Treating Addictive Behaviors: Processes of Change

Treating Addictive
Behaviors
Processes of Change
APPLIED CLINICAL PSYCHOLOGY
Series Editors:
Alan S. Bellack, Medical College of Pennsylvania at EPPI, Philadelphia, Pennsylvania,
and Michel Hersen, University of Pittsburgh, Pittsburgh, Pennsylvania
Current Volumes in this Series
BEHAVIORAL ASSESSMENT AND REHABILITATION OF THE

TRAUMATICALLY BRAIN DAMAGED
Edited by Barry A. Edelstein and Eugene T. Couture
HANDBOOK OF ASSESSMENT IN CHILDHOOD PSYCHOPATHOLOGY

Applied Issues in Differential Diagnosis and Treatment Evaluation
Edited by Cynthia L. Frame and Johnny L. Matson
HANDBOOK OF BEHAVIORAL GROUP THERAPY

Edited by Dennis Upper and Steven M. Ross
ISSUES IN PSYCHOTHERAPY RESEARCH

Edited by Michel Hersen, Larry Michelson, and Alan S. Bellack
A PRIMER OF HUMAN BEHAVIORAL PHARMACOLOGY

Alan Poling
THE PRIVATE PRACTICE OF BEHAVIOR THERAPY

Sheldon J. Kaplan
RESEARCH METHODS IN APPLIED BEHAVIOR ANAL YSIS

Issues and Advances
Edited by Alan Poling and R. Wayne Fuqua
SEVERE BEHAVIOR DISORDERS IN THE MENTALLY RETARDED

Nondrug Approaches to Treatment
Edited by Rowland P. Barrett
SUBSTANCE ABUSE AND PSYCHOPATHOLOGY

Edited by Arthur I. Alterman
TREATING ADDICTIVE BEHAVIORS

Processes of Change
Edited by William R. Miller and Nick Heather
A Continuation Order Plan is available for this series. A continuation order will bring delivery
of each new volume immediately upon publication. Volumes are billed only upon actual ship·
ment. For further information please contact the publisher.
Treating Addictive
Behaviors
Processes of Change
Edited by
William R. Miller
Univmily of New Mexico
Albuquerque, New Merico
and
Nick Heather
Universily of Dundee
Dundee, Scolland
Plenum Press • New York and London

Library of Congress Cataloging in Publication Data
Treating addictive behaviors.
(Applied clinical psychology)

Based on papers presented at the 3rd International Conference on Treatment of Ad-
dictive Behaviors, held at North Berwick, Scotland in Aug. 1984.
Includes bibliographies and index.
1. Compulsive behavior-Treatment-Congresses. 2. Substance abuse-Treatment
-Congresses. I. Miller, William R. II. Heather, Nick. III. International Conference on
Treatment of Addictive Behaviors (3rd: 1984: North Berwick, Scotland) IV. Series.
[DNLM: 1. Compulsive Behavior-therapy-congresses. 2. Substance Dependence-
therapy-congresses. WM 176 T784 1984J
RC533.T73 1986 616.85'227 86-20463
ISBN-13: 978-1-4612-9289-0 e-ISBN-13: 978-1-4613-2191-0
001: 10.1007/978-1-4613-2191-0
1098765
© 1986 Plenum Press, New York

Softcover reprint of the hardcover 1st edition 1986
A Division of Plenum Publishing Corporation
233 Spring Street, New York, N.Y. 10013
All rights reserved

No part of this book may be reproduced, stored in a retrieval system, or transmitted
in any form or by any means, electronic, mechanical, photocopying, microfilming,
recording, or otherwise, without written permission from the Publisher
To
Lillian and Richard
W. R. M.
In loving memory of
Nicholas Heather
N.H.
Contributors
Helen M. Annis, Addiction Research Foundation, and University of

Toronto, Toronto, Ontario, Canada
Claus-Peter Appel, Department of Applied Psychology, University of

Lund, Lund, Sweden
Geir Berg, The Hjellestad Clinic, Hjellestad, Norway
Gerhard Biihringer, Addiction Research Group, Department of Psychol-

ogy, Max-Planck Institute for Psychiatry, Munich, West Germany
Sabine Dehmel, Addiction Research Group, Department of Psychology,

Max-Planck Institute for Psychiatry, Munich, West Germany
Carlo C. DiClemente, Texas Research Institute of Mental Science, Hous-

ton, Texas
Dennis M. Donovan, Department of Psychiatry and Behavioral Sciences,

University of Washington, and Alcohol Dependence Treatment
Program, Veterans Administration Medical Center, Seattle, Wash-
ington
Christopher P. Freeman, Department of Psychiatry, University of Edin-

burgh, and Royal Edinburgh Hospital, Morningside Park, Edin-
burgh, Scotland
Marcus Grant, Division of Mental Health, World Health Organization,

Geneva, Switzerland
vii
viii CONTRIBUTORS
Nick Heather, Department of Psychiatry, Ninewells Medical School, Uni-

versity of Dundee, Dundee, Scotland
Reid K. Hester, Behavior Therapy Associates, Albuquerque, New Mexico
Joanne R. Ito, Department of Psychology, University of Washington, and

Health Services Research and Development, Veterans Administra-
tion Medical Center, Seattle, Washington
Frederick H. Kanter, Department of Psychology, University of Illinois,

Champaign, Illinois
Franz Klett, Addiction Research Group, Department of Psychology,

Max-Planck Institute for Psychiatry, Munich, West Germany
Simonne LeBreton, Addiction Research Foundation, Toronto, Ontario,

Canada
Gloria K. Litman, Addiction Research Unit, Institute of Psychiatry, Lon-

don, England
Barbara S. McCrady, Center of Alcohol Studies, Rutgers University, Pis-

cataway, New Jersey
William R. Miller, Department of Psychology, University of New Mexico,

Albuquerque, New Mexico
Jim Orford, Department of Psychology, University of Exeter, Exeter,

England
James O. Prochaska, Department of Psychology, University of Rhode Is-

land, Kingston, Rhode Island
Howard Rankin, St. Andrew's Hospital, Northampton, England

Martin Raw, Department of Psychology, St. George's Hospital Medical
School, London, England
Bruce Ritson, Department of Psychiatry, University of Edinburgh, An-

drew Duncan Clinic, Royal Edinburgh Hospital, Edinburgh,
Scotland
lan Robertson, Department of Psychology, Astley Ainslie Hospital, Edin-

burgh, Scotland
CONTRIBUTORS ix
David Robinson, Addiction Research Centre, University of Hull, Hull,

England
Arvid Skutle, The Hjellestad Clinic, Hjellestad, Norway
Stephen Sutton, Addiction Research Unit, Institute of Psychiatry, Lon-

don, England
D. Adrian Wilkinson, Addiction Research Foundation, Toronto, Ontario,

Canada
Preface
About a decade ago, psychologists began exploring the commonalities

among alcohol and drug abuse, smoking, and obesity. The term sub-
stance abuse evolved into the current concept of addictive behaviors,
which recognizes similarities with other behaviors that do not involve
consummatory responses (e.g., pathological gambling, compulsions,
sexual deviations). Professional societies and journals now have been
founded in both Britain and the United States with the purpose of focus-
ing on research and treatment in the area of addictive behaviors.
As the field has evolved, new models have emerged to address the
questions and puzzles that face professionals. This volume examines
some of these current issues and, in particular, explores common pro-
cesses of change that seem to cut across the addictive behaviors. The
chapters are based on papers presented at the Third International Con-
ference on Treatment of Addictive Behaviors, which was held at North
Berwick, Scotland, in August of 1984. The conference was organized
around an integrative model of stages and processes of change that has
been useful in organizing new knowledge about how to intervene with
addictive behaviors. This model is set forth by its authors, Jim Prochaska
and Carlo DiClemente, in Chapter 1. In Chapter 2, Fred Kanfer ex-
pounds his own model of self-regulation, which overlaps nicely with the
Prochaska-DiClemente framework and provides a behavioral-theoretical
context.
The remainder of the book is organized around stages at which
clients come into treatment, and at which professionals are called on to
intervene. Part II addresses issues of motivation for change. Marcus
Grant opens this section with a commentary on the roles of the World
Health Organization in engendering change in the addictive behaviors.
Claus-Peter Appel explores applications of experimental social psychol-
xi
xii PREFACE
ogy in general, and decision-making models in particular, to problems

of individual client motivation. Jim Orford discusses research on the
critical minimal conditions for change, and Steve Sutton reviews data on
how smokers decide to quit.
Part III, the largest section of this volume, is a compendium of
theory and research on how people change once they begin taking ac-
tion to alter addictive behaviors. Miller and Hester, in two chapters, first
survey current knowledge on the effectiveness of alternative treatments
for alcohol abuse, then review research on matching clients with optimal
intervention approaches. Geir Berg and Arvid Skutie present the results
of an early intervention program for Norwegian problem drinkers. Chris
Freeman discusses the eating disorders and raises controversial ques-
tions as to whether all of these should be classed with the addictive
behaviors. Two chapters describe and evaluate behavioral treatment
programs for drug abusers in Toronto (Wilkinson and LeBreton) and
Munich (Dehmel, Klett, and Biihringer). The literature on smoking ces-
sation methods is summarized by Martin Raw.
The remaining six chapters in Part III explore important aspects of
treatment and change that transcend particular intervention approaches.
David Robinson traces the history and role of mutual-aid groups. Barbara
McCrady discusses family involvement in addictive behaviors and the
change process, reporting the results of a study and incorporating the
larger literature on social support. Ian Robertson explores the ap-
plicability of cognitive theory and research in understanding and treating
addictive behaviors. Nick Heather expounds on the use of self-help
manuals to assist individuals in changing their behavior, clarifying the
most common method by which people change: self-directed change
without the aid of a therapist. The transtheoretical issue of dependence
remains important for addictive behaviors, and Howard Rankin presents
theoretical perspectives and recent experimental evidence relevant to
treatment. Finally, Bruce Ritson commends "the merits of simple inter-
vention," pointing to data that suggest that we have been making treat-
ment and change altogether too expensive and complicated, bypassing
some relatively simple interventions that suffice for a large percentage of
clients.
Any professional working in this field recognizes that initial change
is only a beginning and is no guarantee of long-range success. One of
the more important insights of the past decade has been increased
awareness of the need for special measures to prevent relapse. Gloria
Litman provides an exegesis of her "survival" model for predicting and
preventing relapse in addictive behaviors, and Helen Annis presents a
comprehensive approach by which relapse risk can be assessed and
reduced. In the final chapter of this volume, Joanne Ito and Dennis
PREFACE xiii
Donovan evaluate research on approaches to aftercare following the

active treatment or change phase.
We are enthusiastic about the progress that is represented in this
volume. The convergence of work from a wide variety of settings is
obvious. Our contributors represent the nations of Canada, England,
Norway, Scotland, Sweden, Switzerland, the United States, and West
Germany. Their professions include psychiatry, psychology, public
health, and sociology. The participants in the conference itself repre-
sented a still broader range of nations and professions. We are encour-
aged not only by the increasing international interest in addictive behav-
iors, but also by the consistency of findings that emerge from well-
designed research across diverse settings. Clear progress is being made
toward more efficacious and cost-effective treatment of these costly,
perplexing, and often devastating problem behaviors.
The editing of an international volume of this kind poses special
challenges. English represents a second or third language for some of
the contributors, and we express our admiration for their multilingual
skills. In conventions of spelling, we have for consistency adhered to
U.S. forms, and referencing in all chapters conforms to the publication
standards of the American Psychological Association.
WILLIAM R. MILLER
NICK HEATHER
Acknowledgments
We wish to express our gratitude to those who made special contribu-

tions to the quality of this volume. We particularly express our apprecia-
tion to Ian Robertson for his help in organizing and arranging the Third
International Conference on Treatment of Addictive Behaviors. We also
thank Eliot Werner, our editor at Plenum Press, for his encouragement
and assistance in the publication process. We acknowledge with appre-
ciation the assistance of the following organizations, which provided
partial funding for the conference while allowing us complete freedom
to determine speakers and content: the Alcohol Education and Research
Fund, the Brewers' Society, and the Scotch Whisky Association. Both
the conference and the preparation of this book were supported in
important ways by the New Directions in the Study of Alcohol Group,
the University of Dundee, and the University of New Mexico. Finally,
we are grateful to the many colleagues who participated in the Scottish
conference, bringing their enthusiasm and expertise to the task of taking
one more step toward the alleviation of the human suffering surround-
ing the addictive behaviors.
xv
Contents
PART I. CHANGE IN THE ADDICTIVE BEHAVIORS
CHAPTER 1. Toward a Comprehensive Model of Change 3

James O. Prochaska and Carlo C. DiClemente
CHAPTER 2. Implications of a Self-Regulation Model of Therapy for

Treatment of Addictive Behaviors 29
Frederick H. Kanfer
PART II. CONTEMPLATION: MOTIVATION FOR CHANGE AND

PREVENTION
CHAPTER 3. From Contemplation to Action: The Role of the World

Health Organization 51
Marcus Grant
CHAPTER 4. From Contemplation to Determination: Contributions

from Cognitive Psychology 59
Claus-Peter Appel
CHAPTER 5. Critical Conditions for Change in the Addictive

Behaviors 91
Jim Orford
xvii
xviii CONTENTS
CHAPTER 6. Trying to Stop Smoking: A Decision-Making

Perspective 109
Stephen Sutton
PART III. ACTION: ASPECTS AND PROCESSES OF CHANGE
CHAPTER 7. The Effectiveness of Alcoholism Treatment: What

Research Reveals 121
William R. Miller and Reid K. Hester
CHAPTER 8. Matching Problem Drinkers with Optimal

Treatments 175
William R. Miller and Reid K. Hester
CHAPTER 9. Early Intervention with Problem Drinkers 205

Geir Berg and Arvid Skutle
CHAPTER 10. Strategies of Change in Eating Disorders 221

Christopher P. Freeman
CHAPTER 11. Early Indications of Treatment Outcome in Multiple

Drug Users 239
D. Adrian Wilkinson and Simonne LeBreton
CHAPTER 12. Description and First Results of an Outpatient Drug-

Free Treatment Program for Opiate Dependents 263
Sabine Dehmel, Franz Klett, and Gerhard Biihringer
CHAPTER 13. Smoking Cessation Strategies 279

Martin Raw
CHAPTER 14. Mutual Aid in the Change Process 289

David Robinson
CONTENTS xix
CHAPTER 15. The Family in the Change Process 305

Barbara S. McCrady
CHAPTER 16. Cognitive Processes in Addictive Behavior

Change 319
Ian Robertson
CHAPTER 17. Change without Therapists: The Use of Self-Help

Manuals by Problem Drinkers 331
Nick Heather
CHAPTER 18. Dependence and Compulsion: Experimental Models of

Change 361
Howard Rankin
CHAPTER 19. Merits of Simple Intervention 375

Bruce Ritson
PART IV. MAINTENANCE: PREVENTING RELAPSE
CHAPTER 20. Alcoholism Survival: The Prevention of Relapse 391

Gloria K. Litman
CHAPTER 21. A Relapse Prevention Model for Treatment of

Alcoholics 407
Helen M. Annis
CHAPTER 22. Aftercare in Alcoholism Treatment: A Review 435

Joanne R. lto and Dennis M. Donovan
Index 457
I
Change in the Addictive Behaviors
1
Toward a Comprehensive Model
of Change
JAMES O. PROCHASKA AND CARLO C. DICLEMENTE
In 1984, a group of researchers, theorists, and therapists gathered at an

international conference in Scotland to contribute to the development of
a more comprehensive model of change for the treatment of addictive
behaviors. The conference and this book that grew out of the conference
are signs of the zeitgeist; they are part of a new attempt to integrate
diverse systems of psychotherapy (Prochaska, 1984). In his classic call
for a rapproachment across competing systems of therapy, Goldfried
(1980) signaled that it is time to move beyond parochial approaches to
treatment. It is time to move toward more comprehensive models of
change.
A comprehensive model of change must meet many competing
demands. A comprehensive model of change in addictive behaviors will
need to be applicable to the broad range of ways that people change-
from maximum interventions of traditional inpatient and outpatient
therapy programs to more minimal interventions, such as a few hours of
therapy for problem drinkers (Miller & Baca, 1983; Orford, this volume)
or self-help manuals for troubled drinkers (Heather, this volume) and
smokers (Glasgow, Schafer, & O'Neil, 1981). A comprehensive model of
JAMES O. PROCHASKA • Department of Psychology, University of Rhode Island, King-

ston, Rhode Island 02881-0801. CARLO C. DICLEMENTE • Texas Research Institute of
Mental Science, Houston, Texas 77030. The research in this chapter was funded by Grant
CA27821 from the National Cancer Institute.
3
4 JAMES O. PROCHASKA AND CARLO C. DICLEMENTE
change will also need to address the fact that with some addictive behav-
iors, like smoking, the vast majority of people change entirely on their
own without the aid of formalized treatment programs (National In-
stitute on Drug Abuse, 1979).
A comprehensive model must also be applicable to the variety of
addictive behaviors that people wish to change. The model will need to
advance our understanding of how people change such diverse behav-
iors as alcohol abuse, cocaine dependence, compulsive gambling, over-
eating, heroin addiction, and smoking. Are there commonalities of
change that can account for how people succeed and fail in their at-
tempts to modify such diverse behaviors?
Furthermore, a comprehensive model should help to serve as a
synthesis for the diverse treatment methods that are currently available
for addictive behaviors. In 1976 ParIoff reported that there were more
than 130 therapies available in the therapeutic marketplace (or "jun-
gleplace," as he more aptly described it). By 1980, ParIoff had docu-
mented more than 250 therapies. People wishing to overcome addictive
behaviors are confronted with the confusion of too many choices with
too few data to decide what should be the treatments of choice for their
particular problems. A comprehensive model can help to integrate a
therapy field that has fragmented into an overwhelming number of
alternative and competing treatments.
A comprehensive model will need to cover the full course of
change, from the time someone becomes aware that a problem exists to
the point at which a problem no longer exists. Most models of change
have been models of action, but there are many changes that precede
and follow a person taking action with addictive behaviors. Trying to
decide how to help someone to change includes taking into account
where in the cycle of change a particular person is.
Just as change is a dynamic and open phenomenon, so too does a
comprehensive model of change need to be open to new developments,
incorporating and integrating additional variables that are discovered to
play important roles in how people change addictive behaviors. We
shall present a model of change as it is currently defined, recognizing
that it is neither complete nor closed.
The model of change that we have been developing over the past
decade is not simple, but it is comprehensible. The days of searching for
simple solutions to complex problems should be behind us. The com-
plexities of changing addictive behaviors require multivariate rather
than univariate solutions. The trans theoretical approach that we have
been developing is a three-dimensional model that integrates stages,
processes, and levels of change.
In a comparative study of self-changers versus smokers participat-
TOWARD A COMPREHENSIVE MODEL OF CHANGE 5
ing in two well-known commercial treatment programs, we discovered

