and post dummy vari- ables, then taking the difference between these probabilities and aver- aging the differences across the sample. This calculation translates the estimated coefficients into a 2-percentage-point greater probability of disabled workers being employed part-time than nondisabled workers, post-legislation relative to pre-legislation. This result is only signifi- cantly different from zero at the 85 percent confidence level, but it does provide some support for the notion that disability legislation, whether by the states or national, influences the hours of work of dis- abled workers.4 Also, as with the preceding wage analysis, these find- ings suggest that the impact of the ADA was dampened somewhat by the adjustments in hours that had already taken place as a result of the state-level legislation. CONCLUSIONS The goal of this chapter was to determine what impact state-level disability legislation has had on the employment, wage, and hours out- State versus Federal Legislation 139 Table 6.4 Employment and Part-Time Employment Bivariate Probit with Selection Results, CPS, 1981–1991 Employment Part-time Regressor equation employment equation disable  1 0.1311*** 0.3406*** (0.0498) (0.0462) post legislation  1 0.0280* 0.0208* (0.0165) (0.0118) disable  post  1 0.0129 0.0828^ (0.0654) (0.0582) Rho 0.7707*** (0.0169) Log-likelihood 54,402 Number of observations 101,584 NOTE: States included in the analysis are Alaska, Arizona, Delaware, Idaho, Massa- chusetts, North Carolina, North Dakota, South Dakota, Texas, and Wyoming. Re- gressors included both in the employment and part-time employment equations (but not reported here) include age, education, region, race, gender, marital status, and central city residence indicator. Regressors unique to the employment equation in- clude the state unemployment rate and the number of weeks worked last year. Re- gressors unique to the part-time employment equation include occupation and industry dummies, nonlabor income, and a government employer indicator. Standard errors are in parentheses. ***  significant at the 99 percent confidence level. *  significant at the 90 percent confidence level. ^  significant at the 85 percent confidence level. comes of disabled workers. The question is whether the ADA is redun- dant with laws passed at the state level. The results indicate that the state-level legislation operates on the labor market in the same way as does the federal ADA. Namely, relative employment probabilities of persons with disabilities are not affected by state-level disability legis- lation. It was also found that labor force participation rates were unaf- fected by the state-level legislation, lending support for the theory that the decline in labor force participation rates observed post-ADA at the national level was not ADA-induced. In addition, the disabled also experienced a relative wage decline and a tentative rise in relative part- time employment at the state level following legislation. The main implication of these results is that the lack of impact of the ADA on employment, while perhaps disappointing to proponents, is consistent with the contention that this type of legislation arrives after society has already adopted its main principles, both at the na- 140 Hotchkiss tional and the state level. On the other hand, observing similar wage and hours effects in states and nationally indicates that the wage and hours impacts of the ADA would likely have been greater in magnitude had the disabled not already partially experienced the impact of protec- tive legislation at the state level. The analyses in this chapter clearly indicate that state-level disability legislation did not fully crowd out the impact of the ADA (not at all regarding employment, and poten- tially only partially regarding wages and hours). Can we conclude, then, that the ADA was redundant? The answer is ‘‘no.’’ There was no employment effect to crowd out (no employment effect at the na- tional level), and there was still a measurable impact on wages and hours at the national level. In addition, the federal legislation served as a mechanism to instill uniformity of expectations of employers (even though some state laws have a broader definition of coverage), and it brought the issue of discrimination against disabled workers to the na- tional forefront. Notes 1. Also see Advisory Commission on Intergovernmental Relations (1989) for addi- tional information. 2. Methodologies that take advantage of differing legislative statuses among states (or,

more generally, across observations) have often been referred to as ‘‘natural experiments’’ and have been applied by a number of researchers. For example, see Chay (1996) and Carrington, McCue, and Pierce (2000). 3. The strategy described here can be likened to a differences-in-differences (DD) methodology but is applied to a nonlinear statistical model. While this type of pooled, cross-sectional analysis has been used by many researchers (for example, Card 1992; Gruber 1994 and 1996; Zveglich and Rodgers 1996; and Hamermesh and Trejo 2000), the technique also has its critics (such as Heckman 1996). The primary criticism of this approach is that it is impossible to control for unobserved changes in the environment that occurred at the same time as the event of interest. The concern is mitigated in the analysis in this chapter, however, by the fact that the post-legislation period varies across states. 4. Also, similar to the approach in Chapter 4, an additional analysis was performed to determine whether, as at the national level, there was a marked increase in voluntary part-time employment, but the results were inconclusive. The coeffi- cient on the interaction disable  post term was not significantly different from zero, and the adjusted R2 was only 0.08. 7 Conclusions and Policy Implications This book has examined and documented the relative labor market experience of workers with disabilities with an eye to evaluating the impact of the ADA. A worker’s labor market experience goes beyond simply whether a person has a job and what he or she is being paid. While these dimensions are fundamental, the quality or characteristics of the worker’s job, the process of obtaining it, and the nature of job separation are also important factors. One intention of the ADA is to break down barriers in the labor market; thus the focus of all analyses in this book is on the experience of the disabled in that environment, not on factors that influence decisions to enter the labor market. Ac- counting for those choices, however, is important in obtaining results generalizable to the disabled population, so measures are taken, where appropriate, to control for the decision to seek employment. In addi- tion to the multiple dimensions of the potential impact of the ADA on disabled Americans in the labor market, there are at least as many more ways in which the ADA affects the lives of all disabled Americans; such issues are not the subject of this work, but they may in fact amount to a much greater overall effect than that experienced by disabled workers alone. Overall, the analyses presented here lead to the conclusion that the labor market experience of disabled workers is quantitatively lower in all dimensions than that of nondisabled workers. In addition, while this relative situation has improved over time in some ways, there is no strong evidence that it has been substantively impacted by the ADA. There are two primary reasons why the ADA may not have had the hoped-for dramatic effect. It could be the case that no one is paying attention to the legislation. In other words, employers may not be complying (either through lack of awareness or refusal) with the man- dates of the ADA, workers may not be aware of their rights under the law, or workers may not be pursuing these rights. Alternatively, it may be the case that the bulk of the experience of disabled workers in the labor market is being defined by factors other than those corrected for by the ADA. 141 142 Hotchkiss It is not likely that the former is the case. For example, cognizance of the ADA is widespread. A 1999 Harris poll indicated that 67 percent of those surveyed were aware of the ADA, and it is likely that even a higher percentage of people with disabilities (and employers) know of the legislation (Roper Center for Public Opinion Research 1999). In addition, the rate of ADA litigation suggests that disabled workers are actively pursuing their rights. After climbing to a height of almost 20,000 in 1995, the number of claims tapered off to about 16,000 in 2001.1 Lastly, there is evidence that at least large employers and mu- nicipalities are complying with the provisions of the ADA (see Scheid 1998; Condrey and Brudney 1998). Additional evidence of compliance is implied by the Equal Employment Opportunity Commission’s ‘‘de- termination of no reasonable cause to believe that discrimination oc- curred based upon evidence obtained in investigation’’ in 54.1 percent (fiscal year 2001) of the ADA charges made.2 This statistic has stayed at 54 percent or higher since 1996. It is probable, therefore, that the lack of notable impact of the ADA on the labor market experience of the disabled implies that, like many other pieces of legislation with a strong social and moral content, it was adopted in an environment that had already embraced its principles and mandates, for the most part. For example, by 1990, every state had adopted some form of legislation granting protection to disabled workers. Results in Chapter 6 indicate that these state-level policies had influences on employment, wages, and hours similar to those found when the federal legislation was implemented. The implication is that some of the anticipated effect of the ADA had already been experi- enced at the state level over a longer period of time, beginning typically in the 1970s. This is not to say that the ADA was an unnecessary piece of legislation from the perspective of the labor market. Even if the ADA merely reflects the environment in which it was passed, it serves to strongly proclaim our social values and to provide a uniform legal mechanism with which to arrest the activities of those who have not yet embraced those values. As such, however, we are left with the nagging question of, ‘‘What do we do now to improve the labor market experience of workers with disabilities?’’ The results of the analyses presented in this book can provide some guidance in answering this question. Conclusions and Policy Implications 143 The dimensions of a worker’s labor market experience evaluated include employment, compensation, hours of work, distribution across and representation in occupations and industries, job separation, unem- ployment, and job search. The CPS is the primary source of data for each analysis. Various repeated cross-sectional and pooled cross- sectional analyses were performed with data spanning the years 1981– 2000. Some of the analyses were supplemented with information in the SIPP. The SIPP is primarily used to identify whether any experience or impact differs across type of disability. The years of analyses with the SIPP are limited to 1986–1997 but generally confirm the conclusions from the longer data series available from the CPS. The results of most notable interest are summarized in the following discussion and are accompanied by policy recommendations. This chapter ends with an overall assessment of the implications of the findings for the ADA and beyond. EMPLOYMENT INCENTIVES Chapter 2 presents evidence that while the joint labor force partici- pation and employment outcome declined among the disabled (driven by a decreasing labor force participation rate), the unconditional em- ployment probability of the disabled did not change, relative to that of the nondisabled. In other words, the employment prospects for the disabled, while not improved by the ADA, were also not harmed. Analysis of the SIPP data set revealed that workers with mental disabil- ities (and disabilities classified as ‘‘other’’) actually experienced an in- crease in employment relative to the nondisabled. This may be the result of easier accommodation of mental impairments or the recent emphasis on providing employment opportunities for the mentally dis- abled by the Presidential Task Force on Employment of Adults with Disabilities (formerly, the President’s Committee on Employment of People with Disabilities).3 In addition, there has been a relative gain in employment in large firms among the disabled, as compared with the nondisabled; of course, large employers are those most able to absorb the cost of accommodation. These results lead to some obvious policy suggestions: providing incentives for the disabled to enter the labor force and facilitating the accommodation process. 144 Hotchkiss Incentives for the Disabled The most recent initiative to encourage labor force participation among the disabled is the Ticket to Work and Work Incentives Im- provement Act of 1999; it applies to recipients of Supplemental Secur- ity Income (SSI) or Social Security Disability Insurance (SSDI) government cash payment programs for people with disabilities. The main provisions that encourage labor force participation under these programs involve reducing the risk and cost associated with ‘‘trying’’ work. One provision is a disregard for impairment-related work ex- penses (e.g., special equipment modified to accommodate a worker’s disability, medical devices, and special transportation needs). These expenses are deducted from a worker’s income before it is evaluated for payment eligibility purposes. In order to encourage and facilitate labor force participation, this provision could be expanded to provide for direct reimbursement of these fixed (out-of-pocket) expenses. The criteria for determining reimbursement could be the same for current deductibility. Another current work incentive provision allows for the exclusion of earned income for SSI recipients. Under this policy, the first $65 a month and half of the remainder of earnings are disregarded in calcu- lating the SSI payment amount. This provision could be made more generous and match that of the Trial Work Period (TWP) available for SSDI recipients. The TWP allows SSDI recipients to work for nine months without any reduction in benefit payments; after that period, payments are discontinued if the person is able to maintain his or her work activity. The continuation of Medicare and Medicaid benefits beyond SSI or SSDI eligibility is an important feature of the current incentives. Many jobs for which disabled workers can qualify might not offer ben- efits, especially if the person must work part-time in order to accom- modate his or her impairment. In addition, the extended period of eligibility (SSDI) and reinstatement of payments without a new appli- cation (SSI) are important safety nets for the disabled worker not sure if he or she should give employment a try. These provisions basically allow recipients to reinstate payments without once again going through the lengthy application process if they discover they are not yet ready to permanently enter or reenter the labor force. SSDI can be Conclusions and Policy Implications 145 reinstated up to 36 months after ineligibility, but former SSI recipients only have 12 months to apply for reinstatement. While the SSI time allowance is considerably shorter than the SSDI time allowance, eligi- bility requirements for the programs likely necessitate this differentia- tion. In addition to these programmatic incentives, strategies encourag- ing labor force participation among all disabled individuals (not just those receiving disability benefits) are warranted, based on the results in Chapter 2. One suggestion, the Disabled Worker Tax Credit (Burk- hauser, Glenn, and Wittenburg 1997), would provide incentives similar to the current Earned Income Tax Credit for the working poor, but be targeted at the disabled worker. Under this program, disabled workers would essentially receive a subsidy to their employment wage. It is not clear whether this program, however, would be effective in improv- ing the labor market experience along compensation or job quality di- mensions. The subsidy may encourage disabled workers to take lower- paying jobs while not holding employers responsible for valuing the skills (with accommodation) of this group equally with those of non- disabled workers. Incentives for Employers There has been a shift of employment among disabled workers toward larger firms. The implication is that cost might be an important factor in the willingness to hire a person with disabilities, and if that cost were reduced, disabled labor force participants would make even greater gains in employment. The federal government currently pro- vides tax incentives to smaller firms to help pay for the expense of accommodating workers with disabilities. Section 44 (Title 26) of the Internal Revenue Service (IRS) Code allows for a tax credit to cover 50 percent (up to $5,000) of an ‘‘eligible access expenditure’’ in one year incurred by a business with total revenues of $1 million or less, or 30 or fewer full-time employees (Hays 1999). Qualifying expenditures under this section include adaptations of existing structures and pur- chase of special equipment and services (such as sign language inter-

preters).4 While this provision allows firms to be reimbursed for out- of-pocket expenses for accommodating disabled workers, the actual physical process of accommodating (i.e., evaluation of need and modi- 146 Hotchkiss fication of the environment) distracts the organization from its primary focus of business. One thing the government could do to ease this situation would be to equip the disabled worker with knowledge and information regarding any specific accommodation he or she might require in the type of job being sought. It is not the intention of the ADA to put the burden of acquiring this background on the worker. However, the more information workers can provide and the easier they can make it for the employer, the better chance the individuals will not be seen merely for the burden they might cause, but for the attributes they provide to the firm. The Work Opportunity Tax Credit (Title 26, Section 51 of the IRS Code) provides a tax credit for hiring individuals from specific target groups, with SSI recipients being one of those groups. The employer can claim 40 percent (up to $6,000) of the hired worker’s first-year wages. The maximum credit applies to individuals employed at least 400 hours during the year, and lesser credit applies to those employed between 125 and 400 hours per year. While it is difficult to find the number of employers who have taken advantage of the tax credit op- portunities associated with hiring the disabled, it is likely that these programs might suffer the same sort of administrative burdens that have resulted in the ineffectiveness of other tax credit and employment subsidy programs (for example, see Tannenwald 1982 and Katz 1998). The Job Accommodations Network (JAN), sponsored by the Office of Disability Employment Policy of the U.S. Department of Labor, is a government resource that provides both employers and workers with valuable information. Network members share experiences in success- ful accommodation strategies; employers who utilize the service are required to join the network (Magill 1997).5 One benefit provided by JAN is Searchable Online Accommodation Resource (SOAR), which allows someone on the Internet to go through a series of steps (select a disability, functional limitation, an affected job function, and an ac- commodation solution) and to obtain a list of vendors (from across the nation) providing the accommodation solution identified that could help a worker with a specified job function. There is also the opportu- nity on the SOAR website to enter specifics about an employer’s or worker’s unique situation and to obtain personal feedback. In addition, JAN provides information services to individuals with disabilities re- garding starting a small business or becoming self-employed. Conclusions and Policy Implications 147 Armed with the data obtained through JAN, the disabled job seeker leaves the employer little excuse for not considering only the qualifica- tions of the disabled applicant (i.e., productivity with accommodations in place). Clearly, if one could document the usefulness and success of these information dissemination efforts, a case might be made for devoting even more resources toward such efforts and perhaps provid- ing regional consultants that could evaluate a worker’s or employer’s situation in person. EDUCATION, TRAINING, AND JOB CHARACTERISTICS One analysis in Chapter 3 indicates that the disabled overall have suffered a cost in terms of lower relative wages post-ADA. With the exception of large firms, wages of disabled workers declined by about 3 percent post-ADA, relative to those of nondisabled workers. People with musculoskeletal disabilities suffered the bulk of the wage loss. Policy suggestions made in the previous section to facilitate worker accommodation should also go toward improving the apparent com- pensation tax on disabled workers for whatever workplace adjustments are required. Further analysis in Chapter 3 indicates that these wage losses may not be directly related to accommodation costs, however, but are suffered by disabled workers whether or not they are covered by the ADA. By facilitating the accommodation process (through in- formation and resources provided by JAN or some other organization), a person with disabilities is in a better position to negotiate a wage comparable to that of nondisabled individuals. Decomposing the wage differential between disabled and nondis- abled workers indicates that only about 30–40 percent of the gap is explained by observed characteristics of the workers themselves. While the remainder cannot all be interpreted as discrimination against the disabled, there is a significant portion of the differential that re- mains unexplained. Within the explained part, the greatest contribu- tions to the wage discrepancy are made by differences in industry, occupation, and educational attainment. In addition, it was found that disabled workers actually typically received a higher return on their education than did nondisabled workers, ceteris paribus. The implica- tion is that investment in disabled workers’ education and training for 148 Hotchkiss high-paying jobs would go a long way to improving their compensation experience in the labor market and is consistent with the results found in Chapter 4. The analyses in Chapter 4 indicate that the distribution of disabled and nondisabled workers across occupations has actually become more dissimilar over time (with some recent improvements), with disabled individuals being significantly more concentrated in the lower-paying occupations. Workers with mental disorders are the most dissimilar in their distribution across occupations, compared with non- disabled workers. The SSI and SSDI programs have a number of provisions that as- sist with training, rehabilitation, and educational attainment. The pri- mary one is that benefit payments will continue while an individual is participating in a rehabilitation program, even if the recipient recovers from his or her disability. Recipients of SSI may also participate in what is called a PASS (Plan for Achieving Self-Support) program, under which a SSI recipient may put aside assets and money toward a plan that helps the recipient become self-sufficient, including rehabili- tation or starting a business. These assets will be ignored in continuing determination of eligibility. In addition, the 1999 Ticket to Work legis- lation provides (as of December 2000) SSI and SSDI recipients with a ‘‘ticket’’ that can be used to obtain vocational rehabilitation, job or other support services from an employment network of the recipient’s choice. This provision makes obtaining vocational rehabilitation more flexible, and thus more accessible. In spite of such provisions, Chatterjee and Mitra (1998) indicate that less than 5 percent of federal spending for SSI and SSDI goes to training and rehabilitation. They suggest that this type of expenditure pattern on the disabled is the result of ‘‘a bias in disability programs in favor of short term equity considerations as opposed to the long run efficiency objectives’’ (p. 360). These authors also show that there is a positive link between enhancing a disabled person’s human capital and his or her chances of participating in the labor market. The impli- cation is that devoting more resources toward education and training would not only improve disabled workers’ labor market experiences, but also help the bottom line of the SSI and SSDI programs by leading to more disabled people becoming self-sufficient. In addition, the focus of SSI and SSDI on vocational rehabilitation may not direct resources where they will provide the greatest boost Conclusions and Policy Implications 149 in compensation and occupational attainment outcomes for disabled workers. The growing earnings gap between those with and without college educations is well documented (for example, see Murphy, Juhn, and Pierce 1993). Training disabled workers for jobs requiring a col- lege degree is an idea that has not received a great deal of attention. According to the results in Chapter 3, workers with musculoskeletal disorders appear to have suffered the most in terms of compensation. One reason for this may be that the cost of accommodating a worker’s musculoskeletal disability in a blue-collar or physically-challenging position is likely much more expensive than if that worker became skilled in a less physically but more mentally demanding, and poten- tially higher-paying, job. Further evidence of the general lack of sup- port for higher education is found in the income exclusion for earnings of disabled students. While attending school, students only under the age of 22 are allowed to earn income that is excluded from calculating benefit eligibility and levels. This income exclusion could be extended to individuals over the age of 22 when they are working toward a termi- nal degree that would improve their occupational outcome.6 One area in which further general education for older SSI recipi- ents is supported is in the development of a PASS; tuition, fees, books, and supplies for school are among those expenses that can be set aside as part of the SSI recipient’s plan for self-support. While laudable, this provision is not likely to be very practical since it requires the SSI recipient, who is subject to strict earnings limitations, to set aside sav- ings and assets that are likely difficult to acquire. An additional provi- sion that allows the recipient to work (without penalty) while attending school may place attainment of higher education within the realm of financial possibility. The federal government does allow educational grants, fellowships, and scholarships used for tuition and fees to be excluded from earnings limitations. Portions of those sources used for room and board, however, are not currently excluded. A report by the Social Security Administration (2000) suggests that all portions of grants, scholarships, and fellowships be excluded from the earnings test.7 Given the importance of general education in improving the em- ployment experience of disabled workers (as demonstrated in the anal- yses of this book), the federal government should explore the possibilities of providing more active support. One could also argue 150 Hotchkiss that if the basic goal is to improve the labor market experience of workers with disabilities, subsidized training and education should be made available to all disabled individuals, not only those affiliated with a government cash payment disability program. Improving the labor market experience of a single disabled worker could also provide exter- nalities to the experience of other disabled workers. As employers and fellow workers become more comfortable working alongside the disabled and become more aware of such individuals’ capabilities, even more doors would open to disabled workers who follow. One policy change that seems to have had an unambiguous effect on the characteristics of jobs held by disabled workers is the extension of Medicaid and Medicare for SSI and SSDI recipients who have be- come ineligible for disability payments. Most part-time jobs do not offer health insurance, but many disabled workers may prefer part-time employment as a means of accommodation for their disability. Due to the extension of Medicaid and Medicare, disabled workers can now consider part-time employment without the risk of losing medical ben- efits. The results in Chapter 4 support this outcome; disabled workers are increasingly more likely to be employed part-time (versus full- time), but they are also increasingly more likely to be voluntarily (ver- sus involuntarily) employed on a part-time basis. On the one hand, the extension of Medicaid and Medicare has opened up work opportunities that may improve a disabled worker’s life. On the other hand, there is some evidence that part-time jobs are more likely to be marginal, temporary, unstable, and lower-paying (Blank 1990). The government might be able to provide assistance to (particularly, small) employers who want to explore flextime in order to accommodate workers’ dis- abilities in occupations that may not be typical candidates for such scheduling, but that pay more and provide for advancement opportuni- ties within the firm. One step that will likely contribute positively to this effort is the President’s ‘‘New Freedom Initiative.’’8 This concept calls for the fulfillment of the promise made by the government to people with disabilities through the passage of the ADA; it includes education, home ownership, access, and employment provisions. Ex- panding telecommuting opportunities is one example: ‘‘The Adminis- tration will provide Federal matching funds to states to guarantee low- interest loans for individuals with disabilities to purchase computers and other equipment necessary to telework from home. In addition, Conclusions and Policy Implications 151 legislation will be proposed to make a company’s contribution of com- puter and Internet access for home use by employees with disabilities a tax-free benefit.’’9 Given the results in Chapter 4, that flexible work arrangements (i.e., part-time jobs) may be an important aspect of greater employment opportunities for people with disabilities, this