that both self-changers and therapy-changers identified common stages
of change that they had experienced in the course of quitting smoking
(DiClemente & Prochaska, 1982). In developing an instrument for as-
sessing the stages of change that clients are in when entering therapy,
McConnaughy, Prochaska, and Velicer (1983) found that four highly
reliable and well-defined components emerged from a study of 150 gen-
eral psychiatric outpatients beginning therapy. The four components
were identified as the precontemplation, contemplation, action, and
maintenance stages of change.
As predicted, the four stages of change formed a simplex pattern in
which adjacent stages were more highly correlated with each other than
with any other stage. These results on the stages of change have been
replicated with 350 general psychiatric patients presenting for outpatient
therapy (McConnaughy, Prochaska, Velicer, & DiClemente, 1984). This
study also found that the patient's stage of change was a better predictor
of progress after 4 months of therapy than were DSM-III diagnoses or
severity of symptoms. The stages of change have also been identified in
150 alcoholics presenting for outpatient therapy (DiClemente & Hughes,
1985).
Figure 1 presents a linear array of the stages of change. It indicates
how successful change involves progressing from precontemplation, to
contemplation, to action, and into the maintenance stage of change. A
major problem in the treatment of addictive behaviors, however, is that
most individuals do not progress linearly through the stages of change.
Figure 2 presents a cyclical pattern that is much more common with
individuals attempting to overcome addictive problems on their own or
in therapy.
In 1971, Hunt, Barnett, and Branch demonstrated that across a
broad range of therapies, between 70% and 80% of alcoholics, heroin
addicts, and smokers relapsed within a year after treatment. Similar
results have been found with obese individuals (Olcott, 1985). That is,
relapse is the rule rather than the exception.
However, most individuals do not give up after relapsing. In a
longitudinal study of 886 self-changers representing the different stages
PRECONTEMPLA TlON - CONTEMPLATION - ACTION - MAINTENANCE
fIGURE 1. A linear pattern of the stages of change.

\
RELAPSE
ADDICTED LIFE
OF
PRECONTEMPLATORS
FIGURE 2. The revolving-door model of the stages of change.
of change, we found that 84% of relapsers moved back into the con-
templation stage and were seriously intending to quit again within a
year. Rather than give up to avoid further failure, most smokers cycle
back into the contemplation stage. On the average, self-changers make
three serious revolutions through the stages of change before they exit
into a life relatively free from temptations to smoke (Marlatt, this vol-
ume; Prochaska & DiClemente, 1983b; Schacter, 1983).
Unfortunately some individuals never get free from their addictive
behaviors. Some individuals get stuck in particular stages of change. Of
a group of 113 individuals contemplating quitting smoking, nearly a
third failed to take action after 2 years of contemplating change (Pro-
chaska & DiClemente, 1983b).
Therapy with addictive behaviors can progress most smoothly if
both the client and the therapist are focusing on the same stage of
change. One type of resistance in therapy occurs when the client and
therapist are working at different stages of change. The more directive,
action-oriented therapist would find a client who is at the contemplation
stage to be highly resistant to therapy. From the client's perspective,
however, the therapist may be seen as wanting to move too quickly. On
the other hand, a therapist who specializes in contemplating and under-
standing the causes of problems will tend to see a client who is ready for
action as resistant to the insight aspects of therapy. The client would be
warned against acting out impulsively. From the client's perspective,
however, the therapist might be warned against moving too slowly.

Therapists, like clients, can get stuck in a favored stage of change.
What do individuals do to progress from one stage of change to the
next? What are the basic processes of change that are used successfully
to complete the cycle of change? A comparative analysis of 29 leading
systems of therapy yielded 10 basic processes of change (Prochaska,
1984). The transtheoretical approach assumes that integration across a
diversity of therapy systems can occur most likely at an analytical level
between theoretical assumptions and therapeutic techniques-the level
of processes of change. Interestingly, Goldfried (1980, 1982), in his call
for a rapproachment, has independently suggested that the principles or
processes of change were the appropriate theoretical starting point at
which integration could occur.
The processes of change represent a middle level of abstraction
between the basic theoretical assumptions of a system of therapy and the
techniques proposed by the theory. A process of change represents a type
of activity that is initiated or experienced by an individual in modifying
affect, behavior, cognitions, or relationships. Whereas there are a large
number of coping activities, there are a limited set of processes that
represent the basic change principles underlying these activities. Con-
sciousness-raising, for example, is the most widely used change process
across diverse therapy systems (Prochaska, 1984). But there are many
therapeutic techniques for increasing consciousness. Educational tech-
niques, confrontational techniques, observational techniques, video-
feedback techniques, and interpretations are just some of the techniques
used to help clients become more aware of themselves and their
problems.
Table 1 presents the 10 processes of change that have received the
TABLE 1.
Ten Change Processes of the
Transtheoretical Approach
1. Consciousness-raising
2. Self-liberation
3. Social liberation
4. Counterconditioning
5. Stimulus control
6. Self-reevaluation
7. Environmental reevaluation
8. Contingency management
9. Dramatic relief
10. Helping relationships
most theoretical and empirical support in our work to date on addictive

behaviors. These basic processes of change have been identified not
only in theoretical and empirical analysis of leading therapy systems
(Prochaska & DiClemente, 1984), they have also been identified in retro-
spective, cross-sectional, and longitudinal studies of self-changers (Di-
Clemente & Prochaska, 1982; Prochaska & DiClemente, 1985).
A common set of change processes has been clearly identified
across such diverse problem areas as psychic distress, smoking, and
weight control (Prochaska & DiClemente, 1985). In each problem area
the set of change processes accounted for nearly 70% of the variance in a
principal component analysis of the Processes of Change Questionnaire.
Not only were a common set of change processes identified across prob-
lem areas, but there were also important similarities in how frequently
the change processes were used across problems. When processes were
ranked in terms of how frequently they were used for each of the three
behavior problems, the rankings of the processes were nearly identical
across problem areas. Helping relationship, consciousness-raising, and
self-liberation, for example, were the top three ranking processes across
problems, whereas reinforcement management and stimulus control
were the lowest ranked processes.
Significant differences do occur, however, in the absolute frequency
of the use of the change processes across problem areas. Individuals rely
more on helping relationships and consciousness-raising for overcom-
ing psychic distress than they do for weight control and smoking cessa-
tion. Weight control subjects rely more on self-liberation and stimulus
control than do distressed individuals. Research to date provides strong
support for the assumption that there is a common set of change pro-
cesses that individuals use in attempts to overcome such problems as
psychic distress and addictive behaviors.
Most major systems of psychotherapy emphasize only two or three
processes of change (Prochaska, 1984). Both clients and self-changers,
however, utilize 8 to 10 processes of change (Norcross & Prochaska, in
press). One of the assumptions of the transtheoretical approach is that
therapists should be at least as cognitively complex as their clients. They
should be able to think and intervene in terms of a more comprehensive
set of change processes.
One of the most helpful findings to emerge from our research with
self-changers and therapy changers is that particular processes of
change are emphasized during particular stages of change (Prochaska &
DiClemente, 1983). The integration of stages and processes of change
can serve as an important guide for therapists. Once it is clear what
stage of change a client is in, the therapist would know which processes
to apply in order to help the client progress to the next stage of change.
Rather than apply change processes in a haphazard or trial-and-error

fashion, therapists could begin to use change processes in a much more
systematic style.
Table 2 presents a diagram showing the integration that was re-
vealed from our research between the stages and processes of change
(Prochaska & DiClemente, 1983). During the precontemplation stage
individuals use the change processes significantly less than people in
any other stage. Precontemplators process less information about their
problems; they spend less time and energy reevaluating themselves;
they experience fewer emotional reactions to the negative aspects of
their problems; they are less open with significant others about their
problems; and they do little to shift their attention or their environment
in the direction of overcoming their problems. In therapy these are
clients who are most resistant to the therapists' efforts to help them
change. Later we will discuss how therapists can help resistant clients
move from precontemplation to contemplation.
Clients in the contemplation stage are most open to consciousness-
raising interventions, such as observations, confrontations, and in-
terpretations (Prochaska & DiClemente, 1983). Contemplators are much
more likely to use bibliotherapy and other educational interventions. As
clients become increasingly more conscious about themselves and the
nature of their problems, they are freer to reevaluate themselves both
affectively and cognitively. Self-reevaluation includes an assessment of
which values clients will try to actualize, to act on, and to make real.
Clients also need to assess which values they will let die. The more
central their problem behaviors are to the core of themselves, the more
will their reevaluation involve changes iI1 their sense of self. Clients ask
themselves, "Will I like myself better as a nondrinker or nonsmoker?
Will others I care about like me better? What if I am a more anxious or
irritable person after I change? If my shared community is primarily with
TABLE 2.
The Stages of Change in which Particular Processes of Change Are
Emphasized the Most and the Least
Precontemplation Contemplation Action Maintenance
Eight processes Consciousness-raising

used the least Self-reevaluation
Self-liberation
Helping relationship
Reinforcement management
Counterconditioning
Stimulus control
10: JAMES O. PROCHASKA AND CARLO C. DICLEMENTE
drinkers or smokers, will I risk rejection? If I fail to change, will I feel

coerced, guilty, or weak?"
During the action stage it is important that clients act from a sense
of self-liberation (Prochaska & DiClemente, 1983). They need to believe
that they have the autonomy to change their lives in key ways. Yet they
also need to accept that coercion is as much a part of life as is autonomy.
Thus, if they slip during action and attribute it all to a lack of willpower,
they can experience considerable guilt or shame that can keep them
from trying to take action again. On the other hand, if clients attribute al~
of their success to a therapist or to a helping relationship, they risk
becoming unduly dependent on a therapist.
Self-liberation is based in part on a sense of self-efficacy (Bandura,
1977, 1982), the belief that one's own efforts playa critical role in suc-
ceeding in the face of difficult situations. Self-liberation, however, can-
not have just an affective and cognitive foundation. Clients must also be
effective enough with behavioral processes, such as countercondition-
ing and stimulus control, to modify the conditional stimuli that can
coerce them into relapsing (Prochaska, DiClemente, Velicer, Ginpel, &
Norcross, 1985). Therapists can assess how adequately clients are able to
apply processes such as contingency management and stimulus control.
Therapists can provide training, if necessary, in behavioral processes to
increase the probability that clients will be successful when they do take
action. As action proceeds, therapists can serve as consultants to the
clients as self-changers, to help clients identify any errors they may be
making in their attempts to modify their behavior and environment in a
freer and healthier direction.
Because action is a partiqtlarly stressful stage of change that in-
volves considerable opportunities for experiencing coercion, guilt,
failure, and the limits of personal freedom, clients are particularly in
need of support and understanding from helping relationships (Pro-
chaska & DiClemente, 1983). For clients, taking action tends to meal'l:
taking risks with rejection. Knowing that there is at least one person
who cares and is committed to helping serves to ease some of the dis-
tress and dread of taking life-changing actions.
Just as preparation for action is essential for success, so too is prepa-
ration for maintenance. Successful maintenance builds on each of the
processes that has come before. Specific preparation for maintenance,
however, involves an open assessment of the conditions under which ~
person is likely to relapse. Clients need to assess the alternatives they
have for coping with such conditions without resorting to self-defeating
defenses and pathological patterns of response. Perhaps most important
is the sense that one is becoming more of the kind of person one wants
to be. Continuing to apply counterconditioning and stimulus control is
most effective when it is based on the conviction that maintaining

change supports a sense of self that is highly valued by oneself and at
least one significant other.
Just as the processes of change can be integrated with the stages of
change, so too can other important change variables be integrated with
the stages of change. Self-efficacy, for example, was presented by Ban-
dura (1977) as the critical variable that can lead toward a unifying theory
of behavior change. In our research, we have found self-efficacy to be ali
important variable in understanding and predicting changes in addictive
behaviors (DiClemente, 1981; DiClemente, Prochaska, & Gibertini,
1985). In our research on smoking, for example, we developed a mea-
sure of self-efficacy that represents the level of confidence individuals
have that they can resist smoking across a broad range of tempting
situations. We also developed a measure that assesses the level of temp-
tation subjects report for these same situations.
A 12-item version of the self-efficacy measure was found to predict
which self-changers and therapy changers would maintain their non-
smoking 5 to 7 months after quitting (DiClemente, 1981). A 31-item
version of self-efficacy was found to differ significantly across the stages
of change. In a cross-sectional analysis, self-efficacy was found to in-
crease from pre contemplation, to contemplation, to action, into mainte-
nance. Self-efficacy did not stabilize until approximately 18 months after
quitting smoking. Temptation levels, on the other hand, fell from pre-
contemplation, to contemplation, to action, and into maintenance;
Temptation did not level off until approximately 3 years after quitting
smoking (DiClemente, Prochaska, & Gibertini, 1985).
These data suggest a working definition of when people successful-
ly terminate from the cycle of change. We assume that individuals suc-
cessfully terminate an addictive behavior when their temptation levels
are zero and their confidence levels are 100% across all problem situa-
tions. Our data suggest that some people are able to terminate an addic-
tive behavior like smoking, whereas others remain in the maintenance
stage even though they have not smoked for 5 years or more.
Decisional balance is another variable that has been presented as a
cornerstone for building a more comprehensive model of change (Janis
& Mann, 1977). In our research on self-change approaches to smoking
cessation, we developed a decisional-balance measure based on Janis
and Mann's (1977) model of decision making. Their model suggested
four separate components of decision making. Principal components
analysis of our 32-item decisional balance questionnaire, however,
yielded only two reliable and well-defined components (Velicer, DiCle~
mente, Prochaska, & Brandenburg, 1985). The components were simply
labeled the "pros of smoking" and the "cons of smoking." Rather than
being polar opposites, these components were independently defined

(i.e., orthogonal). Thus, individuals could be high on both pros and
cons of smoking, low on each, or high on one and low on the other.
Cross-sectional data indicated an interesting pattern in the balance
given to the pros and cons depending on which stage of change people
are in. As expected, precontemplators have a pattern of high pros and
low cons. Contemplators are high on the pros but the cons of smoking
are also high, with the cons slightly outweighing the pros. The people in
the action stage report a pattern in which the cons remain somewhat
higher than the pros, but both are lower than for the contemplators.
Finally, the long-term maintainers show significant reduction of both
the pros and cons of smoking, with the cons still somewhat greater than
the pros. These data suggest that over months and years of not smoking
both the pros and cons of smoking decrease in value until smoking
becomes almost a nonissue for many former smokers (Velicer, DiCle-
mente, Prochaska, & Brandenburg, 1985).
Key change variables like the processes of change, self-efficacy,
temptation, and decisional balance can be used to predict progress from
one stage of change to the next. In a 2-year longitudinal study of 886 self-
changers, six significant discriminative functions predicted movement
for the groups representing the pre contemplation, contemplation, ac-
tion, and relapse stages. (Prochaska, DiClemente, Velicer, Ginpil, &
Norcross, 1985). The long-term quitters representing the maintenance
stage did not produce enough relapse for study. The discriminative
functions involved predicting progress over a 6-month period. The vari-
ables entered into the functions included the 10 processes of change,
self-efficacy, temptation, and the decisional-balance measures.
The six discriminant functions were not only statistically significant
but are also of immense practical significance. These functions were all
defined by variables that are open to change. These functions were not
defined by static variables, such as sex, age, or smoking history, which
are not amenable to psychosocial intervention. Rather, these functions
were defined by processes of self-change, self-efficacy, and decision
making. Not only are these predictor variables capable of modification,
but they can be brought under self-control rather than having to be a
function of professional intervention.
When more static variables, such as age, education, income, smok-
ing history, family's smoking history, reasons for smoking, withdrawal
symptoms, and health problems, were used as predictor variables, the
results were much less significant. Of the 17 predictor variables used in
this research, nearly two thirds demonstrated no significant relationship
to behavior change (Wilcox, Prochaska, Velicer, & DiClemente, 1985).
No significant discriminant function was found for predicting move-
ment out of the contemplation stage. Three variables (greater health

problems, smoking less for pleasure, and fewer years smoked) did pre-
dict movement out of the precontemplation stage. The second signifi-
cant function indicated that following a relapse, individuals with higher
education and income levels are more likely to try again. The third
discriminant function indicated that lighter smokers are more likely to
maintain quitting or take further action than are heavier smokers.
Important patterns of change have been identified in our 2-year
longitudinal study of self-change approaches to smoking cessation (Pro-
chaska, Velicer, & DiClemente, 1985). Cluster analyses determined ty-
pologies for how subjects move through the stages of change. The four
most common patterns of change are (a) a linear profile in which indi-
viduals progress directly from one stage to the next; (b) the more com-
mon cyclical profile in which individuals begin to take action and then
relapse, followed by further contemplation and action before substantial
improvement is maintained; (c) an unsuccessful cyclical profile; and (d) a
non progressing profile in which individuals remain stuck in a stage like
precontemplation or contemplation, without improving over time.
A total of 14 profiles emerged from the cluster analyses. The 14
groups were compared on the processes of change, the pros and cons of
smoking, self-efficacy, and temptation to smoke. Figure 3 presents an
example of how the groups differed in their use of one change process
(self-reevaluation) over the five rounds of the longitudinal study. Figure
3 is not expected to clarify but rather confuse the reader, because confu-
sion is what was produced in the researchers for a considerable period of
time. Clearly, there were group differences in terms of how frequently
the change process was used. But what did these differences mean?
Patterns of change did not become clear until particular profiles
were ipfegrated cross-sectionally and longitudinally across the stages of
changl ,:Prochaska, Velicer, & DiClemente, 1985). Processes of change
were graphed across the following profiles: (a) individuals who re-
mained in precontemplation (Group 1); (b) individuals who progressed
from precontemplation to contemplation (Group 14); (c) individuals who
progressed from contemplation to action (Group 12); (d) individuals
who took repeated action during the 2 years (Group 3); (e) individuals
who progresc;ed from action to maintenance (Group 5); and, (f) indi-
viduals who progressed from maintenance to termination (Group 4).
Figure 4 presents the pattern that emerged for the utilization of self-
reevaluation across the stages of change. This pattern was dubbed "Mt.
Change." The pattern indicates that the utilization of self-reevaluation
peaks during contemplation and action and then gradually reduces dur-
ing maintenance until it returns to levels comparable to those used dur-
ing precontemplation. Similar patterns were found for almost all of the
58
56 10
54
52
>- 50 13
U
Z 13
W II 14
::J 48
0
W
a:::
I.J.. 46
,~~:
44
42
40
2 3 4 5
ROUND
FIGURE3. Comparison of 14 profile groups on frequency of use of self-reevaluation across 5
rounds of self-change.
FREQUENCY
62 r-S-E-L--F---R-E-E-V--A-L-U-A-T--IO~N------------------------------------~
58
54
50
46
42
38
34
STAGE~p_C~~~P~C_P~C~-L-L_C~C~~~~A~A-L-L-L~A~A~~~_M~M~~~~T~
ROUND 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5
'------...----J '------...----J '------...----J ' - - - - - - ' ' - - - - - - ' ' - - - - - - '
GROUP 14 12 3 5 4
FIGURE 4. Frequency of use of self-reevaluation of 6 profile groups across 4 stages of

change.
change processes, with the biggest difference being the stage during
which particular processes would peak. Consciousness-raising, for ex-
ample, is at a very low level during precontemplation, as subjects resist
becoming more fully aware of a potential problem or a solution to the
problem. Consciousness-raising increases dramatically for individuals
who progress to contemplation, peaks in contemplation, and then de-
clines through action and maintenance to precontemplation levels. Pro-
cesses like stimulus control and counterconditioning, on the other hand,
remain relatively low during contemplation but peak in action. Rather
than declining to prechange levels, however, these processes level off at
higher levels as individuals rely on these processes as relapse preven-
tion strategies. These patterns of change generated the Mt. Change
metaphor, which has been extremely useful in creating a generation of
contemplation and action self-help manuals. The metaphor encourages
the users to conceptualize overcoming smoking as being similar to
climbing a mountain-they need to be adequately prepared; they need
adequate guides to find their way; they may not make it the first time;
but when they succeed, they have a tremendous sense of accomplish-
ment.
Whereas the change processes appear to follow a pattern analogous
to a mountain, other variables reveal a different pattern across stages.
Figure 5, for example, indicates that self-efficacy or confidence across
smoking situations shows a rather steady increase across the stages of
change. Temptation, on the other hand, demonstrates a steady decrease
in Figure 6. More importantly, if the two figures were superimposed, it
would become clear that levels of confidence and temptation are about
equal throughout the action stage. It is not until individuals are moving
into maintenance that self-efficacy becomes greater than temptations to
LEVELr-________~------------------------------~~~,
62 CONFIDENCE
58
54
50
46
42
38
STA~: PC PC PC C C
~-L~~~-L~~~~L-~-L~~~-L~~-L-L~~-L~
A A A A MM
ROUND 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5
"----..------/ "----..------/ "----..------/ "----..------/ "----..------/ "----..------/
GROUP 1 14 12 3 5 4
FIGURE 5. Levels of confidence or self-efficacy for 6 profile groups integrated across 4

stages of change.
INTENSITY
62 ,......"T:-;E;-;:M7.P~T;-A-;-:;:-T-:-::IO:-cN:-:-----------·----------·----· --",
58 1
-<
54 ~
-1
50 j
46
42
38
34
STAGE PC PC PC C C
~-L~~~-L~~~~~LJ-L~~~-L~~-L~~LJ-L~
A A A A M M T
ROUND I 2 3 4 5 I 2 3 4 5 I 2 3 4 5 1 2 3 4 5 I 2 3 4 5 1 2 3 4 5
'----..-----/ '----..-----/ '----..-----/ '----..-----/ '----..-----/ '----..-----/
GROUP 1 14 12 3 5 4
FIGURE 6. Intensity of temptation for 6 profile groups integrated across 4 stages of change.
smoke. Because increases in self-efficacy over temptation are associated

mainly with movement from action to maintenance, self-efficacy exer-
cises are being emphasized in our action materials rather than in our
contemplation programs.
Figures 7 and 8 suggest that the decisional-balance variables are
associated much more with movement from precontemplation to con-
templation and from contemplation to action. The pros of smoking clear-
ly outweigh the cons of smoking until subjects move into the contempla-
tion stage. During this stage, however, the cons begin to surpass the
pros, even though both are quite important for the smoker. Increasing
the cons of smoking beyond the pros is one of the goals for our con-
templation manual, as it prepares individuals to take more effective
action.
Thus far we have been discussing only how to approach a single,
well-defined addictive problem. We have been discussing a two-dimen-
STRENGTH
62~~nr~~nr.~nm~----------------,
58
54
50
46
42
38
34
STAGE PC PC PC C C A A A A MM T
~-L~~~-L~~~-L~LJ-L~~~-L~~~-L~~~~
ROUND 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5
'----..-----/ '----..-----/ '----..-----/ '----..-----/ '----..-----/ '----..-----/
GROUP 14 12 3 5 4
FIGURE 7. The strength of the pros of smoking for 6 profile groups integrated across 4
stages of change.
STRENGTH
62~C-O-N-S--O-F--S-M-O-K-IN-G------------------------------~
58
54
50
46
42
38
34
S T" GE L..;.PC---l.--'---.LJ.PC_P.L...C-L-.l..--L_C-'---1C---l.--'---.l..--LA_A.l..--L-,--,-A.l..--LA-'--'--.l....-.LM--,--M-,--,---,---,-T-.J
ROUND 123451234512345123451234512345
~~~~~~
GROUP 1 14 12 3 5 4
FIGURE 8. The strength of the cons of smoking for 6 groups integrated across 4 stages of
change.
sional model involving stages and processes of change as they are ap-
plied to a single addictive behavior. However, reality is not so accom-
modating and human behavior change is not so simple a process.
Although we can identify and isolate certain addictive behaviors, these
often occur in the context of complex, interrelated levels of human func-
tioning. The third basic dimension of the transtheoretical approach ad-
dresses this issue. The levels-of-change dimension represents a hier-
archical organization of five distinct but interrelated levels of psychologi-
cal problems which are addressed in treatment. These levels are:
1. Symptom/situational
2. Maladaptive cognitions
3. Current interpersonal conflicts
4. Family/systems conflicts
5. Intrapersonal conflicts
Historically, systems of psychotherapy have attributed psychologi-
cal problems primarily to one or two levels and focused their interven-
tions to address these levels. Behaviorists have focused on the symptom
and situational determinants; cognitive therapists on maladaptive cogni-
tions; family therapists on the family/systems level; and psychoanalytic
therapists on intra personal conflicts. It appears to us to be critical in the
process of change that both therapists and clients be in agreement as to
which level they attribute the problem to and at which level or levels
they are willing to work to change the problem behavior. Once again it is
extremely important that the therapist engage the client at an appropri-
ate and at least implicitly agreed upon level or levels for the work of
therapy to progress smoothly.
In the trans theoretical approach we prefer to intervene initially at
the symptom/situational level because change tends to occur more

quickly at this more conscious and contemporary level of problems and
because this level often represents the primary reason for which the
individual entered therapy. Furthermore, most self-changers seem to
prefer to intervene initially at the symptom/situational level of addictive
problems. The further down the hierarchy we focus, the further re-
moved from awareness are the determinants of the problem likely to be.
Moreover, as we progress down the levels, the further back in history
are the determinants of the problem and the more interrelated the prob-
lem is with the sense of self. Thus, we predict that the deeper the level
that needs to be changed, the longer and more complex the therapy is
likely to be and the greater the resistance of the client (Prochaska &
DiClemente, 1984). In addition, these levels are not completely sepa-
rated from one another. Change at anyone level is likely to produce
change at other levels. Symptoms often involve intrapersonal conflicts;
maladaptive cognitions often reflect family/system beliefs or rules. In
the transtheoretical approach, the therapist is prepared to intervene at
any of the five levels of change, though the preference is to begin at the
most conscious and contemporary level that clinical assessment and
judgment can justify.
TABLE 3.
Levels x Stages x Processes of Change
Stages
Levels Precontemplation Contemplation Action Maintenance
Symptom/ situational Consciousness-

raising
Self-
reevaluation
Self-
liberation
Contingency
management
Helping
relationship
Counter-
conditioning
Stimulus
~~~~~contrOI
:
Maladaptive cognitions
Interpersonal conflicts
Family/systems conflicts
Intrapersonal conflicts :
In summary, the transtheoretical approach views comprehensive

treatment as the differential application of the processes of change at the
four stages of change according to the problem level being addressed.
Integrating the levels with the stages and processes of change provides a
model for intervening hierarchically and systematically across a broad
range of therapeutic content. Table 3 presents an overview of the inte-
gration of levels, stages, and processes of change.
Three basic strategies can be employed for intervening across multi-
ple levels of change. The first is a shifting-levels strategy. Therapy would
typically focus first on the client's symptoms and the situations support-
ing the symptoms. If the processes could be applied effectively at the
first level and the client could progress through each stage of change,
therapy could be completed without shifting to a more complex level of
analysis. If this approach were not effective, therapy would shift to
other levels in sequence in order to achieve the desired change. The
strategy of shifting from a higher to a deeper level is illustrated in Table 3
by the arrows moving first across one level and then down to the next
level.
The second is the key-level strategy. If the available evidence points
to one key level of causality of a problem and the client can be effectively
engaged at that level, the therapist would work almost exclusively at
this key level.
The third alternative is the maximum-impact strategy. With many
complex clinical cases, it is evident that multiple levels are involved as a
cause, an effect, or a maintainer of the clients' problems. In this case,
interventions can be created that attempt to affect clients at multiple
levels of change in order to establish a maximum impact for change in a
synergistic rather than a sequential manner.
What moves people from precontemplation into the contemplation
stage of change? What facilitates or forces people to become aware that
previously acceptable patterns of behavior are now problematic or
pathological? To respond to these important questions we have had to
go beyond research data and rely more on clinical experience and theory
(Prochaska & DiClemente, 1984).
We propose that progress from precontemplation into the con-
templation stage appears to be due to either developmental changes or
environmental changes that occur in peoples' lives. Many individuals
begin to contemplate changing particular aspects of their lives because
of developmental processes that move them into a new stage of life. As
Levinson and his colleagues (1978) suggest in their work, The Seasons of a
Man's Life, many men find themselves quite satisfied with a particular
spouse during their twenties. When they enter the transition into the
thirties, however, they begin to contemplate radical changes in their
marriages. Similarly, many smokers seriously begin to contemplate

stopping smoking as they approach age 40 and feel pressured to face the
finiteness of their lives. It is not coincidental that the self-changers in our
research who have been most successful in quitting smoking took action
at a mean age of 39. Developmentally, facing 40 is a key time for many
people to reevaluate their lives to determine where changes are needed.
Other individuals appear ready for change not because of internal
developmental changes but because their external environment has
changed. Perhaps a spouse or a child has reached a new developmental
stage and asks or insists that they stop drinking. Or they begin to realize
that their environment no longer reinforces their drinking or smoking as
it once did, but now responds with subtle and not so subtle punish-
ments for their old habits. Other changes occur in the environment that
mayor may not be related to people's personal behavior and yet these
events can cause them to contemplate seriously a change in their behav-
ior. A poignant example of such an enviornmental event occurred with a
married couple who participated in our self-change research. Both
spouses were heavy smokers for over 20 years. Then their dog died from
lung cancer. The husband quit smoking. The wife bought a new dog.
The important theoretical issue here is that intentional change, such
as occurs in therapy, is only one type of change that can move people.
Developmental and environmental changes are other events that can
cause people to alter their lives. The transtheoretical approach focuses
primarily on facilitating intentional change, but it recognizes and, at times,
relies on other types of change when working with clients. It is as-
sumed, however, that unless developmental or environmental changes
produce intentional change as well, then clients will feel coerced and
will be likely to revert to previous patterns once the coercion is removed.
It is all too common, for example, that alcohol-troubled people quit
drinking when their spouses threaten divorce. Once their spouses are
safely back into the marriage again and coercion from the threat of
divorce is lifted, these individuals are likely to relapse back into troubled
drinking.
Under what conditions are we likely to be open to the developmen-
tal processes within us or to the environmental processes outside us as
freeing influences that enable us intentionally to change our lives? Un-
der what conditions do we experience these same processes as coercive
forces imposing change on us that we must resist with our best de-
fenses? Similarly, under what conditions do clients experience therapy
as a freeing influence that enables them intentionally to change their
lives? Under what conditions do they experience therapy as a coercive
force imposing changes on them that they must resist with their best
defenses?
Therapists can help clients progress more freely into the contempla-
tion stage of change if they can help their clients identify with the devel-
opmental or environmental forces that are pressuring them to change.
Clients may, for example, have difficulty identifying with the develop-
mental process of aging even though it comes from within. Whether
entering a new age becomes a life crisis or an opportunity, for growth
may be determined by whether we experience aging as imposed on us
or as part of us. Most of us, for example, identify with aging when we
become 21. Our sense of self includes becoming more independent,
mature, and adult. Becoming 40 or 50, on the other hand, is more often
experienced as an imposition in a society that identifies with youth.
Clients may resist a coercive aging process in self-defeating ways.
They may deny any potential health or mental health problems, so that
they do not have to contemplate changing their depressing drinking
habits. They may turn to stimulants to regain the energy of youth they
feel slipping away. They may then turn to barbiturates as a way of
sleeping through the night. They may turn to meaningless affairs to
deny that their sexual drives are decreasing. They may spend money
recklessly to deny that their lives are limited.
The same self-defeating defenses can occur against environmental
pressures to change. A client named Harold was in marital therapy for 3
months when he said, "You know, I still don't know why I am coming
here. I am coping perfectly fine with all the stresses in my life. It's my
wife who can't cope, and yet she insists that I come to therapy or she
will leave." Of course, it did not help his wife to repeat for the ump-
teenth time that Harold was spending money until they were nearly
bankrupt; that he was at risk of losing his job for the third time in 4
years; that the children were afraid to be around him because of his
violent temper; and that she was seriously considering separating be-
cause all his energy was going into his skiing club and none into their
marriage. But Harold could no longer identify with his wife, or with her
reasons for changing. He experienced her as a manipulating mother
trying to take away his freedom and fun.
How can therapists intervene in a manner that allows them to be
experienced by defensive clients as freeing influences rather than as
coercive forces? Obviously, the more clients can identify with the thera-
pist and the elements of therapy, the more therapy can be experienced
as a freeing influence. With precontemplators, in particular, the thera-
peutic relationship becomes a precondition for further change. Identifi-
cation with the therapist is more likely to occur if the client feels that the
therapist genuinely cares. Identification is also more likely to occur if the
client feels that the therapist is truly trying to understand the client's
unique experience, including the client's need to be defensive as well as
the client's desire to be open. Identification is also more likely to occur if
the client believes that the therapist is committed to helping the client
change in ways that are best for the client and not some other agent,
such as the courts, the schools, the employer, or the mental health
center.
Caring, understanding, and commitment to the well-being of cli-
ents are, of course, values that should be essential elements in the iden-
tity of a therapist. Therapists feel most free as therapists when they are
able to care about their clients, understand their clients, and be commit-
ted to the well-being of their clients. There are, of course, times when
therapists can be coerced by countertransference or other forces so that
they are not really caring about the client but rather about their own
needs. Or they are not really understanding this client but rather are
responding to their projections onto the client. Most therapists recog-
nize how fine a line there can be between projection and empathy,
because empathy is accurate projection. But therapists are committed to
putting their own needs and their problematic projections aside so that
they can identify with their particular patients.
Ironically, clients need first to feel that the therapist is free to identi-
fy with the client before the client is free to identify with the therapist. If
the client feels that the therapist cannot identify with the client's predi-
cament in life because the therapist does not care, does not understand,
or is not committed, then the client is likely to terminate therapy before
it begins. Clients need to believe that the therapist can identify with
them as if they were friends and family, not foreigners who are alien to
the therapist's sense of self. If clients believe the therapist cannot identi-
fy with them because the therapist is of the wrong gender, ethnic back-
ground, social class, or sexual orientation, then clients will not feel free
in therapy. Clients are likely to avoid such therapists lest they risk coer-
cion to change according to sterotypes of gender, ethnicity, social class,
or sexual orientation.
As clients and therapists begin to develop a shared identity that is
the essence of a therapeutic relationship, clients become much more
open to influence from therapists. Clients are much freer to respond to
feedback and education about the alienated aspects of their lives. Clients
are particularly free to process information from therapists or others
with whom they have a helping relationship. Therapists also become
more open to influence from their clients, such as to have a favored
formulation invalidated by further information from the client. But our
focus will remain centered on how clients change in therapy rather than
on how therapists change over the course of therapy.
A helping relationship, such as a therapeutic relationship, proVides
people with the freedom to process developmental or environmental
events in a friendly rather than coercive atmosphere. Easing up on their
defenses, they can begin to see themselves more clearly. They can begin
to contemplate making intentional changes in their lives without feeling
that they are entirely coerced" by developmental or environmental