pro- posal holds promise. SCREENING AND MATCHING The evidence of job separation and job search experiences of dis- abled workers presented in Chapter 5 is ambiguous. On the one hand, disabled workers search three weeks longer, on average, than similar nondisabled individuals before finding a job. On the other hand, job separations are less likely to be for involuntary reasons among disabled than among nondisabled workers, implying that disabled workers are not likely the ‘‘marginal’’ employees that some have speculated they are. While longer search spells are consistent with discriminatory hir- ing practices on the part of employers, the finding that most of the observed longer search spell is explained by individual characteristics suggests that endowments of disabled and nondisabled workers are being valued equally, but that employers go to greater lengths to dis- cern the fit of a disabled worker’s traits with a particular job. This care in hiring on the part of the employer would also lead to the lower probability that a separation is involuntary. In addition, the firm may have made some expenditures in accommodating the worker’s disabil- ity and thus will be reluctant to lose that investment by laying off or firing the worker. The lower probability that a separation is involuntary means that disabled workers experience a higher likelihood that a separation is voluntary. While this may be interpreted as disabled workers having a fair amount of mobility in the labor market, it may also mean that they have more difficulty than nondisabled individuals in finding a good job match. What is called for is a policy that assists employers in discern- ing the qualifications and fit of a disabled worker more quickly and that helps disabled workers in determining the appropriateness of a particular job. Measures for improving the efficacy and speed of job matches would include a clearing house at which employers could post 152 Hotchkiss job openings and workers could advertise skills (with appropriate ac- commodations); third-party certification of worker skills; and assis- tance with information (such as JAN) and with cost (such as tax credits) for accommodating a particular disability for the job to be filled. The National Easter Seal Society provides many services that would facilitate matching of workers with employers. Skills evaluation and screening, employment skills training, and job placement services are among the programs offered through Easter Seals. JAN could also be expanded to provide job placement services, which it currently does not do.10 The U.S. Office of Disability Employment Policy, however, has some initiatives that do assist in employment on a limited basis.11 The Employer Assistance Referral Network (EARN) is designed to help employers in locating and recruiting qualified individuals with disabilities. The Workforce Recruitment Program is involved with se- curing summer employment and internships for students with disabili- ties, and Project EMPLOY is designed to expand and enhance job opportunities for people with cognitive disabilities by, primarily, pro- viding funding to other organizations to provide employment services. One thing that might be difficult for an employer to assess is the productivity of a disabled worker with accommodating equipment or services in place. If the government, or some private entity, could certify a particular skill (such as typing or editing) when a disabled individual has access to facilitating equipment, the employer would not be forced to bear the risk of hiring the worker and installing the neces- sary equipment without knowing what the outcome will be. There could be testing centers set up regionally, or mobile testing centers, that would contain the most common equipment needed for the worker to do the job in question. THE ADA AND BEYOND The labor market provisions of the ADA comprise a small part of the overall goals of the legislation. Furthermore, given the relatively low proportion of the disabled population that is actually in the labor force, and smaller yet that is employed, the potential impact of these provisions is not nearly as widespread as the effect of other elements Conclusions and Policy Implications 153 of the legislation that cover aspects of a disabled person’s life outside the labor market. As such, measuring the influence of the ADA on the overall quality of life of the disabled goes far beyond the potential impact on the labor market experience. Indeed, the key contributions of the ADA may be beyond quantification in economic terms; it is difficult to put a price on the dignity and respect that proponents might argue are among the most important dividends of the ADA. Nevertheless, learning that the ADA did not result in dramatic, or even notable, improvement across multiple dimensions of the labor market experience of the disabled must be quite disappointing for the proponents of the legislation. In light of these findings, we are left with the question of why no such impact was forthcoming. One possi- ble reason for legislation not having an effect on the intended benefi- ciaries is that there was nothing to improve: the disadvantaged are not really as disadvantaged as they might appear (because of factors the researcher may not be able to observe, for example). If this is the case, then the ADA is directed at a nonexistent target. A second possibility for finding no influence of the ADA on labor market outcomes among the disabled might be that the labor market provisions of the legislation are focused on the wrong things. Title I of the ADA is couched in terms of eliminating discriminatory behavior on the part of employers. It could be the case that discrimination is not the culprit determining inferior labor market outcomes for the disabled; in other words, the ADA is aiming at the wrong target. A third potential explanation for finding no impact is that the ADA is focused on the right target, but just missed. In other words, it may be the case that the legislation is ineffective, that employers are finding ways to get around the provi- sions and that workers are not able to exercise their rights under the law. With the amount of evidence presented in this book, as well as that provided elsewhere, it is not likely that the ADA is aiming at a nonexis- tent target. One advantage of examining labor market outcomes from multiple dimensions, as was done here, is to be able to rule out numer- ous explanations for inferior labor market outcomes. The disabled clearly have further to go before enjoying labor market outcomes com- parable to those experienced by the nondisabled. The possibility that the ADA has missed the target and that employers are able to somehow get around complying with the legislation is also not likely. The inabil- 154 Hotchkiss ity of employers to ignore the ADA is evidenced by the fact that dis- ability claims made up 20 percent of all claims made to the EEOC during fiscal year 2001. This percentage was not too far behind the ratio of claims filed based on race (36 percent), sex (31 percent), and age (22 percent) during the same year and suggests that workers are aware of their rights and are holding employers accountable. The chance, then, that the ADA is aiming at the wrong target is still a possibility. While individual cases of discrimination (as evidenced by the number of EEOC claims) indicate that discrimination against dis- abled workers is likely occurring in the labor market, it still may not be the overwhelming determinant of inferior labor market outcomes. As suggested earlier in this chapter, other policies, such as promoting education and training, may go further toward improving labor market outcomes than a policy outlawing discrimination (which may only touch a small portion of the disabled). Regardless of why the ADA does not seem to have affected the labor market experience of the disabled, this absence of impact begs the question of whether the ADA is necessary. Clearly, it is possible to argue the merits of the ADA on the ways in which it has likely improved the quality of life among the disabled beyond its labor market or quantifiable influence. However, even with regard to its labor mar- ket provisions, the ADA does serve as a statement of our social values, and it provides a legal mechanism with which to stem the activities of those who have not yet accepted those values. In addition, there is no question that it does serve to set the stage and to provide a labor market environment in which effective reforms, more narrowly focused on the needs of the disabled for improving outcomes, can be introduced. Notes 1. See table of statistics titled, ‘‘Americans with Disabilities Act of 1990 (ADA) Charges FY1992–FY2001,’’ found at http://www.eeoc.gov/stats/ada-charges .html. 2. See table of statistics titled, ‘‘Americans with Disabilities Act of 1900 (ADA) Charges FY1992–FY2001,’’ found at http://www.eeoc.gov/stats/ada-charges .html. 3. Other possible explanations include the potential for the ADA to have a greater impact on overcoming negative social attitudes against people with mental dis- abilities (versus other types of disabilities) or improvements in medications occur- Conclusions and Policy Implications 155 ring during the same time period, which facilitated the labor market performance of people with mental disabilities. 4. More substantial architectural and transportation adaptation expenses can be de- ducted from tax liability by all businesses under IRS Code Section 190. 5. The web site for JAN is http://www.jan.wvu.edu/. Additional government re- sources for people with disabilities can be found at http://www.disabilitydirect .gov. The National Easter Seal Society (http://www.easter-seals.org/) also of- fers free information referral and suggestions for technological devices to help workers do their jobs. 6. This recommendation is also made by the Social Security Administration’s March 2000 report to the Congress on income and resource exclusions. That report also recommends that the amount of earnings excluded be increased and then indexed to the CPI. 7. The report also suggests that grants, fellowships, and scholarships be excluded from resource limitations for nine months, based on the reality that many forms of financial aid are received at the beginning and paid out over the balance of the school year. 8. See the U.S. Health and Human Services Web site, ‘‘New Freedom Initiative: Fulfilling America’s Promise to Americans with Disabilities,’’ found at http://www .hhs.gov/newfreedom/. 9. See the U.S. Health and Human Services Web site, ‘‘New Freedom Initiative: Fulfilling America’s Promise to Americans with Disabilities,’’ found at http://www .hhs.gov/newfreedom/. Also see Chen (2001). 10. See note 5. 11. These job search and placement efforts (and others) can be located from the http:// www.disabilitydirect.gov Web site. This Page Intentionally Left Blank Appendix A CPS Sample Construction This appendix provides information regarding the use of the Current Pop- ulation Survey data set for the analyses in this book. The lessons learned might be useful to others creating successive cross-sections across many years using the CPS. The notes and recommendations reflect the experience of the author only. 1. The complete set of outgoing rotation groups from the CPS was ob- tained from Unicon Research Corporation (http://www.unicon.com). The out- going rotation group in the CPS consists of individuals in their 4th and 8th month of eight monthly interviews. A CPS respondent is interviewed for four consecutive months, not interviewed for four months, then interviewed again for four consecutive months. This source is highly recommended for not only outgoing rotation groups, but for all of the CPS data sets one might want. The data arrive on CDs with easy-to-use extraction software. The documentation is excellent; the coding across years is consistent; technical support is accessi- ble and helpful; and the documentation also makes note of known data anoma- lies or errors. Unicon makes these data available for a fee. 2. The March supplemental files for each year were obtained from the Inter-University Consortium for Political and Social Research (ICPSR), http:// www.icpsr.umich.edu/. We identified tremendous (un-correctable) prob- lems with the 1994 March CPS obtained from ICPSR and ultimately obtained the data needed for 1994 from Unicon. 3. A variable of crucial importance to the analyses in this book is the indicator of disability status. That indicator is not available in the CPS public use file between 1981 and 1987; we contracted with the Census Bureau to extract the necessary variable and individual identifiers essential for matching with the rest of the CPS file for those years. 4. Creation of the CPS data sets for each year

required matching individ- uals in each of the outgoing rotation groups from March, April, May, and June with the supplemental questionnaire in March. The match rate was approxi- mately 90 percent for each month, except March, where the match rate, of course, was 100 percent. 5. The coding of the variable in the March supplement indicating whether an individual worked last year changed over the entire time period. Although this coding change is well documented, it could confound analyses if the dif- ference is not noticed. Specifically, for 1981–1987, a ‘‘1’’ indicates that the 157 158 Hotchkiss person did not work during the previous year, and, for 1988–1999, a ‘‘1’’ indicates that the person did work during the previous year. 6. Prior to 1994, the question of usual hours worked per week was asked of those who were earner eligible (outgoing rotation groups) and employed during the previous week. Starting in 1994, this variable was coded as zero for workers with variable hours, and the question regarding how many hours were worked at all jobs last week was changed to read, ‘‘Last week, how many hours did you actually work at your (main) job?’’ Unfortunately, the Census code book (which accompanies data from ICPSR) does not reflect the change in this question since 1994 (it still indicates that the question refers to hours worked at all jobs, when in fact, it now only reflects hours at the main job). Appendix B SIPP Sample Construction In the terminology of the Survey of Income and Program Participation, a Panel refers to a group of interviewees and the year in which that group was first interviewed. A Wave refers to an interview within a Panel. Each Panel has anywhere from 3 (1989 Panel) to 12 (1996 Panel) interviews. Further details of the SIPP can be found on the Internet at http://www.bls.census .gov/sipp/. Core data from each Wave within a Panel used for sample con- struction were merged with the topical module for Wave 2 from the same Panel. Wave 2 topical modules were used for identification of a work disabil- ity for consistency over as many consecutive years as possible. When avail- able, core data from Wave 2 in one Panel were combined with core data from Wave 5 in the previous Panel and with core data from Wave 8 in the Panel before that in order to construct a larger sample year. Table B.1 presents the layout of the SIPP sampling structure. The Waves merged across Panels (col- umns in Table B.1) are combined for illustration using a bold box outline. Since the goal was to create multiple cross sections comparable to the CPS, only Waves 2, 5, and 8 in each Panel were exploited. Combining data across Panels was not possible for sample years 1986, 1989, 1990, 1995, 1996, and 1997 due to the lack of availability of overlap- ping Panels for which disability information is available. The labor market information, due to the rotation of sets of questions in the SIPP, corresponds to June of the year referenced. The exception to this pattern is the 1997 sample labor market information, which came from Wave 5 of the 1996 panel, which corresponds to August of 1997. Of course, these waves were merged with the disability topical module for that panel. The disability check in the topical module was the variable used to desig- nate a work-limiting disability. There are two checks in topical modules for the 1986–1993 Panels, but only one check in the 1996 Panel, resulting in the slightly smaller incidence of individuals with work-limiting disabilities in the 1996 and 1997 samples. The match rate across Waves within a Panel ranged from 81 to 89 percent success. Labor status refers to activity during the previ- ous month (as opposed to the previous week, as in the CPS), since job infor- mation corresponds to activity over the month. The primary usefulness of the SIPP derives from an ability to identify the nature of a disabled worker’s disability. The categories identified are too numerous for all of them to be included in the analysis, so they are grouped into broad headings based on the classifications used by the Social Security Administration. Table B.2 shows how specific disabilities are classified. 159 160 Hotchkiss Table B.1 SIPP Data Structure Waves Panel Year Month 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1996 1983 Oct. 1 1984 Feb. 2 June 3 Oct. 4 1985 Feb. 5 1 June 6 2 Oct. 7 3 1986 Feb. 8 4 1 June 9 5 2 <- 1986 sample Oct. 6 3 1987 Feb. 7 4 1 June 8 5 2 <- 1987 sample Oct. 6 3 1988 Feb. 7 4 1 June 5 2 <- 1988 sample Oct. 6 3 1989 Feb. 7 4 1 June 1989 sample -> 5 2 <-Wave 2 (1989) top. mod. not avail. Oct. 6 3 1990 Feb. 1 June 2 <- 1990 sample Oct. 3 SIPP Sample Construction 161 1991 Feb. 4 1 June 5 2 <- 1991 sample Oct. 6 3 1992 Feb. 7 4 1 June 8 5 2 <- 1992 sample Oct. 6 3 1993 Feb. 7 4 1 June 1993 sample -> 8 5 2 Oct. 6 3 1994 Feb. 7 4 June 1994 sample -> 8 5 Oct. 9 6 1995 Feb. 10 7 June 1995 sample -> 8 Oct. 9 1996 Apr.. 1 Aug. 1996 sample -> 2 Dec. 3 1997 Apr. 4 Aug. 1997 sample -> 5 Dec. 6 NOTE: Wave 2 topical module (number in bold) is where disability information is located. Waves bordered in bold are merged to create the sample year indicated. Each wave used for sample construction was merged with the Wave 2 topical module, then combined across panels as indicated. The 1996 panel continues for 12 waves through 1999, but only six were available at the time of analysis. 162 Hotchkiss Table B.2 Classification of Disabilities in the SIPP Data Set 1. Musculoskeletal Arthritis or rheumatism Back or spine problems Broken bone/fracture Head or spinal cord injury Hernia or spinal injury Missing legs, feet, arms, hands, or fingers Paralysis of any kind Stiffness or deformity of the foot, leg, arm, or hand Blindness or vision problems Deafness or serious trouble hearing Speech disorder 2. Internal systems AIDS or AIDS-related condition Cancer Diabetes Heart trouble, hardening of the arteries High blood pressure Kidney stones or chronic kidney trouble Lung or respiratory problems Stomach trouble Thyroid trouble or goiter 3. Mental disorder Alcohol or drug problem or disordera Learning disability Mental or emotional problem or disorder Mental retardation Senility/dementia/Alzheimer’s disease Cerebral palsy Epilepsy Stroke Tumor, cyst, or growth 4. Other NOTE: Complete classification of disabilities by the Social Security Administration can be found at http://www.ssa.gov/OP_Home/cfr20/404/404-ap09.htm. aDrug addiction is excluded from protection by the ADA. Appendix C Supplemental Tables This Page Intentionally Left Blank Supplemental Tables 165 Table C.1 Trends in the Percentages of Total Sample and of Workers Indicating a Work-Limiting Disability, CPS Combined Outgoing Rotation Groups, 1981–2000 Year % of sample % of workers 1981 9.79 2.68 1982 10.19 2.69 1983 9.73 2.53 1984 10.20 2.74 1985 10.34 2.66 1986 9.89 2.68 1987 9.62 2.69 1988 9.15 2.63 1989 9.26 2.67 1990 9.34 2.74 1991 9.34 2.60 1992 9.44 2.70 1993 9.82 2.87 1994 10.95 2.68 1995 11.14 2.53 1996 10.88 2.73 1997 10.75 2.81 1998 10.05 2.51 1999 9.84 2.37 2000 9.99 2.70 166 Hotchkiss Table C.2 Percentages of Disabled and Nondisabled Individuals Employed, CPS, 1981–2000 Subsample employed (%) Year Disabled Nondisabled 1981 17.76 62.62 1982 17.27 61.39 1983 16.34 60.45 1984 17.21 62.93 1985 16.76 63.57 1986 17.70 63.79 1987 18.32 64.49 1988 19.11 63.98 1989 19.43 65.74 1990 19.35 65.98 1991 18.41 64.74 1992 18.88 64.39 1993 19.05 64.42 1994 16.50 66.16 1995 15.44 66.32 1996 16.41 66.71 1997 17.30 67.40 1998 17.09 67.88 1999 16.60 68.05 2000 17.15 68.70 Supplemental Tables 167 Table C.3 Predicted Joint Probability of Labor Force Participation and Employment and Predicted Unconditional Employment Probability, by Disability Status, CPS, 1981–2000 Average predicted probability Nondisabled Disabled Being in the labor Being in the labor force and employed Employment force and employed Employment (1) (2) (3) (4) Year Prob(lfp1, emp1) Prob(emp1) Prob(lfp1, emp1) Prob(emp1) 1981 0.64271 0.84244 0.47840 0.79271 1982 0.62609 0.81519 0.47742 0.77384 1983 0.61727 0.79255 0.47474 0.76856 1984 0.64824 0.85387 0.50693 0.82802 1985 0.65675 0.85665 0.50837 0.82475 1986 0.66467 0.85364 0.51167 0.81465 1987 0.67498 0.86818 0.51993 0.83475 1988 0.68827 0.87833 0.53576 0.82908 1989 0.69070 0.88346 0.53884 0.85364 1990 0.69150 0.87427 0.52609 0.82427 1991 0.67530 0.84653 0.52123 0.81285 1992 0.67426 0.82397 0.51998 0.76743 1993 0.67738 0.83780 0.53733 0.80476 1994 0.68448 0.84973 0.49882 0.81405 1995 0.68820 0.87372 0.49520 0.83877 1996 0.69305 0.87742 0.50769 0.84971 1997 0.70309 0.87734 0.51408 0.84607 1998 0.71169 0.89380 0.51334 0.86027 1999 0.71529 0.89326 0.50755 0.83657 2000 0.72085 0.91141 0.52832 0.86493 NOTE: Estimates obtained from a bivariate probit model with selection. Regressors for both labor force and employment determination included the state unemployment rate; age; age squared; female, nonwhite, education, and regional dummies; and a disability dummy. The labor force participation equation also included nonlabor income, marital status, and a worked-last-year indicator. The employment equation included the number of weeks worked last year. The probabilities for each column are the average across the entire sample obtained by calculating the probability for each person (varying the disability dummy variable between 0 and 1), then averaging across the sample. 168 Hotchkiss Table C.4 Relative Predicted Probabilities of Working in Each Firm Size, CPS 1987–1999 Ratio of predicted probabilities of a disabled worker relative to the predicted probability for a nondisabled coworker being employed by a firm by size Year Small firm Medium firm Large firm 1987 1.1842 1.0756 0.9189 1988 1.1836 1.1063 0.9134 1989 1.1524 1.0403 0.9409 1990 1.1372 1.0174 0.9507 1991 1.1682 0.9417 0.9493 1992 1.2127 0.8759 0.9501 1993 1.0873 1.0410 0.9512 1994 1.0567 0.9999 0.9787 1995 1.1772 0.9384 0.9399 1996 1.1811 0.8598 0.9575 1997 1.1341 0.9900 0.9511 1998 1.0110 0.9319 1.0106 1999 1.0963 0.9840 0.9675 Change over time period 0.0879 0.0917 0.0487 NOTE: Small firms have fewer than 25 employees; medium firms have 25–99 em- ployees; and large firms have 100 or more employees. Regressors (in addition to a disability dummy variable) in the multinomial logit regression included age; age squared; and regional, education, female, nonwhite, and central city dummy variables. Supplemental Tables 169 Table C.5 Observed and Selectivity-Corrected Wage Differentials, CPS, 1981–2000 Wage differential Observed corrected for wage selectivity into Endowment Coefficient Selection Year differentiala the labor marketb effectc effectd effecte 1981 0.1587 0.1315 0.0386 0.0929 0.0272 1982 0.1294 0.1318 0.0307 0.1011 0.0024 1983 0.1717 0.1991 0.0056 0.1935 0.0275 1984 0.1885 0.1902 0.0312 0.1590 0.0017 1985 0.1706 0.2037 0.0476 0.1561 0.0332 1986 0.1948 0.2144 0.0634 0.1510 0.0197 1987 0.2008 0.2342 0.0591 0.1751 0.0334 1988 0.2089 0.1954 0.0436 0.1518 0.0135 1989 0.2512 0.2009 0.0766 0.1243 0.0503 1990 0.2469 0.2605 0.0807 0.1798 0.0136 1991 0.2649 0.2494 0.0769 0.1725 0.0154 1992 0.2644 0.2901 0.0890 0.2011 0.0258 1993 0.2723 0.3114 0.0852 0.2262 0.0391 1994 0.2733 0.3337 0.0911 0.2427 0.0605 1995 0.2400 0.2768 0.0792 0.1976 0.0368 1996 0.2638 0.2642 0.0929 0.1713 0.0004 1997 0.2117 0.2090 0.0791 0.1299 0.0027 1998 0.2843 0.2776 0.0884 0.1892 0.0066 1999 0.2941 0.3028 0.0881 0.2147 0.0087 2000 0.2809 0.2566 0.0674 0.1892 0.0243 NOTE: The first-stage probit estimation included the following regressors: age; age squared; nonwhite, female, disabled, single household, education, and worked-last- year dummy variables; and nonlabor income. Regressions for 1981 and 1982 do not include a union dummy. Second-stage wage estimations included the following regressors: hour of work; age; age squared; and union, female, single household, nonwhite, education, region, industry, occupation, and government dummy variables. Regressions for 1983 do not include dummies for the service; farming, fishing, and forestry; or the craft occupations due to the absence of representation of disabled workers in these occupations in the sample. ̂nd (coefficients from the nondisabled estimation) was used to represent the ‘‘nondiscriminatory’’ world since the disabled make up such a small proportion of the whole. a ln Wnd  ln Wd b ln Wnd  ln Wd  [(ĉnd nd)  (ĉd d)] c ̂nd(Xnd  Xd) d Xd(̂nd  ̂d) e [(ĉnd nd)  (ĉd d)]. 170 Hotchkiss Table C.6 Marginal Effect of Being Disabled on the Probability of Employer-Provided Health Insurance and of Being Included in the Employer’s

Israel Deaconess Medical Cen- ter, Harvard Medical School, Division of Endocrinology, Diabe- tes and Metabolism, Boston, MA, USA Margo Hudson, MD Brigham and Women’s Hospital, Department of Endocrinology, Hypertension and Diabetes, Boston, MA, USA https://www.facebook.com/groups/2202763316616203 Contributors xv Gwendolyne Anyanate Jack, MD, MPH Weill Cornell Medical Center-New York Presbyterian Hospital, Department of Medi- cine, Division of Endocrinology, Diabetes and Metabolism, New York, NY, USA Mark Anthony Jara, MD University of Miami, Miller School of Medicine, Division of Endocrinology and Metabolism, Miami, FL, USA Ursula B. Kaiser, MD Division of Endocrinology, Diabetes and Hypertension, Department of Medicine, Brigham and Women’s Hospital and Harvard Medical School, Boston, MA, USA Melissa G. Lechner, MD, PhD David Geffen School of Medi- cine, University of California at Los Angeles, Division of Endo- crinology, Diabetes and Metabolism, Los Angeles, CA, USA Alan Ona Malabanan, MD, FACE, CCD Beth Israel Deacon- ess Medical Center, Harvard Medical School, Division of Endo- crinology, Diabetes and Metabolism, Boston, MA, USA Roselyn Cristelle I. Mateo, MD MSc Joslin Diabetes Center, Department of Endocrinology, Beth Israel Deaconess Medical Center, Boston, MA, USA J. Carl Pallais, MD, MPH Brigham and Women’s Hospital, Division of Endocrinology, Department of Medicine, Boston, MA, USA Johanna A. Pallotta, MD Harvard Medical School, Beth Israel Deaconess Medical Center, Department of Medicine, Endocrinol- ogy and Metabolism, Boston, MA, USA Daniela V. Pirela, MD Jackson Memorial Hospital/University of Miami Hospital, Division of Endocrinology, Diabetes and Metabolism, Miami, FL, USA Megan Ritter, MD Weill Cornell Medicine, New York Presbyte- rian, New York, NY, USA Jeena Sandeep, MD St. Elizabeth Medical Center, Department of Medicine, Division of Endocrinology, Brighton, MA, USA xvi Contributors Julian L. Seifter, MD Brigham and Women’s Hospital, Depart- ment of Medicine, Boston, MA, USA Antonia E. Stephen, MD Harvard Medical School, Massachu- setts General Hospital, Department of Surgery, Boston, MA, USA Catherine J. Tang, MD Beth Israel Deaconess Medical Center, Harvard Medical School, Division of Endocrinology, Diabetes and Metabolism, Boston, MA, USA Elena Toschi, MD Adult Clinic, Joslin Diabetes Center, Harvard Medical School, Beth Israel Deaconess Medical Center, Depart- ment of Endocrinology and Diabetes, Boston, MA, USA Nicholas A. Tritos, MD, DSc Harvard Medical School, Massa- chusetts General Hospital, Neuroendocrine Unit and Neuroendo- crine & Pituitary Tumor Clinical Center, Boston, MA, USA Anand Vaidya, MD, MMSc Harvard Medical School, Center for Adrenal Disorders, Division of Endocrinology, Diabetes, and Hypertension, Brigham and Women’s Hospital, Boston, MA, USA Maria Vamvini, MD Adult Clinic, Joslin Diabetes Center, Har- vard Medical School, Beth Israel Deaconess Medical Center, Department of Endocrinology and Diabetes, Boston, MA, USA Gregory P. Westcott, MD Beth Israel Deaconess Medical Cen- ter and Joslin Diabetes Center, Department of Endocrinology, Diabetes and Metabolism, Boston, MA, USA https://www.facebook.com/groups/2202763316616203 Pituitary Apoplexy 1 Ana Paula Abreu and Ursula B. Kaiser Contents Definition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 Precipitating Factors/Patients at Risk . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 Diagnosis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 Obtain Detailed Clinical History . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 Perform Detailed Physical Exam Including Cranial Nerves and Visual Fields . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 Evaluation of Endocrine Dysfunction/Laboratory Assessment . . . . . . 4 Imaging . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 Management . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 Consider Initiation of Corticosteroid Treatment . . . . . . . . . . . . . . . . . 7 Acute Intervention: Surgery vs. Conservative Treatment . . . . . . . . . . 8 Postoperative Care . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10 Monitor for Signs and Laboratory Abnormalities Suggestive of Diabetes Insipidus (DI) . . . . . . . . . . . . . . . . . . . . . . 10 Assess Pituitary Reserve . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11 Visual Assessment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12 A. P. Abreu (*) · U. B. Kaiser Division of Endocrinology, Diabetes and Hypertension, Department of Medicine, Brigham and Women’s Hospital and Harvard Medical School, Boston, MA, USA e-mail: apabreu@bwh.harvard.edu; ukaiser@bwh.harvard.edu © Springer Nature Switzerland AG 2020 1 R. K. Garg et al. (eds.), Handbook of Inpatient Endocrinology, https://doi.org/10.1007/978-3-030-38976-5_1 2 A. P. Abreu and U. B. Kaiser Follow-Up After Discharge . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12 Check Electrolytes After 1 Week . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12 Reassess Pituitary Function After 4 to 8 Weeks . . . . . . . . . . . . . . . . . 12 Suggested Reading . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13 Definition Apoplexy means “sudden attack” in Greek. Classical pituitary apoplexy (PA) is a clinical syndrome characterized by abrupt hemorrhage and/or infarction of the pituitary gland. Severe head- ache of sudden onset is the main symptom, sometimes associated with visual disturbances or ocular palsy. Apoplexy usually occurs in patients with preexisting pituitary adenomas and evolves within hours or days. Subclinical PA is defined as asymptomatic or unrecognized pituitary hemorrhage and/or infarction. It may be detected on rou- tine imaging or during histopathological examination. The fre- quency of subclinical hemorrhagic infarction in pituitary tumors is around 25%. Precipitating Factors/Patients at Risk The precise pathophysiology of PA is not completely understood. Since most cases occur in preexisting pituitary adenomas, it has been hypothesized that a reduction in blood flow or abnormal vas- cularity of the tumor could be mechanisms contributing to PA. The underlying process can be simple infarction, hemorrhagic infarc- tion, or mixed hemorrhagic infarction. The pituitary gland is enlarged in pregnancy and prone to infarction from hypovolemic shock. Pituitary necrosis that occurs in the setting of large-volume obstetric hemorrhage post- partum is referred to as Sheehan syndrome. It is a rare but poten- tially life- threatening complication that can result in postpartum hypopituitarism. https://www.facebook.com/groups/2202763316616203 1 Pituitary Apoplexy 3 The clinical symptoms of PA mimic other common neurologi- cal disorders such as subarachnoid hemorrhage, migraine, bacterial meningitis, or stroke, which can lead to delayed or even missed diagnosis. A high degree of clinical suspicion is needed to diagnose pituitary apoplexy, as most patients do not have a previ- ous history of known pituitary adenoma. Precipitating risk factors have been identified in 10–40% of cases of PA, and it is important to recognize them. Hypertension has been considered a precipitat- ing factor for PA, although recent studies question this associa- tion. Surgery, particularly coronary artery surgery, and angiographic procedures have been reported to be associated with PA. Dynamic testing of pituitary function, using growth hormone- releasing hormone, gonadotropin-releasing hormone, thyrotropin- releasing hormone, and corticotrophin-releasing hormone (less commonly), or an insulin tolerance test, is also associated with PA. Initiation or withdrawal of dopamine receptor agonists, estro- gen therapy, radiation therapy, pregnancy, head trauma, and coag- ulopathy are some other factors known to induce pituitary apoplexy. A diagnosis of PA should be considered in all patients with these risk factors who present with acute severe headache, with or without neuro-ophthalmologic signs. Diagnosis Obtain Detailed Clinical History The clinical presentation can be acute or subacute and is highly variable, determined by the extent of hemorrhage, necrosis, and edema. Headache is present in more than 80% of patients. It is usually retro-orbital but can be bifrontal or diffuse. Nausea and vomiting can be associated. As most patients have an underlying macroadenoma, signs and symptoms of hypopituitarism may have been present prior to the episode of PA. As discussed ear- lier, most patients do not have a history of a known prior pitu- itary adenoma and therefore do not carry a diagnosis of hypopituitarism. 4 A. P. Abreu and U. B. Kaiser Sheehan syndrome usually presents with a combination of fail- ure to lactate postdelivery and amenorrhea or oligomenorrhea, but any of the manifestations of hypopituitarism (e.g., hypotension, hyponatremia, hypothyroidism) can occur at any time from the immediate postpartum period to years after delivery. Perform Detailed Physical Exam Including Cranial Nerves and Visual Fields More than half of patients with PA have some degree of visual field impairment, with bitemporal hemianopsia being the most com- mon. About half of patients have oculomotor palsies due to func- tional impairment of cranial nerves III, IV, and/or VI. Cranial nerve III is most commonly affected, resulting in ptosis, limited eye adduction, and mydriasis due to nerve compression. Extravasation of blood or necrotic tissue into the subarachnoid space can result in meningismus and an altered level of consciousness. Evaluation of Endocrine Dysfunction/Laboratory Assessment Most patients will have dysfunction of one or more pituitary hor- mones at the time of initial presentation. The most clinically important hormone deficiency is adrenocorticotrophic hormone (ACTH), which can be life-threatening. It is present in 50–80% of patients and can cause hemodynamic instability and hyponatre- mia. Of note, hyponatremia is a consequence of cortisol deficiency, with loss of feedback inhibition of arginine vasopressin/antidi- uretic hormone (ADH) release despite hypoosmolality and a direct water excretion defect. Additionally, hypothalamic irritation in the setting of PA can result in the syndrome of inappropriate antidi- uretic hormone. Nausea/vomiting and hypoglycemia (secondary to GH and/or ACTH deficiency) are also stimuli for ADH secretion. Secondary hypothyroidism can also contribute to hyponatremia. Patients with suspected PA should have electrolytes, renal function, kidney function, coagulation, and CBC checked to assess for risk factors and for the general condition of the patient. 1 Pituitary Apoplexy 5 Pituitary endocrine evaluation is necessary to diagnose secretory pituitary adenomas as well as hypopituitarism. An initial random cortisol, ACTH, LH, FSH, testosterone or estradiol, FT4, TSH, IGF-1, and prolactin should be measured immediately upon the diagnosis of PA to screen for a hyperfunctioning pituitary ade- noma. Low serum prolactin at presentation is seen in patients with the highest intrasellar pressure, who are less likely to recover pituitary function. It is important to emphasize that blood sam-