events. Movement into the contemplation stage, like many changes in
life, is usually experienced as a combination of coercion and personal
freedom.
Once clients begin to move into the contemplation stage, their in-
sight and understanding are critical for further progress. Whether the
insight is historical-genetic, interactive, cognitive, or situational de-
pends on the level of change that is needed. For clients working at the
symptom/situational level, a functional analysis of the immediate ante-
cedents and consequences of troubled behavior may be all the under-
standing that is needed. Clients attempting to change troubled rela-
tionships, however, will need insight into the interactive nature of their
problems. Clients who are not free enough from their family of origin or
who are plagued by intrapersonal conflicts are more likely to need in-
sight into the historical-genetic causes of their conflicts.
Insight and understanding can become an endless process of con-
sciousness-raising, however, if clients wish to have a complete grasp of
all that influences them. Some personalities have a propensity to be-
come bogged down in prolonged contemplation of a problem. Ob-
sessive personalities in particular prefer to believe that if they keep
thinking enough about an issue, eventually the problem will go away or
enough understanding will be gained that points to a perfect solution to
a complex problem. The obsessive does not like to admit that there are
serious limits to thinking and that many personal problems can only be
resolved by commitments that go beyond reason. The fear of facing the
irrational can keep obsessives seeking for years for sufficient insights,
moving from one book to another or from one therapist to another. Of
course, some therapists are also afraid of making commitments to action
without an obsessive understanding of their client's problems.
Moving from contemplation to action involves both consciousness-
raising and self-reevaluation processes. Consciousness-raising interven-
tions, like observations, confrontations, and interpretations, are most
important during the contemplation stage. Value-clarification tech-
niques are also important in preparing clients for taking effective action.
Helping clients to work through a decisional balance, for example, can
clarify which course of action is most likely to reflect the type of person
the client wants to become. Balancing the pros and cons of a particular
course of action also prepares clients to pay the price that comes with
any major change in life.
When it comes to action, skill acquisition and/or utilization are most
important for therapeutic progress. If a therapist is skilled only in con-
sciousness-raising interventions, like interpretations, then the con-
templation stage can become excessively and obsessively long. Apply-
ing such behavioral skills as desensitization, assertion, communication,
or negotiation are important aspects of the action stage. Which skills are
utilized depends on the client's level of change. Desensitization, for
example, is used most often at the symptom/situational level whereas
communication training is much more important for the interpersonal
level. Renegotiating dysfunctional family rules can be particularly liber-
ating at the family/systems level. Assertiveness based on existential val-
ues can be one of the most liberating means for expressing the enhanced
sense of self that emerges when intrapersonal conflicts are being
resolved.
From a transtheoretical perspective, the therapeutic relationship,
interpretations, and skill acquisition and utilization are all fundamen-
tally important to producing change. Their relative importance varies
from stage to stage, with therapeutic relationships most important for
facilitating movement from pre contemplation to contemplation, con-
frontation and interpretation most important during contemplation, and
skill acquisition and utilization most important during action and
maintenance.
We will conclude this chapter by indicating how the transtheoretical
model of change addresses the comprehensive questions of who
changes, what changes, and when, where, why, and how changes oc-
curs. The transtheoretical approach that we have been developing has
focused on when changes occur, how changes occur, and what changes
occur when addictive behaviors are modified. The stages-of-change di-
mension indicates when people make particular changes in modifying
addictive behaviors. The processes of change address how people make
particular changes when progressing from one stage to the next. The
levels reveal what people need to change in order to overcome their
particular addictive problems.
Where people change has been assumed not to be a critical dimen-
sion of change. Whether people change in residential treatment pro-
grams, in outpatient therapy, in self-help groups, with self-help man-
uals or at home working entirely on their own can have important
practical implications, but does not appear to be a critical dimension for
developing a comprehensive model of change.
Why people attempt to overcome addictive problems relates to the
important issue of motivation that many practitioners believe is a key to
successful treatment. To date, we have not paid adequate attention to
the question of why some people attempt to change whereas others
avoid change. Decisional balance has been the variable we have studied
that is most closely related to the issue of why some individuals change
whereas others continue with their addictive behaviors. Data on deci-
sional balance and our informal observations suggest that motivations to
change often vary with the stage an individual is in. Why s0l!leone
begins to contemplate quitting smoking, for example, can be different
from why that person eventually takes action to quit smoking. Why
someone makes the continued efforts to maintain nonsmoking can be
different from why someone tries again after failing to maintain an
addiction-free life-style. An individual may, for example, begin con-
templating quitting smoking because a friend or relative develops lung
cancer. The same person may take action because there is a stop-smok-
ing campaign at the office. Struggles to maintain nonsmoking may be
based in part on a motivation to avoid failure. Returning to contempla-
tion rather than giving up may be based on a motivation to be in control
of one's life rather than to believe that an addictive behavior is beyond
one's self control. From a transtheoretical perspective, we assume that a
comprehensive analysis of motivation to change will include an analysis
of motivation at each stage of change and how motivation can be a
dynamic phenomenon that fluctuates from one stage of change to the
next.
Who changes in treatment and who fails to change has been ad-
dressed traditionally by the study of client characteristics. Such variables
as age, gender, socioeconomic level, duration, frequency, and intensity
of the problem, intellectual level, psychological mindedness, and degree
of psychopathology have been some of the client characteristics that
have received considerable attention in the therapy outcome literature
(Luborsky, Chandler, Auerback, Cohen, & Bachrach, 1971; Meltzoff &
Kornreich, 1970). We have not paid as much attention to such variables
for several reasons. The most important reason is that such demograph-
ic, personality, and psychopathology variables tend to be trait-like vari-
ables that are not particularly open to change and are not likely to be
under the client's control or the therapist's control. Decades of psycho-
therapy research on who benefits from treatment has done little to ad-
vance our knowledge of how we can help more people to change. This
same research has often added to clinicians' pessimisim about their abili-
ties to help whole classes, cultures, and communities of people over-
come destructive problems, like addictive behaviors. Until we develop
more adequate models of treatment based on more comprehensive mod-
els of change, we really will not know the answers to who can change
with maximum treatment, who can change with minimal treatment,
'who can change on their own, and who cannot or will not change
regardless of what they or we try to do.
REFERENCES
Bandura, A. (1977). Self-efficacy: Toward a unifying theory of behavior change. Psychologi-

cal Review, 84, 191-215.
Bandura, A. (1982). Self-efficacy mechanism in human agency. American Psychologist, 37,
122-147.
DiClemente, C. C. (1981). Self-efficacy and smoking cessation maintenance: A preliminary

report. Cognitive Therapy and Research, 5, 175-187.
DiClemente, C. c., & Hughes, S. (1985). Stages of change in alcoholism treatment: Measure-
ment and profiles. Manuscript in preparation.
DiClemente, C. c., & Prochaska, J. O. (1982). Self-change and therapy change of smoking
behavior: A comparison of processes of change of cessation and maintenance. Addic-
tive Behaviors, 7, 133-142.
DiClemente, C. c., Prochaska, J. 0., & Gibertini, M. 1985. Self-efficacy and the stages of
self-change of smoking. Cognitive Therapy and Research, 9, 181-200.
Glasgow, R. E., Schafer, 1., & O'Neill, N. K. (1981). Self-help books and amount of
therapist contact in smoking cessation programs. Journal of Consulting and Clinical
Psychology, 49, 659-667.
Goldfried, M. (1980). Toward the delineation of therapeutic change principles. American
Psychologist, 35, 931-950.
Goldfried, M. (1982). Converging themes in psychotherapy. New York: Springer.
Hunt, W., Barnett, 1., & Branch, 1. (1971). Relapse rates in addiction programs. Journal of
Clinical Psychology, 27, 455-456.
Janis, 1. 1., & Mann, 1. (1977). Decision making: A psychological analysis of conflict, choice and
commitment. New York: The Free Press.
Levinson, D., Dorrow, c., Klein, E., Levinson, M., & McKee, B. (1978). The seasons of a
man's life. New York: Knopf.
Luborsky, 1., Chandler, M., Auerback, A., Cohen, J., & Bachrach, H. (1971). Factors
influencing the outcome of psychotherapy: A review of quantitative research. Psycho-
logical Bulletin, 75, 145-185.
McConnaughy, E. A., Prochaska, J. 0., & Velicer, W. F. (1983). Stages of change in
psychotherapy: Measurement and sample profiles. Psychotherapy: Theory, Research, and
Practice, 20, 368-375.
McConnaughy, E., Prochaska, J. 0., Velicer, W., & DiClemente, C. (1984). Replication of the
stages of change in psychotherapy. Unpublished manuscript, University of Rhode Island.
Meltzoff, J., & Kornreich, M. (1970). Research in psychotherapy. New York: Atherton Press.
Miller, W., & Baca, 1. (1983). Two-year follow-up of bibliotherapy and therapist-directed
controlled drinking training for problem drinkers. Behavior Therapy, 14, 441-448.
National Institute on Drug Abuse. (1979). Cigarette smoking as a dependence process (National
Institute on Drug Abuse Research Monograph series, DHEW, Monograph 23, 194
pg.). Washington, D.C.: U.S. Government Printing Office.
Norcross, J., & Prochaska, J. O. (In Press). Psychotherapist heal thyself: 1. The psychologi-
cal distress and self-change of psychologists, counselors and laypersons. Psycho-
therapy: Theory, Research and Practice.
Olcott, P. (1985). Perceived change processes in successful weight reduction and maintenance of
weight 1055. Unpublished doctoral dissertation, University of Rhode Island.
Parloff, M. (1976, February). Shopping for the right therapy. Saturday Review, pp. 14-16.
Parloff, M. (1980). Psychotherapy and research: An anaclitic depression. Psychiatry, 43,
279-293.
Prochaska, J. O. (1984). Systems of Psychotherapy: A Transtheoretical Analysis. Homewood,
Illinois: The Dorsey Press.
Prochaska, J. 0., & DiClemente, C. C. (1982). Transtheoretical therapy: Toward a more
integrative model of change. Psychotherapy: Theory, Research and Practice, 19, 276-278.
Prochaska, J. o. & DiClemente, C. (1983a). Stages and processes of self-change of smok-
ing: Toward an integrative model of change. Journal of Consulting and Clinical Psychol-
ogy, 51, 390-395.
Prochaska, J. 0., & DiClemente, C. (1983b). Self change processes, self efficacy and deci-
sional balance across five stages of smoking cessation. In P. Engstrom (Ed.), Advances
in Cancer Control (pp. 131-140). New York: Alan R. Liss.
Prochaska, J. a., & DiClemente, C. (1984). The transtheoretical approach: Crossing the tradi-
tional boundaries of therapy. Homewood, IL: Dow Jones/Irwin.
Prochaska, J. a., & DiClemente, C. (1985). Common processes of change for smoking,
weight control, and psychological distress. In S. Schiffman & T. Wills (Eds.), Coping
and Substance Abuse. New York: Academic Press.
Prochaska, L Velicer, W., & DiClemente, C. (1985). Patterns of self-change in smoking cessa-
tion. Unpublished manuscript, University of Rhode Island.
Prochaska, J. a., DiClemente, C. c., Velicer W. F., Ginpil, S., & Norcross, J. (1985).
Predicting change in smoking status for self-changers. Addictive Behavior, 10,395-406.
Schacter, S. (1982). Recidivism and self cure of smoking and obesity. American Psychologist,
37, 436-444.
Velicer, W., DiClemente, c., Prochaska, L & Brandenburg, N. (1985). A decisional bal-
ance measure for predicting smoking cessation. Journal of Personality and Social Psychol-
ogy, 48, 1279-1289.
Wilcox, N., Prochaska, J. a., Velicer, W., & DiClemente, C. (1985). Client characteristics
as predictors of self-change in smoking cessation. Addictive Behavior, 10, 407-412.
2
Implications of a Self-Regulation
Model of Therapy for Treatment of
Addictive Behaviors
FREDERICK H. KANFER
Social systems have developed out of the need for the regulation of
individual behaviors in order to facilitate communal living. A major goal
of the control developed by social systems is the subordination of indi-
vidual needs to the larger goals of the survival of the group. What is
beneficial for an individual is often a satisfaction attained at the expense
of pain or harm to others or to oneself at a future time. It is for such
reasons that a Skinnerian analysis has viewed society as a giant mecha-
nism for the enforcement of self-regulation. Social and cultural evolution
has developed elaborate agencies of religion, education, government,
family, and law. But they leave many loopholes in the control of
individuals.
SELF-CONTROL AND ADDICTION
A particularly heavy burden that societies place on the individual is

the demand for self-control. By self-control we mean the exercise of a
controlling response or strategy that reduces the probability of executing
FREDERICK H. KANFER • Department of Psychology, 603 East Daniel Street, University

of Illinois, Champaign, Illinois 61820.
29
30 FREDERICK H. KANFER
a behavior that is either very firmly established as a long time habit or

momentarily attractive because it fulfills biological or acquired needs
and desires. These actions are usually easy to execute but disadvan-
tageous in the long run. Many of these behaviors are not easily prevent-
able by the salient social or physical environment. The control of these
acts, oriented toward immediate personal satisfactions, forms part of the
social contract and is incorporated in the socialization of children in all
cultures. It is the study of these self-regulatory processes and their invo-
cation for the purpose of therapeutic change that has concerned me for
over two decades.
In the area of alcohol abuse (as in eating disorders, sexual disorders,
and other areas) the individual engages in behaviors that are biologically
detrimental to the person in the long run. They also reduce the abuser's
contributions to society. The intensity of the self-control conflict is
heightened because of conflicting messages from different sources in the
social environment. Moderation is prescribed by society for many be-
haviors that have incentive value, because the long-term consequences
of these behaviors have aversive consequences for the person's physical
well-being, and/or social and psychological status. Diverse interest
groups, however, flourish in our societies and some thrive on producing
or exalting behaviors and products that tempt the flesh, the mind, and
the palate. Even within the same social infrastructures inconsistent rules
of conduct are given. Social rules and etiquettes guide proper timing,
frequency, and quantity of alcohol consumption. Social structures and
laws attempt to ease the individual's temptation by regulating such
factors as drinking age, sources of supply, and advertising. Neverthe-
less each person is ultimately held responsible for monitoring their judi-
cious use of alcohol, and expected to control drinking within defined
ranges and on specified occasions. Attractive social settings, cultural
rituals, and social approval increase the conflict for those individuals
whose biological and psychological makeup leave them dependent on
conflicting external messages rather than on their own internal feedback
cues for guidance about how to handle the temptation to initiate or
continue drinking.
The introduction to this presentation provides the central theme of
our model of self-management therapy. Personal or self-control is a
complex process. It is not a pervasive characteristic of the person but
involves the total context of a behavioral episode at the biological, social,
and psychological level of a person in a social system (Kanfer, 1977). It is
not a simple skill that can be learned and generalized, nor does it occur
in isolation. It is the result of an interplay between a person's urges,
opportunities, and social demands.
Self-control situations are defined as situations in which a person is
IMPLICATIONS OF A SELF-REGULATION MODEL OF THERAPY 31
faced with the task of engaging in or stopping behaviors that are initially
less motivated, less enjoyable, and less skilled than the automatically
processed acts that we carry out easily from moment to moment. Per-
sons set goals and believe that they can achieve them. They must engage
in controlled processing by making decisions and generating their own
incentives. Frequently they must also defy concurrent social or internal
aversive cues to get the nonpreferred behavior started. This means that
the person needs to engage in a deliberate change from a habitual, easy,
and often pleasant behavioral sequence. Self-control problems do not
occur unless strong biological motives or behavioral dispositions must
be altered.
OBSTACLES TO CHANGE
When the addicted client contemplates a change, he or she usually

has no clear understanding or certainty about how or whether this effort
will benefit him or her. In the treatment of addiction problems, as in any
therapy or change program, change is frightening. Especially in ad-
dicted clients, the problematic behavior itself is frequently the result of
earlier failures to find a satisfactory life pattern. Frustration and self-
depreciation associated with these failures, lower positive expectations
for change and lowered self-efficacy often cause demoralization. These
factors represent powerful obstacles to the optimistic attitude required
for acceptance of the heavy burden of withdrawal and for investing
effort in a change program. The removal of these factors and an infusion
of some hope for, or at least a neutral attitude about, therapy outcome is
a critical initial goal of treatment.
At the outset of change, two central questions for all client are:
"Will I be better off if I do?" and "Will I be able to do it or will I fail?".
These questions are often not answered positively by the client. Further-
more, many therapists fail to begin treatment by helping the client to
answer the "Why should I?" question in a clear way. Rational answers,
persuasion or horror stories of the long-term consequences, as we know
from experience with educational programs on which billions have been
spent, are not sufficient to counter other influences. The client's commit-
ment must be verbalized, felt, experienced, and acted on to effect an
enduring change. The commitment rarely is based initially on a genuine
desire to stop drinking or engaging in any addictive behaviors. If a client
is eager and motivated at all to change, he or she would usually most
prefer to change only the consequences of drinking behavior rather than
the behavior itself. It is the task of therapy first to strengthen commit-
ment by helping the client to increase the attractiveness of the new
behavior pattern and life-style associated with it. A second task is to
help the client to change the required behaviors by making the change
as easy and rewarding as possible.
A PROCESS MODEL OF CHANGE
Considerations and experiences with clients have led me to concep-

tualize treatment for psychological problems, including addictions, in a
general systems approach. The levels of analysis are interrelated and
operate iteratively and recursively. The social and biological systems
define and influence the boundaries of the psychological system that we
usually seek to alter. At the psychological level, the core processes of the
client's self-regulatory system, its emotional and cognitive components
and its relationship to the execution of specific behaviors in the pursuit
of a desired end state (objective), become the focus of the therapeutic
approach. The technological aspect of therapy has two major interre-
lated goals: (a) to alleviate the current problem, and (b) to strengthen the
self-regulatory system to cope with future problem situations. The latter
includes the person's ability to behave differently. But in addition, a
person must learn to recognize and avoid, alter, or confront external
(social) and internal (biological) stressors or conflict-producing settings
and events. An effective person uses the self-regulatory process to influ-
ence emotional, motivational, and behavioral sequences toward attain-
ment or maintenance of an appropriate, effective and desirable style of
life.
We now turn to a brief overview of our model (Kanfer & Grimm,
1980), shown in Table 1. It should be clear that the phases in the therapy
process overlap. They are recursive and iterative and the need for inten-
sive work in each phase varies with the client and the specific treatment
objectives. The conceptual model serves as a guideline for the therapist
in setting priorities for different issues during the course of therapy.
Further, it indicates how dealing with these issues in proper sequence
facilitates the progression in treatment. Each phase is preparatory for
the following phases, though its theme may need to be carried over
throughout therapy. The model (Kanfer & Grimm, 1980) is similar to
that described by Prochaska & DiClemente (1982). Both emphasize the
dynamic and recursive nature of the change process. Both models pro-
pose commitment to change as the prerequisite to modification of the
target behaviors and stress the reciprocal effects of intentions (and other
verbal-symbolic processes) and actions. Both models note the fragility of
therapeutic gains, and require that treatment focus on maintenance of
effects. Finally, both note the importance of the client as the critical
agent of change.
TABLE 1.
A 7-Phase Process Model of Therapy
Phase Primary goals
1. Role structuring and creating a 1. Facilitate the person's entry to the

therapeutic alliance role of client
2. Formation of a working relationship
3. Establish motivation to work with
therapist
2. Developing a commitment for 1. Motivate client to consider positive
change consequences of change
2. Activate client toward change of
status quo
3. Reduce demoralization
3. The behavioral analysis 1. Refine client's problem definition
2. Identify relevant functional
relationships
3. Motivate client toward specific
changes
4. Negotiating treatment 1. Seek agreement of target areas
2. Establish priorities for change pro-
gram and initiate specific
procedures
3. Accept responsibility for engaging
in planned therapy program
5. Treatment execution and motivation 1. Conduct treatment program
maintenance 2. Assess collateral and radiating ef-
fects of change in target
behaviors
3. Evaluate and, if necessary, enhance
motivation to change and comply
with treatment requirements
6. Monitoring and evaluating progress 1. Assess behavior change
2. Assess client's use of general
coping skills
3. Introduce new therapy objectives, if
necessary
4. Motivate program completion
7. Treatment generalization and termi- 1. Evaluate and foster self-manage-
nation ment skills for meeting future
problems
2. Phase out contact with client
Adapted from Kanfer & Grimm (1980), pp. 440-441.

No model of a therapy process is complete without some attention

to processes that lead clients to define themselves (or to be defined by
others) as in need of change. The analysis of these variables would have
to include a theory of etiology as well, to guide us toward selection of
the critical historical, genetic, and sociocultural factors that combine to
produce a client or patient.
The scope of this task is overwhelming; in fact it spans several
disciplines and still baffles experts. One way to organize our knowledge
and raise research questions is to set our model in the context of the
usual sequence of events that defines the transition of a person to client
status and back again to a person (though some authors hold that a
former patient is never again a member of the set of nonpatient citizens).
Table 2 presents a simple flow chart that reflects these critical transition
points. Of the 10 steps listed, the first 4 are pre therapy events, the last
step extends beyond therapy. The reader will recognize the contribu-
tions of recent research in psychology, biology, and sociology to the
extratherapy steps. For example, social norms and physiological factors
determine perception and evaluation of an addictive behavior. Social
networks, and personal experiences with health-care delivery systems,
and the person's self-confidence affect the decision to seek help. How-
ever, lack of time, the complexity of the total constellation, and the
TABLE 2.
Common Sequences in Therapy: A Flow Chart of
Clinical Interventions
Examples of
Event exit reasons
1. Person notices problem

Defines as trivial, defines
2. Evaluates problem as not due to self or
not solvable by action
3. Decides to seek help No confidence or no
resources
4. Seeks help Conditions for treatment
5. Is diagnosed or advised not favorable
6. Decides to accept treatment See 2. or 4.
7. Responds to treatment Expectations not met
Considers task completed
8. Makes needed changes External events change or
7.
9. Stops treatment Improvement sufficient at
this time
10. Maintains change pattern Distress relieved, or new
crises arise
IMPLICATIONS OF A SELF·REGULATION MODEL OF THERAPY 35
speculative nature and dearth of evidence on the integration of the

psychological processes with the biological and sociocultural features
limit our treatment of this fascinating problem. Suffice it to note that
some of the following events influence later steps: (a) the particular
event or subjective reaction to it that leads a client to notice a problem;
(b) the client's evaluation of these events and definition of the problem;
(3) the factors (persons and events) involved in timing the decision to
seek help and where to seek it. The understanding of these factors
therefore is of great importance in planning a change program.
As shown in Table I, the first four phases in our model represent
the foundation for any effective intervention program. They are not
target specific. They are designed to establish a basis for therapeutic
interaction, motivate the client toward a commitment for change, devel-
op goals and incentives and involve the client in a behavioral analysis
that, in collaboration with the therapist, refines and clarifies the targets
and goals of treatment. In Phase 4, the last phase that focusses on
creating the conditions most favorable for change, specific objectives are
negotiated and commitment to a particular program is contracted. The
phases that precede an action program have usually not been discussed
in books on treatment methods; yet these preparatory phases are at least
as important as the choice of techniques in Phase 5. Further, they are not
arranged haphazardly. Social and learning psychology provides back-
ground heuristics for them, as they do for the later stages. Clearly,
tentative therapy strategies and methods are applied from the first meet-
ing on. But their function is to facilitate the commitment to change at
first. Only later can other methods be utilized to initiate a long-term and
lasting behavior change. The early phases are conducted in a context in
which the client is stimulated and encouraged-and often taught-to
assume increased responsibility for the content, direction, and speed of
the change process. For many addictive clients, the early stages may
require systematic practice even for such rudimentary tasks as imagin-
ing (and later, sampling) a change in their daily living routine that elimi-
nates substance abuse, yet provides some satisfactions. Systematic exer-
cises may have to be provided which, with much social support, demon-
strate that the client is capable of control. Positive self-reactions and
willingness to accept the challenge of change need to be heavily rein-
forced in these early phases.
Although symptoms and discomforts are dealt with, the model is
future oriented; that is, it is designed to prepare, anticipate, and "pre-
hearse" coping techniques for future situations. The intent of treatment
is not only to alleviate the current state but also to work toward a clearly
defined goal that is more satisfactory to the client (and acceptable to the
therapist and to society).
Phase 5 represents the conduct of a specific intervention program,

as widely described in texts on treatment methods. Initially, intense
efforts focus on the development of new behaviors by techniques rang-
ing from strengthening controlled information processing, such as self-
monitoring or prep lanning, to the establishment of contingencies in
artificial environments. Frequent practice, task assignments, and similar
methods are used to speed up the acquisition of a new behavior pattern.
As treatment progresses it is gauged to be successfui if new acquired
behaviors or environment changes become established and automatic;
the scope and intensity of treatment is then gradually reduced. Phase 6
overlaps with the preceding one. It is during treatment rather than at its
termination that preparations are made for the generalization of effects
and for coping with unanticipated difficulties or relapse. As clients ap-
pear ready to attempt a new life-style, reassessment is needed to ascer-
tain that (a) they have the necessary skills to adopt, for example, to a
living pattern without alcohol, and (b) that the setting to which they will
return is not incompatible with this pattern. In addictive disorders this is
the time to plan for specific life changes, to experiment with them on a
provisional basis, and to clarify the options in leading a life free of the
addictive substances and their consequences. Frequently, changes in
vocations, social settings, or even geographic location, as well as in
social and intra personal behavior patterns, may be required to assure
continued abstinence or control.
INDIVIDUAL DIFFERENCES REQUIRE

INDIVIDUALIZED PROGRAMS
The dynamic-recursive nature of our model suggests a problem-

solving rather than a diagnostic approach. Only after treatment has
progressed into what Prochaska and DiClemente have called the action
stage is it possible to evaluate the extent to which skill-training or en-
vironmental changes are needed to reduce the probability of future diffi-
culties. In this phase, as in the preceding and following one, the client is
helped to recognize and avoid or to cope with signals of increased stress
or hazard, in order to interrupt as early as possible sequences of events
and actions that could precipitate renewed conflicts. Particularly in alco-
holic and drug-addicted clients, possessing the skills to handle a new
alcohol- and drug-free life are essential. For many clients this means
learning not only new behaviors but a new life-style.
Clients with alcohol and drug problems do not present a universal
and unitary picture of etiology, personality, or prognosis. A common
task in therapy is to establish self-regulatory or self-control skills, that is,
IMPLICATIONS OF A SELF-REGULA nON MODEL OF THERAPY 37
to interpose personal control in critical situations. During the initial

phases such general skills as self-observation, problem solving, goal
setting, and anxiety reduction are strengthened in order to enrich the
client's repertoire for assuming responsibility for his life and dealing
with his problems more effectively. During Phases 5 and 6 attention
shifts from rehabilitating or teaching these basic skills to dealing with
the patient's unique problem situation.
Patients differ widely with regard to the variables that precipitate
drinking or drug-taking behavior and their sensitivity to the feedback
from the altered biological or social consequences of drinking or drug
taking. Further, their social environments present different demands and
vary in tolerance of behavioral inefficiencies resulting from the habit.
Consequently, for each person treatment requires consideration of the
specific factors that resulted in drinking or drug taking, the motivational
resources that can be used to maintain a change, the specific functions
served by the addictive habit, and the conflicts and unfulfilled obligations
or demands that results from the habit. These individual needs must be
met in therapy to attain long-term success. The last phase represents the
gradual withdrawal of the "therapeutic umbrella," as the client is reinte-
grated in his or her family and work setting. Preparation for follow-up is
an essential theme of this phase, particularly for inpatients.
Our model, arrived at through the incorporation of self-regulation,
general systems, and motivational principles into the basic framework of
behavioral therapy, is similar to that presented by Prochaska and DiCle-
mente in Chapter 1. It stresses the need for helping the client recognize
and accept the existence of a problem. Parenthetically, however, this
does not mean resignation to suffering from alcoholism as a disease, or
drug addiction as a given fact. On the contrary, it means acceptance of
the problem as a development of a life pattern that has had serious
harmful consequences and requires drastic change. Both models empha-
size the integration of cognitive and behavioral changes, with stronger
emphasis on the former at the beginning of treatment. We do, however,
believe that cognitive and behavioral components are nearly inseparable
in all phases. In fact, it is the confirmatory evidence from behavioral
changes that strengthens the patient's positive self-reactions and alters
further intentions and actions.
We stress the central goal of early therapy to develop the motivation
and commitment for change, followed by strengthening the skills to
change one's own behavior and one's environment. Prochaska's model
and our own emphasize the transitory nature of therapy and differenti-
ate between patients who achieve the level of change that obviates fur-
ther need for life-long efforts at control and continuing social support
and those who recycle after relapse and require long-term self-control
and therapeutic support to maintain abstinence. Our model suggests the

need for helping the client to cope with collateral personal problems,
antecedent to or resulting from addictive behaviors. Last, but not least,
we stress a future orientation toward the therapeutic target, not only in
order to eliminate the undesirable habit but also to anticipate and cope
with future situations that can facilitate a recurrence of the habit. To this
end, the client must be helped (a) to set positive goals, (b) to acquire a
skill repertoire for reaching them, and (c) to continue a commitment to
these goals during and after therapy. To do this, a client must experi-
ence success and learn that these goals can be achieved and maintained.
To focus purely on drinking or other addictive behaviors is to lift
one element arbitrarily out of a complex system that encompasses the
social context, the psychological status, and the biology of a person.
Neither alcohol nor drug addictions nor excessive eating are unitary
pathological processes-their contexts and consequences vary from per-
son to person. And none of the addictive problems respond to solutions
or programs that fail to prepare the client to assume the heavy responsi-
bility for undertaking and maintaining a change. Thus, instead of focus-
ing solely on the inhibition of excessive alcohol drinking or drug taking,
self-management therapy stresses the need for a wider change in per-
sonal motivation, values, and living patterns. Initially, substitutions or
"positive addictions" may be helpful as transitional patterns. The for-
mer drug addict who becomes an "exercise nut" or a crusader for some
cause builds up a new set of behaviors and a new social context that is
incompatible with the drug scene. However, the ultimate goal is not to
help the patient become an antialcohol, antidrug, or antismoking cru-
sader, or to be a "diet freak," but to achieve a stable life pattern in which
there is no longer any preoccupation with resisting temptations.
The ease with which this can be accomplished varies also with the
nature of the particular addiction. For example, whereas drug addictions
usually result in a living pattern that centers almost exclusively on the
drug habit, smoking, or some overeating patterns permit maintenance
of a socially acceptable style of life. Thus, although a psychological anal-
ysis may reveal similarities, social and biological consequences are quite
different in alcohol, drug, or tobacco abuse. It would be an error not to
address these differences and similarities during therapy.
RESEARCH SUPPORT FOR THE MODEL

SELF-MANAGEMENT FOR SUBASSERTIVE CLIENTS
The model of therapy that I have presented has been used with a
variety of populations, among them clients with eating disorders and
alcohol problems. In a recent study, Schefft (1983) compared a number

of outcome and process variables in three groups of subassertive women
clients to examine the specific effects that differentiate self-management
therapy from other forms of cognitive behavior therapy. The implemen-
tation of our model, as can be noted from the foregoing description,
does not lie in novel treatment methods but in the context that is created
for the use of various techniques. The major features contributing to the
context are the sequential structure of themes, the focus on client re-
sponsibility for goal-setting and self-regulation, and the extensive and
continuing attention to client motivation. In this comparative study of
treatment outcome and processes we selected subassertiveness as the
criterion for treatment, not because of a primary interest in this problem
per se but because of practical considerations, such as available popula-
tions, relative ease of measurement of the dependent variables, and a
clean comparison with a standardized and widely used program. Our
intent was to study the process and outcome rather than the specific
content of the different therapies.
College females were recruited via public announcements of an
offer to obtain free training in "interpersonal communication and self-
expression." Participants were screened to eliminate persons who had a
previous history of psychiatric hospitalizations, or who were currently
in therapy for psychological problems and, on the basis of an interview,
were judged to be suicidal or psychotic. Final acceptance of nonassertive
women was based on obtaining a raw score of 7 or below on the Rathus
scale (1973). Fifty-five non assertive women were selected for participa-
tion. In a blocked design, subjects were assigned randomly to balance
the three treatment groups for unassertiveness (Rathus Scores), overall
level of psychopathology (MMPI-168 Scores), and level of motivation
(Motivation Questionnaire scores). The three treatment conditions con-
sisted of administration of the standard cognitive-behavioral assertion
treatment program by Lange and Jakubowski (1976), a self-management
treatment, consisting of the same content as the cognitive-behavioral
assertion treatment but with administration of it in the therapeutic struc-
ture specified by Kanfer and Grimm's (1980) self-management model,
and relationship treatment, consisting of group therapy described by
Patterson (1974) and Rogers (1957, 1961). Two replications were con-
ducted, each consisting of nine consecutive, weekly 2-hour sessions.
Assessment included behavioral and self-report measures of therapeutic
processes and outcome. These data were obtained every 3 weeks during
treatment and at 6-week and 3-month follow-ups after termination.
It was hypothesized that the treatments would differ with respect to
outcome measures and to measures of the therapy process. Differences
between the cognitive-behavioral assertion treatment and the self-man-
agement treatment would be attributed to the differences in the struc-

ture and context prescribed by the Kanfer and Grimm model because the
substantive content, that is, the program exercises and subject matter,
were identical in both groups. Because all three treatment approaches
were deliberately selected for their demonstrated effecfiveness in pro-
ducing change, it was not primarily the extent of change in assertiveness
but the pattern of process and outcome differences that were the focus
of the study. Independent judges were asked to determine the identity
of each of the therapies on the basis of audiotape segments of sessions in
order to check whether the therapist indeed conducted each treatment
group in accordance with prepared manuals. The ratings of the judges,
made on treatment-specific criteria, indicated that the therapist adhered
closely to the manual. Judges identified the segments accurately for the
three treatment conditions in all 54 rated samples. In a second validity
check 15-minute videotape samples were randomly selected from the
first 90 minutes of Sessions 1, 3, 6, and 9. The judges rated four dimen-
sions of therapy for presence or absence, based on distinguishing crite-
ria for the three different treatments. The very high accuracy of rating
for the primary features in each treatment indicated that the therapies
were executed according to the specification of the treatment manuals.
Each treatment method was found to be effective in increasing as-
sertiveness and lowering the overall pathological level reflected by the
MMPI-168 scores. The self-management treatment resulted in the great-
est change on scores of interpersonal difficulty, nonassertive behavior,
assertive refusal ability, self-evaluation of assertive skills, and level of
assertion standards. On the Rathus Scale the self-management and cog-
nitive behavior groups improved more than the relationship group. The
greatest advantage of the self-management treatment was found on self-
report measures of assertive behavior and self-perception about the ade-
quacy of assertiveness. The self-management treatment also produced
the highest level of attendance at meetings and of client participation
and involvement in treatment as measured by a questionnaire devel-
oped by Yalom, Houts, Zimerberg, and Rand (1967). It also produced
the lowest level of resistance on the Vanderbilt Negative Indicators Scale
(Strupp, 1979). This treatment was the most effective in enhancing per-
ceived control and perceived confidence. Measures of rate and du-
rability of change further indicated that self-management therapy led to
the most rapid rate of change in positive self-reactions, as measured by
subject reports on their belief in controlling life events, their potential for
change (Schefft, 1983), and their resistance level and perceived involve-
ment in therapy. Self-management also resulted in the greatest mainte-
nance of therapeutic gains. The cognitive behavioral treatment was
found to be more effective than relationship therapy in increasing
IMPLICA nONS OF A SELF-REGULATION MODEL OF THERAPY 41
positive self-view and assertive behavior. The main findings of this

study, supporting the rationale of our conceptual model, are shown in
Table 3.
First, client motivation during treatment and completion of therapy
tasks was considerably higher in the self-management group than in the
other two treatment groups. Second, it is in the self-management group
that the specific training for interpersonal assertiveness also generalized
TABLE 3.
Summary of Results of Assertiveness-Training Study
I. Validity of therapy methods