ples for ACTH and cortisol measurements should be obtained prior to the administration of steroids. The hypothalamic-pitu- itary-adrenal axis usually responds to critical illness with an increase in serum cortisol levels, and it is expected that a random cortisol level will similarly be elevated during the acute phase of PA without hypopituitarism. There is no clearly agreed-upon cut- off for random cortisol levels for the diagnosis of adrenal insuf- ficiency during the acute phase of PA, but studies have shown that in patients with PA and proven central adrenal insufficiency, cortisol levels are very low. Approximately 40–70% of patients with PA have thyrotropin or gonadotropin deficiency. Hormone replacement of these defi- ciencies can begin when the patient has recovered from the acute illness. GH deficiency is seen in almost all patients but is not always tested or treated. Diabetes insipidus is present in less than 5% of patients with PA and may be further masked by the development of secondary adrenal failure and/or hypothyroidism. Imaging CT is usually the initial imaging modality performed for patients with sudden onset of headache. CT is useful to rule out subarach- noid hemorrhage and can detect a sellar mass in up to 80% of cases. In 20–30% of cases, the CT scan will detect hemorrhage into the pituitary mass, confirming the diagnosis of PA. Magnetic resonance imaging (MRI) is the imaging procedure of choice and has been found to identify an underlying tumor, if present, in over 90% of the cases. Therefore, MRI is more sensitive for the diag- nosis of PA and should be done in all patients with suspected 6 A. P. Abreu and U. B. Kaiser PA. MRI can detect hemorrhagic and necrotic areas and can show the relationship between a tumor and adjacent structures such as the optic chiasm, cavernous sinus, and hypothalamus (Fig. 1.1). However, conventional MRI sequences may not detect an infarct a b c d Fig. 1.1 MRI of a patient with pituitary apoplexy. Images were obtained approximately 24 hours after onset of symptoms (sudden headache, nausea, vomiting, and fatigue). Images show enlargement of the pituitary gland, which contains a fluid hematocrit level. (a) T1 sagittal, the upper margin of the pituitary gland is contacting and slightly displacing the optic chiasm supe- riorly. (b) Axial pre-contrast. (c) T1 axial post gadolinium. (d) T2 axial 1 Pituitary Apoplexy 7 for up to 6 hours after the acute event. Diffusion-weighted imag- ing (DWI) is a commonly performed MRI sequence for the detec- tion of small infarcts and initial hemorrhage and can be very helpful in the early phases of PA. Thickening of the sphenoid mucosa in the sphenoid sinus beneath the sella turcica has been reported during the acute phase of PA and corresponds with marked mucosal swelling from increased pressure in the venous sinuses draining the sinus area. Such mucosal thickening has been shown to correlate with worse neurological and endocrinological outcomes. Management Patients with PA should be managed by neurosurgeons and endo- crinologists in a hospital with an acute care neurosurgical unit available and with access to ophthalmological evaluation. Consider Initiation of Corticosteroid Treatment PA can be a true medical emergency. The course of PA is variable and management will depend on a patient’s clinical condition. The first intervention is hemodynamic stabilization and correction of electrolyte disturbances. As corticotropin deficiency is present in the vast majority of patients and may be life-threatening, corti- costeroids should be administered intravenously as soon as the diagnosis is confirmed and blood is collected for cortisol and ACTH measurement. A bolus of 100 mg (some studies recom- mend 200 mg) of hydrocortisone followed by 50–100 mg IV every 6–8 hours is given; alternatively, 2–4 mg/h by continuous administration should be given. There are no randomized trials comparing different doses, so the ideal dose of hydrocortisone administration is not known. Dexamethasone may be used instead of hydrocortisone to reduce edema as a part of a conservative approach for treatment of PA. Although the majority of the litera- ture recommends empiric corticosteroid treatment for all patients with diagnosis of PA, the UK guidelines for the management of 8 A. P. Abreu and U. B. Kaiser pituitary apoplexy recommend steroid therapy in patients with hemodynamic instability, altered level of consciousness, reduced visual acuity, and severe visual field defects, or if a 9:00 am corti- sol is less than 18 mcg/dL. Acute Intervention: Surgery vs. Conservative Treatment Most cases of PA improve with either surgical or expectant man- agement, but the most appropriate approach in the acute phase is controversial. Studies comparing the two modalities are retro- spective and suffer from selection bias. The ideal surgical treat- ment is via the transsphenoidal approach. One important factor to consider is the risk of surgery, and in the acute setting, the opera- tion may be performed by the on-call neurosurgeon rather than by skilled pituitary surgeons; this may increase the risk of complica- tions. Studies suggest that the posttreatment prevalence of pitu- itary deficiency is similar after either treatment modality. The endocrine prognosis is poorer in patients with pituitary adenoma and PA than in uncomplicated pituitary adenoma, as pituitary damage more commonly occurs during the acute apoplectic event. Studies suggest that visual field defects improve or normalize in most patients regardless of the treatment modality. However, it is the general consensus – and is the recommendation by the UK guidelines for the management of PA – to consider surgical treat- ment in patients with severe neuro-ophthalmologic signs such as severely reduced visual acuity or severe and persistent or deterio- rating visual field defects. A deteriorating level of consciousness is also an indication for surgical treatment. Studies suggest, although one should keep selection bias in mind, that patients treated conservatively have better outcomes with regard to ocular palsies. Resolution of ocular paresis resulting from involvement of cranial nerve III, IV, or VI is usually seen within days to weeks, and it is not an indication for surgery. Surgery should be per- formed within 7 days of the onset of the symptoms. One study showed that the prognosis of visual defects is less favorable when surgery is done more than a week after onset. A Pituitary Apoplexy 1 Pituitary Apoplexy 9 Table 1.1 Pituitary Apoplexy Score Variable Points Level of consciousness Glasgow Coma Scale 15 0 Glasgow Coma Scale 8–14 2 Glasgow Coma Scale <8 4 Visual acuity Normal 0 Reduced unilateral 1 Reduced bilateral 2 Visual field defects Normal 0 Unilateral defect 1 Bilateral defect 2 Ocular paresis Absent 0 Present unilateral 1 Present bilateral 2 Reprinted with permission from Rajasekaran S, Vanderpump M, Baldeweg S, et al. UK guidelines for the management of pituitary apoplexy. Clin Endocri- nol;74(1):9–20, © 2011, with permission from John Wiley and Sons Score (see Table 1.1) was designed by the UK guidelines for the management of PA to enable more uniform clinical description of PA and enable better comparison between different management options. It is rare to change from conservative treatment to an opera- tive course, but urgent imaging should be done in the presence of a new or deteriorating visual field deficit or neurological dete- rioration. Reduction in tumor size is frequent after apoplexy, and follow- up imaging can show empty sella, partially empty sella, or even normal pituitary. The tumor recurrence rate is similar with both treatment modalities, and it has been shown to be approximately 6%. Therefore, long-term surveillance is recommended. Patients with simple infarction on MRI typically have less severe clinical features and better outcomes than those with hemorrhage or hem- orrhagic infarction. 10 A. P. Abreu and U. B. Kaiser Postoperative Care Postoperative management of patients following surgery for PA is similar to that of elective pituitary surgery for pituitary tumors. In some cases, patients may not have had a complete evaluation prior to surgery, and the pituitary function status will not be known. An early postoperative CT or sellar MRI should be performed in any patient with a new or worsened neurological deficit such as visual deterioration or diplopia and in anyone with significant rhinorrhea and a suspected CSF leak. Monitor for Signs and Laboratory Abnormalities Suggestive of Diabetes Insipidus (DI) Alterations in sodium and fluid balance are relatively common in the early postoperative phase. The classic reported triphasic response, in which patients initially develop DI in the first 24 to 48 hours, followed by transient SIADH developing 4–10 days postoperatively, followed by the return of DI in a matter of weeks, is not the most common pattern seen postpituitary surgery but can occur. More often, patients present with DI within the first days postsurgery and then either recover completely or develop SIADH about 5 days postsurgery or later. Fluid balance, serum electro- lytes, urea, creatinine, and plasma and urine osmolality should be monitored closely during the first week postsurgery. During the first 2 days after surgery, fluid balance, electrolytes, and urine and serum osmolality should be checked every 8–12 hours; thereafter, further monitoring will depend on the patient’s clinical status. DI is present in about 5% of the patients after PA but can be seen in up to 25% of patients undergoing transsphenoidal pitu- itary surgery. In most cases, the patients may develop transient DI but do not require any therapy. They should be allowed to drink to thirst and their serum sodium should be monitored closely. When treatment is needed, desmopressin (DDAVP) should be given sub- 1 Pituitary Apoplexy 11 cutaneously or intravenously (0.5–2 mcg every 24 hours as needed), or alternatively, an oral formulation can be given (often starting with 0.1–0.2 mg orally as a single evening dose, with doses up to 0.3 mg orally three times daily sometimes needed). Intranasal DDAVP is not generally used acutely in patients who have undergone transsphenoidal surgery until after the nose has healed and nasal congestion has improved. SIADH, when it occurs, usually presents 4 to 10 days postoperatively and can often be treated with fluid restriction and close monitoring. Assess Pituitary Reserve As discussed above, most of the patients will receive corticoste- roid treatment during the acute phase of PA. The dose should be tapered to replacement doses when the patient is clinically stable. In patients without a previous diagnosis of adrenal insufficiency, a morning fasting cortisol should be checked on day 2 or 3 after surgery to assess residual postpituitary infarction and post-steroid treatment reserve after the acute event of PA and postoperatively. Hydrocortisone should be held for at least 24 hours prior to mea- suring cortisol levels. In patients with known and documented cortisol deficiency before surgery, a morning cortisol level should be checked within 4 to 8 weeks to determine if they will need long-term steroid treatment. TSH and free T4 (FT4) should be checked on day 3 or 4 post- operatively, and thyroid hormone replacement should be consid- ered if deficient. The interpretation of thyroid function tests postsurgically should be careful as “sick euthyroid syndrome” can alter TSH and FT4 hormone levels and affect the interpretation of these tests. Thyroid deficiency may take several weeks to be diag- nosed given thyroid gland reserve and the half-life of T4, so thy- roid function can be normal in the immediate postoperative period; hence it is important to test it again ~4–8 weeks postop- eratively or if symptoms of hypothyroidism develop. 12 A. P. Abreu and U. B. Kaiser Visual Assessment Visual fields, eye movements, and visual acuity should be exam- ined at the bedside as soon as the patient can cooperate with the examination, ideally within 48 hours. A formal visual field assess- ment using a Humphrey analyzer or Goldmann perimetry should be performed within a few weeks after the acute event. Follow-Up After Discharge Check Electrolytes After 1 Week Patients should be seen in follow-up within 1 week of surgery to have sodium, thyroid function, ACTH, morning cortisol, and urine osmolality tested. As discussed earlier, patients may develop SIADH up to 10 days after surgery or after PA and should be monitored closely. Reassess Pituitary Function After 4 to 8 Weeks Hypopituitarism (discussed separately) is one of the complica- tions of PA and may not be detected during the acute phase of PA. Thyroid deficiency may take several weeks to be diagnosed given thyroid gland reserve and T4 half-life. All patients should be seen 4–8 weeks after presenting with PA for

evaluation of pitu- itary function. On the other hand, some pituitary hormonal defi- ciencies may recover postoperatively, and such recovery can also be assessed as part of this evaluation. Studies have shown partial or complete recovery of pituitary function in up to 50% of patients. In most cases, patients will be treated with glucocorticoids during the acute episode of PA; the long-term need for glucocorticoid replacement therapy should be determined at this time. Thyroid, adrenal, gonadal, and GH axes may be assessed at this visit. Patients should also have formal visual field, visual acuity, and eye movement assessment. 1 Pituitary Apoplexy 13 Patients treated for apoplexy should have at least annual bio- chemical assessment of pituitary function, which should usually include FT4, TSH, LH, FSH, testosterone in men, estradiol in women, prolactin, IGF-1, and dynamic tests of cortisol and growth hormone secretion if clinically appropriate. Suggested Reading Bonicki W, Kasperlik-Zaluska A, Koszewski W, Zgliczynski W, Wislawski J. Pituitary apoplexy: endocrine, surgical and oncological emergency. Incidence, clinical course and treatment with reference to 799 cases of pituitary adenomas. Acta Neurochir. 1993;120(3–4):118–22. Briet C, Salenave S, Bonneville JF, Laws ER, Chanson P. Pituitary apoplexy. Endocr Rev. 2015;36(6):622–45. https://doi.org/10.1210/er.2015-1042. Briet C, Salenave S, Chanson P. Pituitary apoplexy. Endocrinol Metab Clin N Am. 2015;44:199–209. Loh JA, Verbalis JG. Diabetes insipidus as a complication after pituitary sur- gery. Nat Clin Pract Endocrinol Metab. 2007;3(6):489–94. Rajasekaran S, Vanderpump M, Baldeweg S, et al. UK guidelines for the management of pituitary apoplexy. Clin Endocrinol. 2011;74:9–20. Randeva HS, Schoebel J, Byrne J, Esiri M, Adams CB, Wass JA. Classical pituitary apoplexy: clinical features, management and outcome. Clin Endocrinol. 1999;51:181–8. Semple PL, Jane JA Jr, Laws ER Jr. Clinical relevance of precipitating factors in pituitary apoplexy. Neurosurgery. 2007;61:956–61; discussion 61-2. Sibal L, Ball SG, Connolly V, et al. Pituitary apoplexy: a review of clinical presentation, management and outcome in 45 cases. Pituitary. 2004;7:157–63. Panhypopituitarism 2 Ana Paula Abreu and Ursula B. Kaiser Contents Definition and Significance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16 Identify Causes of Hypopituitarism . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16 Mass Lesions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16 Traumatic Brain Injury . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18 Medications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18 Hypophysitis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18 Systemic Diseases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20 Genetic Causes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20 Vascular . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21 Diagnosis of Hypopituitarism in the Hospital . . . . . . . . . . . . . . . . . . . . 21 Assess Anterior Pituitary Function . . . . . . . . . . . . . . . . . . . . . . . . . . 21 Assess Posterior Pituitary Function . . . . . . . . . . . . . . . . . . . . . . . . . . 24 Imaging . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24 Management of Hypopituitarism in the Hospital . . . . . . . . . . . . . . . . . . 25 Management of Hypopituitarism at the Time of Discharge . . . . . . . . . . 26 Suggested Reading . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26 A. P. Abreu (*) · U. B. Kaiser Division of Endocrinology, Diabetes and Hypertension, Department of Medicine, Brigham and Women’s Hospital and Harvard Medical School, Boston, MA, USA e-mail: apabreu@bwh.harvard.edu; ukaiser@bwh.harvard.edu © Springer Nature Switzerland AG 2020 15 R. K. Garg et al. (eds.), Handbook of Inpatient Endocrinology, https://doi.org/10.1007/978-3-030-38976-5_2 16 A. P. Abreu and U. B. Kaiser Definition and Significance Hypopituitarism is the inability of the pituitary gland to provide sufficient hormones for the needs of the individual. It is the result of the failure in either the production or secretion in one or more pituitary hormones. The diagnosis of hypopituitarism is important in the hospital because some hormone deficiencies, such as ACTH, pose significant risk to the patient’s life and need to be treated. Also, it is crucial to diagnose and treat diabetes insipidus as it can cause hypernatremia, severe dehydration, coma, and death. The diagnosis of central hypothyroidism is challenging in the hospital, but thyroid hormone should be replaced in patients with secondary hypothyroidism. On the other hand, other pitu- itary hormone deficiencies do not pose acute risk to patient’s life, and replacement may be postponed to the outpatient setting. Identify Causes of Hypopituitarism A high diagnostic suspicion is necessary to identify patients not previously diagnosed with hypopituitarism. Therefore, it is impor- tant to know what causes hypopituitarism in order to detect it (Table 2.1). Insults in the regulation, production, or secretion of any pituitary hormones can result in pituitary insufficiency. Physiological secretion of pituitary hormones relies on intact function of the hypothalamus. Mass Lesions Any structural disruption of the hypothalamic-pituitary region can cause decreased production or secretion of the hormones. Pituitary tumors are the most common cause of hypopituitarism, but any other tumor occupying the region can also cause pituitary dysfunc- tion (Table 2.1). Mechanical compression of portal vessels and the 2 Panhypopituitarism 17 Table 2.1 Causes of hypopituitarism Structural causes: Mass lesions Pituitary adenoma Craniopharyngioma Rathke’s cleft cyst Metastatic disease Lymphomas, germinomas, and other tumors Infiltrative diseases Hypophysitis (lymphocytic and others) Sarcoidosis Hemochromatosis Tuberculosis and other infections Syphilis Vascular events Pituitary apoplexy Sheehan’s syndrome (infarction of the pituitary gland after postpartum hemorrhage) Intra-sellar carotid artery aneurysm Traumatic injury Traumatic brain injury Perinatal trauma Neurosurgery Radiation Functional causes: Medications Glucocorticoids Megestrol acetate Immunotherapy – CTLA-4 inhibitors/PDL1 antibodies Opioids GnRH agonists Systemic diseases Chronic illness Anorexia nervosa Developmental and Several genetic defects can cause isolated or inherited genetic causes combined pituitary deficiency pituitary stalk, with resulting ischemic necrosis, is thought to be the predominant mechanism by which mass lesions cause hypo- pituitarism. Hyperprolactinemia in non-prolactin producing tumors is common with pituitary macroadenomas, given the dis- ruption of the normal suppressive effects of dopamine from the hypothalamus. https://www.facebook.com/groups/2202763316616203 18 A. P. Abreu and U. B. Kaiser Traumatic Brain Injury This is an underestimated cause that can cause hypopituitarism even years after the trauma. Given the frequency of traumatic brain injury in the general population, it is important not to over- look this important cause of hypopituitarism. Medications Several medications can cause hypopituitarism. Chronic use of systemic corticosteroids inhibits the hypothalamic-pituitary- adrenal axis and is a common cause of central adrenal insuffi- ciency. This is very relevant for patients admitted to the hospital for acute disorders or for procedures, as they will need to receive higher doses of steroids (usually called stress doses of steroids) to compensate for the lack of an endogenous increase in the amount of cortisol production during stress. Chronic administration of some opioids such as fentanyl and hydromorphone has the poten- tial to cause secondary adrenal insufficiency as well as secondary hypogonadism. Immunotherapy causes hypopituitarism second- ary to hypophysitis that is discussed below. Hypophysitis Autoimmune hypophysitis and medication-induced hypophysitis can cause pituitary deficiency. Lymphocytic hypophysitis is more common in females, and more than half of the cases (57%) pres- ent during pregnancy or postpartum. Immunotherapy, used to treat melanoma, renal cell carcinoma, and other malignancies, can also cause hypophysitis. The pathophysiology and clinical pre- sentation of immunotherapy-induced hypophysitis are different from those of lymphocytic hypophysitis. The monoclonal anti- bodies, ipilimumab and tremelimumab, which bind and inhibit cytotoxic T-lymphocyte antigen-4 (CTLA-4), are reported to cause hypophysitis in 1–18% of treated patients. Most cases are caused by ipilimumab. The anti-programmed cell death protein 2 Panhypopituitarism 19 antibodies (anti-PD-1 Abs), nivolumab and pembrolizumab, rarely cause hypophysitis. Patients with hypophysitis present with headache, pituitary enlargement, and hypopituitarism (Fig. 2.1). In most patients, the pituitary enlargement eventually resolves, a b c d Fig. 2.1 MRI of a patient with lymphocytic hypophysitis. A pregnant woman presented to the emergency room at 30 weeks of gestational age with a 2-week history of worsening headaches and blurred vision in the last 24 hours: (a) coronal and (b) sagittal images without contrast showing enlargement of the pituitary gland, which measures approximately 1.4 cm in craniocaudal dimension. The gland has a convex superior border and is protruding into the suprasellar cistern. Since patient was pregnant, no contrast was given. Coro- nal images 3 months after the initial images and 1 month postpartum: (c) pre-gadolinium and (d) post-gadolinium. Images show interval decrease in size of anterior pituitary, which is no longer enlarged and now measures approximately 6 to 7 mm and demonstrates a flat superior surface 20 A. P. Abreu and U. B. Kaiser but hypopituitarism is usually permanent. The most common hor- mone deficiency in classical autoimmune lymphocytic hypophysi- tis is ACTH deficiency, seen in approximately 32% of the patients. Eighteen percent present with hyperprolactinemia and 31% develop diabetes insipidus. In anti-CTLA-4-induced hypophysitis, around 85% of the patients presented thyrotroph and gonadotroph deficiency, while 73% had corticotroph deficiency. In these cases, pituitary function recovered in approximately 25% of cases only. Systemic Diseases Infiltrative systemic diseases such as sarcoidosis, hemochromato- sis, and rarely infiltrative infections such as tuberculosis can cause hypopituitarism. Neurosarcoidosis typically presents with DI. Systemic diseases can also cause functional hypopituitarism, but the significance of the disruption of pituitary function in this context is not always completely understood. Functional hypopi- tuitarism in some cases is thought to be an appropriate response of the organism to insults. It is important to understand that some hormonal changes seen in admitted patients do not require treat- ment. One example frequently seen in the hospital is euthyroid sick syndrome, in which critically ill patients have functional sec- ondary hypothyroidism with low TSH, T4, and T3 levels. Based on current knowledge, there is no indication for thyroid hormone replacement in these patients, and the thyroid hormone levels nor- malize when they recover from the acute phase of their illness. Chronically ill and malnourished patients frequently have central hypogonadism that also does not require treatment in the acute setting. For these reasons, it is not

useful to measure gonadal and thyroid function in acutely ill patients. Similarly, the growth hor- mone axis is not assessed in hospitalized patients. Genetic Causes Mutations in several genes involved in pituitary development and differentiation, and hormone production and secretion are 2 Panhypopituitarism 21 associated with isolated or combined pituitary hormone defi- ciency. Genetic defects in genes associated with pituitary gland development can also cause pituitary hypoplasia, aplasia, or other midline defects. Most of these patients have a known diagnosis of hypopituitarism upon admission to the hospital. Vascular Vascular events, including pituitary apoplexy and Sheehan’s syn- drome as discussed elsewhere, can also disrupt pituitary function. Diagnosis of Hypopituitarism in the Hospital The clinical manifestations of hypopituitarism can vary greatly depending on the axis affected, age, gender, and clinical status of the patient. Symptoms of hypopituitarism in acutely ill patients can be particularly challenging to recognize, given the masking of some subtle symptoms by other complications. Assess Anterior Pituitary Function Adrenal Axis As stated above, central adrenal insufficiency must not be missed in a hospitalized patient given the risk that it poses to the patient’s life. Patients with adrenal insufficiency may have longstanding nonspecific symptoms. In comparison with primary adrenal insuf- ficiency, patients with central adrenal insufficiency have relative sparing of aldosterone secretion due to the preservation of renin and angiotensin control of aldosterone production. With this resid- ual aldosterone production, severe hypotension and hyperkalemia are less common. Nonetheless, they can still present with hemody- namic instability despite adequate fluid resuscitation, most often associated with a hyperdynamic circulation and decreased systemic vascular resistance. This is an important diagnostic clue and should trigger investigation for adrenal insufficiency. 22 A. P. Abreu and U. B. Kaiser The combined occurrence of hypoglycemia, hyponatremia, and eosinophilia should alert the clinician to the possibility of hypoad- renalism. Patients with central adrenal insufficiency do not have the characteristic hyperpigmentation that classically occurs in patients with primary adrenal insufficiency, resulting from accu- mulation of proopiomelanocortin (POMC). The diagnosis of adrenal insufficiency in the hospitalized patients is challenging not only because of the lack of specific symptoms or clinical signs but also due to the difficulties estab- lishing cutoff values for cortisol levels in acutely ill patients. Expected cortisol levels vary with the type and severity of disease, making it difficult to define normal ranges. Patients admitted to the hospital do not have the classical circadian rhythms with higher cortisol levels early in the morning and lower levels at night. Many threshold levels have been proposed for the definition of an insufficient cortisol level (measured at any time of day) dur- ing acute illness, but none is entirely satisfactory. In critically ill patients, cortisol levels are usually elevated, and a serum cortisol value of ≥18 mcg/dL (497 nmol/L) effectively rules out adrenal insufficiency. Patients with central adrenal insufficiency have low cortisol levels with inappropriately normal or low ACTH levels. Given the lack of a cutoff value for random cortisol levels, an ACTH stimulation test can be performed to confirm the diagnosis of adrenal insufficiency if basal levels are equivocal. Cosyntropin, a synthetic ACTH corresponding to amino acids 1–24 of ACTH that has full biologic potency, is used to evaluate the capacity of the adrenal gland to produce cortisol. The ACTH stimulation or cosyntropin test consists of measuring serum cortisol immedi- ately before and 30 and 60 minutes after intravenous or intramus- cular injection of 250 mcg (85 nmol or 40 international units) of cosyntropin. Serum cortisol concentration ≥18 to 20 mcg/dL (500 to 550 nmol/L) after the injection indicates normal adrenal func- tion. It is important to highlight that patients with new onset of central adrenal insufficiency may have an appropriate response to cosyntropin stimulation because the adrenal gland will respond to an ACTH stimulus normally. Therefore, a normal response to 2 Panhypopituitarism 23 cosyntropin test does not rule out central adrenal insufficiency, and physicians will have to rely on basal cortisol and ACTH levels and clinical judgment. Patients with adrenal atrophy resulting from chronically low stimulation by endogenous ACTH will have an abnormal response in the cosyntropin test. It is also important to note that hydrocortisone, prednisone, and several other corticosteroids cross-react with the assays used to measure cortisol and interfere with the assay results. Dexamethasone is not measured by the cortisol assays, but it is a strong inhibitor of the hypothalamic-pituitary axis and has a bio- logical effect for almost 54 hours. For these reasons, cortisol lev- els should be interpreted with caution in patients who are currently receiving or recently received corticosteroids. Thyroid Axis As discussed above, the diagnosis of central hypothyroidism is challenging in the hospital, but thyroid hormone should be replaced in patients with secondary hypothyroidism in the hospi- tal. If there is a clinical suspicion of pituitary dysfunction, TSH and FT4 should be measured. Patients with central hypothyroid- ism have low FT4 levels combined with inappropriately normal TSH levels. While this scenario can be seen in patients with “sick euthyroid syndrome,” the presence of any known cause of hypo- pituitarism (see Table 2.1) would corroborate the diagnosis of central hypothyroidism. Other Axes Growth hormone and gonadotropin deficiencies usually do not pose risk to the patient’s life, and thus there is usually no indica- tion for testing these axes in hospitalized patients. Prolactin mea- surement can help diagnosing hypophysitis, as the levels can be low in this condition. All individuals with mechanical compres- sion of the pituitary stalk can present with elevated prolactin lev- els; however, levels above 100 ng/dL are more suggestive of a prolactinoma. It is also important to keep in mind that several medications can increase prolactin levels. 24 A. P. Abreu and U. B. Kaiser Assess Posterior Pituitary Function Patients with central DI are unable to concentrate urine. The diag- nosis is particularly challenging in the hospital when patients fre- quently receive significant amounts of intravenous fluids for resuscitation and have increased volumes of dilute urine when fluid is being redistributed. True hypernatremia (plasma sodium concentration greater than 150 meq/L) is rare in adults with DI and no cognitive impairment, because the initial loss of water stimulates thirst, resulting in an increase in fluid intake to match the urinary losses. However, debilitated patients may not have free access to water, and some patients may have impaired thirst mechanisms. In this setting, the plasma sodium concentration can be elevated. Elevated serum sodium associated with low urine osmolality, particularly when urine osmolality is less than the plasma osmolality, is indicative of DI. The response to DDAVP treatment will differentiate between central and nephrogenic DI. Imaging Unless the patient has an unequivocal cause for a specific hor- monal deficiency, patients should have imaging of the hypothalamic- pituitary region. Pituitary adenomas do not usually cause diabetes insipidus; this is usually caused by suprasellar lesions. Pituitary microadenomas do not usually cause hypopitu- itarism. Empty sella is a radiological term. It can be seen in asso- ciation with pituitary hypoplasia/aplasia or can be a consequence of a previous insult to the pituitary gland such as apoplexy or hypophysitis. However, it can also be seen in patients with normal pituitary function. Absence of the posterior white spot may have no clinical significance or may be secondary to absence of ADH storage in the posterior hypophysis in patients with diabetes insip- idus. Hypophysitis can present as enlargement of the pituitary gland and thickening of the stalk. However, normal imaging of the pituitary does not rule out hypophysitis. 2 Panhypopituitarism 25 Management of Hypopituitarism in the Hospital Patients with a known diagnosis of hypopituitarism should continue their hormone replacement when admitted to the hospital, but some adjustments to the therapy are frequently necessary. The most important treatment is corticosteroid therapy. Critically ill patients should receive stress dose steroids. A bolus of 100 mg of hydrocortisone followed by 50–100 mg IV every 6–8 hours is given; alternatively, 2–4 mg/h by continuous admin- istration should be given. Dose should be tapered down to main- tenance oral dose as the clinical condition improves. Patients undergoing surgical procedures or immobilized patients at higher risk of deep venous thrombosis should discon- tinue estrogen therapy. Growth hormone and testosterone therapy are often discontinued during hospitalization, depending on spe- cific situations. Thyroid replacement should be given to patients in the hospital, and special attention should be given to patients on enteral diets, given the need for it to be taken on an empty stom- ach for proper absorption. Patients with significant edema may have impaired absorption, and patients with proteinuria have increased wasting of thyroid hormones and may need adjustments in their dose of levothyroxine. Patients on DDVAP treatment should receive it while in the hospital and should be carefully monitored. Mental status alterations due to anesthesia or current disease can affect the thirst mechanism and interfere with appro- priate water intake, requiring adjustment of the DDAVP dose. Patients are often NPO or have nasal tubes, and the DDAVP route of administration may need to be adjusted. If DDAVP cannot be administered intranasally or orally, it can be given subcutaneously or intravenously. A usual antidiuretic dose is 0.5 to 2 mcg admin- istered subcutaneously or intravenously; the duration of action, as judged by increased urine osmolality, will be 12 hours or more. Some patients do not respond well to subcutaneous DDAVP due to inadequate absorption. 26 A. P. Abreu and U. B. Kaiser Management of Hypopituitarism at the Time of Discharge Patients with hypopituitarism should be discharged on hormone replacement therapy. Patients on stress doses of steroids should be tapered to a replacement dose (around 3–4 mg of prednisone or 15 mg of hydrocortisone daily) as soon as they are clinically sta- ble. Those with a new diagnosis of hormone deficiency should be reevaluated for the need of replacement therapy. If the diagnosis of hypopituitarism was made in an acute setting, after brain trauma or surgery, for example, patients may recover pituitary function and may not need long-term replacement or may need adjustment of dose of hormonal replacement. The appropriate time for post-discharge follow-up will depend on the specific hor- monal deficiency and the cause of hypopituitarism. A follow-up visit 6 weeks after discharge to evaluate pituitary status may be ideal in most cases. Suggested Reading Benvenga S, Campenni A, Ruggeri RM, Trimarchi F. Clinical review 113: Hypopituitarism secondary to head trauma. J Clin Endocrinol Metabol. 2000;85:1353–61. Caturegli P, Newschaffer C, Olivi A, Pomper MG, Burger PC, Rose NR. Autoimmune hypophysitis. Endocr Rev. 2005;26:599–614. Schneider HJ, Aimaretti G, Kreitschmann-Andermahr I, Stalla GK, Ghigo E. Hypopituitarism. Lancet. 2007;369:1461–70. Thodou E, Asa SL, Kontogeorgos G, Kovacs K, Horvath E, Ezzat S. Clinical case seminar: lymphocytic hypophysitis: clinicopathological findings. J Clin Endocrinol Metabol. 1995;80:2302–11. Postoperative Management 3 After Pituitary Surgery Anna Zelfond Feldman and Pamela Hartzband Contents Assess Hormone Status Before Surgery if Possible . . . . . . . . . . . . . . . . 28 Intraoperative/Postoperative Steroids . . . . . . . . . . . . . . . . . . . . . . . . . . . 28 Patient with Unknown Adrenal Function Prior to Surgery . . . . . . . . . 28 Patients Known to Have Preexisting Adrenal Insufficiency . . . . . . . . 29 Patients Known to Have Normal Adrenal Function Preoperatively and Patients with Cushing’s Disease . . . . . . . . . . . . . . . . . . . . . . . 29 Diabetes Insipidus/Sodium Management . . . . . . . . . . . . . . . . . . . . . . . . 30 Arrange Endocrine Follow-Up Within 1–2 Weeks of Discharge . . . . . . 31 Suggested Reading . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32 A. Z. Feldman (*) Harvard Medical School, Beth Israel Deaconess Medical Center, Department of Endocrinology, Boston, MA, USA e-mail: afeldma1@bidmc.harvard.edu P. Hartzband Harvard Medical School, Beth Israel Deaconess Medical Center, Department of Endocrinology and Metabolism, Boston, MA, USA e-mail: phartzba@bidmc.harvard.edu © Springer Nature Switzerland AG 2020 27 R.

K. Garg et al. (eds.), Handbook of Inpatient Endocrinology, https://doi.org/10.1007/978-3-030-38976-5_3 28 A. Z. Feldman and P. Hartzband Assess Hormone Status Before Surgery if Possible – Ideally patient should be seen by endocrinology as outpatient prior to surgery. Intraoperative/Postoperative Steroids Patient with Unknown Adrenal Function Prior to Surgery – Day of surgery (pre- or intra-op): Hydrocortisone 100 mg IV q 8 h (alternative methylprednisolone 60 mg IV × 1). – POD 1: Hydrocortisone 100 mg IV × 1 at 8 am and hydrocor- tisone 50 mg IV × 1 at 4 pm (alternative methylprednisolone 30 mg IV × 1), and then hold steroids. – POD 2: Draw 8 am fasting cortisol level. After blood is drawn (while results pending), give hydrocortisone 50 mg IV q 12 h (alternative methylprednisolone 20 mg IV × 1). – POD 3: If a.m. cortisol drawn on POD 2 is >/= 10, stop steroids. If a.m. cortisol drawn on POD 2 is <10 (or results pending, or patient received steroids within 12 hours prior to cortisol assessment), start oral hydrocortisone 15 mg morning and 5 mg afternoon (alternative prednisone 5 mg morning). Higher doses of steroids may be needed depending on clinical status. – Discharge: Discharge patient on hydrocortisone or prednisone if indicated as above. – Additional outpatient assessment of HPA axis should be done in approximately 2 weeks at endocrine follow-up. 3 Postoperative Management After Pituitary Surgery 29 Patients Known to Have Preexisting Adrenal Insufficiency – Day of surgery (pre-op): hydrocortisone 100 mg IV q 8 h (alternative methylprednisolone 60 mg IV × 1). – POD 1: Hydrocortisone 50 mg IV × q 8 h (alternative methyl- prednisolone 30 mg IV × 1). – POD 2: Hydrocortisone 50 mg IV q 12 h (alternative methyl- prednisolone 20 mg IV × 1). – POD 3: Start oral hydrocortisone 15 mg morning and 5 mg afternoon (alternative prednisone 5 mg morning). Higher doses of steroids may be needed depending on clinical status. – Discharge: No assessment of adrenal function in the hospital; patient should be discharged on steroids as above. Patients Known to Have Normal Adrenal Function Preoperatively and Patients with Cushing’s Disease – Day of surgery: Do not give pre-, peri-, or postoperative glucocorticoids. – Monitor patient closely for adrenal insufficiency. – POD 1: Check 8 a.m. fasting cortisol. After blood is drawn (while results pending), give hydrocortisone 50 mg IV × q 8 h (alternative methylprednisolone 30 mg IV × 1). – POD 2: If a.m. cortisol drawn on POD 2 is >/= 10, stop steroids. If a.m. cortisol drawn on POD 2 is <10 (or results pending or patient received steroids within 12 hours prior to cortisol assessment), give hydrocortisone 50 mg IV q 12 hours (alter- native methylprednisolone 20 mg IV × 1). 30 A. Z. Feldman and P. Hartzband – POD 3: If patient still on steroids: Start oral hydrocortisone 15 mg morning and 5 mg afternoon (alternative prednisone 5 mg morning). Higher doses of steroids may be needed depending on clinical status. – Discharge: Endocrinologist to follow up 8 a.m. cortisol drawn on POD1 prior to discharge and decide on steroid management. Diabetes Insipidus/Sodium Management – Day of surgery: Check serum Na and urine osmolality or spe- cific gravity post-op. – POD 1–2: Check serum Na and urine osmolality or specific gravity bid. Increase frequency if clinically indicated. Reduce to daily if clinically stable. – Strict ins/outs with special attention to urine output (UOP): Urine output should be measured hourly (if cathe- ter) or Q 2 hours (if no catheter) postoperatively. – If UOP is >250 cc/h × 2 or more hours: Send serum Na and urine osmolality or specific gravity. – If labs consistent with DI with Na >145 with urine specific gravity <1.005 or osmolality <300: – Give DDAVP 1 mcg IV × 1. – If patient getting NS IVF, change IVF to one-half NS (alter- native D5W). – Monitor serum Na, urine osmolality, or specific gravity q 4–6 hours until either DI resolves or Na normalizes and is stable, and then decrease frequency of labs. – Repeated doses of DDAVP IV may be needed. However, DI post-transsphenoidal surgery is often transient. If DI is per- sistent, patient may need oral daily, BID, or rarely TID doses of DDAVP. Intranasal DDAVP is contraindicated in the immediate postoperative period. 3 Postoperative Management After Pituitary Surgery 31 – If discordant labs with Na >145 with urine specific gravity >1.005 or osmolality >300 or Na normal with urine specific gravity <1.005 or osmolality <300: – Monitor closely, but DDAVP is not necessarily indicated. – Clinical judgment must be used in these cases about DDAVP administration. – Discontinue IVF as quickly as possible and allow patient to drink to thirst. – If thirst mechanism is not intact, match PO intake to urine output. – Discharge: If DI is persistent and patient is discharged on oral DDAVP, a serum Na level should be checked within 1–2 days of discharge, as some patients will develop transient syndrome of inappropriate diuretic hormone secretion (SIADH). – These patients need close follow-up postoperatively, ideally within 1 week of discharge. Arrange Endocrine Follow-Up Within 1–2 Weeks of Discharge – Order 8 a.m. cortisol and Na to be done fasting 1–2 days prior to endocrine follow-up. – If the patient is discharged on hydrocortisone, advise patient to hold this medication the afternoon prior and the morning of the lab draw and to take it immediately after labs are drawn that morning. – If the patient is discharged on prednisone, advise patient to hold this medication on day of lab draw and take it immedi- ately after labs are drawn that morning. – Urine osmolality or specific gravity may also be ordered if clinically indicated. 32 A. Z. Feldman and P. Hartzband Suggested Reading Fleseriu M, Hashim IA, Karavitaki N, Melmed S, Murad MH, Salvatori R, Samuels MH. Hormonal replacement in hypopituitarism in adults: an endocrine society clinical practice guideline. J Clin Endocrinol Metabol. 2016;101(11):3888–921. Woodmansee WW, Carmichael J, Kelly D, Katznelson L. AACE Neuroendocrine and Pituitary Scientific Committee. American Association of Clinical Endocrinologists and American College of Endocrinology Disease State Clinical Review: Postoperative Management Following Pituitary Surgery. Endocr Pract. 2015;21(7):832–8. Ziu M, Dunn IF, Hess C, Fleseriu M, Bodach ME, Tumialan LM, et al. Congress of neurological surgeons systematic review and evidence-based guideline on posttreatment follow-up evaluation of patients with nonfunc- tioning pituitary adenomas. Neurosurgery. 2016;79(4):E541–3. Severe Thyrotoxicosis 4 and Thyroid Storm Melissa G. Lechner and Trevor E. Angell Contents Performing the History for Thyrotoxicosis . . . . . . . . . . . . . . . . . . . . . . 34 Assess Symptoms of Thyrotoxicosis . . . . . . . . . . . . . . . . . . . . . . . . . 34 Assess for the Etiology of Thyrotoxicosis . . . . . . . . . . . . . . . . . . . . . 34 Assess for Medications That Affect Thyroid Status . . . . . . . . . . . . . 35 Performing the Physical Exam for Thyrotoxicosis . . . . . . . . . . . . . . . . 35 Key Findings in Thyrotoxicosis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35 Specific Exam Findings in Different Causes of Thyrotoxicosis . . . . 35 Assessing for Thyroid Storm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36 Making a Diagnosis of Thyroid Storm . . . . . . . . . . . . . . . . . . . . . . . . . 36 Obtain Thyroid Function Tests . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37 Biochemical Findings in Thyrotoxicosis . . . . . . . . . . . . . . . . . . . . . . . . 37 M. G. Lechner David Geffen School of Medicine, University of California at Los Angeles, Division of Endocrinology, Diabetes and Metabolism, Los Angeles, CA, USA e-mail: mlechner@mednet.ucla.edu T. E. Angell (*) Keck School of Medicine, University of Southern California, Division of Endocrinology and Diabetes, Los Angeles, CA, USA e-mail: trevor.angell@med.usc.edu © Springer Nature Switzerland AG 2020 33 R. K. Garg et al. (eds.), Handbook of Inpatient Endocrinology, https://doi.org/10.1007/978-3-030-38976-5_4 34 M. G. Lechner and T. E. Angell Obtain Testing to Identify Underlying Illnesses . . . . . . . . . . . . . . . . . . 37 Initial Emergent Therapy for Thyroid Storm or Severe Thyrotoxicosis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38 Provide Aggressive Supportive Care . . . . . . . . . . . . . . . . . . . . . . . . . 38 Order β-Blocker Therapy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39 Order Antithyroid Drug Therapy . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39 Order Iodine Therapy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40 Order Steroid Therapy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40 Use of Adjunct Treatments in Refractory Cases . . . . . . . . . . . . . . . . 40 Treatment of Thyrotoxic Patients Without Thyroid Storm . . . . . . . . . . 41 Monitoring Clinical Response . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41 Planning Outpatient Follow-Up . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42 Suggested Reading . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42 Performing the History for Thyrotoxicosis Assess Symptoms of Thyrotoxicosis Symptoms of thyrotoxicosis frequently include sweating, heat intolerance, palpitations, fatigue, and dyspnea on exertion. Cognitive symptoms may include anxiety, hyperactivity, or diffi- culty with concentration. Reported weight loss may be modest or substantial. Note any previous history of Graves’ disease, other thyroid disorders, thyroid surgery, and radioactive iodine treat- ment. In apathetic hyperthyroidism, overt symptoms are absent or often limited to weight loss, failure to thrive, fatigue, or lethargy. In patients already taking β-blockers, thyrotoxic symptoms may be blunted. Assess for the Etiology of Thyrotoxicosis Patients with longer duration of symptoms (>3 months) likely have persistent hyperthyroidism such as Graves’ disease or auton- omous nodule(s). Patients may report diffuse thyroid enlargement 4 Severe Thyrotoxicosis and Thyroid Storm 35 in Graves’ disease, a tender thyroid in subacute (painful) thyroid- itis, or a history of thyroid nodules suggesting possible autono- mous function. Eye complaints (protrusion, inflammation, or visual changes) are seen only in Graves’ disease. Assess for Medications That Affect Thyroid Status Note use of