Near 100% accurate rater match of videotaped session excerpts with manuals for
groups.
II. Outcome Variables·
Assertive Behaviors:
1. Improvement on interpersonal difficulties: SM> CB,RTb
2. On Conflict Resolution Inventory (CRI):
a. Assertive refusal ability: SM > CB,RT
b. Nonassertive behavior reduction: SM > CB,RT
c. Self-evaluation of assertive skills: SM > CB,RT
d. Magnitude of goal-setting: SM > CB,RT
e. On all other CRI measures: SM = CB > RT
3. On Rathus Scale: SM = CB > RT
4. All groups improved on assertiveness from pre- to posttreatment
Nontargeted behaviors:
1. MMPI- all groups improved
2. Self-Reports:
a. Self-efficacy: all groups improved: SM> CB,RT
b. Perceived control: all groups improved from
pre- to posttreatment SM >RT
Only SM maintained gains at follow-up
c. Belief in change: all groups improved: SM >RT
Only SM maintained gains at follow-up.
d. Self-esteem: all groups improved: SM = CB> RT
Only SM increased gains at follow-up.
III. Process Variables
Perceived involvement in therapy: SM> CB,RT
Perceived treatment value: SM> CB > RT
Vanderbilt Negative Indications Scale: SM decreased
RT unchanged
CB increased
Completed assignments: SM > CB
Attendance: SM > RT > CB
• Treatment groups: SM = Self-management; CB = Cognitive-behavior therapy; and RT = Relationship

therapy.
b All results in this table are significant at p < .05.
From Schefft, B. (1983).
most widely to nontargeted behaviors after treatment. Finally, the enact-

ment of self-regulatory processes enhanced the durability of treatment
effects after termination. These findings tend to support the utility of the
model presented here, particularly in enhancing motivation, increasing
client participation and commitment, and extending therapeutic gains.
It is interesting to note that the model is consistent with the ap-
proach proposed by Marlatt in stressing the need for techniques that
enhance and maintain an individual's compliance and adherence to pro-
gram requirements and for utilizing both specific behavioral techniques
and cognitive intervention procedures (Marlatt & Parks, 1982). The re-
lapse model proposed by Marlatt (1979) and Marlatt and Gordon (1980)
also stresses the importance of preparing the client for future difficulties
in viewing problem behaviors as a probabalistic function of the concur-
rence of numerous variables. The findings of the study presented here
are also consistent with the results reported by Miller, Hedrick, and
Taylor (1983), who used behavioral self-control training with problem
drinkers and found changes in non targeted life problems on follow-up
as long as 24 months. Addictive clients are notorious for creating prob-
lems in therapy by their resistance to treatment requirements, non-
compliance, and other behaviors indicative of low motivation in addic-
tive disorders. Therefore this group of clients is particularly appropriate
for a treatment approach that emphasizes the development of moti-
vation toward a different living pattern, the participation of the client in
selecting treatment objectives, and assuming responsibility for carrying
out therapeutic tasks and a future orientation that stresses preparation
for enduring maintenance of treatment gains.
SELF-MANAGEMENT IN RESIDENTIAL TREATMENT
The model that I have presented has also been applied in various
settings for individual patients and group programs. One example of its
application to addictive behaviors is a residential treatment program for
alcoholics, developed and directed by Ralph Schneider and his col-
leagues in Germany. It was recently described in a book, edited by
Schneider (1982). The program is consistent with our model in its aim to
change problematic drinking behavior. But the program is also based on
the assumption that understanding the context of this behavior, in-
creased self-confidence and feelings of competence (self-efficacy), reduc-
tion of anxiety and skill deficits, and motivated changes in life-style are
necessary for a satisfactory adjustment to a life without substance abuse.
In the program at the Furth Clinic, West Germany, these general
objectives are pursued by a dual approach. First, all patients participate
in therapeutic activities that relate to the common factors associated with
alcohol abuse, for example, group therapy that deals with understand-
ing the biological and emotional context and consequences of alcohol,
applied to each person's situation; methods of utilizing self-control and
relapse training, relaxation training, problem solving, physical fitness,
and leisure-time organization, etc. Second, patient groups, with mini-
mal guidance by the staff, also meet to discuss problems, evaluate pro-
gress, and deal with any items that are put on the agenda by group
members. These group meetings practice the assumption of responsibil-
ity by the patients for their own activities, therapy progress, and future
welfare. Third, individual therapy sessions allow specification of the
individual patient's problems and planning of the combination of vari-
ous specific treatment components that are available in addition to the
obligatory components; for example, preparation for occupational pur-
suits, social skills training, etc. Finally, family therapy is scheduled as a
required component of the program.
Because of the need for a gradual shift from initially high structure
to increasing personal responsibility, patients are first assigned to highly
structured groups. After about 6 to 8 weeks, they spend several days in
a relatively free period of "individual deliberation." Therapy meetings
are reduced in frequency and each patient is responsible for engaging in
intensive individual work on developing a "life balance-sheet." Follow-
ing the goal- and value-clarification phase in group work, this period
allows patients to work intensively on assessing their personal life goals
and developing plans for achieving them. Several daily therapeutic con-
tacts with individual staff members are of short duration (10 minutes).
They are designed to offer patients assistance and to monitor their pro-
gress. The patient then joins the more fully structured groups. The last
stages in our model, the preparation for generalization and transfer, are
represented by emphasis on return to the community or home environ-
ment for increasing durations. Patients prepare for contact either with
self-help groups or psychological service centers and plan how to
follow-up these contacts after discharge from the clinic. Contacts with
employers and employment agencies are also made at this time. A re-
cent extension of this approach has been proposed by Schneider in what
he calls "interval therapy" (in press). Its central feature is that patients
return to the clinic for "booster" treatments for periods of 2 weeks, at
intervals of 10 to 12 weeks, 6 months, and 22 months to strengthen
progress achieved and to deal with new problems. If the patient is in
outpatient treatment, clinic residence is not needed after the 6 month
readmission. The concept of realistic coordination of treatment, life
events, and social context underlies the interval-therapy program,
which is currently in the planning stage.
The Furth Clinic has an extensive system for data collection in order
to evaluate and improve treatment strategies and methods. Approx-

imately 500 patients have been treated yearly since 1978, primarily for
alcohol abuse. A one-year follow-up (for patients from the year 1979)
indicated an 80% abstinence rate for those who returned the mail ques-
tionnaires. About 35% of the patients did not return the questionnaire.
Of these 120 persons, a sample of 65 was visited in person by a psychol-
ogist. On the basis of this information it was established that 40% of this
group had remained abstinent. A 4-year follow-up revealed essentially
the same findings. It is interesting to note that patients who had re-
lapsed after 6 months were also those who had relapsed at the 4-year
follow-up.
A FEW UNRESOLVED ISSUES
The model that I have presented stresses the blending of social and
personal control in therapy. It emphasizes the need for helping the
client to experience, not just verbalize, his or her potentials for change,
to set clear goals, and to accept responsibility for the change process.
Although each element requires the use of some techniques, our view is
closer to a general-systems approach than to a model that highlights
only one or a few limited principles or mechanisms of change, such as
reconditioning, removal of barriers to self-realization, or extending con-
scious awareness of the origin and nature of central emotional conflicts.
Recently, Smith, Glass, and Miller (1980), Shapiro and Shapiro
(1982) and others reported the generality of therapeutic effects, re-
gardless of the type of treatment, therapist experience, and other fac-
tors. These findings have created considerable emotional reactions for
several reasons that are well described by Parloff (1984). Nevertheless,
even if the research methods are flawed or the samples biased, the
findings do suggest that there may be critical components in the therapy
process without which successful outcome cannot be achieved. Some of
these factors were noted by Jerome Frank over 20 years ago (1973). The
presence of such common features and the report of similar proportions
of success across widely differing schools of psychotherapy has led
some authors to attribute the effects of therapy entirely to the rituals of
treatment (e.g., Fish, 1973), or to nonspecific factors. In the treatment of
addictions, as in other life problems, I believe that the effects are the
results of a combination of the unique constellation of factors that in-
clude the nature and severity of the client's problem (e.g., schizophrenia
vs. subassertiveness), the social context in which the client lives and the
therapist and patient operate, (e.g., a court-referred middle-aged mar-
ried male problem drinker vs. a self-referred, single adolescent client),

and the clinical approach used.
Although various meta models of the therapy process are useful in
definition of the problem, in activation of the client, and in the creation
of favorable conditions for change, several tasks remain: first, the fur-
ther development of a repertoire of methods by which various key ele-
ments of the therapy process are best facilitated for a given client; sec-
ond, the wider use and application of data from other branches of
psychology, such as learning, motivation, information-processing, so-
cial processes, and others, in the development of these methods. The
issue of client motivation and commitment is particularly focal. Recent
work on compliance, the reexamination of the concept of resistance, and
the closer analysis of client motivation that I have proposed here, all
point to the need for further research in this area and for translating
concepts and laboratory procedures into clinical methods. Finally, we
need to recognize that the heuristics for application of a body of scien-
tific knowledge to a specific life problem can be developed only by close
collaboration between the theoretician-researcher and the practitioner.
Although general principles are needed to guide conceptualizations and
methods, specific experience and/or data in the client's problem area are
required to know what principles to apply how, with whom, when and
by whom for greatest effectiveness. This does not mean individual ther-
apy in the sense of uniqueness and artistry but rather the planned com-
bination of various methods, objectives, and therapy contexts corre-
sponding to the capacities, potentials, and limitations that each client
presents.
SUMMARY
It is now possible to summarize the implications for the treatment of

addictive behaviors, the issue that is of central concern here. I have
presented a conceptual framework that may be useful mainly because it
calls attention to the diversity of factors that determine successful treat-
ment. But at the same time this perspective points to critical contextual
components that may be indispensible for any treatment that desires
durable effects.
It is widely accepted among therapists that addicted persons do not
change by enforced deprivation of the abused substance, nor do most of
them view the abandonment of the habit or substance as a virtue per se.
Therapy therefore cannot aim solely toward temporary compliance with
abstinence rules nor toward breaking the behavior sequence associated
with the consummatory act. Although physical and psychological bene-
fits can be derived from such procedures, the key to long-term therapeu-
tic effects lies in helping patients to develop goals and incentives that are
based on their acceptance of the inherent advantages of an addiction-
free life. This motivational source cannot be the desire to avoid social
punishment or to please others but it must originate in goals that the
patient generates. To this end we have suggested a model for the thera-
py process that helps the patient
1. To develop goals and incentives that are potentially attainable
and fit with his or her life experiences and sociocultural milieu
2. To experience feedback over time that reflects a balance of great-
er satisfactions and/or lesser distress in favor of the drug- or
alcohol-free life-style over the addictive life pattern
3. To acquire a skill repertoire (a) for attaining some of the same
positive outcomes as previously achieved by the addictive habit
(e.g., stress-reduction, social contacts, and support) by other be-
haviors and achievements, and (b) for handling temptation,
seeking or creating environments that support and model non-
addictive life-styles, and for coping with new problems or cues
that prompt the old behavior pattern
For expository purposes these therapeutic tasks can be put in overly
optimistic and simple terms: the therapist must help the patient to
dream new dreams that are achievable and move the patient to action, to
taste some success on the road to making the dream a reality, and to
offer professional help to make the transition as easy and painless as
possible.
REFERENCES
Fish, J. M. (1973). Placebo therapy. San Francisco: Jossey-Bass.

Frank, J. D. (1973). Persuasion and healing. (rev. ed.) Baltimore, MD: Johns Hopkins Univer-
sity Press.
Kanfer, F. H. (1977). The many faces of self-control, or behavior modification changes its
focus. In R. B. Stuart (Ed.), Behavioral self-management (pp. 1-48). New York:
Brunner/Maze!.
Kanfer, F. H., & Grimm, L. G. (1980). Managing clinical change: A process model of
therapy. Behavior Modification, 4, 419-444.
Lange, A. J., & Jakubowski, P. (1976). Responsible assertive behavior: cognitive-behavioral
procedures for trainers. Champaign, IL: Research Press.
Marlatt, G. A. (1979). Alcohol use and problem drinking: A cognitive-behavioral analysis.
In P. C. Kendall & S. D. Hollon (Eds.). Cognitive-behavioral interventions: Theory, research
and procedures (pp. 319-356). New York: Academic Press.
Marlatt, G. A., & Gordon, J. R. (1980). Determinants of relapse: Implications for the
maintenance of behavior change. In P. O. Davidson & S. M. Davidson (Eds.), Behav-

ioral Medicine: Changing health life styles. (pp. 410-452). New York: Bruner/Maze\.
Marlatt, G. A., & Parks, G. A. (1982). Self-management of addictive disorders. In P. Karoly
& F. H. Kanfer (Eds.), Self-Management and behavior change: From theory to practice (pp.
443-488). New York: Pergamon Press.
Miller, W. R., Hedrick, K. E., & Taylor, C. A. (1983). Addictive behaviors and life prob-
lems before and after behavioral treatment of problem drinkers. Addictive Behaviors, 8,
403-412.
ParJoff, M. B. (1984). Psychotherapy research and its incredible credibility crisis. Clinical
Psychology Review, 4, 95-109.
Patterson, C. H. (1974). Relationship counseling and psychotherapy. New York: Harper & Row.
Prochaska, J. 0., & DiClemente, C. O. (1982). Transtheoretical therapy: Toward a more
Rathus, S. A. (1973). A 30-item scale for assessing assertive behavior. Behavior Therapy, 4,
398-406.
Rogers, C. R. (1961). On becoming a person. Boston, MA: Houghton-Mifflin.
Rogers, C. R. (1957). The necessary and sufficient conditions of therapeutic personality
change. Journal of Consulting Psychology, 21, 95-203.
Schefft, B. (1983). Self-management therapy vs. cognitive restructuring plus behavior rehearsal vs.
relationship psychotherapy: A controlled study of process and outcome. (Doctoral disserta-
tion, University of Wisconsin-Milwaukee, 1983. University Microfilms No. 1848)
Schneider, R. (Ed.). (1982). Stationiire Behandlung von Alkoholabhiingigen. Munich, West
Germany: Gerhard Roettger.
Schneider, R. (in press). Interval therapy. Suchtgefahren.
Shapiro, D. A. & Shapiro, D. (1982). Meta-analysis of comparative outcome studies: A
replication and refinement. Psychological Bulletin, 92, 581-604.
Smith, M. L., Glass, G. V., & Miller, T. 1. (1980). The benefits of psychotherapy. Baltimore,
MD: Johns Hopkins University Press.
Strupp, H. H. (1979). The Vanderbilt Negative Indicators Scale: Rater Manual. Mimeographed
manuscript, Nashvile, TN: Department of Psychology, Vanderbilt University.
Yalom, I., Houts, P., Zimerberg, S., & Rand, K. (1967). Prediction of Improvement in
group therapy. Archives of General Psychiatry, 17, 159-168.
II
Contemplation
Motivation for Change and Prevention
3
From Contemplation to Action
The Role of the World Health Organization
MARCUS GRANT
The purpose of this chapter is to review current activities within the

World Health Organization's (WHO) global program on the prevention
of alcohol-related problems and to explore the extent to which these
activities reflect the comprehensive model of change that underpins this
volume. In particular, because the majority of the other chapters focus
on the enhancement of clinical practice, work in two linked areas will be
emphasized-advocacy of the public health interest, and the develop-
ment of national policies.
To an extent, of course, it is stretching its limits to take Prochaska
and DiClemente's model of change, which was derived from processes
in individuals, and apply it to larger social concerns. Nevertheless, there
are interesting lessons to be learned, particularly in relation to the stage
of decision making. Countries, like individuals, do not change ar-
bitrarily and it is possible to see the work of an organization like WHO as
assisting in the movement from contemplation to action.
WHO AS A FOCUS FOR CONCERN
Ever since the First World Health Assembly in 1948, WHO has
recognized that it has a role as the focus for international concern about
MARCUS GRANT • Division of Mental Health, World Health Organization, 1211 Geneva
27, Switzerland. The views expressed in this chapter are those of the author and do not
necessarily reflect those of the World Health Organization.
51
52 MARCUS GRANT
alcohol-related problems. The range and severity of these problems vary

considerably from country to country as well as within countries. Nev-
ertheless, the accumulated research evidence of recent years demon-
strates that there is generally a positive association between trends in
alcohol consumption and trends in alcohol-related problems. This needs
to be understood from an individual and a social perspective. The rela-
tionship between consumption and problems is certainly complex, be-
cause even in a single country it cannot be assumed that drinking behav-
ior is evenly spread throughout the population. It is also important to be
aware that there may be groups who are particularly at risk.
Recent decades have witnessed considerable increases in alcohol
consumption and in alcohol-related problems in countries in all regions
of the world (WHO, 1980). Within the WHO European Region, the
number of countries with an annual per capita intake of more than 10
litres of pure alcohol increased from 3 in 1950 to 18 in 1979. Countries in
the WHO Western Pacific Region report sharp increases in alcohol-relat-
ed health damage, in alcohol-related crimes, and in alcohol-related acci-
dents during the 1970s. Similar reports have emerged from countries in
other WHO regions, including those with long traditions of abstinence
from alcohol. Although some countries in Western Europe and North
America are now reporting a leveling-off and even a modest decline in
alcohol consumption, the global trend is still that of continuing growth,
with particularly sharp increases in commercially produced alcoholic
beverages in some developing countries in Africa, Latin America, and
the Western Pacific (Walsh & Grant, 1985).
Not only the alcohol dependence syndrome itself, but also a wide
range of disabling and sometimes fatal physical and psychological con-
ditions, can be attributed either wholly or in part to excessive drinking.
In addition, alcohol-related traffic accidents account for significant pro-
portions of deaths in many countries, especially among young people.
Other accidents, including accidents in the work setting, are more fre-
quently related to alcohol consumption than is widely recognized. In
more general terms, the disruption to family life caused by the excessive
drinking of one or more family members causes distress and can also
result in violence and neglect. Other areas of concern include drinking
by young people and drinking by pregnant women, because both pose
questions about the possible harm that may be caused to vulnerable
populations. Drinking practices in some developing countries, which do
not have long historical traditions of consuming commercially produced
beverages of the range and strength available in most developed coun-
tries, may lead to a concentration of alcohol-related problems among
technicians and professionals, who are the scarcest resource, or among
young people, who represent the country's investment in its future. In
ROLE OF THE WORLD HEALTH ORGANIZATION 53
such circumstances, the real cost to the community is greater than

would be apparent from a simple statement of alcohol-related mortality.
CURRENT WHO ACTIVITIES
Following the recognition, at the 32nd World Health Assembly