thyroid hormone preparations, antithyroid medication (methimazole or propylthiouracil [PTU]), exposure to iodinated contrast or iodine supplements, weight loss supplements, or other medications, including amiodarone, lithium, tyrosine kinase inhib- itors, and immune checkpoint inhibitors for cancer treatment. Performing the Physical Exam for Thyrotoxicosis Key Findings in Thyrotoxicosis Findings include sweating, skin warmth, fine tremor, low body weight and loss of muscle mass, hyperactivity, and poor attention. Heart rate (HR) may be mildly elevated (80–100 bpm) or clearly tachycardic. Supraventricular tachyarrhythmias, particularly atrial fibrillation, or signs of heart failure may be present. There is often mild systolic hypertension with widened pulse pressure. Patients may be mildly tachypneic from thyrotoxicosis alone. Rarely, patients may present with episodes of paralysis (termed paroxysmal periodic paralysis), affecting the lower before the upper extremities, proximal more than distal muscle groups, and usually sparing the diaphragm. Specific Exam Findings in Different Causes of Thyrotoxicosis A diffusely enlarged non-tender thyroid gland suggests Graves’ disease, and the presence of a thyroid bruit in combination with ophthalmopathy, pretibial myxedema, or digital clubbing is 36 M. G. Lechner and T. E. Angell pathognomonic. Exquisite thyroid tenderness indicates subacute (painful) thyroiditis. The presence of a large thyroid nodule on palpation may suggest an autonomously functioning nodule. The presence of a normal or small thyroid gland without any abnormal characteristics in the appropriate clinical setting may suggest accidental or surreptitious patient use of thyroid hormones. Assessing for Thyroid Storm The critical physical exam findings include hyperthermia, altered mentation (e.g., confusion, lethargy, seizures, coma), tachyar- rhythmias (most commonly atrial fibrillation), or congestive heart failure (e.g., elevated jugular venous pressure, lower extremity swelling, pulmonary edema, congestive hepatopathy). Other features of thyrotoxicosis described above will often be present in patients with thyroid storm but are not specific and frequently present in thyrotoxic patients without thyroid storm as well. Making a Diagnosis of Thyroid Storm Thyroid storm is a clinical diagnosis. Thyroid storm has been recognized traditionally as a clinical syndrome of thyrotoxicosis, hyperthermia, altered mentation, and a precipitating event. Other manifestations of thyrotoxicosis are often present but are not spe- cific for thyroid storm. The Burch-Wartofsky Score assigns points for dysfunction of the thermoregulatory, central nervous, gastrointestinal (GI)-hepatic, and cardiovascular systems. A score of >45 is considered highly suspicious for thyroid storm. However, this cutoff is not specific, indicating that some thyro- toxic patients without thyroid storm have a score greater than 45. The numerical score should not supplant physician judgment in making the diagnosis. 4 Severe Thyrotoxicosis and Thyroid Storm 37 Obtain Thyroid Function Tests Biochemical assessments of thyroid function are the most perti- nent laboratory results to consider. A suppressed thyroid- stimulating hormone (TSH) helps establish thyrotoxicosis. Thyroid hormone measurement, such as free thyroxine (FT4), should also be performed to confirm thyroid hormone excess, since mild TSH suppression in hospitalized patients occurs with non-thyroidal illnesses. Biochemical confirmation of thyrotoxico- sis is not necessary to diagnose and begin treatment for thyroid storm. The degree of thyroid hormone elevation or other tests are not helpful in determining which thyrotoxic patients have thyroid storm. Biochemical Findings in Thyrotoxicosis Mild hyperglycemia, hypercalcemia, normocytic anemia, and elevation in alkaline phosphatase and transaminase concentra- tions are all commonly seen in thyrotoxicosis. Serum creatinine is lower in thyrotoxicosis, leading to overestimating glomerular fil- trate rate. An elevated bilirubin is a particularly important finding since it has been correlated with adverse outcomes in thyroid storm. Obtain Testing to Identify Underlying Illnesses It is most critical to identify concurrent illnesses that may compli- cate thyrotoxicosis or be precipitants of thyroid storm. In addition to a thorough physical exam, pregnancy test should be performed if relevant, and sources of infection should be assessed through urinalysis, blood cultures, and chest imaging. Consideration of concomitant adrenal insufficiency should be assessed with a ran- dom serum cortisol, but doing so should not delay delivery of ste- https://www.facebook.com/groups/2202763316616203 38 M. G. Lechner and T. E. Angell roid treatment (discussed below). If adrenal insufficiency is suspected, more formal evaluation, such as a cosyntropin (ACTH 1–24) stimulation test, may need to be deferred until recovery from the patient’s acute presentation. Further testing, such as abdominal imaging or lumbar puncture, should be performed when clinically indicated. Other testing should evaluate for poten- tial acute coronary syndrome, hyperglycemia and diabetic keto- acidosis, and drug use (especially cocaine and methamphetamines). Initial Emergent Therapy for Thyroid Storm or Severe Thyrotoxicosis The treatment of thyroid storm should be initiated as early as pos- sible after recognition of the diagnosis. For patients with severe thyrotoxicosis who are considered to have “impending” thyroid storm based on clinical evaluation or a BWS of 25–45, similar treatment to thyroid storm may be considered. Provide Aggressive Supportive Care • Hemodynamic stabilization is critical. Management in the intensive care unit (ICU) is usually needed for patients with thyroid storm. Invasive hemodynamic monitoring is often appropriate. Intravenous fluid is typically necessary to improve perfusion in the setting of absolute or effective hypovolemia. If hypotension is not responsive to fluid resuscitation, vasopres- sors should be used. • Sedatives, narcotics, and diuretics should be used carefully because they may lower blood pressure and worsen hypoperfu- sion. • Hyperthermia can initially be treated with cooling measures and acetaminophen. Salicylates may increase free hormone levels by lowering protein binding. 4 Severe Thyrotoxicosis and Thyroid Storm 39 • Underlying illnesses should be treated. Specifically, broad- spectrum empiric antibiotics should be considered given the frequency with which infections precipitate thyroid storm. Order β-Blocker Therapy β-Adrenergic blockade improves tachycardia, cardiac workload, oxygen demand, and thyrotoxic symptoms. The HR goal is approximately 90–110 bpm rather than slower rates until thyro- toxicosis resolves. Carefully monitoring and use of shorter-acting agents may reduce the risk of cardiovascular insufficiency arising from excess β-blockade. Initial regimens include the following: • Intravenous propranolol 0.5–1.0 mg and then continuous infusion (5–10 mg/hour) • Oral propranolol 60–80 mg every 4 hours • Intravenous esmolol 0.25–0.50 mg/kg loading dose and then continuous infusion (0.05–0.1 mg/kg/minute) Order Antithyroid Drug Therapy Start antithyroid drugs at least 1 hour before iodides to prevent iodine incorporation into additional thyroid hormone. Initial regi- mens include the following: • PTU (oral loading dose of 500–1000 mg and then 250 mg every 4 hours) is favored for thyroid storm, because it decreases peripheral conversion of T4 to T3. • Methimazole (oral 60–80 mg daily) is an alternative when there is endogenous hyperthyroidism (e.g., Graves’ disease), but does not inhibit peripheral convention. 40 M. G. Lechner and T. E. Angell • When a patient is unable to take medication orally, nasogastric tube (NGT) administration may be employed. If there are con- traindications to NGT use or other issues that limit upper GI function, an intravenous reconstitution of methimazole in 0.9% saline solution given as a slow IV push has been reported. Per rectum regimens also have been employed (see also Chap. 7). Order Iodine Therapy Because thyroidal exposure to excess iodine acutely attenuates thyroid hormone secretion, consideration should be given to initi- ating an inorganic iodine preparation, such as saturated solution of potassium iodide [SSKI] (250 mg [0.25 ml/drop] every 6 hours). Iodines produce rapid decrease in thyroidal hormone release and can lower circulating thyroid hormone levels to near normal within 4–5 days. Again, to prevent incorporation of iodine in newly formed thyroid hormone and potentially prolong or exacerbate hyperthyroidism, antithyroid medication should be given before initiation of iodine therapy (see above). Order Steroid Therapy Give glucocorticoid therapy (intravenous hydrocortisone 300 mg and then 100 mg every 8 hours) to reduce T4 to T3 conversion and potentially treat coexisting adrenal insufficiency. Use of Adjunct Treatments in Refractory Cases Other treatments have been used in cases where patients are unable to receive traditional treatment or remain critically ill despite therapy. The data supporting these measures are limited. • Calcium channel blockers. Verapamil and diltiazem, which are not dihydropyridines and therefore do not cause vasodilator- 4 Severe Thyrotoxicosis and Thyroid Storm 41 induced reflex tachycardia, have been used for rate control in lieu of, but not in combination with, β-blockers. • Cholestyramine. When used with thionamides and a β-blocker, serum T4 and T3 levels drop faster during the first 2 weeks of therapy. • Lithium carbonate (300 mg every 6 hours and titration to lith- ium level of 0.8–1.2 mEq/L) causes inhibition of thyroid hor- mone release from the thyroid. • Plasmapheresis (2.5– 3 L volume of combined fresh-frozen plasma and 5% albumin) can remove excess thyroid hormone from circulation. • L-carnitine (1–2 g twice daily) may inhibit T3 action in the nucleus. • Thyroidectomy may be necessary in rare patients to treat hyperthyroidism when medical therapy does not adequately control thyrotoxicosis. Treatment of Thyrotoxic Patients Without Thyroid Storm • Hospitalized thyrotoxic patients usually require prompt but not emergent therapy. Providing β-blockers, such as propranolol (orally 10–40 mg every 8 hours), ameliorates adrenergic symp- toms. Initial dosing should be based on blood pressure and heart rate tolerability and the presence of CHF. Initiation of methimazole for patients diagnosed with Graves’ disease dur- ing hospitalization may be appropriate. Monitoring Clinical Response After initiation of therapy, supportive care measures should be adjusted to treat hemodynamic status as needed. Titration of β-blocker therapy should be performed to goal heart rates while avoiding hypotension. Repeat assessment of circulating thyroid hormone levels may be helpful to assure improvement starting 2–3 days after treatment initiation. 42 M. G. Lechner and T. E. Angell Planning Outpatient Follow-Up Discharge plans should include a timely follow-up visit with an endocrinologist for continued treatment. In Graves’ disease, inter- ruption of therapy can result in recurrence of symptoms. Patients should be advised about common adverse effects, which are most often rash, itching, GI upset, taste change, and joint pain. Agranulocytosis, vasculitis, hepatic inflammation, and cholestasis are rare but potentially life-threatening complications of antithy- roid drugs, and patients should be informed to report relevant symptoms. Repeat thyroid hormone testing should be performed 2–4 weeks after discharge to assure improvement in thyrotoxico- sis. Once stable, definitive treatment (radioactive iodine ablation or thyroidectomy) may be warranted for patients with Graves’ dis- ease to prevent recurrent hospitalization for thyrotoxicosis. Disclosure Statement All authors have no financial disclosures. Suggested Reading Alfadhli E, Gianoukakis AG. Management of severe thyrotoxicosis when the gastrointestinal tract is compromised. Thyroid. 2011;21(3):215–20. Angell TE, Lechner MG, Nguyen CT, Salvato VL, Nicoloff JT, LoPresti JS. Clinical features and hospital outcomes in thyroid storm: a retrospec- tive cohort study. J Clin Endocrinol Metab. 2015;100(2):451–9. Burch HB, Wartofsky L. Life-threatening thyrotoxicosis. Thyroid storm. Endocrinol Metab Clin North Am. 1993;22(2):263–77. Klein I, Danzi S. Thyroid disease and the heart. Curr Probl Cardiol. 2016;41(2):65–92. Klubo-Gwiezdzinska J, Wartofsky L. Thyroid emergencies. Med Clin North Am. 2012;96(2):385–403. Ross DS, Burch HB, Cooper DS, Greenlee MC, Laurberg P, Maia AL, Rivkees SA, Samuels M, Sosa JA, Stan MN, Walter MA. American thyroid association guidelines for diagnosis and management of hyperthyroidism and other causes of thyrotoxicosis. Thyroid. 2016; 26(10):1343–421. Myxedema Coma 5 Gwendolyne Anyanate Jack and James V. Hennessey Contents Assess Preadmission Thyroid Status and Thyroid Treatment . . . . . . . . . 44 Evaluate for Risk Factors Associated with Myxedema Coma . . . . . . . . 45 Identify Cardinal Features of Myxedema Coma . . . . . . . . . . . . . . . . . . . 45 Use a Systems-Based Approach to Identify Multi-organ Dysfunction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46 General . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46 Neurologic . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46 Cardiovascular . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46 Respiratory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47 Gastrointestinal . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48 Genitourinary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48 Order Thyroid Function Tests. Do Not Wait for Results . . . . . . . . . . . . 48 Order Cortisol for Adrenal Insufficiency Evaluation. Do Not Wait for Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49 G. A. Jack (*) Weill Cornell Medical Center-New York Presbyterian Hospital, Department of Medicine, Division of Endocrinology, Diabetes and Metabolism, New York, NY, USA e-mail: gwj9003@med.cornell.edu J. V. Hennessey Beth Israel Deaconess Medical Center, Harvard Medical School, Division of Endocrinology, Diabetes and Metabolism, Boston, MA, USA e-mail: jhenness@bidmc.harvard.edu © Springer Nature Switzerland AG 2020 43 R. K. Garg et al. (eds.), Handbook of Inpatient Endocrinology, https://doi.org/10.1007/978-3-030-38976-5_5 44 G. A. Jack and J. V. Hennessey Admit Critically Ill Patients to Intensive Care Unit for Close Monitoring . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49 Provide Emergency Supportive Care . . . . . . . . . . . . . . . . . . . . . . . . . . . 49 Administer Stress Dose Glucocorticoids . . . . . . . . . . . . . . . . . . . . . . . . 50 Manage Underlying Precipitating Factors . . . . . . . . . . . . . . . . . . . . . . . 50 Administer Levothyroxine +/− LT3 . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51 Follow Up the Results of Thyroid Function Tests and Cortisol . . . . . . . 51 Reassess Clinical Status of Patient . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52 Hospital Discharge Plan . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52 Suggested Reading . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52 Abbreviations AST aspartate aminotransferase ATA American Thyroid Association GCS Glasgow Coma Scale GFR glomerular filtration rate LDH Lactate dehydrogenase LDL Low-density lipoprotein LT3 Liothyronine LT4 Levothyroxine MC Myxedema coma NTI Nonthyroidal illness T3 Triiodothyronine T4 Thyroxine TSH Thyroid-stimulating hormone Assess Preadmission Thyroid Status and Thyroid Treatment Since patients present with altered mentation, obtaining a history may be difficult. Therefore, it is also important to obtain collateral information from family, friends, and outpatient medical records. 5 Myxedema Coma 45 A detailed history may reveal underlying hypothyroidism, previ- ous history of thyroidectomy (with thyroidectomy scar on exam), radioactive iodine ablation (RAI), and medication noncompli- ance/inappropriate discontinuation of thyroid hormone therapy. Signs and symptoms such as lapses in memory, slowness of thoughts, disorientation, fatigue, cold intolerance, weight gain, edema, constipation, brittle nails, and thin/coarse hair may be present. Vital signs may show hypothermia, hypotension, and bra- dycardia. Some patients may have undiagnosed hypothyroidism; therefore, other physical exam findings typical of hypothyroidism may provide clues such as the presence of a goiter, cold dry skin, delayed reflex relaxation phase, periorbital edema, facial puffi- ness, non-pitting edema in the upper and lower extremities, and enlarged tongue. Evaluate for Risk Factors Associated with Myxedema Coma In addition to determining if the patient has not been taking thy- roid hormone or taking it improperly, it is important to evaluate for other precipitating factors such as infection, cold exposure, heart failure, myocardial infarction, cerebrovascular accident, trauma, surgery, gastrointestinal bleed, substantial iodine intake such as from chronic raw bok choy consumption, and several cul- prit medications such as amiodarone, lithium, phenytoin, seda- tives, antidepressants, and anesthetics. Identify Cardinal Features of Myxedema Coma Myxedema coma should be suspected in a patient with decreased mental status, hypothermia (which can be as low as core tempera- ture < 80 °F), and bradycardia, in addition to clinical signs/symp- toms of hypothyroidism, especially in older women during the winter months. 46 G. A. Jack and J. V. Hennessey Use a Systems-Based Approach to Identify Multi-organ Dysfunction General As discussed, hypothermia is one of the key features of myx- edema coma. It is important to note that in the setting of underly- ing infection, patients may be normothermic due to blunted ability to mount a febrile response in the setting of thermal dysregula- tion. Typical signs of infection (fever, tachycardia, diaphoresis) may not be evident. Therefore in a patient with suspected myx- edema coma, normal temperature should warrant more in-depth investigation for underlying infection. Complete blood count with differential, urinalysis, urine culture, and blood culture should be obtained (Table 5.1). Neurologic Neurocognitive disturbances observed in myxedema include a reduced level of consciousness, confusion, psychomotor slowing, cerebellar ataxia, memory deficits, dementia, depression, seizure, lethargy that devolves into stupor, and ultimately coma. Sensory and motor peripheral neuropathy, psychosis, and hallucinations (“myxedema madness”) have also been described. Cardiovascular Cardiovascular disturbances include reduction in myocardial con- tractility, low cardiac output, hypotension, cardiogenic shock, and bradycardia. Mucopolysaccharide accumulation in the pericar- dium can lead to pericardial effusions and subsequent cardiac tam- ponade physiology. Patients may endorse dyspnea. Physical exam would reveal jugular venous distention, muffled heart sounds, tachycardia, and pulsus paradoxus and can be further confirmed by chest X-ray, EKG, and cardiac echocardiogram. Confirmatory tests include chest X-ray, EKG, and cardiac echocardiogram. 5 Myxedema Coma 47 Table 5.1 Signs and symptoms of myxedema coma Signs/symptoms General Fatigue, weakness Hypothermia without shivering Neurologic Memory deficits Delayed relaxation phase of DTR Seizure Psychomotor slowing Decreased level of consciousness (lethargy, stupor, coma) Cardiac Bradycardia (tachycardia if cardiac tamponade is present) Hypotension Diastolic dysfunction Tamponade Respiratory Dyspnea Renal Decreased urine output Bladder atony Increased creatinine Decreased GFR Metabolic Hypoglycemia Hyponatremia Elevated creatine kinase, elevated LDH Elevated LDL Gastrointestinal Constipation, fecal impaction Paralytic ileus, megacolon Ascites Elevated AST Dermatologic Puffy face and extremities Cool, dry skin Cold intolerance Brittle nails Thin, sparse, dry hair Musculoskeletal Myalgia, easy fatigability Respiratory Respiratory manifestations including hypoventilation, hypoxia, and hypercapnia can also occur in myxedema coma patients. An arterial blood gas analysis to assess for hypercapnia and hypoxemia should be considered. Also, laryngeal edema and macroglossia, resulting 48 G. A. Jack and J. V. Hennessey in airway narrowing, can pose a challenge during endotracheal tube placement. Pleural effusions and underlying pneumonia can also contribute to diminished respiratory function. Chest X-ray should be obtained to evaluate for underlying pneumonia and pleu- ral effusion. Gastrointestinal Bowel wall edema can result in reduced intestinal motility, atony, paralytic ileus, and toxic megacolon. Patients may present with nausea/vomiting, abdominal distension, constipation, and fecal impaction. Also, absorption of medications can be diminished, and dose adjustment of oral medications may be needed. Impaired gluconeogenesis, infection, and concomitant adrenal insuffi- ciency may contribute to hypoglycemia; therefore, serum blood sugar should be obtained. Genitourinary Severe hypothyroidism results in a decrease in renal glomerular filtration rate and renal perfusion and rhabdomyolysis. Obtain a complete metabolic panel, which may reveal hyponatremia, elevated creatinine kinase, creatinine, and aspartate aminotrans- ferase. Urine output should also be monitored for decreased urine output from acute renal injury and bladder atony. Order Thyroid Function Tests. Do Not Wait for Results The diagnosis of myxedema coma is in large part based on clin- ical suspicion and confirmed with biochemical testing. Thyroid function tests including thyroid-stimulating hormone (TSH), free T4 (fT4), should be obtained prior to administration of thy- roid replacement therapy. Initial evaluation includes an assess- ment of underlying precipitating factors; however, if a high 5 Myxedema Coma 49 index of suspicion is obvious, do not wait until the results of the evaluation are definitive before initiating treatment. Order Cortisol for Adrenal Insufficiency Evaluation. Do Not Wait for Results Given the overlap in the presentation of MC with adrenal insuffi- ciency such as fatigue, hyponatremia, hypothermia, and hypogly- cemia, evaluation for adrenal insufficiency is prudent. Adrenal insufficiency can be from hypopituitarism or can be part of poly- endocrine syndrome in a patient with underlying Hashimoto’s thyroiditis with autoimmune primary adrenal insufficiency (i.e., Schmidt syndrome). Also, after starting thyroid replacement ther- apy, there is concern for increased metabolism of cortisol leading to adrenal crisis; therefore, obtaining baseline serum cortisol level is essential. In a stressed patient with normal serum albumin lev- els, a serum cortisol level greater than 18 mg/dl would rule out adrenal insufficiency and would permit rapid taper of hydrocorti- sone after limited initial exposure. Admit Critically Ill Patients to Intensive Care Unit for Close Monitoring Myxedema crisis is an emergency that warrants frequent monitor- ing of patient’s clinical status; therefore, it should be managed in a critical care setting. Provide Emergency Supportive Care Maintain a low threshold for intubation and mechanical ventila- tion in the setting of worsening hypoxemia and hypercapnia and concern for airway protection in the setting of reduced Glasgow Coma Scale, macroglossia, or suspicion for laryngeal edema. Cardiovascular collapse and hypotension may require volume resuscitation with isotonic normal saline or 5–10% dextrose in 50 G. A. Jack and J. V. Hennessey half-normal saline if hypoglycemia is also present. Hypotension might be refractory to intravenous fluids without thyroid replace- ment therapy and in the setting of adrenal insufficiency. Vasopressors may be added if fluid resuscitation is inadequate in providing cardiovascular support; however, it should be weaned off as soon as clinically indicated. Regarding management of hypothermia, external warming techniques with blankets are suitable, though this may worsen hypotension through vasodilation. Aggressive external rewarming and central warming can potentiate cardiovascular collapse; therefore, it is generally not recommended. Treatment with thy- roid hormone should restore thermoregulation. Administer Stress Dose Glucocorticoids Due to the concern for underlying adrenal insufficiency, and the risk of inciting adrenal crisis with initiation of thyroid hormone therapy, it is recommended that stress dose steroids be adminis- tered prior to thyroid supplementation. Hydrocortisone 50–100 mg can be administered intravenously every 6–8 hours until clinical improvement and quickly tapered off if labs obtained prior to glu- cocorticoid initiation ultimately do not demonstrate adrenal insuf- ficiency. Manage Underlying Precipitating Factors Based on clinical presentation, laboratory analysis, and imag- ing studies, treat for any precipitants such as myocardial infarc- tion, gastrointestinal bleeding, and other underlying medical conditions. If suspicious for infection, draw cultures and start empiric antibiotics. Metabolic derangements such as hypo- glycemia and hyponatremia should be monitored and treated accordingly. Caution must be exercised to not rapidly correct serum sodium, given the risk for osmotic demyelination syn- drome. This is focused on the classic definition; however, a similar approach can be adopted in patients with features of profound hypothyroidism. 5 Myxedema Coma 51 Administer Levothyroxine +/− LT3 Controversy exists regarding optimal thyroid hormone regimens including the

type, dose, route, frequency of administration, and duration of therapy. An approach adopted by earlier studies is administration of an intravenous L-thyroxine 300–600 μg loading dose to replete the deficit in the total body thyroid hormone pool, followed by maintenance doses of 50–100 μg LT4 daily by intra- venous or oral route (if mentally alert). The ATA 2014 guideline recommends an intravenous loading dose of 200–400 μg of L- thyroxine, followed by daily oral L-thyroxine dose of 1.6 μg/kg body weight or 75% of this dose if intravenous route. The under- lying principle behind levothyroxine monotherapy is that it allows for restoration to near-normal levels. In myxedema coma, T3 is low and concomitant nonthyroidal illness further decreases T4 to T3 conversion. Therefore, another strategy is the addition of LT3 to L-thyroxine therapy. LT3 has a quicker onset on action, crosses the blood-brain barrier readily, increases core temperature within 2–3 hours (as opposed to 14 hours from LT4 intravenously), and possibly improves neuropsychiatric manifestations more rapidly. According to the ATA 2014 guide- lines, clinicians can consider coadministration of LT3, with an ini- tial intravenous loading dose of LT3 5–20 μg, followed by a maintenance dose of LT3 IV 2.5–10 μg every 8 hours, which can be continued until patient’s clinical status has improved and mainte- nance oral LT4 can be administered. Alternatively, LT3 therapy can be added if clinical status does not improve after 24–48 hours of L-thyroxine alone. Due to the potential risk of cardiac arrhythmias and myocardial infarction, especially in older patients, it is recom- mended to use lower doses of L-thyroxine and LT3. Follow Up the Results of Thyroid Function Tests and Cortisol Marked TSH elevation is consistent with primary hypothyroid- ism. In patients with central hypothyroidism, TSH is not a reliable measure. Also, TSH may not be substantially elevated in the set- ting of nonthyroidal illness (NTI) and glucocorticoid or dopamine 52 G. A. Jack and J. V. Hennessey administration. In myxedema coma, free T4 is low, and with con- comitant nonthyroidal illness, these parameters may be even lower. In those known to be hypothyroid or with well-established hypothyroidism, if thyroid function tests and cortisol are normal, glucocorticoids can be promptly discontinued, and depending on clinical status, one may consider transitioning to the preadmission LT4 dose. Reassess Clinical Status of Patient It is important to monitor the patient frequently for clinical improvement. Once stabilized, frequent monitoring is no longer necessary, and patient can be transferred to medical floors for fur- ther management. Hospital Discharge Plan Patients with underlying hypothyroidism should be instructed to administer levothyroxine on an empty stomach and wait at least 1 hour before eating/drinking, in order to optimize absorption of LT4. If dose adjustments of LT4 were done during hospitalization, repeat TSH can be obtained in 6–8 weeks to determine if further dosage adjustments are needed. Outpatient follow-up with the patient’s endocrinologist and primary care physician should be arranged to reassess thyroid status at an appropriate interval. Suggested Reading Chiong YV, Bammerlin E, Mariash CN. Development of an objective tool for the diagnosis of myxedema coma. Transl Res. 2015;166(3):233–43. Fliers E, Wiersinga WM. Myxedema coma. Rev Endocr Metab Disord. 2003;4(2):137–41. Holvey DN, Goodner CJ, Nicoloff JT, Dowling JT. Treatment of myxedema coma with intravenous thyroxine. Arch Intern Med. 1964;113(1):89–96. Jordan RM. Myxedema coma: pathophysiology, therapy, and factors affect- ing prognosis. Med Clin N Am. 1995;79(1):185–94. 5 Myxedema Coma 53 Kasid N, Hennessey JV. Myxedema Coma. In: Endocrine and metabolic medical emergencies: a clinician’s guide; 2018. p. 252–61. Klubo-Gwiezdzinska J, Wartofsky L. Thyroid emergencies. Med Clin N Am. 2012;96(2):385–403. Liamis G, Filippatos TD, Liontos A, Elisaf MS. Management of endocrine disease: Hypothyroidism-associated hyponatremia: mechanisms, implica- tions and treatment. Eur J Endocrinol. 2017;176(1):R15–r20. Mathew V, Misgar RA, Ghosh S, Mukhopadhyay P, Roychowdhury P, Pandit K, et al. Myxedema coma: a new look into an old crisis. J Thyroid Res. 2011;2011:493462. Osborn LA, Skipper B, Arellano I, MacKerrow SD, Crawford MH. Results of resting and ambulatory electrocardiograms in patients with hypothyroid- ism and after return to euthyroid status. Heart Dis. 1999;1(1):8–11. Popoveniuc G, Chandra T, Sud A, Sharma M, Blackman MR, Burman KD, et al. A diagnostic scoring system for myxedema coma. Endocr Pract. 2014;20(8):808–17. Sorensen JR, Winther KH, Bonnema SJ, Godballe C, Hegedus L. Respiratory manifestations of hypothyroidism: a systematic review. Thyroid. 2016;26(11):1519–27. Wartofsky L. Myxedema coma. Endocrinol Metab Clin N Am. 2006;35(4):687–98, vii-viii. Abnormal Thyroid 6 Stimulating Hormone Values That Are Not due to Common Causes of Primary Hypothyroidism or Thyrotoxicosis Zsu-Zsu Chen and James V. Hennessey Contents Physiologic Causes of Variations in TSH Levels in Asymptomatic Patients . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56 Laboratory Detection of Isoforms or Assay Interference Leading to Abnormal TSH Values in Patients Who Are Asymptomatic . . . . . . . . 58 Medications That Can Affect TSH Levels . . . . . . . . . . . . . . . . . . . . . . . 59 Medications Associated with Low TSH Levels . . . . . . . . . . . . . . . . . 61 Medications Associated with High TSH Levels . . . . . . . . . . . . . . . . . 62 Z.-Z. Chen (*) Beth Israel Deaconess Medical Center, Department of Endocrinology, Diabetes and Metabolism, Boston, MA, USA e-mail: zchen5@bidmc.harvard.edu J. V. Hennessey Beth Israel Deaconess Medical Center, Harvard Medical School, Division of Endocrinology, Diabetes and Metabolism, Boston, MA, USA e-mail: jhenness@bidmc.harvard.edu © Springer Nature Switzerland AG 2020 55 R. K. Garg et al. (eds.), Handbook of Inpatient Endocrinology, https://doi.org/10.1007/978-3-030-38976-5_6 56 Z.-Z. Chen and J. V. Hennessey Immune Checkpoint Inhibitor Effects on TSH in Patients Treated with These Therapies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63 Tyrosine Kinase Inhibitor Effects on TSH in Patients Treated with These Therapies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63 Nonthyroidal Illness . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64 Thyroid Function Tests That Are Consistent with Nonthyroidal Illness . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64 Treatment Considerations in a Patient with Nonthyroidal Illness . . . 65 Central Hypothyroidism Is Associated with Variable TSH Levels . . . . . 65 Consider Genetic Causes of Central Hypothyroidism . . . . . . . . . . . . 67 Consider Acquired Causes of Central Hypothyroidism . . . . . . . . . . . 68 Rare Causes of Abnormal TSH Levels . . . . . . . . . . . . . . . . . . . . . . . . . . 69 Resistance to TSH Causes High TSH Levels . . . . . . . . . . . . . . . . . . . 69 Resistance to TRH Is a Rare Cause of Central Hypothyroidism . . . . 69 Resistance to Thyroid Hormone Is a Rare Cause of Abnormal TSH Levels . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70 Consider TSH Secreting Tumor as a Rare Cause of High TSH and High Peripheral Thyroid Hormones . . . . . . . . . . . . . . . . . . . . 70 Suggested Reading . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71 Physiologic Causes of Variations in TSH Levels in Asymptomatic Patients Thyroid stimulating hormone (TSH) levels can vary up to 40–50% within a single day. Levels are typically lowest in the late afternoon and highest at bedtime. Normal ranges for TSH laboratory values are also determined based on where 95% of TSH values fall within a carefully screened population of healthy euthyroid volunteers. This means that 5% of people with normal thyroid function could still have TSH values that are considered abnormal. Also, these normal ranges can differ based on which reference population was studied. This chapter will give an over- view of the different causes of abnormal serum TSH values not due to primary thyroid dysfunction (several of which are outlined in Fig. 6.1). 6 Abnormal Thyroid Stimulating Hormone Values That Are Not… 57 Variable TSH Genetic mutations Physiologic Variation (Euthyroid) Laboratory Artifact (Euthyroid) - Isolated or combined congenital hypothyroidism - - - Resistance to TRH, TSH Central hypothyroidism Diurnal variation Detection of biologically inactive TSH (variable TSH and T4/T3, TSH is not an - Normal population outliers isoformsAcquired hypothyroidism appropriate surrogate marker of - Age - Antibodies- Mass lesion thyroid function in these cases) - Ethnicity - Heterophile antibodies - latrogenic - IgG complexes - Traumatic brain injury/cerebral vascular event - Biotin - Infiltrative/infectious process - Autoimmunity ↑ TSH ↓TSH TSH secreting pituitary adenoma Medications - Dopamine Hypothalamus ↑ TSH secretion Medications - Dopamine agonist TRH - Dopamine receptor blockers - Bromocriptine - Amphetamines - Cabergoline (+) Nonthyroidal Illness - Somatostatin analogues Pituitary - Octreotide ↓ TSH TSH Antiepiletic medications - - Pasireotide secretion - Valproate ↓T3, nl/↓T4,nl/↓TSH Unknown etiology Carbamazepine - - Lanreotide - Sometimes transient ↑ - High dose glucocorticoids (-) - OxcarbazepineTSH with recovery - Opiates Failure to - Metformin lodine/lodine contrast escape Wolff- (+) - Retinoid × receptor (Bexarotene) Amiodarone Chaikoff (-) lodine/lodine Contrast ↑ T3/T4 Ritonavir Amiodarone production St. John’s Wort Possibly Negative feedback Positive feedback Selective serotonin reuptake alters T3/T4 to hypothalamus inhibitors (SSRIs) ClearanceThyroid hormone analogues to hypothalamus - Levothyroxine and pituitary and pituitary (↓ T3/T4, ↑ TSH) lodine/lodine Contrast - Liothyronine (↑ T3/T4, ↓ TSH) Interferon alpha Amiodarone Lithium ↓ T3/T4 Tyrosine Kinase Inhibitors Acute thyroiditis Thyroid Thionamides production Immune checkpoint inhibitors - initial ↑ T3/T4, ↓ TSH, - Propyllthiouracil or secretion - Pembrolizumab - subsequent ↓ T3/T4, ↑ TSH - Methimazole - Nivolumab - resolution or chronic - lpilimumab hypothyroidism with ↑ TSH T3/T4 ↑ T3/T4, Thyroid Hormone Resistance Autoimmunity nl/↑ TSH 58 Z.-Z. Chen and J. V. Hennessey Fig. 6.1 Causes of abnormal serum TSH values not due to primary thyroid dysfunction. This figure demonstrates the normal hypothalamic-pituitary- thyroid axis. Thyrotropin-releasing hormone (TRH) is secreted by the hypo- thalamus that provides positive feedback to the pituitary that secretes thyroid stimulating hormone (TSH). TSH provides positive feedback to the thyroid gland causing increased production of thyroxine (T4) and triiodothyronine (T3) that both in turn provide negative feedback to the hypothalamus and pitu- itary. Causes of TSH abnormalities discussed in this book chapter are grouped as those that cause elevated TSH, decreased TSH, and variable TSH levels Age can also affect TSH values with an estimated 0.3 mIU/L increase in value for every 10-year increase in age after 30–39 years old. TSH normal ranges can also vary based on eth- nicity. In the NHANES III reference population study, African Americans aged 30–39 years old had the lowest TSH values, while Mexican Americans that were 80 years or older had the highest TSH values. Laboratory Detection of Isoforms or Assay Interference Leading to Abnormal TSH Values in Patients Who Are Asymptomatic Issues with laboratory detection of TSH

levels can lead to abnor- mal values in a euthyroid patient. Several TSH isoforms can be expressed in humans that are not biologically active. Some labo- ratory assays may detect these isoforms leading to the reporting of abnormal TSH values that, however, do not correlate with hypothalamic- pituitary-thyroid dysfunction. Antibodies can also interfere with immunoassays used for the detection of TSH levels. Patients with high heterophile antibodies (HAb) can have falsely elevated levels. Manufactures have refined their assays to over- come this issue, but a small percentage of patients can still have high enough titers to cause interference. The most common het- erophile antibodies in humans are those targeting animal antigens, specifically human anti-mouse antibodies (HAMA). Human anti- bodies targeting human antigens, such as rheumatoid factor, can also interfere. Suspicion should be raised in patients that have a 6 Abnormal Thyroid Stimulating Hormone Values That Are Not… 59 clinical picture inconsistent with their lab values. People at high risk for development of these antibodies are those who have had recent vaccines, blood transfusions, and monoclonal antibody treatments, veterinarians, or those who have jobs that require fre- quent animal contact. If there is suspicion that there is a hetero- phile antibody, the TSH can be measured with a different manufacturer’s assay. The presence of TSH autoantibodies can cause TSH immuno- globulin G (IgG) complexes that also interfere with TSH immu- noassays. This could lead to either falsely elevated or lower TSH values. These autoantibodies are immunoreactive, causing the erroneous lab values, but are not biologically active and therefore do not cause pituitary-thyroid axis dysfunction. There are assays available for removal of these IgG complexes prior to processing the serum sample. High-dose oral biotin supplementation (>5000–10,000 μg daily) can also cause interference with the detection assays of TSH, total thyroxine (T4), and total triiodothyronine (T3). It can have a negative effect on TSH and positive effect on T3 and T4, biochemically mimicking thyrotoxicosis. Clinicians should ask about biotin supplementation so that biotin interference can be considered if the labs are inconsistent with the clinical picture. Biotin supplementation can be held prior to the blood draw to prevent interference, and the literature regarding the duration of holding biotin varies from at least 2 days to 7 days. Medications That Can Affect TSH Levels Medications, some of which are detailed in Table 6.1, can also affect TSH secretion and/or interrupt the hypothalamic-pituitary- thyroid axis. Often, abnormal TSH levels due to medications will normalize once the offending agent is stopped. However, some patients may require long-term antithyroidal treatment or thyroid hormone replacement therapy due to persistent thyroid dysfunc- tion. Of note, there are several medications including iodine, iodine contrast agents, and amiodarone that can cause both hypo- thyroidism and thyrotoxicosis. These medications typically cause 60 Z.-Z. Chen and J. V. Hennessey Table 6.1 Medications that affect thyroid stimulating hormone (TSH) Low TSH Dopamine Dopamine agonists Bromocriptine Cabergoline Somatostatin analogues Octreotide Pasireotide Lanreotide Glucocorticoids Opiates Retinoid X receptor (i.e., bexarotene) Metformin Thyroid hormone analogues High TSH Dopamine receptor blockers Amphetamines Ritonavir St. John’s wort Selective serotonin reuptake inhibitors (SSRIs) Thionamides (i.e., antithyroid medications) Propylthiouracil Methimazole Lithium Antiepileptic medications Valproate Carbamazepine Oxcarbazepine Interferon alpha High or low TSH Immune checkpoint inhibitors Pembrolizumab Nivolumab Ipilimumab 6 Abnormal Thyroid Stimulating Hormone Values That Are Not… 61 Table 6.1 (Continued) Tyrosine kinase inhibitors Amiodarone Iodine/iodine contrast agents Table derived in part from Table 10 of Garber JR, Cobin RH, Gharib H, Hen- nessey JV, Klein I, Mechanick JI, et al. Clinical practice guidelines for hypo- thyroidism in adults: cosponsored by the American Association of Clinical Endocrinologists and the American Thyroid Association. Endocr Pract Off J Am Coll Endocrinol Am Assoc Clin Endocrinol. 2012:18(6);988–1028. See Suggested Reading thyroid dysfunction in patients who already have underlying thy- roid disease or who have the propensity to develop dysfunction (i.e., antithyroid antibody positivity or nodular goiter). The devel- opment of hypo- or hyperfunction is dictated by the underlying thyroid disorder. Medications Associated with Low TSH Levels Dopamine infusions and agonists, high-dose glucocorticoids, somatostatin analogues including octreotide, oral bexarotene (a retinoid X receptor-selective ligand used to treat cutaneous T-cell lymphoma), metformin, and opiates have been linked to inhibi- tion of TSH release and therefore low TSH levels. Thyroid hor- mone medications, including liothyronine (LT3) and levothyroxine (LT4), also lower TSH due to hypothalamic-pituitary suppression. Iodine, iodine contrast agents, and amiodarone (which is rich in iodine) can cause hyperthyroidism due to increased thyroid hor- mone synthesis resulting in a low TSH. The normal thyroid gland has mechanisms to prevent increased thyroid hormone production when there is excess substrate (i.e., iodine) and prevent hyperthy- roidism. However, in patients who have autonomous function of the whole or part of the thyroid gland (i.e., with Graves’ disease 62 Z.-Z. Chen and J. V. Hennessey or toxic nodule), these mechanisms may be bypassed and lead to hyperthyroidism with exposure to increased iodine. Amiodarone can also cause hyperthyroidism through the mechanism outlined above due to its high iodine content. This is known as type 1 amiodarone-i nduced thyrotoxicosis (AIT). Type 2 AIT is a destructive thyroiditis that leads to excessive release of T3 and T4. The thyrotoxicosis can resolve or persist and eventually lead to long-term hypothyroidism. Medications Associated with High TSH Levels Medications including dopamine receptor blockers and amphet- amines have been linked with increased TSH secretion and therefore a high TSH. Other medications such as ritonavir, St. John’s wort, and selective serotonin uptake inhibitors (SSRIs) could affect the clearance of T3 and T4 leading to elevated TSH levels. Patients on long-term lithium are also at increased risk for development of hypothyroidism. This increased risk is thought to be due to decreased T4 and T3 secretion or due to a thyroiditis that may be transient. Other antiepileptic medications including valproate, carbamazepine, and oxcarbazepine have been associated with increased risk of hypothyroidism but with unclear mechanism. Interferon alpha, a treatment for hepatitis C, has been associated with hypothyroidism likely due to a destructive thyroiditis that can lead initially to transient hyper- thyroidism, then to hypothyroidism, and eventually to either resolution or permanent hypothyroidism. Thionamides, used as antithyroid therapy in acute thyrotoxicosis, also causes hypothy- roidism and an elevated TSH due to the blocking of thyroid hor- mone synthesis. Iodine, iodine contrast agents, and amiodarone can enhance the inhibitory effect iodine has on the thyroid gland and cause hypothyroidism. As discussed above, the thyroid gland can tran- siently decrease thyroid hormone synthesis when there is excess iodine. This prevents the development of hyperthyroidism and is known as the Wolff-Chaikoff effect. Eventually, however, the thyroid gland will return to normal function and “escape” the Wolff- Chaikoff effect. In thyroid glands that are already dam- 6 Abnormal Thyroid Stimulating Hormone Values That Are Not… 63 aged by preexisting autoimmune thyroiditis, the gland may not escape from the Wolff-Chaikoff effect leading to persistent hypothyroidism. Immune Checkpoint Inhibitor Effects on TSH in Patients Treated with These Therapies Immune checkpoint inhibitors used for cancer therapy, including anti-programed cell death protein 1 (anti PD-1) immunotherapy pembrolizumab and nivolumab and cytotoxic T-lymphocyte- associated antigen (CTLA-4) therapy with ipilimumab, have been linked with thyroid dysfunction. Clinically, patients can present with thyrotoxicosis due to thyroiditis (initially low TSH and ele- vated T4 that can resolve or progress to overt hypothyroidism). Patients who already have hypothyroidism can develop worsening of their hypothyroidism (elevated TSH with low T4) requiring higher doses of thyroid hormone replacement. Patients can also develop centrally mediated hypothyroidism that usually occurs with panhypophysitis. This diagnosis is made with a low free T4 levels. TSH levels can be low, inappropriately normal, or elevated. Patients with transient thyroiditis typically do not require treat- ment therapy. A short course of beta-blockers can be considered when there are significant clinical symptoms of thyrotoxicosis, but this is typically not required. Patients who develop overt hypothy- roidism or panhypophysitis will need long-term thyroid hormone supplementation. Tyrosine Kinase Inhibitor Effects on TSH in Patients Treated with These Therapies Tyrosine kinase inhibitors (TKI) have been linked to hypothyroid- ism in euthyroid patients causing an elevated TSH. In patients who already have hypothyroidism, there can be increasing thyroid hormone replacement therapy requirements. This could be due to a transient destructive thyroiditis similar to the immune check- point inhibitors or possibly due to altered set points in the hypothalamic- pituitary-thyroid axis. 64 Z.-Z. Chen and J. V. Hennessey Nonthyroidal Illness Serum TSH can be suppressed in patients with acute illnesses, especially those that are hospitalized and in the intensive care units. There is a distinctive pattern of serum thyroid hormone level derangements that is known as nonthyroidal illness syn- drome. It is also referred to in the literature as sick euthyroid and low T3 syndrome. This entity has been described in healthy fasting patients as well as in a wide range of patients with acute and chronic illnesses including starvation, infec- tion, trauma, surgery, sepsis, heart disease, cerebral vascular accidents, renal failure, and malignancy. It is believed that these changes help reduce energy expenditure and cellular catabolism which could have protective effects while fasting. It is unclear if these changes are an adaptive or maladaptive process in acute i llness. Thyroid Function Tests That Are Consistent with Nonthyroidal Illness The diagnosis of nonthyroidal illness is usually obvious in acutely ill patients. Thyroid function tests show low T3 levels. T4 levels can be normal or low. TSH is typically normal but can be low. If TSH is undetectable (<0.01 mU/L), there is an increased likelihood that the patient has true hyperthyroidism. TSH levels can rise in parallel with normalization of serum T4 and T3 levels and could suggest recovery from the pituitary- thyroid axis suppression of nonthyroidal illness. If the TSH is very elevated (> 20 mU/L), there is a far higher likelihood that the patient will have persistent hypothyroidism. Serum rT3 levels can occasionally be used to help differentiate central hypothyroidism from nonthyroidal illness since it is elevated in the latter. However, its diagnostic utility is limited since it can also be slightly elevated in patients with mild hypothy- roidism. 6 Abnormal Thyroid Stimulating Hormone Values That Are Not… 65 Treatment Considerations in a Patient with Nonthyroidal Illness Several small randomized controlled trials studied treatment of nonthyroidal illness with LT3 and/or LT4 but did not show benefit in patient outcomes. Also, theoretically, if these thyroid hormone changes are adaptive, then attempts to correct the transient low T3 state could cause harm. Current recommendations are not to treat thyroid function test abnormalities likely due to nonthyroidal ill- ness with thyroid hormone replacement unless the patient has overt clinical signs of hypothyroidism. Thyroid function tests should be checked after sufficient time has elapsed following resolution of the illness to confirm normalization of thyroid hormone levels. Central Hypothyroidism Is Associated with Variable TSH Levels Central hypothyroidism is characterized by an impaired TSH response. This can be caused by defects in thyrotropin-releasing hormone (TRH) – including defects in the TRH receptor – or in defects in TSH. These defects can either be due to congenital or acquired causes. Most congenital causes are due to genetic muta- tions and can lead to isolated central hypothyroidism or combined pituitary hormone deficiencies. With the advent of newborn screen- ing, these are usually diagnosed during infancy. Acquired causes are typically due to processes that disrupt or destroy TRH produc- ing cells in the hypothalamus or TRH sensing cells in the pituitary due to invasive or compressive lesions, trauma, vascular accidents, autoimmune or infectious diseases, infiltrative processes, or iatro- genic causes. Acquired causes typically lead to combined pituitary hormone deficiencies, and isolated central hypothyroidism is less common. Some of the causes of central hypothyroidism are detailed in Table 6.2. In these patients, the TSH level can either be high or low or even appear normal. However, the peripheral thy- roid hormone levels (including free T4 and total T3) will be low. 66 Z.-Z. Chen and J. V. Hennessey Table 6.2 Causes of central hypothyroidism Genetic causes (gene mutations) Isolated central hypothyroidism TSHB TRHR TSHR IGSF1 TBL1X Combined congenital hypothyroidism LHX3, LHX4 HESX1 SOX3 OTX2 PROP1 POU1F1 LEPR Acquired causes Mass lesion Pituitary adenoma (functional/nonfunctional) Craniopharyngioma Meningioma Rathke’s cleft cyst Empty sella Metastasis Iatrogenic Intracranial surgery Radiation Traumatic brain injury Cerebral vascular events Cerebral infarct Intracranial hemorrhage Sheehan’s syndrome Autoimmune disease Lymphocytic hypophysitis Polyglandular autoimmune disease Infiltrative

process Sarcoidosis 6 Abnormal Thyroid Stimulating Hormone Values That Are Not… 67 Table 6.2 (Continued) Histiocytosis X Iron overload (hemochromatosis, blood transfusions) Infectious diseases Tuberculosis Toxoplasmosis Fungal infections Table derived in part from Table 2 of Beck-Peccoz P, Rodari G, Giavoli C, Lania A. Central hypothyroidism - a neglected thyroid disorder. Nat Rev. Endocrinol. 2017:13;588–598. See Suggested Reading Consider Genetic Causes of Central Hypothyroidism The most frequent cause of inheritable isolated central hypothy- roidism are mutations of the TSHB gene that encodes for the β-subunit of the TSH molecule. This leads to decreased levels of functional TSH and increased circulating levels of the glycopro- tein hormone α-subunit (α-GSU). The α-GSU is also a subunit for follicle stimulating hormone (FSH), luteinizing hormone (LH), and human chorionic gonadotropin (hCG) and is frequently referred to as the α-subunit. A loss of function mutation in the immunoglobulin superfamily member 1 (IGSF1) gene leads to central hypothyroidism and macroorchidism. A missense m utation in the TBL1X gene is associated with central hypothyroidism and hearing loss. Mutations in pituitary transcription factors (includ- ing LHX3, LHX4, HESX1, SOX3, OTX2, PROP1, POU1F1, and LEPR genes) are the most common genetic causes of combined congenital hypothyroidism. For both genetic causes of either isolated or combined congeni- tal hypothyroidism, TSH levels can be low or normal in the setting of low free T4. These patients should be treated with l evothyroxine hormone replacement. Free T4 levels should be used to monitor the adequacy of hormone replacement since TSH levels can be unreli- able. Goal free T4 levels in these individuals are the same as those with central hypothyroidism, which are to be in the upper half of the normal laboratory free T4 range. It is important to remember that 68 Z.-Z. Chen and J. V. Hennessey free T4 should be checked in the morning prior to the ingestion of a patient’s levothyroxine dose because that can cause transient eleva- tion in the free T4 level. In the case of combined congenital hypo- thyroidism, hormone replacement should also be initiated for any other identified pituitary hormone deficiencies. Consider Acquired Causes of Central Hypothyroidism Masses in the hypothalamus and/or pituitary are the most com- mon cause of acquired central hypothyroidism. These lesions can lead to both qualitative and quantitative dysfunction of TSH. The most common masses found are nonfunctioning pituitary mac- roadenomas. Other lesions include craniopharyngiomas, menin- giomas, Rathke’s cleft cysts, metastases, as well as empty sella syndrome. Intracranial surgeries, especially for resection of a pituitary lesion, can lead to panhypopituitarism with associated central hypothyroidism. The risk of developing hypothyroidism is associated with the size and position of the tumor as well as the experience of the surgeon. Radiation therapy, especially in or around the sella, can also cause hypopituitarism with associated central hypothyroidism. Patients with traumatic brain injury are also at risk for central hypothyroidism with an estimated disease prevalence of 15–68%. Vascular events including cerebral infarcts, subarachnoid hemor- rhage, as well as Sheehan’s syndrome are more rare causes of central hypothyroidism. Other less common causes of acquired central hypothyroidism include autoimmune diseases (such as lymphocytic hypophysitis which may be associated with poly- glandular autoimmune disease), infiltrative processes (such as sarcoidosis, histiocytosis X, or iron overload from hemochroma- tosis or blood transfusions), and infectious diseases (tuberculosis, toxoplasmosis, and fungal infections). Patients with acquired central hypothyroidism typically have low free T4. TSH can be low, inappropriately normal, or even high. High TSH levels likely reflect a qualitative defect in the cir- culating TSH, hence the development of hypothyroidism despite elevated TSH levels. Treatment should be with thyroid hormone https://www.facebook.com/groups/2202763316616203 6 Abnormal Thyroid Stimulating Hormone Values That Are Not… 69 replacement. Free T4 levels should be measured to monitor ade- quacy of hormone replacement since the TSH is unreliable. Of note, adrenal insufficiency should be ruled out in any cases where there is concern for hypopituitarism. Adrenal insufficiency should be treated prior to initiation of thyroid hormone replacement to prevent precipitation of an adrenal crisis. Rare Causes of Abnormal TSH Levels Resistance to TSH Causes High TSH Levels Several point mutations in the TSH receptor gene can cause resis- tance to TSH. These patients typically have elevated levels of TSH with low or normal T4 and T3 concentrations. They are usually iden- tified with newborn screens. Unlike other causes of congenital hypo- thyroidism, these patients do not have goiters due to the lack of TSH stimulation of the thyroid gland. Clinical symptoms and treatment are dependent on the type of mutation and percentage of functional TSH receptors present. Some patients may be euthyroid with hyperthyro- tropinemia alone. Others develop severe congenital hypothyroidism requiring thyroid h ormone replacement. Patients who have hyperthy- rotropinemia can be differentiated from patients with autoimmune hypothyroidism because their TSH levels remain stable over time while the TSH levels change over time in autoimmune hypothyroid- ism. TSH levels will normalize with thyroid hormone replacement. Resistance to TRH Is a Rare Cause of Central Hypothyroidism A mutation in the thyrotropin-releasing hormone receptor (TRHR) has also been identified as a cause of isolated central hypothyroid- ism. However, this TRH resistance typically does not clinically manifest until childhood or early adulthood with delayed growth. Of note, the diagnosis of TRH resistance was made in a 33-year- old woman after her second successful pregnancy. She had nor- mal height and IQ and had no difficulties breastfeeding, and her children had normal pre- and postnatal growth without thyroid 70 Z.-Z. Chen and J. V. Hennessey hormone supplementation. TSH levels are typically normal, but there is a blunted TSH response to an infusion of thyrotropin- releasing hormone (TRH). Resistance to Thyroid Hormone Is a Rare Cause of Abnormal TSH Levels There are also mutations associated with dysfunction of thyroid hormone action. This resistance to thyroid hormone can be caused by malfunction of the thyroid hormone nuclear receptors, cell membrane transport of the hormone, or hormone metabolism. The most common mutation is associated with THRB which encodes for thyroid hormone receptor β. There is phenotypic variability even in patients with the same mutation. Typically in these syn- dromes, free T4 and T3 are high with normal or slightly elevated TSH. Patients can have goiters or develop attention deficit disor- der and tachycardia. However, some may not have obvious symp- toms of clinical thyrotoxicosis. Patients typically do not require thyroid hormone treatment because they will compensate for the insensitivity with increased T4 and T3 production. The most important thing to avoid in these patients is surgical thyroidec- tomy or radioactive iodine ablation of the thyroid. Mutation of the THRA gene that encodes for thyroid hormone receptor alpha is associated with low free T4, normal or slightly elevated T3, and normal TSH. The TSH is normal because the beta receptors that control TSH output remain intact. Clinically, patients exhibit signs of hypothyroidism in the peripheral tissues. These can lead to significant bony abnormalities, gastrointestinal tract dysmotil- ity, bradycardia, and mental disabilities. Consider TSH Secreting Tumor as a Rare Cause of High TSH and High Peripheral Thyroid Hormones TSH secreting tumors are a very rare cause of hyperthyroidism. They represent 0.5–3% of functional pituitary adenomas, even though this could be an underestimation, and they are typically 6 Abnormal Thyroid Stimulating Hormone Values That Are Not… 71 benign. Patients are usually diagnosed in their fifth to sixth decades. The majority of the lesions only secrete intact TSH, but they can also secrete growth hormone or prolactin. Sometimes there is also increased α-GSU (α-subunit). Patients develop clinical symptoms of hyperthyroidism. If it is a mac- roadenoma, compressive symptoms including headache and visual field defects can develop as well as adrenal and gonadal axis dysfunction. In patients who co-secrete growth hormone or prolactin, the development of acromegaly or galactorrhea may be seen. Labs are consistent with elevated levels of T4 and T3. TSH can be inappropriately normal or mildly elevated. Patients should also be screened for hyper- and hyposecretion of other pituitary hor- mones, and α-GSU levels should be checked. MRI of the pituitary is recommended for imaging. Once the diagnosis of a mass is made, the patient should be evaluated by neurosurgery for resec- tion. Interval treatment prior to surgery for hyperthyroid symp- toms can include the use of beta-blockers. Somatostatin analogues or dopamine agonists can also be used for lesions that co-secrete growth hormone and prolactin. Suggested Reading Barroso-Sousa R, Barry WT, Garrido-Castro AC, Hodi FS, Min L, Krop IE, et al. Incidence of Endocrine Dysfunction Following the Use of Different Immune Checkpoint Inhibitor Regimens: A Systematic Review and Meta- analysis. JAMA Oncol. 2018;4(2):173–82. Beck-Peccoz P, Rodari G, Giavoli C, Lania A. Central hypothyroidism - a neglected thyroid disorder. Nat Rev Endocrinol. 2017;13:588–98. Bonomi M, Busnelli M, Beck-Peccoz P, Costanzo D, Antonica F, Dolci C, et al. A Family with Complete Resistance to Thyrotropin-Releasing Hormone. N Engl J Med. 2009;360(7):731–4. Demir K, van Gucht ALM, Büyükinan M, Çatlı G, Ayhan Y, Baş VN, et al. Diverse Genotypes and Phenotypes of Three Novel Thyroid Hormone Receptor-α Mutations. J Clin Endocrinol Metab. 2016;101(8):2945–54. Estrada JM, Soldin D, Buckey TM, Burman KD, Soldin OP. Thyrotropin iso- forms: implications for thyrotropin analysis and clinical practice. Thyroid. 2014;24(3):411–23. Fliers E, Bianco AC, Langouche L, Boelen A. Endocrine and metabolic con- siderations in critically ill patients 4. Lancet Diabetes Endocrinol. 2015;3(10):816–25. 72 Z.-Z. Chen and J. V. Hennessey Garber JR, Cobin RH, Gharib H, Hennessey JV, Klein I, Mechanick JI, et al. Clinical practice guidelines for hypothyroidism in adults: cosponsored by the American Association of Clinical Endocrinologists and the American Thyroid Association. Endocr Pract. 2012;18(6):988–1028. Piketty M-L, Polak M, Flechtner I, Gonzales-Briceño L, Souberbielle J-C. False biochemical diagnosis of hyperthyroidism in streptavidin- biotin- based immunoassays: the problem of biotin intake and related interferences. Clin Chem Lab Med. 2017;55(6):780–8. Refetoff S, Weiss RE, Usala SJ. The syndromes of resistance to thyroid hor- mone. Endocr Rev. 1993;14(3):348–99. Tenenbaum-Rakover Y, Almashanu S, Hess O, Admoni O, Hag-Dahood Mahameed A, Schwartz N, et al. Long-term outcome of loss-of-function mutations in thyrotropin receptor gene. Thyroid. 2015;25(3):292–9. Management 7 of a Hospitalized Patient with Thyroid Dysfunction Megan Ritter and James V. Hennessey Contents Hyperthyroidism 74 Antithyroid Medications 74 Beta-Adrenergic-Blocking Drugs 75 Glucocorticoids 80 Iodine 80 Hypothyroidism 80 Suggested Reading 83 Abbreviations MMI Methimazole NPO Nil per os PO Per os M. Ritter (*) Weill Cornell Medicine, New York Presbyterian, New York, NY, USA e-mail: mer9114@med.cornell.edu J. V. Hennessey Beth Israel Deaconess Medical Center, Harvard Medical School, Division of Endocrinology, Diabetes and Metabolism, Boston, MA, USA e-mail: jhenness@bidmc.harvard.edu © Springer Nature Switzerland AG 2020 73 R. K. Garg et al. (eds.), Handbook of Inpatient Endocrinology, https://doi.org/10.1007/978-3-030-38976-5_7 74 M. Ritter and J. V. Hennessey PTU Propylthiouracil T3 Triiodothyronine T4 Thyroxine TSH Thyroid stimulating hormone Hyperthyroidism Antithyroid Medications Thyroid hormone affects virtually all organ systems. Thyroxine, or T4, can be viewed as the prohormone, while triiodothyro- nine, or T3, is the active form of thyroid hormone. Thyrotoxicosis can be manifested in many ways, including atrial fibrillation, weight loss, neuropsychiatric symptoms, or muscle weakness. The consequences of untreated hyperthyroidism include osteo- porosis as well as frank thyroid storm which can lead to cardio- vascular collapse and death. Further, untreated hyperthyroidism is associated with an increase in mortality, whereas treated hyperthyroidism negates this increase in risk. Longer duration of TSH suppression is associated with an increased hazard of mortality as well. Thus, it is prudent in both the outpatient and inpatient setting to establish the etiology of and to manage hyperthyroidism. The 2016 American Thyroid Association guidelines recom- mend methimazole (MMI) as the first-line antithyroid drug except during the first trimester of a hyperthyroid woman’s pregnancy when propylthiouracil (PTU) should be considered. PTU can also be used in select situations including MMI allergy or thyroid storm. Additional options for the treatment of hyperthyroidism include radioactive iodide or thyroidectomy and are tailored to patient preference in addition to concurrent medical conditions, including pregnancy or heart failure. It is reasonable to continue ambulatory doses of MMI or PTU upon hospital admission in patients who can continue to take medications orally, have no contraindications to continuing the medication, and are well- controlled outside the hospital. 7 Management of a Hospitalized Patient with Thyroid Dysfunction 75 Situations may arise where oral medications cannot be contin- ued upon a patient’s admission to the inpatient setting. The table reviews alternative methods physicians