(WHA) in 1979, "that problems related to alcohol, and particularly to its
excessive consumption rank among the world's major public health
problems ... " (WHO, 1985, pp. 103-104), the technical discussions at
the WHA 3 years later brought together participants from more than 100
countries. In the report of their discussions, they emphasized four basic
messages: that alcohol-related problems are health problems; that action
to reduce them is urgent; that there is sufficient consensus on priorities;
and that explicit commitment must replace token action (WHO, 1982a).
These themes were echoed in a further resolution at the 36th World
Health Assembly, which urged member states "to formulate compre-
hensive national policies, with prevention as a priority, and with atten-
tion to populations at special risk, within the framework of the strategy
of health for all" (WHO, 1985). This resolution demonstrates the impor-
tance of the alcohol program to WHO's global strategy of health for all
and expresses the collective decision of all member states of WHO to
deal with major risks to health through a resolute program of action
involving all sectors of government.
An international meeting on alcohol and health held in Geneva in
November 1983 identified effective advocacy approaches. This meeting
involved media practitioners and communication scientists. The recom-
mendations are now being implemented through the Division of Mental
Health and the Division of Information and Education for Health. Na-
tional and international meetings and workshops are being organized to
develop, promote, and test specific approaches to the prevention of
alcohol-related public health problems, with special emphasis on the
needs of developing countries.
A review was undertaken of the various documents on alcohol pro-
duction, consumption, and related health problems issued or drafted by
the Organization during the past few years. This review revealed that
the information available had been insufficiently exploited for advocacy
purposes. It also highlighted some discrepancies between various sets of
data. Action has now been initiated for improving our data base and for
producing a series of documents on alcohol production, consumption,
and health-related problems. These include an analysis of alcohol pro-
duction and trade that documents major trends, discusses their public
54 MARCUS GRANT
health implications, and suggests areas for future work (Walsh & Grant,
1985).
Thus, within the terms of Prochaska and DiClemente's comprehen-
sive model of change, it can be seen that activities in the advocacy area
have the double purpose of providing substance for contemplation and
providing a stimulus for translating the results of contemplation into
specific plans for action. This is, in itself, an important component of
international efforts in the area of alcohol-related problems, because
there are still many countries that have been reluctant to accept the
seriousness of the public health risks associated with particular life-
styles that include increasing rates of alcohol consumption.
A second priority area within the WHO alcohol program is collab-
oration with countries in the development and evaluation of the effec-
tiveness of national policies on alcohol-related problems, within the con-
text of national health planning and development. Following a compre-
hensive review of the world literature (Moser, 1980), a basic document
has been prepared (Farrell, 1985) on policy options for decision-makers.
It distinguishes between those policy measures for which there is now
sufficient objective evidence of effectiveness, those for which the evi-
dence is mixed, and those for which there is widespread popular sup-
port but little objective information. It is hoped that this document will
be of use in a wide range of countries, and that opportunities will arise
for working with countries to help test the impact of different ap-
proaches to policy development and implementation.
In the meantime, intercountry and national workshops are being
organized in a number of countries, particularly in the WHO African
Region. Simultaneously, the European Regional Office is coordinating an
eight-country project on community response to alcohol-related prob-
lems, as part of the development of more effective national programs.
The lack of adequate statistical information to support work in this
area has been repeatedly deplored in World Health Assembly resolu-
tions and in requests from member states. Following detailed planning,
it has now been established that collaborating centers are especially well
placed to playa leading role in developing activities in this area. The
Addiction Research Foundation (Canada) and the National Institute on
Alcohol Abuse and Alcoholism (USA) are bringing together groups of
scientists from a range of countries to review the existing situation and
to plan for specific international collaborative work. The first meeting
will concentrate on the improvement of the measurement of the alcohol
component in casualty statistics.
All these activities in the area of national policy development rely
on the involvement of sectors other than health, as well as health au-
thorities themselves. They therefore serve as a stimulus to the creation
of a truly integrated approach to national policy formulation and imple-
mentation. Indeed, it is as the process moves from formulation to imple-

mentation that the change from contemplation to action achieves partic-
ularly vivid relevance. Policy formulation is itself an active process,
involving the participation of representatives of agriculture, tourism,
law and order, finance, and many others. Then, as the policy is imple-
mented, so the quality of action becomes increasingly essential. It is, in a
sense, comparatively undemanding to create a theoretical policy; but
when that policy is actuaIly going to change the legal framework and the
sociocultural context in which people live, then it becomes an enterprise
demanding the most careful and the most courageous of approaches.
This is true, also, although in a more restricted sense, of the third
priority area for the WHO alcohol program-namely, the development of
techniques for identification, prevention, and management in primary
health care settings. Because this area of work is more familiar to readers
of this volume, it is sufficient to note that a series of activities are devoted
to the development of measures to reduce alcohol-related problems in
family settings and in employment settings. Of particular interest is work
on the early detection of alcohol problems and the evaluation of simple
treatment interventions in the primary health care context.
The work on early detection has involved the comparison of exist-
ing screening methods and their testing in Australia, Bulgaria, Kenya,
Mexico, Norway, and the United States. The results of this project are
being fed directly into work on the development evaluation of measures
to be used for treatment and management of alcohol problems in the
primary health care setting. The project on treatment emphasizes the
development of effective and simple low-cost methods, designed to in-
crease (a) the number of people that will be reached, (b) the likelihood
that health systems can incorporate the relevant knowledge, and (c) the
likelihood that countries can afford the cost.
A simple advice session,will be compared to a counseling session
and self-help manual, in order to test their effectiveness in reducing
consumption and/or alcohol-related problems in problem drinkers.
Efforts in this area are supplemented by a descriptive survey of the role
of general medical practitioners in the management of alcohol-related
problems in a range of developing and developed countries. Reports
based on studies from 14 centers in 12 countries have been prepared in
order to achieve a better understanding of the current practice and the
future potential of general practitioners in identification, treatment, and
prevention of alcohol-related problems.
CONCLUSION
As is apparent from this brief description of the current WHO ac-

tivities on the prevention and control of alcohol-related problems, the
56 MARCUS GRANT
intention is to establish a program that makes the most cost-effective use

of scarce resources, that will attract the support of relevant national and
international interests, and that will contribute significantly to the
efforts of the Organization to work with countries to achieve the goal of
health for all.
Because these activities fall within the work of the Division of Men-
tal Health in relation to the prevention of control of alcohol and drug
abuse, it is important to be aware of the strong links that exist between
efforts to alleviate alcohol-related problems and efforts to alleviate drug-
related problems. For many of the activities described above in terms of
the alcohol program, there exist parallel activities in the field of other
psychoactive drugs. Some issues are of particular relevance to one area
or the other, but many benefit from the development of a common
approach to alcohol and drug abuse.
In either case, the relevance of descriptions of change processes in
individuals to change processes at community, national, and interna-
tionallevels is more than accidental. As one moves from contemplation
to action, so the consequences of particular policies become actual and
presume commitments of their own. Although this is sometimes a dis-
tressing experience in which cherished approaches turn out to be less
effective than had been hoped, it is of course best viewed as a challenge.
The conclusions of scientific studies and careful analyses of policy op-
tions have to be applied in the world of political and social realities.
Either they improve the health of people and of nations, or they do not.
And if not, then at least they should be able to provide experience that
enables one to plan more effectively for the future. The work of WHO,
in this as in other areas of technical expertise, has to do with gathering
the best available data, marshalling those data into useful strategies, and
then disseminating the strategies globally. This is an active process,
consisting of a judicious blend of scientific credibility and persuasive
advocacy. It also involves a watchdog function, in which WHO can act
as a collective conscience for its member states.
All these interrelated tasks are apparent in the current efforts to
prevent and control alcohol- and drug-related problems. These, too,
need to be seen within the context of the complete program of work in
the Division of Mental Health (WHO, 1982b). In order to achieve its
goals, the program gives central attention to the development of ways
that can help in the preservation and enhancement of mental health at
all ages and in the specific sociocultural contexts of member states. Alco-
hol-related problems impinge on many areas of physical and mental
health as well as on social functioning. National and international efforts
to prevent and control alcohol abuse are best seen as part of the broader
health concerns that are reflected in the strategies to achieve health for
all by the year 2000.
REFERENCES
Farrell, S. (1985). Review of national policy measures to prevent alcohol-related problems. Un-
published World Health Organization Document, MNH/PAD/85.14. Geneva: World
Health Organization.
Moser, J. (Ed.). (1980). Prevention of alcohol-related problems: An international review of preven-
tive measures, policies and programmes. Toronto: Alcohol and Drug Addiction Research
Foundation.
Walsh, B., & Grant, M. (1985). Public health implications of alcohol production and trade. World
Health Organization Offset Publication No. 88. Geneva: World Health Organization.
World Health Organization. (1980). Problems related to alcohol consumption. Report of a WHO
Expert Committee. Technical Report Series No. 650. Geneva: Author.
World Health Organization. (1982a). Alcohol consumption and alcohol-related problems: Devel-
opment of national policies and programmes. Report of the Technical Discussions. A/35 Tech-
nical Discussions/6. Geneva: Author.
World Health Organization. (1982b). Seventh general programme of work covering the period
1984-1989. Health for All Series No.8. Geneva: Author.
World Health Organization. (1985). Handbook of resolutions and decisions of the World Health
Assembly and the Executive Board (Vol. 2). Geneva: Author.
4
From Contemplation to
Determination
Contributions from Cognitive Psychology
CLAUS-PETER APPEL
HUMUNCULUS ALCOHOLICUS: A CASE OF ATTRIBUTIONAL

FAILURE?
Before beginning, I would like to comment on my choice of approach. I

have chosen to treat the subject matter entirely from a cognitive, infor-
mation-processing point of view. This choice emanates from the impres-
sion that cognitive aspects (except for descriptions of cognitive impair-
ment; Jones, Jones, & Hatcher, 1980; Mello, 1972) have not been given
proper attention in the treatment of addictive behaviors, especially with
regard to the client's propensity to delay decision making with respect to
drinking. Yet, it is this very delay that seems to foster the therapist's
resentment that is sometimes reported in the literature (Baekelund &
Lundwall, 1975, 1977).
The patient behaves in what seems to be an erratic manner. Hence,
patients are often described in a pejorative fashion, as recently demon-
strated in a paper by Einstein (1982). Reflecting on the concept of drug
user he concluded
CLAUS-PETER APPEL • Department of Applied Psychology, University of Lund, Para-

disgatan 5, 223 50 Lund, Sweden.
59
60 CLAUS-PETER APPEL
the concept is a pejorative misnomer that prevents adequate treatment plan-

ning, negatively affects the treatment process and is irrelevant to treatment
outcome research. The drug user is associated with categories of pathology,
irresponsible behaviour or deviance that serves to limit adequate treatment
planning. (p. 205)
The type of attributional failure pointed out by Einstein (1982) is by

no means new in the history of clinical psychology. My feeling is that
such misattribution to a large extent is due to a lack of communication
between those researching in clinical and general psychology. For the
sake of illustration I would like to remind you of a discussion that went
on in the beginning of the 1950s. At that point, psychiatrists reported a
phenomenon that was judged to be typical of the pathology of reasoning
in schizophrenic patients: patients were described as incapable of logical
reasoning. This judgment was based on the observation that schizo-
phrenic patients frequently would judge items as belonging to the same
class if they shared a common quality (Arieti, 1955). This notion was
vividly rebutted by some researchers in cognition (Chapman & Chap-
man, 1959) on the grounds that research on reasoning with normal
individuals showed that this manner of lumping things together was
also characteristic of normal individuals. Chapman & Chapman (1959),
in summing up their observations, concluded:
If those writers are reporting a partially valid observation, the validity must
exist in a greater error tendency among schizophrenics, or the appearance of
the error tendency in contexts in which normals would not show it ... in
exacerbations of normal error tendencies. (p. 225)
In a recent article by Heather & Robertson (1983), a parallel to the

above may be found, in that it represents a refusal to accept prematurely
all drinking behavior as pathological:
But what can be asserted with reasonable confidence is that the drinking of
persons labelled alcoholic is subject to the same laws as drinking in persons
not so labelled and that alcoholic drinking is modifiable in principle. (p. 140)
For the success of a model of change, like the one proposed by

Prochaska and DiClemente (1982), it seems important to me that the
kind of misattributions reported above be avoided. This requires that we
disentangle the specific from the general, the pathological from the nor-
mal, in our understanding of the process that leads from contemplation
to determination. I therefore would like to seize this opportunity to
address some aspects of behaviors displayed by normal individuals
when solving problems and making decisions.
Were I to summarize in advance the main point of my presentation,
I would paraphrase Heather and Robertson (1983) by saying that the
shortcomings in decisional behavior often observed in the addictive cli-
FROM CONTEMPLATION TO DETERMINATION 61
ent are governed by the same rules as the decisional shortcomings in

people not so labeled, and that much can be done to understand, and,
we might hope, facilitate the transition from contemplation to determin-
ation by utilizing results from research on normal problem solving and
decision making.
CONTEMPLATION: COMMON SHORTCOMINGS IN THINKING

AND REASONING
The process of contemplation may, to a large extent, be understood

as reasoning with oneself. From my reading of the literature on thinking
and reasoning, much of the behavior displayed by alcoholic clients can
be understood at least equally well by applying a frame of reference
based on normal psychology as by applying a frame of reference based
on assumptions of pathology. It is my belief that, unless we learn how to
handle the shortcomings normally displayed, little progress will be
made in counseling in general and in the counseling of alcoholic clients
in particular.
PROBLEM SOLVING
In reviewing classic research on problem solving, I was struck by

the rates of success in a number of studies. Maier's (1931) by now classi-
cal work on problem solving may serve as an illustration. He asked
subjects to tie together two ropes that were suspended from the ceiling
but too far apart to be reached simultaneously. In order to tie them
together, subjects had to realize that a pendulum could be produced by
making use of tools lying about in the laboratory. The following rates of
success were reported: about 40% of the subjects solved the problems
without aid given by the experimenter, 38% with some help, and 22%
failed even after assistance.
These figures are intriguing because they are very similar to out-
comes reported in studies on treatment in general and alcoholism in
particular (Luborsky, Singer, & Luborsky, 1975; Menaghen, 1983). They
are also in line with findings regarding the percentage of people who
successfully handle addictive problems on their own without coming
into treatment (Pearlin & Schooler, 1978, Schachter, 1981). Perhaps the
similarity between these observations is not a matter of pure coinci-
dence, but rather is due to the fact that the process of problem solving
studied in the laboratory has much in common with the processes en-
countered in the natural environment and in the therapeutic context.
Maier (1931) also reported the behavior of his subjects when failing
to arrive at a satisfactory solution. They would, when finally realizing

that a pendulum could solve the problem at hand, blow at the cords,
throw things at them, or talk about a magnetic force that might draw the
cord. The resemblance to behavior that alcoholic patients reportedly
display as described by Jacobs (1981) is striking: when failing to achieve
their objectives, people will often revert to what is called "wishful think-
ing" as if engaged in a "brainstorming session," where much of the
result will end in the waste paper basket once it is scrutinized for feasi-
bility.
Maier (1931) also reported the amount of aid (or counseling if you
will) necessary in order to maximize success: 80% of those solving the
problem did so within 10 minutes without any aid, 49% of the subjects
in need of a hint solved the problem within one minute after the hint
was given. Finally, it was concluded that the probability of solving the
problem decreased as a function of time the subjects spent on the prob-
lem and of the number of hints given.
Again, the results reported by Maier remind me vividly of the liter-
ature on counseling, where more counseling does not necessarily yield a
better result. Maier also made some interesting observations on how
people progressed towards the solution and how they reported their
progress: there was a marked tendency to repeat variations of pre-
viously unsuccessful solutions. Solutions often appeared suddenly as a
complete idea. When hints are given, subjects usually were unable to
point out what it was that made them come up with a solution.
The results reported by Maier (1931) again are similar to findings
reported by researchers studying the therapeutic process. Clients will
often repeat previously unsuccessful behavior (like handling emotional
problems by means of drinking), suddenly change their pattern of be-
havior, and when asked what it was that affected the change, more
often than not, both the client and the counselor are unable to pinpoint
anything in particular.
In cases like these, Kohler (1947) would probably have offered the
concept of insight as an explanation. Although stimulating, it offers little
in the way of practical suggestions for the counselor. Other researchers,
like Maier (1931) would explain the kind of behavior discussed here in
terms of concepts like productive and reproductive thinking: most of the
everyday problems we are confronted with can be managed by the
implimentation of problem-solving behavior readily available to us.
Thus, we are required only to think reproductively. It is only when we are
faced with novel situations that we cannot rely on means readily avail-
able to us, and thus are required to think productively. In such instances,
experience has to be restructured, and this seems to be a task that is
more difficult to accomplish than learning from scratch. Birch & Rabino-
witz (1951) have pointed out that the importance of past experience for
the solution of a novel problem lies not in whether relevant information
has been acquired, but rather in what context the information or skill has
been acquired, because this seems to shape the perception of cues and
stimuli. More recently, Pearlin & Schooler (1978) have shown that cop-
ing efforts that are successful in one domain might have no effect, or
might be detrimental, in others. Hence it may prove difficult to perceive
a pair of scissors as a means of constructing a pendulum, or, more
relevant to the present issue, to learn to use a colleague as a resource for
getting rid of an alcohol problem: in order to do that, you have to
perceive him or her as a person not primarily to impress, but also as a
person who can help. The problem is to get rid of what Duncker (1945)
would have termed "functional fixedness," which is known to interfere
with problem solving by limiting the generation of alternatives.
When scrutinizing the literature for remedies for the kind of short-
coming discussed thus far, little is offered that may directly aid counsel-
ing. However, the point I am trying to make is that this situation will not
prevail once we start to see the similarity between the shortcomings of
supposedly abnormal behavior and normal behavior.
REASONING
One phenomenon discussed in the literature on reasoning is the

effect of context on reasoning. Henle (1962) points out that failures to
reason logically often emanate from the fact that subjects have a tenden-
cy to slip additional premises into the problem. Poor results are thus
often due to a misunderstanding of the problem rather than to a failure
to reason logically. It is interesting to note that subjects, capable of
solving simple syllogisms, drastically reduced their performance when
syllogisms were reissued in a form that had a highly emotional content,
as when pregnant women were asked to solve syllogisms about infant
care. Kopp (1960), analyzing data from schizophrenic patients, found
the same shortcomings with regard to reasoning as did Henle in her
normal subjects. A recent paper on alcoholics' decisional and reasoning
processes (Samsonowitz & Sjoberg, 1981) described the reasoning pro-
cess of addicts as "twisted" when analyzing factors that addicts held to
be responsible for their relapse into drinking. I have the feeling that a
concept like "twisted reasoning" will add little to our understanding of
the process at hand and runs the risk of being used in the pejorative
manner pointed out by Einstein (1982).
In the previous example we have been dealing with deductive rea-
soning. Inductive reasoning has also been intensively studied (Bartlett,
1958). A paradigm often employed in this context is sectional map read-
ing. It requires subjects to make a plan for getting from point A to B.