have used to administer antithyroid medications. Studies range from healthy volunteers without thyroid dysfunction to case reports of critically ill patients with thyroid disease (Tables 7.1 and 7.2). These studies show that rectal administration of PTU and MMI in either enema or suppository is readily absorbed and well- tolerated for up to several days’ duration. Several options exist for making a rectal application of MMI or PTU. Pharmacy availabil- ity of the different materials will determine which formulation is ultimately used to treat an individual patient. Further, although suppositories might be better tolerated based on smaller size and potentially higher degree of retention, enemas have been shown to have more rapid and robust absorption. Intravenous (IV) medications can be advantageous when both oral and rectal administrations of medications are not possible. Given PTU’s properties, methimazole has more commonly been used as an intravenous medication. Although the use of these medications IV is not widespread, the above case studies support that they can be administered IV in a safe and an effective manner. Beta-Adrenergic-Blocking Drugs Beta-blockers are useful in managing symptoms related to hyper- thyroidism and are important initial tools in the treatment of all forms of hyperthyroidism, while diagnosis is established and con- trol of thyroid hormone levels is being established. Atenolol, esmolol, propranolol, and metoprolol are commonly used beta- blockers but others exist. Intravenous formulations of beta- blockers are generally readily available. No studies have been found comparing efficacy of IV propranolol, IV esmolol, or IV metoprolol in hyperthyroid patients. Pharmacokinetics of both IV esmolol and IV propranolol have been studied. IV esmolol has an elimination half-life of 2 minutes and a duration of action of 9 minutes. IV propranolol has an elimi- nation half-life of 10 minutes and a duration of action of 2.3 hours. 76 M. Ritter and J. V. Hennessey Table 7.1 Alternative strategies for methimazole administration Patient and study Dose Preparation characteristics Source (a) Intravenous administration 10–30 mg q 6–12 hours 500 mg of MMI USP powder 1. A 76-year-old man with Hodak SP, Huang C, Clarke D, reconstituted with pH-neutral biochemical Burman KD, Jonklaas J, 0.9% sodium chloride to attain hyperthyroidism and an ileus Janicic-Kharic N. Intravenous 10 mg MMI/mL was filtered and Clostridium difficile methimazole in the treatment through a 0.22 μm filter. Two diarrhea treated with IV of refractory hyperthyroidism. mL aliquots were transferred MMI 10 mg every 12 hours, Thyroid. 2006;16(7):691–695 into 10 mL sterile vials and increased to 10 mg every refrigerated. MMI was pushed 8 hours. Serum FT4 intravenously over 2 minutes decreased from 2.9 ng/dL to then followed by a normal 2.1 ng/dL saline flush 2. A 42-year-old male with end-stage liver disease had recurrent gastrointestinal bleeding. He was treated with IV MMI for 1 week (tapered from 30 mg IV MMI every 6 hours to 2.5 mg IV MMI every 12 hours); serum FT4 decreased from 5.6 ng/dL to 1.6 ng/dL 7 Management of a Hospitalized Patient with Thyroid Dysfunction 77 10 mg, one-time dose MMI powder was dissolved in Normal and hyperthyroid Okamura Y, Shigemasa C, 1 mL physiologic salt solution. patients were given a one-time Tatsuhara T. Pharmacokinetics Solution was enclosed in dose of methimazole in normal ampule and autoclaved to There was no difference in subjects and hyperthyroid sterilize pharmacokinetics of MMI patients. Endocrinol Jpn. between normal and 1985;33(5):605–615 hyperthyroid patients (b) Rectal administration 60 mg (one-time dose) Suppository: 1200 mg MMI One suppository was Nabil N, Miner DJ, Amatruda dissolved in 12 mL of water. administered to euthyroid JM. Methimazole: an Two drops Span 80® added to volunteers alternative route of 52 mL cocoa butter. Solution Peak serum MMI levels were administration. J Clin placed in 2.6 mL suppository not statistically different Endocrinol Metab. molds among groups. No thyroid 1982;54(1):180–181 outcomes were measured 78 M. Ritter and J. V. Hennessey Table 7.2 Alternative strategies for PTU administration Dose Preparation Patient characteristics Source (a) Intravenous 50 mg PTU tablets were dissolved in A 27-year-old woman, with a Gre Gregoire, G. Presented at the alkalinized 0.9% normal saline and history of multiple small bowel 77th annual meeting of the then administered in 50 mg/mL resections, developed Endocrine Society doses hyperthyroidism secondary to Graves’ disease (b) Rectal 400 mg (one- time Suppository: 200 mg of PTU Patients with biochemical Jongjaroenprasert W, Akarawut W, dose) dissolved into an unspecified hyperthyroidism were given a Chantasart D, Chailurkit L, amount of polyethylene glycol one-time dose of either Rajatanavin R. Rectal base suppository or enema before Administration of Propylthiouracil Enema: eight, 50 mg tablets of transition to PO medication in hyperthyroid patients: ground PTU dissolved in 90 mL Enema group was found to have comparison of suspension Enema sterile water higher peak levels of and suppository form. Thyroid. PTU. Concentration rT3 increased 2004;12(7):627–631 and serum FT3 decreased 7 Management of a Hospitalized Patient with Thyroid Dysfunction 79 400 mg q 6 hours Suppository: 50 mg PTU tablets A 47-year-old male with Zweig S, Schlosser JR, Thomas solubilized in light mineral oil. thyrotoxicosis and a perforated SA, Levy CJ, Fleckman This was mixed in 36 g of cocoa gastric ulcer AM. Rectal administration of butter solid suppository base. One Serum FT4 levels decreased from propylthiouracil in suppository gram suppository molds were 5.6 to 2.5 ng/dL during 5 days of form in patients with made administration thyrotoxicosis and critical illness: case report and review of literature. Endocr Pract. 2006;12(1):43–47 400 mg q 6 hours Suppository: eight, 50 mg PTU A 49-year-old woman with thyroid Walter RM Jr., Bartle WR. Rectal tables were dissolved in 60 mL storm and perforated viscus treated administration of propylthiouracil Fleets’ mineral oil or 60 mL with IV methylprednisolone, IV in the treatment of graves’ disease. Fleet’s phospho soda propranolol, and rectal PTU Am J Med. 1990;88(1):69–70 Serum thyroxine decreased from 26 μg/dL to 8.1 μg/dL after 3 days PTU is largely insoluble at physiologic pH, so its use intravenously is limited but has been reported 80 M. Ritter and J. V. Hennessey Onset of drug action is similar. Given the rapidly changing clinical course that often accompanies thyrotoxicosis, IV esmolol may be beneficial since effects wear off rapidly; however, caution should be observed since a drop in blood pressure may occur. Propranolol does have the effect of reducing plasma T3 concentrations. However, the clinical relevance of this is uncertain as the doses of propranolol that cause reductions in T3 are larger than doses used clinically. Metoprolol can be dosed intravenously every 4–6 hours for heart rate control and is beta-1 selective, which can be beneficial in patients with heart failure. Ultimately, esmolol, propranolol, and metoprolol are all viable options in managing hyperthyroidism, and the dose will be titrated according to the patient’s hemodynamics. Glucocorticoids Glucocorticoids decrease peripheral conversion of thyroxine (T4) to triiodothyronine (T3) and can be used to treat thyrotoxicosis and thyroid storm. Hydrocortisone, methylprednisolone, and dexamethasone are available in intravenous forms and a review of their characteristics is discussed below (Table 7.3). Iodine SSKI and Lugol’s solution can be used to treat thyrotoxicosis and are indicated in the treatment of thyroid storm. Inorganic iodine reduces release of preformed T3 and T4. Five drops of SSKI, 0.25 mL, is equivalent to 250 mg iodine and can be dosed every 6–8 hours. Five to seven drops of Lugol’s solution can be used every 6–8 hours as well. SSKI has been administered rectally after diluting with 20 to 60 mL of sterile water. Hypothyroidism Upon admission to an inpatient setting, levothyroxine (LT-4) ther- apy should be continued in order to maintain a patients’ euthyroid state. It is commonly known that LT-4 absorption is impaired by 7 Management of a Hospitalized Patient with Thyroid Dysfunction 81 Table 7.3 Characteristics of glucocorticoids Equivalent dose (mg) Anti-inflammatory activity Mineralocorticoid activity Duration of action (hours) Hydrocortisone 20 1 1 8–12 Prednisone∗ 5 4 0.8 12–36 Prednisolone 5 4 0.8 12–36 Methylprednisolone 4 5 0.5 12–36 Dexamethasone 0.75 25 0 36–72 ∗Prednisone is administered orally and there is no intravenous form Adapted From: Goodman LS, Brunton LL, Chabner B, Knollmann BC, editors. Goodman & Gilman’s pharmacological basis of therapeutics. 13th ed. New York: McGraw-Hill; 2017 82 M. Ritter and J. V. Hennessey food and that the ideal way to ensure a stable degree of absorption from day-to-day is to take LT-4 fasting, at least 60 minutes prior to eating or 3–4 hours after eating. There are several medications known to impair levothyroxine absorption. Clinically, this can be manifested as an increase in TSH and development of a frank hypothyroid state after previously being euthyroid. Medications that can reduce PO LT-4 absorption include calcium carbonate, cholestyramine, aluminum hydroxide, sevelamer, raloxifene, and ferrous sulfate. The data regarding proton pump inhibitor (PPI) impact on PO LT-4 absorption is mixed. But, LT-4 dosing should be separated from any PPI administration since increases in TSH with initiation of PPI therapy have been reported. During a patient’s hospitalization, LT-4 should not be combined with any other medications at time of administration and should be given while the patient is fasted (i.e., first thing in the morning or 3–4 hours after the last PO intake). In contrast to treating NPO hyperthyroid patients, treating NPO hypothyroid patients is simpler. Options for treating NPO patients with hypothyroidism include intravenous LT-4 or sublin- gual LT-4, or subcutaneous and intramuscular LT-4 injections have been reported in the literature, but these are not approved routes for administration. Pharmacokinetic studies of intramuscu- lar LT-4 have not been done, so it is difficult to determine if a dose change is required. The absorption of oral LT-4 is incomplete; thus, transitioning to intravenous dosing can be challenging. Understanding oral absorption can help better tailor intravenous LT-4 dosing. Hays and Nielson (1994, see suggested readings) analyzed LT-4 absorp- tion in patients based on age. In subjects between 21-year-olds and 69-year-olds, PO LT-4 absorption did not differ with age and was 69.3 ± 11.9%. In subjects over 70 years, PO LT-4 absorption was reduced at 62.8 ± 13.5% with p < 0.001. When transitioning to intravenous LT-4, a dose reduction of 30% for patients less than 70 years and 40% in patients over 70 years is reasonable. The American Thyroid Association recommends a dose reduction of 25% in a patient’s LT-4 dose in hospitalized patients with compro- mised enteral absorption. If intravenous LT-4 therapy is pro- longed, reassessing thyroid function may be indicated. 7 Management of a Hospitalized Patient with Thyroid Dysfunction 83 Levothyroxine can be prepared in liquid and soft gel forms. The liquid form of LT-4 is LT-4 dissolved in glycerol and etha- nol, while soft gel formulation LT-4 is dissolved in glycerin sur- rounded by a layer of gelatin. Both formulations have been used to treat hypothyroidism, particularly in patients with mal- absorption. Some patients with hypothyroidism are treated with a combi- nation of levothyroxine and liothyronine. It is reasonable to con- tinue an oral outpatient regimen in a euthyroid patient with an intact gastrointestinal tract and no impairments in absorption. Liothyronine is also available intravenously and can be used in the treatment of myxedema coma. There is extremely limited data on intravenous liothyronine in hospitalized patients who are NPO, so it would be reasonable to continue an outpatient regimen or tran- sition to LT-4 monotherapy at an appropriate increased dose. Suggested Reading Burch HB, Cooper DS. Management of graves disease: a review. JAMA. 2015;314(23):2544. Carroll R, Matfin G. Review: endocrine and metabolic emergencies: thyroid storm. Ther Adv Endocrinol Metab. 2010;1(3):139–45. De Leo S, Lee SY, Braverman LE. Hyperthyroidism. Lancet. 2016;388(10047):906–18. Goodman LS, Brunton LL, Chabner B, Knollmann BC, editors. Goodman & Gilman’s pharmacological basis of therapeutics. 13th ed. New York: McGraw-Hill; 2017. Hays MT, Nielsen KRK. Human thyroxine absorption: age effects and meth- odological analyses. Thyroid. 1994;4(1):55–64. Hodak SP, Huang C, Clarke D, Burman KD, Jonklaas J, Janicic-Kharic N. Intravenous methimazole in the treatment of refractory hyperthyroid- ism. Thyroid. 2006;16(7):691–5. Jongjaroenprasert W, Akarawut W, Chantasart D, Chailurkit L, Rajatanavin R. Rectal Administration of propylthiouracil in hyperthyroid patients: comparison of suspension enema and suppository form. Thyroid. 2004;12(7):627–31. Jonklaas J, Bianco AC, Bauer AJ, Burman KD, Cappola AR, Celi FS, et al. Guidelines for the treatment of hypothyroidism: prepared by the american thyroid association task force on thyroid hormone replacement. Thyroid. 2014;24(12):1670–751. 84 M. Ritter and J. V. Hennessey Liwanpo L, Hershman JM. Conditions and drugs

interfering with thyroxine absorption. Best Pract Res Clin Endocrinol Metab. 2009;23(6):781–92. Nabil N, Miner DJ, Amatruda JM. Methimazole: an alternative route of administration. J Clin Endocrinol Metab. 1982;54(1):180–1. Okamura Y, Shigemasa C, Tatsuhara T. Pharmacokinetics of methimazole in normal subjects and hyperthyroid patients. Endocrinol Jpn. 1985;33(5):605–15. Reilly CS, Wood M, Koshakji RP, Wood AJ. Ultra-short-acting beta-b lockade: a comparison with conventional beta-blockade. Clin Pharmacol Ther. 1985;38(5):579–85. Ross DS, Burch HB, Cooper DS, Greenlee MC, Laurberg P, Maia AL, et al. American thyroid association guidelines for diagnosis and management of hyperthyroidism and other causes of thyrotoxicosis. Thyroid. 2016;26(10):1343–421. Walter RM Jr, Bartle WR. Rectal administration of propylthiouracil in the treatment of graves’ disease. Am J Med. 1990;88(1):69–70. Wiersinga WM. Propranolol and thyroid hormone metabolism. Thyroid. 1991;1(3):273–7. Zweig S, Schlosser JR, Thomas SA, Levy CJ, Fleckman AM. Rectal admin- istration of propylthiouracil in suppository form in patients with thyro- toxicosis and critical illness: case report and review of literature. Endocr Pract. 2006;12(1):43–7. Perioperative 8 Management of Patients with Hyperthyroidism or Hypothyroidism Undergoing Nonthyroidal Surgery Catherine J. Tang and James V. Hennessey Contents Assess the Preoperative Patient 86 Evaluate the Type of Surgery 86 The Hypothyroid Patient Undergoing Nonthyroid Surgery 86 Elective Surgery 89 Urgent Surgery 91 The Hyperthyroid Patient Undergoing Nonthyroid Surgery 93 Elective Surgery 94 Urgent Surgery 95 References 98 C. J. Tang (*) · J. V. Hennessey Beth Israel Deaconess Medical Center, Harvard Medical School, Division of Endocrinology, Diabetes and Metabolism, Boston, MA, USA e-mail: ctang@bidmc.harvard.edu; jhenness@bidmc.harvard.edu © Springer Nature Switzerland AG 2020 85 R. K. Garg et al. (eds.), Handbook of Inpatient Endocrinology, https://doi.org/10.1007/978-3-030-38976-5_8 86 C. J. Tang and J. V. Hennessey Assess the Preoperative Patient • History and physical: Focus on comorbid cardiopulmonary disease and other endocrine disorders. • Labs: thyroid function test (TFT) including TSH and free T4 if hypothyroid or TSH, free T4, and total T3 if hyperthyroid; complete blood count (CBC) and basic metabolic panel (BMP). • Radiology: chest X-ray to look for tracheal deviation and com- pression. • Other testing: indirect laryngoscopy if thyroid gland is enlarged to look for vocal cord dysfunction (if present may indicate a difficult intubation). • Anesthesia: airway management. Evaluate the Type of Surgery Once a patient is determined to have a thyroid function abnormal- ity, the next step is to assess the urgency of the surgery. If the surgery can wait until thyroid hormone levels become normal, which may take several weeks or longer, then it is considered elective. If the surgery must be done within a few days, then it is considered urgent (see Table 8.1). The Hypothyroid Patient Undergoing Nonthyroid Surgery Thyroid hormone acts on nearly every tissue in the body and regu- lates essential metabolic pathways, including energy balance, thermogenesis, normal growth, and development [1, 2]. About 90% of free thyroid hormone circulating in the blood is in the form of thyroxine (T4), which is converted to the more potent triiodothyronine (T3) by deiodinase enzymes in the target tissues. The effect of thyroid hormone on the cardiopulmonary system is the primary concern in surgical outcomes, including decreases in 8 Perioperative Management of Patients with Hyperthyroidism… 87 Table 8.1 Summary table Elective surgery Urgent surgery Postoperative Hypothyroidism Subclinical Proceed Proceed Reassess thyroid hypothyroidism status as clinically indicated Moderate Wait until Age <60 years and no CVD: Start PO Continue PO LT4 hypothyroidism euthyroid LT4 at 1.6 mcg/kg daily at time of diagnosis Age >60 years or CVD: Start PO LT4 at 12.5–75 mcg daily at time of diagnosis Severe Wait until Immediately start IV LT4 200–400 mcg Switch to PO LT4 hypothyroidism euthyroid loading dose, followed by maintenance monotherapy IV given at 75% of oral dose 1.6 mcg/kg Taper glucocorticoid daily. If patient not responding, optional as tolerated addition of IV T3 at a loading dose of Reassess pituitary- 5–20 mcg, followed by a maintenance adrenal axis dose of 2.5–10 mcg every 8 h) to outpatient normalize thyroid function If hemodynamically unstable or pituitary-adrenal axis is unknown, start IV hydrocortisone 50–100 mg every 6–8 h before the administration of thyroid hormone (continued) 88 C. J. Tang and J. V. Hennessey https://www.facebook.com/groups/2202763316616203 Table 8.1 (continued) Elective surgery Urgent surgery Postoperative Hyperthyroidism Subclinical Start a BB Start a BB (preferably beta-1 selective Can stop BB hyperthyroidism (preferably blocker such as atenolol or metoprolol beta-1 selective succinate) and proceed blocker such as atenolol or metoprolol succinate) and proceed Overt hyperthyroidism Wait until BB + ATD ± inorganic iodine (add at Continue BB euthyroid least 1 h after thionamide is given) Continue thionamide GD and unable to tolerate ATD: Stop inorganic BB + inorganic iodine iodine TNG and unable to tolerate ATD: BB Taper glucocorticoid alone over 3 postoperative If high risk for thyroid storm regardless days of underlying thyroid etiology, also add Stop cholestyramine glucocorticoid ± cholestyramine Abbreviations: CVD cardiovascular disease, LT4 levothyroxine, T3 triiodothyronine or liothyronine, BB beta-blocker, ATD antithyroid drug (thionamide), GD Graves’ disease, TNG toxic nodular goiter, inorganic iodine – iopanoic acid, SSKI, or Lugol’s solution 8 Perioperative Management of Patients with Hyperthyroidism… 89 systemic vascular resistance, cardiac output, cardiac contractility, heart rate, blood volume, and blood pressure. Clinically, this may present as hypotension, bradycardia, hypoventilation, narrowed pulse pressure, cardiomyopathy, pericardial effusion, and tam- ponade. Other systems can also be adversely affected, resulting in constipation (decreased gastrointestinal motility), hyponatremia (increased antidiuretic hormone), anemia, hypoglycemia, and drug toxicity (reduced renal and hepatic clearance). However, there are no randomized controlled studies that eval- uated the surgical outcomes of hypothyroid versus euthyroid patients in nonthyroidal surgeries, though there are data from ret- rospective and observational studies. Clinical judgment pertain- ing to each individual case is thus crucial. Elective Surgery Subclinical Hypothyroidism Generally, if the surgery is elective, patients should be rendered euthyroid before proceeding to surgery. The exception may be sub- clinical hypothyroidism, where the thyroid-stimulating hormone (TSH) is elevated above reference range but typically lower than 10 μIU/ml and a normal free thyroxine level (FT4). A South Korean observational study found that subclinical hypothyroidism was associated with an increased incidence of transient postopera- tive atrial fibrillation (45.5% vs. 29%) in patients who underwent coronary artery bypass grafting (CABG) [3]. However, the sample size was small (N = 36 subclinical hypothyroid patients), and the subclinical hypothyroid group had a higher rate of preoperative acute myocardial infarction (within 3 months of CABG) compared with the euthyroid group, raising the possibility of acute myocar- dial damage that predisposed to atrial fibrillation. The study did not find any other differences in cardiopulmonary outcomes, including other types of arrhythmias, myocardial infarction, stroke, or respiratory complications. A Boston, Massachusetts, retrospec- tive study found that patients with subclinical hypothyroidism who underwent percutaneous transluminal angioplasty (PTCA) showed no differences in success of the procedure, hospital discharge 90 C. J. Tang and J. V. Hennessey destination, hospital costs, or in-hospital mortality [4]. Given the lack of strong evidence that subclinical hypothyroidism has a sig- nificant negative impact on surgical outcomes, the clinician must consider the patient’s other comorbidities and anticipate possible minor surgical complications. Thus, the decision to proceed with an elective surgery should be individualized, but generally it is rea- sonable to do so. Overt Hypothyroidism In contrast, in patients with overt hypothyroidism, where the TSH is above and the FT4 is below the reference range, elective sur- gery should be deferred until euthyroidism has been achieved. A retrospective study at Massachusetts General Hospital in Boston, MA, compared surgical outcomes in 40 hypothyroid patients matched with 80 euthyroid controls [5]. The hypothyroid patients had a median TSH of 99 (reference 0.5–3.5 μIU/ml) and T4 of 2.0 (reference 4.0–12.0 μg/dl). The study found that in noncardiac surgery, hypothyroid patients were more likely to have intraopera- tive hypotension, though it was corrected quickly with no associ- ated myocardial infarction or cerebrovascular accident. There was no difference in rate of intraoperative arrhythmias or the amount of blood loss. Among patients undergoing cardiac surgery, hypo- thyroid patients were more likely to have perioperative heart fail- ure, but no differences in myocardial infarction or arrhythmias were observed. The authors acknowledged that there might be an inherent bias in these observations in a retrospective study. Other notable findings in the hypothyroid patients included greater prev- alence of postoperative gastrointestinal (constipation, ileus) and neuropsychiatric (confusion, psychosis) occurrences in hypothy- roid patients. Hypothyroid subjects were also more likely to have experienced a difficult endotracheal intubation and were less likely to manifest postoperative fever in response to infection. However, there was no difference in pulmonary complications, hyponatremia, length of hospitalization, or death rates. Another retrospective study at the Mayo Clinic in Rochester, MN, compared surgical outcomes in 59 hypothyroid patients and 59 matched euthyroid controls [6]. The study found that hypothyroid patients had more preoperative risk factors including 8 Perioperative Management of Patients with Hyperthyroidism… 91 lower hemoglobin levels and a higher rate of hypertension, but no difference in surgical outcomes were observed, including intraop- erative blood pressure, arrhythmias, fluid and electrolyte imbal- ances, myocardial infarction, pulmonary complications, bleeding complications, sepsis, or length of hospitalization. There was a trend toward longer time to extubation in hypothyroid patients, though it was not statistically significant. Postoperative gastroin- testinal and neuropsychiatric outcomes were not assessed in this study. Admittedly, these studies are older, had small sample sizes, and may therefore not conclusive. Nonetheless, if a surgery is elective, it is prudent to render the overtly hypothyroid patient euthyroid to avoid any potential perioperative complications such as intraoperative hypotension and prolonged time to extubation. In addition to medical management, careful attention should also be paid to airway management. Obstructive goiters may be present in either hypo- or hyperthyroidism and can cause mechan- ical difficulties for the anesthesiologist. Retrosternal goiters may obstruct the inferior vena cava, and vocal cord dysfunction may cause a difficult intubation. For these reasons, patients with a goi- ter may need additional preoperative assessment including a chest X-ray to look for tracheal compression and deviation, as well as an indirect laryngoscopy to look for vocal cord dysfunction [7]. Urgent Surgery Subclinical Hypothyroidism In patients with subclinical hypothyroidism, it is generally fine to proceed with an urgent surgery, for the reasons stated above in “elective surgery.” Overt Hypothyroidism However, overtly hypothyroid patients should initiate thyroxine (T4) replacement as soon as possible as to minimize the delay in proceeding with an urgent surgery. Both the American Thyroid Association (ATA) and the American Association of Clinical Endocrinologists (AACE) recommend levothyroxine (LT4) to be the drug of choice in the treatment of hypothyroidism [8, 9], 92 C. J. Tang and J. V. Hennessey which has largely replaced the previously favored desiccated thy- roid. Due to the variable bioequivalence of several levothyroxine tablet and one gelatin formulations on the market, it is best to maintain the same preparation, whether brand or generic, in order to minimize fluctuations in thyroid hormone levels. The initial LT4 dosage depends on the severity of the hypothyroidism, etiol- ogy of the hypothyroidism, age, and comorbidities. In general, a full replacement dose of 1.6 mcg/kg of actual body weight per day can be initiated in younger patients (<60 years old) who are oth- erwise healthy with no cardiovascular comorbidities. On the other hand, older patients (>60 years) or those with known cardiovascu- lar disease should start at a lower dose, from 12.5 to 75 mcg daily, erring on the lower range in patients with cardiovascular disease and higher range if TSH >12 mIU/L [8, 9]. Repeat TSH monitor- ing should be done every 4–6 weeks. However, given that it usu- ally takes several weeks for the TSH to normalize, it is not always necessary to wait until euthyroidism if surgery must be done urgently. As such, the surgery and anesthesia teams should antici- pate possible mild to moderate perioperative complications as detailed above under “elective surgery” and prepare accordingly. In the case of severe hypothyroidism where the patient has or is suspected to have myxedema coma, surgery should be delayed until patient is adequately treated. However, if surgery is urgent and cannot be delayed, then intravenous (IV) replacement with LT4 should be instituted immediately, at a loading dose of 200– 400 mcg, with the lower end of the range for patients who are older and have smaller body habitus and in the presence of cardio- vascular comorbidities. Maintenance levothyroxine should be given intravenously at 75%

(approximately the fraction of oral LT4 that is absorbed) of the oral dosing at 1.6 mcg/kg of body weight per day. The administration of intravenous liothyronine (T3) in addition to IV LT4 is optional. The rationale is that periph- eral T4 to T3 conversion is decreased in acutely ill patients, and IV T3 may accelerate clinical improvement. If used, the loading dose of IV T3 is 5–20 mcg, followed by a maintenance dose of 2.5–10 mcg every 8 h. Again, one should aim for the lower end of the range for patients who are older and have smaller body habitus and in the presence of cardiovascular comorbidities [9]. 8 Perioperative Management of Patients with Hyperthyroidism… 93 Concurrent adrenal insufficiency must be considered in myx- edematous patients, and if they are hemodynamically unstable or the function of their pituitary-adrenal axis is unknown, IV stress dose corticosteroids should be given before the administration of thyroid hormones, to avoid the precipitation of an adrenal crisis. A typical regimen is hydrocortisone 50–100 mg every 6–8 h. Once clinical improvement ensues, LT4 can be switched to an oral formulation. Generally, T3 is not continued orally and LT4 mono- therapy is preferred for maintenance. Note that even with immedi- ate treatment, patients in myxedema coma remain high surgical risk and should have close perioperative hemodynamic monitor- ing. Thus, a multidisciplinary approach with the surgeon, anesthe- siologist, and endocrinologist is crucial in caring for these patients who must undergo urgent surgery. The Hyperthyroid Patient Undergoing Nonthyroid Surgery Excess thyroid hormone produces classic features of hyperthy- roidism such as weight loss, tremor, heat intolerance, and hyper- activity. In the cardiovascular system, too much thyroid hormone increases cardiac contractility, heart rate, cardiac output, and sys- tolic blood pressure, while decreases in diastolic blood pressure and systemic vascular resistance are observed. In the pulmonary system, excess thyroid hormone increases oxygen consumption, respiratory rate, and minute ventilation while decreasing vital capacity and lung compliance. Cardiovascular symptoms can include palpitations, shortness of breath, tachycardia, widened pulse pressure, cardiac murmurs, and chest pain. Of particular interest to the anesthesiologist are atrial fibrillation, ischemic heart disease, and congestive heart failure [7]. Surgery in patients with poorly controlled thyrotoxicosis has been associated with a mortality rate as high as 20% [10], primar- ily due to the precipitation of thyroid storm. However, the actual surgical risk is probably lower in recent decades due to better perioperative management. Again, as in hypothyroidism, there are few randomized controlled studies that have evaluated the surgi- 94 C. J. Tang and J. V. Hennessey cal outcomes of hyperthyroid versus euthyroid patients in nonthy- roidal surgeries. Clinical judgment pertaining to each individual case is thus crucial. Elective Surgery Subclinical Hyperthyroidism Patients with subclinical hyperthyroidism, where the TSH may be slightly suppressed but the levels of T3 and T4 are normal, can generally proceed with an elective surgery, after the initiation of a beta-blocker if no contraindications are evident. A randomized, prospective Swedish study compared surgical outcomes in 30 hyperthyroid patients undergoing thyroid surgery (hemithyroid- ectomy or subtotal resection) and preoperatively managed with either methimazole and thyroxine or metoprolol alone [11]. The methimazole and thyroxine group was treated for 12 weeks and was rendered clinically and biochemically euthyroid prior to thy- roid surgery. The metoprolol group was treated for 5 weeks and remained biochemically hyperthyroid but appeared clinically euthyroid prior to thyroid surgery. The two groups did not differ in anesthetic or cardiovascular complications, nor did anyone suf- fer from thyroid storm. The authors’ conclusion was that meto- prolol alone may be a reasonable choice for preoperative management for hyperthyroid patients needing thyroid surgeries, with the advantage of a shorter preoperative treatment period and without suffering any apparent serious complications. One limita- tion of the study was that the dose of metoprolol was 200–400 mg per day in divided doses, which is much higher than is typically used today. But it is worth noting that this study particularly looked at overtly hyperthyroid patients who underwent thyroid surgeries to treat their hyperthyroidism, so it is likely that the dose of beta-blocker requirement is actually much lower in subclinical hyperthyroid patients. Although some clinicians still prefer pro- pranolol for its reduction of peripheral T4 to T3 conversion, a beta-1 selective blocker such as atenolol or metoprolol succinate is probably better given its longer duration of action and greater safety in patients with obstructive pulmonary disease. A starting 8 Perioperative Management of Patients with Hyperthyroidism… 95 dose may be atenolol or metoprolol succinate 25–50 mg daily and uptitrate as needed for a target heart rate of less than 80 bpm. Overt Hyperthyroidism Because of the risk of precipitating thyroid storm, elective surger- ies should always be postponed in patients with overt hyperthy- roidism, until the patient is rendered euthyroid [10, 12]. Moreover, atrial fibrillation occurs in 10–15% of patients with overt hyper- thyroidism with the prevalence higher in older individuals [12]. Urgent Surgery Subclinical Hyperthyroidism Patients with subclinical hyperthyroidism may proceed with urgent surgeries after the initiation of a beta-blocker, for reasons as discussed above. Overt Hyperthyroidism Overtly hyperthyroid patients should wait until euthyroid before proceeding with surgery. However, if surgery cannot wait and is urgent or emergent, immediate action must be taken to stabilize thyrotoxicosis to reduce the risk of perioperative mortality. For all thyrotoxic patients regardless of etiology, beta-blockers should be initiated immediately. Calcium channel blockers such as diltia- zem and verapamil can be used if beta-blockers are contraindi- cated. There is no general consensus on the superiority of any particular beta-blocker, though each may offer its advantages. Nonspecific beta-blocker propranolol has the additional benefit of blocking 5′-mono deiodinase activity, thus decreasing peripheral T4 to T3 conversion at higher doses, and can be started at 40–80 mg PO every 4–8 h and titrated for a target heart rate less than 80 bpm [10]. Alternatively, beta-1 selective blocker such as atenolol or metoprolol succinate may be used, at an initial dose of 25–50 mg. Though they are longer acting than propranolol and the conventional once daily dosing is more convenient, realistically they may still need to be given twice daily due to the accelerated clearance seen in hyperthyroidism [12]. IV administration may be 96 C. J. Tang and J. V. Hennessey achieved through metoprolol tartrate, propranolol, or esmolol. Esmolol has the shortest half-life of only a few minutes and thus the advantage of fast adjustment of hemodynamic parameters; an initial loading dose is 250–500 mcg/kg, followed by maintenance infusion of 50–100 mcg/kg/min [10, 13]. Beta- blockers should be continued postoperatively in nonthyroidal surgeries, possibly in lower doses, for as long as the patient remains clinically thyro- toxic or until the underlying cause of the hyperthyroidism is addressed. If the underlying etiology of hyperthyroidism is Graves’ dis- ease or toxic nodular goiter, thionamide therapy should be insti- tuted as soon as an urgent nonthyroid surgery is deemed necessary. Thionamide is a class of antithyroid drugs (ATD) which includes propylthiouracil (PTU), methimazole (MMI), and carbimazole. Only PTU and MMI are available in the United States, whereas carbimazole is available in Europe and elsewhere. PTU and MMI can be given either orally or rectally. They block new thyroid hor- mone synthesis by inhibiting the enzyme thyroid peroxidase, which is responsible for the organification of iodine and the cou- pling of mono- and diiodotyrosines to make T3 and T4. Since thionamides largely affect new thyroid hormone synthesis but not secretion of preformed thyroid hormones, they usually take 3–8 weeks to achieve euthyroidism [12]. In the case of an urgent surgery that may take place within a matter of days or hours, thi- onamide therapy alone is not adequate, and additional treatment should be instituted to stabilize the thyrotoxicosis, which are dis- cussed below. PTU is shorter acting and has the additional benefit of decreased T4 to T3 conversion. A typical starting dose of PTU is 100–150 mg every 6–8 h [12]. However, if the patient is severely hyperthyroid or if thyroid storm is suspected, the 2016 ATA guidelines recommend a PTU loading dose of 500–1000 mg, fol- lowed by 250 mg every 4 h [14]. MMI is longer acting and is generally preferred over PTU for its lesser degree of toxicities, except during the first trimester of pregnancy, when PTU is less teratogenic. A typical starting dose of MMI is 20–40 mg daily, though in a severely hyperthyroid patient, the dose is increased to 60–80 mg per day [14], which may be divided into two to three doses daily due to the increased clearance seen in hyperthyroid- 8 Perioperative Management of Patients with Hyperthyroidism… 97 ism. Both PTU and MMI have similar side effect profiles and have 50% cross-reactivity. Minor toxicities include rash, urticaria, and arthralgia, which occur in 1–5% of patients [10, 12]. A more seri- ous complication is hepatotoxicity, which is more common in PTU (2.7%) than in MMI (0.4%), though liver failure remains rare in either (0.03–0.05%) [14]. Perhaps the most dreaded toxic- ity is agranulocytosis, which occurs in 0.1–0.5% of patients [10, 14], with the vast majority of cases occurring within 60–90 days of starting therapy. Although the effect is conventionally thought to be dose-related in MMI (rare at doses below 40 mg daily) but not in PTU, a more recent Danish study demonstrated that the average dose of MMI at the onset of agranulocytosis was 25 mg daily in patients with Graves’ disease [15]. Postoperatively, ATD should be continued at the same dose until thyroid hormone levels are no longer elevated. As mentioned earlier, since ATD takes 3–8 weeks to achieve euthyroidism, it alone is not sufficient in preparing a patient for urgent surgery. In such cases, inorganic iodine should be added to ATD to decrease the production of new thyroid hormone, which is also known as the Wolff-Chaikoff effect and can be seen within 24 h of administration. In addition, inorganic iodine also decreases the release of preformed thyroid hormone. A normal thyroid gland will eventually escape the Wolff-Chaikoff effect and resume thy- roid hormone production, but the effect may persist in those with autoimmune thyroid disease. In contrast, a toxic nodular goiter (TNG) may use the excess iodine as substrate to make more thy- roid hormone (known as the Jod-Basedow effect), further exacer- bating the thyrotoxicosis [10, 12]. For this reason, while iodine can be used as monotherapy in Graves’ disease, it should never be used as monotherapy in TNG. In fact, it is crucial that iodine should be given at least an hour after the administration of a thionamide. In this setting where thyrotoxicosis must be urgently stabilized, iodine can be administered as saturated solution of potassium iodide (SSKI) five drops (50 mg of iodide per drop) every 6 h [12, 14]. Inorganic iodine should be stopped after surgery. Other agents that may be used in the acutely thyrotoxic patient regardless of the underlying cause in preparation of urgent surgery include glucocorticoids and cholestyramine. Glucocorticoids 98 C. J. Tang and J. V. Hennessey reduce peripheral T4 to T3 conversion within hours and can be tapered over 72 h postoperatively [10, 12]. Choice of glucocorti- coids includes hydrocortisone 100 mg every 8 h, dexamethasone 2 mg every 6 h, or betamethasone 0.5 mg every 6 h, which all can be given either IV or PO (betamethasone can also be given as IM) [10, 12]. Postoperatively, glucocorticoids should be tapered over the course of 72 h [10, 12]. Cholestyramine, a bile acid seques- trant, binds to thyroid hormone in the intestine and reduces its reabsorption, thus decreasing its enterohepatic circulation. It is not a first- or second-line agent but is potentially useful in situations where it is not possible to render the patient completely euthyroid prior to surgery or if the patient is intolerant of ATD [14]. A typical dose of cholestyramine used in this setting is 4 g four times daily [12]. Cholestyramine is generally stopped postoperatively. Patients who are intolerant of ATD and those who have Graves’ disease as the underlying etiology may be treated with beta- blockers and iodine, with the addition of glucocorticoid and pos- sibly cholestyramine if hyperthyroidism is severe. In ATD-intolerant patients whose underlying etiology is TNG, preoperative manage- ment may consist of beta-blockers alone, with the addition of glu- cocorticoid and possibly cholestyramine if hyperthyroidism is severe. Iodine is not used in this

scenario due to concern of exacer- bating thyrotoxicosis, for the reasons stated previously. As in the case of severe hypothyroidism, patients with severe hyperthyroidism remain high surgical risk despite optimal periop- erative management. A multidisciplinary approach with the sur- geon, anesthesiologist, and endocrinologist is vital in the caring of these patients who require urgent surgeries, and careful attention must be paid to airway management (due to goiters, as discussed in the previous section) and hemodynamic monitoring. References 1. Lin JZ, Martagón AJ, Cimini SL, Gonzalez DD, Tinkey DW, Biter A, et al. Pharmacological activation of thyroid hormone recep- tors elicits a functional conversion of white to brown fat. Cell Rep. 2015;13(8):1528–37. 8 Perioperative Management of Patients with Hyperthyroidism… 99 2. Mullur R, Liu YY, Brent GA. Thyroid hormone regulation of metabolism. Physiol Rev. 2014;94(2):355–82. 3. Park YJ, Yoon JW, Kim KI, Lee YJ, Kim KW, Choi SH, et al. Subclinical hypothyroidism might increase the risk of transient atrial fibrillation after coronary artery bypass grafting. Ann Thorac Surg. 2009;87(6):1846–52. 4. Mantzoros CS, Evagelopoulou K, Moses AC. Outcome of percutaneous transluminal coronary angioplasty in patients with subclinical hypothy- roidism. Thyroid. 1995;5(5):383–7. 5. Ladenson PW, Levin AA, Ridgway EC, Daniels GH. Complications of surgery in hypothyroid patients. Am J Med. 1984;77(2):261–6. 6. Weinberg AD, Brennan MD, Gorman CA, Marsh HM, O’Fallon WM. Outcome of anesthesia and surgery in hypothyroid patients. Arch Intern Med. 1983;143(5):893–7. 7. Farling PA. Thyroid disease. Br J Anaesth. 2000;85(1):15–28. 8. Garber JR, Cobin RH, Gharib H, Hennessey JV, Klein I, Mechanick JI, Pessah-Pollack R, Singer PA, Woeber KA, American Association of Clinical Endocrinologists and American Thyroid Association Taskforce on Hypothyroidism in Adults. Clinical practice guidelines for hypothy- roidism in adults: cosponsored by the American Association of Clinical Endocrinologists and the American Thyroid Association. Endocr Pract. 2012;18(6):988–1028. 9. Jonklaas J, Bianco AC, Bauer AJ, Burman KD, Cappola AR, Celi FS, et al. Guidelines for the treatment of hypothyroidism: prepared by the American Thyroid Association task force on thyroid hormone replace- ment. Thyroid. 2014;24(12):1670–751. 10. Langley RW, Burch HB. Perioperative management of the thyrotoxic patient. Endocrinol Metab Clin North Am. 2003;32(2):519–34. 11. Adlerberth A, Stenstrom G, Hasselgren PO. The selective beta 1-blocking agent metoprolol compared with antithyroid drug and thyroxine as preop- erative treatment of patients with hyperthyroidism. Results from a pro- spective, randomized study. Ann Surg. 1987;205(2):182–8. 12. Palace MR. Perioperative management of thyroid dysfunction. Health Serv Insights. 2017;10:1178632916689677. 13. Buget MI, Sencan B, Varansu G, Kucukay S. Anaesthetic management of a patient with thyrotoxicosis for nonthyroid surgery with peripheral nerve blockade. Case Rep Anesthesiol. 2016;2016:9824762. 14. Ross DS, Burch HB, Cooper DS, Greenlee MC, Laurberg P, Maia AL, et al. 2016 American Thyroid Association guidelines for diagnosis and management of hyperthyroidism and other causes of thyrotoxicosis. Thyroid. 2016;26(10):1343–421. 15. Andersen SL, Olsen J, Laurberg P. Antithyroid drug side effects in the population and in pregnancy. J Clin Endocrinol Metab. 2016;101(4): 1606–14. Thyroid Problems 9 Encountered Specifically in Inpatients with Cardiac Disease Jeena Sandeep and James V. Hennessey Contents Hyperthyroidism/Thyrotoxicosis 102 Hypothyroidism 103 Hyperthyroidism/Thyrotoxicosis 104 Effects on the Cardiovascular System 104 Management of Thyrotoxicosis in the Hospitalized Patient 105 Establishing the Etiology of Thyrotoxicosis 105 Treatment Recommendations Based on Etiology 106 Subclinical Hyperthyroidism 107 Treatment with Amiodarone in Patients with Cardiac Arrhythmias 108 J. Sandeep (*) St. Elizabeth Medical Center, Department of Medicine, Division of Endocrinology, Brighton, MA, USA e-mail: jeena.sandeep@steward.org J. V. Hennessey Beth Israel Deaconess Medical Center, Harvard Medical School, Division of Endocrinology, Diabetes and Metabolism, Boston, MA, USA e-mail: jhenness@bidmc.harvard.edu © Springer Nature Switzerland AG 2020 101 R. K. Garg et al. (eds.), Handbook of Inpatient Endocrinology, https://doi.org/10.1007/978-3-030-38976-5_9 102 J. Sandeep and J. V. Hennessey Hypothyroidism and Cardiac Disease 108 Impact of Hypothyroidism on Cardiovascular Risk Factors 108 Cardiovascular Hemodynamics in Hypothyroidism 109 Heart Failure and Arrhythmias in Hypothyroidism 109 Diagnosis of Hypothyroidism 110 Treatment of Overt Hypothyroidism 111 Thyroid Hormone Treatment in Cardiac Failure and Patients Undergoing Cardiac Surgery 112 Suggested Reading 113 Hyperthyroidism/Thyrotoxicosis • Thyrotoxicosis – excessive thyroid hormone regardless of eti- ology – affects the cardiovascular system resulting in cardiac arrhythmias, EKG changes, congestive heart failure (CHF) and angina, and/or myocardial infarction. • Patients with subclinical thyrotoxicosis may be asymptomatic, manifest resting tachycardia, or may develop atrial fibrillation and be at risk for embolic stroke. • Establishing the etiology of thyrotoxicosis is critical in deter- mining appropriate therapeutic interventions. The most com- mon conditions are Graves’ disease, toxic multinodular goiter (toxic MNG) or toxic adenoma (TA), iatrogenic thyrotoxico- sis, and subacute thyroiditis (SAT). • Amiodarone can cause either hypothyroidism or thyrotoxico- sis. There are two types of amiodarone-induced thyrotoxico- sis (AIT) which may be difficult to differentiate from one another. • Diagnostic evaluation includes the history, medication (iatro- genic) and supplement ingestion (factitia), viral infection, neck pain [suggestive of subacute thyroiditis]), presence of eye and pretibial symptoms (Graves’ disease), and presence of thyroid nodules (toxic nodular goiter or adenoma). Lab testing includes TSH, assessment of free thyroxine, T3 testing, thyrotropin receptor antibodies (TRab), and 123-I uptake and/or scanning. 9 Thyroid Problems Encountered Specifically in Inpatients… 103 • Treatments include beta-blockers, antithyroid drugs (methima- zole generally preferred), 131-I ablation, and/or surgery. Treatment for type 1 AIT includes antithyroid drugs, potas- sium perchlorate (not available in the United States), or sur- gery, and type 2 AIT may either be observed, or if mild, may be treated with glucocorticoids, and if inadequately controlled, may require surgery. Hypothyroidism • Hypothyroidism may affect cholesterol metabolism and other CV risk factors such as C-reactive protein and homocysteine that may promote CAD and can also predispose patients to atrial fibrillation (Afib). • Hypothyroidism increases systemic vascular resistance (SVR) and diminishes cardiac output, stroke volume, and heart rate resulting in lower cardiac output. • Diminished erythropoietin production results in blood volume decrease. Capillary permeability allows pericardial and pleural effusions further compromising cardiac (tamponade) and pul- monary function. • Prolongation of the QT interval predisposes to ventricular arrhythmia and AV nodal dysfunction. • Diagnosis requires measurement of TSH and free T4 (FT4) levels. Overt hypothyroidism is recognized when TSH is clearly elevated (over 10 mIu/L) and FT4 is low, while sub- clinical hypothyroidism is characterized by a sustained TSH level between the upper reference range and 10 mIu/L and a normal FT4. • Treatment with levothyroxine (LT4) in overt hypothyroidism improves LDL cholesterol metabolism, diastolic hypertension, and cardiac dysfunction while accelerating heart rate and delaying progression of atherosclerosis. Due to potential underlying CAD, caution in reestablishing euthyroidism may be warranted. Gradual increasing doses at intervals of 6–8 week allow equilibration of thyroid hormone levels before retesting. 104 J. Sandeep and J. V. Hennessey • Coronary bypass surgery is generally considered safe in patients with hypothyroidism. Concerns regarding periopera- tive care and the risk of postoperative atrial fibrillation in hypothyroid patients undergoing cardiac surgery have been raised. Hyperthyroidism/Thyrotoxicosis Effects on the Cardiovascular System Increased levels of circulating thyroid hormones alter the function of the cardiovascular system and produce significant clinical effects. Increased systolic blood pressure due to increased con- tractile ventricular force and decreased diastolic pressure due to lower peripheral vascular resistance result in wide pulse pressure and enhanced cardiac output. Tachycardia reflects a positive chro- notropic action of thyroid hormone on the heart. Increased stroke volume and volume expansion result in increased circulating blood volume. The drop in DBP leads to activation of renin angio- tensin aldosterone system increasing sodium reabsorption and consequent volume expansion. Thyroid hormone directly increases erythropoietin synthesis increasing the red cell mass increasing the preload. Resting tachycardia results from thyroid hormone acting on the sinus node to increase heart rate. The net result of decreased afterload, increased left ventricular contractil- ity, increased preload, wide pulse pressure, and increased heart rate is an overall increase in cardiac output. Cardiac Arrhythmias and EKG Changes Arrhythmias include resting tachycardia and atrial fibrillation which may convert to sinus rhythm within 8–12 weeks after effec- tive antithyroid treatment in the absence of valvular heart disease and of recent onset. EKG changes include the supraventricular arrhythmias and nonspecific ST segment and T-wave changes as well as a short PR interval due to an increased rate of conduction through the AV node. 9 Thyroid Problems Encountered Specifically in Inpatients… 105 Congestive Heart Failure (CHF) and Angina Hyperthyroidism may cause or worsen preexisting cardiac dis- ease by increasing myocardial oxygen demand, contractility, and heart rate. These changes may lead to silent ischemia, angina, or compensated heart failure and even endothelial dysfunction. Heart failure in the thyrotoxic patient occurs in the setting of underlying coronary heart disease and/or atrial fibrillation. Cardiac failure can occur even in young patients without known underlying heart disease suggesting a cardiomyopathy associated with thyrotoxicosis which may well be reversible. Recent reports suggest an increased prevalence of pulmonary hypertension in the setting of thyrotoxicosis suggesting that thyroid hormones increase pulmonary vascular resistance in contrast to its effects on systemic vascular resistance. Management of Thyrotoxicosis in the Hospitalized Patient Establishing the Etiology of Thyrotoxicosis Symptomatic patients with thyrotoxicosis, especially patients with heart rates above 90 beats per minute, elderly patients, and patients with underlying cardiovascular disease should be treated with beta-blockers while determining the etiology of thyrotoxicosis. The most frequent endogenous diagnoses of thy- rotoxicosis include Graves’ disease, toxic multinodular goiter (toxic MNG) or toxic adenoma (TA), and subacute thyroiditis (SAT). Graves’ disease patients may be recognized by the pres- ence of pretibial myxedema and eye involvement. Those with toxic MNG and toxic adenoma may have a prior history of thy- roid nodules. The history may point to thyroiditis as the etiol- ogy. Some may report recent viral infection followed by pain and tenderness over the thyroid or have recently given birth prior to developing thyrotoxicosis. A radioactive iodine uptake is useful to distinguish exogenous sources of thyrotoxicosis and thyroiditis from Graves’ disease and toxic nodular disorders. 106 J. Sandeep and J. V. Hennessey Iodine uptake may be near zero in exogenous thyrotoxicosis, painless, postpartum or painful subacute (deQuervains) thyroid- itis. In addition, low or near zero uptake can also be seen in cases of ectopic thyroid disorders such as Struma Ovarii and after iodine contamination such as post CT with iodine contrast or while on treatment with medications such as amiodarone. Radioactive iodine uptake is typically elevated and diffuse in Graves’ disease while focal nodular patterns are seen in patients with toxic nodular goiters. Treatment Recommendations Based on Etiology Thyrotoxicosis due to exogenous thyroid hormone application is best treated with beta-blocker and discontinuation of the thyroid hormone source. Severely toxic individuals may respond quickly to plasmapheresis. Patients with subacute thyroiditis are treated with beta-blockers, nonsteroidal anti-inflammatory drugs for pain, and if needed corticosteroids. Once Graves’ disease is con- firmed, treatment may include antithyroid drugs (ATDs), radio- active iodine, or surgery. Patient preference is an important factor in deciding on a treatment modality. However, patients with acute cardiopulmonary disease who are clinically unstable are usually treated with ATDs followed by radioactive iodine treatment, particularly if they are still at high surgical risk. Patients who are pregnant, lactating mothers, those with coexist- ing thyroid cancer, and those unable to comply with radiation safety guidelines or planning a pregnancy within 4–6 months should not receive radioactive iodine. On the other hand, definite contraindications to antithyroid medication usage and switching include major adverse reactions such as agranulocytosis and hepatotoxicity. Methimazole is the preferred ATD with the potential exception of thyroid storm, the first trimester of pregnancy, and minor reac- tions to methimazole when propylthiouracil (PTU) is preferred. Starting dose of methimazole is usually 10–20 mg once daily and is titrated based on the clinical and biochemical response. PTU is shorter acting and is administered multiple times a day with start- 9 Thyroid Problems Encountered Specifically in Inpatients… 107 ing dose from 50 to 150 mg three times daily. Patients are moni- tored for side effects including rash, agranulocytosis (fever, sore throat), and cholestasis (pale stools, jaundice, abdominal pain, nausea, vomiting). Thyrotoxicosis due to toxic multinodular goiter (TMNG) or toxic adenoma (TA) should be treated with radioactive iodine or surgery. Chronic low-dose methimazole can be considered for those with contraindications to definitive treatment. In patients with cardiovascular disease or elderly patients or in situations where there is high risk of worsening of hyperthyroidism post- radioactive iodine treatment, beta-blockers should be initiated prior to administering radioactive iodine treatment.