They are fed relevant bits of information about the geography as they
proceed with the task. A consistent finding reported by Bartlett (1958)
was that subjects, once they had decided about a route, were not easily
convinced to abandon their course, no matter how much evidence was
piled up against it. Furthermore, subjects seemed to prefer routes that
gave them a variety of later choices, rather than routes that led to binary
choices. The subjects in Bartlett's study are, to illustrate the results, like
people on the way to their weekend cottage who decide to take a short-
cut. After a while it becomes clear to everybody involved that the so-
called shortcut has turned out to prolong the journey considerably.
Meanwhile everybody gets hungry, tired, and irritated, yet the road that
seems to lead no where is not abandoned. This behavior, in my mind,
differs little from the behavior of a client who decides to take a shortcut
to comfort by buying yet another bottle of liquor, although it has
dawned on him that the shortcut is a mistake. The fact that the severity
of consequences differs in these two examples does not imply that dif-
ferent types of concepts need to be employed in order to understand
these two decisional behaviors.
Rayner (1958) has pOinted out that people generally have great diffi-
culties when planning ahead, especially developing alternative plans in
case the original plan does not work out. This goes even for very easy
tasks. Combining Rayner's (1958) results with those of Bartlett (1958),
one would expect that a treatment program that extends over a limited
span of time and that offers many alternatives, thus postponing final
choices, would be more attractive to clients contemplating treatment
than would programs that span over long periods of time and offer very
little in terms of alternatives. Similarly, it seems that a counselor would
do well in helping the client to develop alternative ways of handling
various problems, thus delaying premature steps towards closure. Once
a premature solution is attempted, it may be very difficult to stop the
process, let alone to reverse it.
Another commonly observed shortcoming in normal thinking is
due to the fact that we usually look for confirmation rather than nega-
tion about hypotheses or beliefs we hold to be true. A major source
regarding this topic is the work done by Bruner, Goodnow, and Austin
(1956). Their results suggest that only a small minority of subjects would
use a strategy where they tried to refute their own hypothesis. The
mainstream behavior is to try to gather evidence for one's own beliefs.
This is so common in everyday life, that we tend to overlook its role as a
major source of distress. Indeed, it is the common way in which science
(as well as this presentation) tends to proceed, as has been pointed out
by Kuhn (1962) and Popper (1959). The result of such self-confirming
behavior may very well be to slow down progress, be it on the level of

progress in science or on the level of recovery in our alcoholic patients,
when they reason that "there simply is no other alternative" to handle
the problem at hand.
For the sake of illustration I would like to describe one of my clients.
He had come for counseling because he recently had fallen and injured
himself. This worried him but he did not link his lack of balance to
drinking. Quite to the contrary: his own observation was that these
symptoms were reduced after taking a couple of drinks. What worried
him now was that drinking did not help any longer. In order to accom-
plish any change, my client would have to do something that few people
ever do, even when sober: challenge his own experience, not unlike the
subjects in Bartlett's (1958) and Bruner's (1956) experiments, by conduct-
ing an experiment of a special kind: quit alcohol for a couple of weeks
and compare the resultant state to the present one!
DETERMINATION: COMMON SHORTCOMINGS

IN DECISION MAKING
When faced with the task of decision making, the shortcomings

commonly observed in the thinking and reasoning process are likely to
come into play. However, research on decision making has exposed a
number of shortcomings that are particular to the context of decision
making. There is little reason to believe-a priori-that the principles of
the decisional processes in addictive behaviors differ from the principles
normally observed in people, even though the consequences may have a
quality of their own.
Much of the research done on the topic of decision making has
focused on model building, an enterprise with its roots in the 18th
century. Basically there are two classes of models. One, usually called
"normative," is based on some criterion of rationality and prescribes
what one should do in order to be rational, like "maximizing outcome"
in a choice situation or alternatively "minimizing loss." The other class
of models, often referred to as "descriptive," represents an attempt at
building models based on behavioral deviations from the normative
models of decision making.
OPTIMIZING
The optimizing type of model comes with various labels, depending

on how its various parameters are defined. In its basic form it would
include the two parameters expectation (E) and value (V). Hence, the
model may be referred to as the EV model. Other forms of the model

have essentially the same structure and rationale. (For readers interested
in the formal aspects of these models, I would recommend consulting
Lee, 1971).
Basically, these models hold that people would choose the alter-
native with the highest payoff in a set of options, like choosing the best
apple one can get for a given amount of money, which seems to be
reasonable by anyone's standards. The problem is, however, that you
would have to select the best apple on display, which means that you
would have to check each of the apples available with regard to all
aspects that make an apple a good bargain: size, color, taste, texture,
firmness, and price. My guess is that you will not attend to all aspects, at
least not when buying an apple, and it is a good bet that you will never
do so, no matter what you have to decide. Yet, this is exactly what the
optimizing model would expect you to do.
Simon (1979), in reviewing the literature on optimizing, concluded
that people simply lack the "wits" to maximize. Part of the problem is
that checking on all the available alternatives would make unreasonable
demands on our information processing capacity and exceed by far that
of our memory (Miller, 1956), which seems to work within the limits of
seven plus or minus two chunks of information (provided we are un-
affected by drugs). Thus, the frequently made observation that people
will behave in an erratic fashion when faced with decisions where many
variables have to be considered, is understandable.
Another problem involved is the tendency to frequently shift the
values attached to the aspects that enter the decision (like deciding
whether color or texture is more important in an apple). As a result,
drastic changes with regard to preferences are observed, and it does not
matter in this context whether we are talking about decisions concerned
with short- or long-lasting consequences. If we are dealing with long-
range consequences, there is, however, yet another problem that may
turn out to be detrimental to good decision making: usually we seek to
solve immediate problems, sacrificing long-range solutions to problems,
especially when under time pressure and real or imagined stress
(Young, 1966).
Thus, when scrutinized through the optics of this normative model,
it is uncommon for people to behave rationally, whether drunk or sober.
But, if we can not optimize, what can we do?
SATISFICING
The satisficing model has been formulated by, among others, Simon
(1976). Its attractiveness is due to the fact that it successfully handles
some of the shortcomings found in the optimizing model.
Instead of having you optimize outcomes, for example, looking for

the best apple available for a given price, this model assumes that you
only continue searching until you have found one that is "good enough."
In other words, a decision is made when an alternative is found that
meets a minimum set of requirements. One of the major advantages of
this model is that it requires a less thorough search of information as well
as less storage of information than the previous model. The payoff for this
reduction of effort would be less cognitive strain, but at the price of
making suboptimal decisions.
Numerous studies have corroborated this model in both real life
settings and laboratory experiments. For instance, in one study on exec-
utive decision making it was found that executives would lean toward
more conventional decisions, choosing the second best alternative, as
soon as uncertainty was involved in the decision (Young, 1966). In an-
other study (Tversky, 1972) it was shown that the tendency to subop-
timize will be stronger when greater uncertainty must be dealt with,
which often means sacrificing long-term gain for short-term accept-
ability. According to Etzioni (1968) this holds true for consumers in the
supermarket as well as for governmental officials. Thus, it seems reason-
able that it also will encompass the behaviour of our clients when mak-
ing a decision about sobriety.
What people seem to be looking for is an improvement over prevail-
ing states, and typically only two alternatives are compared at a time. If
neither of the alternatives will do, one is likely to repeat the whole
decisional sequence once more until one is satisfied. The model thus
does not preclude the contemplation of a larger number of alternatives,
but the process of scrutinizing is most likely to be done in a sequential
rather than in a parallel fashion. Thus a great deal of the information
gathered is usually lost along the way, which is one good reason why
record keeping and progressive balance sheets are of value in coun-
seling.
The satisficing model comes in various shapes. In one of them,
subjects will use one criterion only in their decision. Using only a single
criterion in the decisional process is by no means a sign of cognitive
impairment or inebriety. This "single mindedness" is frequently ob-
served when people are faced with a major or a difficult decision. In the
information gathering phase of the decisional process, this is illustrated
by asking only one close friend for advice, one lawyer, or one physician
rather than a number of people; a simple decision rule has also been
found to satisfactorily describe choice behavior of consumers (Hansen,
1972).
A special case arises when morality is involved (Schwartz, 1970). In
such cases, people very often felt that they have not been suboptimizing,
but have chosen "the only possible alternative." (In a similar vein,
political decisions are often characterized by a simple decisional rule, like

the consensus rule.) As a result, the likelihood increases that the quality
of the decision may be impaired. This phenomenon is particularly pro-
nounced when disapproval is anticipated either from others or from
oneself. Utilitarian considerations that might be important are likely to be
set aside for the sake of adhering to an overriding moral criterion.
It is a common observation that moral issues often are raised in the
context of drinking problems. If we are to trust the literature on decision
making, in such instances a situation is created that promotes impaired
decision making.
It is safe to say that this model of decision making is closer to
what people actually do. It is equally clear, however, that the decisional
process understood in terms of this model is still far from rational,
because people will not make use of all the information available. Thus,
a bystander as well as a researcher or a counselor may very well despair
when watching what goes on.
EUMINATION BY ASPECTS
One of the major assumptions of elimination-by-aspects model is

that the decisional process has much in common with the game known
as "Twenty Questions" (Tversky, 1972). People are not single minded as
in the previous model. Instead a considerable number of criteria may be
used. What happens is that people are expected to engage in a suc-
cessive narrowing-down process. Usually one would start with the most
important criterion or aspect and check available alternatives with re-
gard to this aspect. Alternatives that do not meet this criterion are thus
rejected as unacceptable. If we stick to the previous example of the
purchasing of an apple, it is like eliminating all green apples from con-
sideration, keeping the red ones only. Next, one would proceed by
scrutinizing the remaining red apples with regard to the criterion second
in importance, like size.
One of the major findings when researching this model is that peo-
ple frequently behave erratically with regard to what they consider the
most important aspect in their decision. After having scrutinized half of
all the apples on display, people may discover that the color of apples
really is not as important as their size or texture. Thus the process has to
start all over again.
Even thuugh this process often is quite slow, it will usually progress
toward an optimal course of action in terms of "incremental improve-
ments" geared to alleviating present shortcomings (Miller & Starr, 1967).
It is also a type of process that is regarded as a safeguard against drastic
changes (Popper, 1959). In the context of addictive behavior, the incre-
mental-decision-making model would fit the situation where a person,

beginning with an occasional drink for the purpose of relaxation, would
end up with what euphemistically is called a "drinking problem" with-
out ever making an active decision about drinking behavior.
Evidence for major life decisions being the result of an incremental
decisional procedure comes from various studies. Ginsberg, Ginsburg,
Axelrad, and Herma (1951) noted that this type of model adequately
described the choice of career even in people with skilled occupations.
Similar results were obtained by Matza (1964) on the "careers" of delin-
quents, and by QaIler (1938) with regard to the way in which people
chose partners for marriage.
It is unlikely that people will adopt but one decisional strategy, and
Etzioni (1967) has proposed a model combining aspects of all three mod-
els previously reviewed. He claims that subjects would attempt optimiz-
ing when confronted with major difficulties, whereas satisficing would
be more common when simple decisions are asked for. However, os-
cillation between different approaches is likely to present a problem of
its own, especially when these oscillations occur unsystematically in the
context of counseling.
DETERMINATION: COMMON SHORTCOMINGS IN JUDGMENT
Judgment may be considered an integral part of all contemplation

and decision making and many researchers in decision making have also
turned their attention to various aspects of this topic. Some of their
findings may be of interest when reflecting on the normalcy of client
behavior.
One phenomenon described in the literature is called the "conjunc-
tion fallacy." The probability of a conjunction, such as P(A&B), cannot
exceed the probability of its constituents, P(A) and P(B). Judgments
under uncertainty, however, are often mediated by intuitive heuristics
that do not adhere to the conjunction rule. This means that in everyday
life, a conjunction can be more "representative" in our minds than any
of its constituents. Such violations of the conjunction rule have been
demonstrated in various contexts; for example, personality judgment,
decision under risk, suspicion of criminal acts, etc. Tversky and Kahn-
eman (1983, p. 313) conclude:
In cognition, as in perception, the same mechanisms produce both valid and
invalid judgments. Indeed, the evidence does not seem to support a "truth
plus error" model, which assumes a coherent system of beliefs that is per-
turbed by various sources of distortion and error.
One such normal mechanism is the influence of affect on the judg-

ment of the probability of events. Reading about an undesirable event
increases people's estimates of the frequency of risks (Johnes & Tversky,
1983), even for events totally unrelated to the event read about. Con-
versely, reading about a positive event can change affect accordingly.
Thus, the way in which information influences judgment about
future events may be systematic, but not necessarily adequate. In this
respect, alcoholic clients confirm the rule, rather than being an
exception.
Our normally somewhat disturbed relation to information is also
confirmed by research on the Bavesian model. This research deals prom-
inently with the impact of information on probabilities involved in deci-
sional problems. (For the mathematics of the model, see Lee, 1971). The
question is, how are people influenced by new eivdence when making a
decision?
What has been found in many studies is that people generally, at
least in single-stage decisions, undervalue new information; that is, they
behave more conservatively than is prescribed by the Bayesian model.
In multistage decisions the tendency is reversed: people tend to be more
extreme than they ought to. In any case, people tend to be highly confi-
dent in their judgment. Cohen, Chesnick, and Haran, describing what
they call the "inertial-phi effect" conclude: "the phenomenon which our
experiments appear to have identified ... may turn out to be a charac-
teristic feature of human judgment, while pin-pointing one of its
Achilles heels" (1972, p. 46).
Tversky and Kahneman (1974) have suggested that simple heuris-
tics, which they have labeled representativeness, availability, and ad-
justment, might be responsible for our tendency to produce erroneous
and overconfident judgments.
The heuristic of representativeness implies, among other things,
that judgment regarding the likelihood that one event will generate
another, depends on the similarity between these events. (Similar
events are judged to be representative of one another.) In the process,
small samples are usually judged as being typical for large populations.
Finally, error due to unreliability in the data judgments are based on is
usually underestimated.
Research on the heuristic of availability shows that an event is
judged likely or frequent if it is easy to imagine or recall in relevant
instances. (For an alcoholic it may be neither easy to recall nor to imag-
ine what things are like when sober.) Because availability is affected by
factors unrelated to likelihood, such as familarity, recency, and emo-
tional salience, heavy reliance (as possibly done in satisficing) on this
heuristic is likely to result in a systematic bias in the decisional process.
Finally, similar to results in psychophysics and social psychology

(e.g., anchoring, halo effect) the heuristics researched by Tversky &
Kahneman (1974) show that people are unable to adjust their judgment
to new information, which usually results in overconfidence. This is so
common in judgments involving chance and skill that some researchers
have talked about the "illusion of control" (Langer & Roth, 1975; Slovic,
Fischoff, & Lichtenstein, 1976).
DEALING WITH OVERCONFIDENCE
One prominent line of reasoning to account for the phenomenon of

overconfidence (often found to be typical in the first steps of recovery in
clients) has been that the environment is often insufficiently structured
to provide adequate feedback to make changes in one's judgments.
Considering the power and persistence of this tendency, it would proba-
bly require equally persistent and powerful feedback to achieve correc-
tion, something that is unlikely to happen in ordinary social conduct
without being regarded as a violation of rules for conduct. It seems thus
plausible that an important aspect of counseling would be to provide the
kind of feedback necessary, which is unlikely to occur in the natural
environment. Technically, this could imply the design of tasks a patient
would be asked to carry out between sessions.
However, even if adequate feedback is provided, there is still room
for distortion, because it is known that people are more likely to re-
member favorable outcomes than unfavorable ones in a number of dif-
ferent tasks (Langer & Roth, 1975). Furthermore, it is also well docu-
mented that people exaggerate their confidence and ability to predict
behavior, especially the extent to which others should have predicted an
outcome and their own ability to predict future events (Fischoff, 1975).
Inspected with this in mind, the client's prediction, his confidence in his
ability to "quit the habit" must be understood for what it is: a declara-
tion of an intention rather than a promise, one that is subject to the same
fallacy of all "normal predictions." Thus, what often is perceived as a
moral issue, or a character deficiency, can be understood otherwise.
SITUATIONS IN WHICH FEEDBACK WILL NOT WORK
One way of looking at the dilemma faced by the addict when trying
to change is to understand it in terms of a model called "Multiple Cue
Probability Learning" (Slovic & Lichtenstein, 1971), a type of inferential
learning. In essence, this means handling a number of cues (like situa-
tions or moods when one drinks) that come with different probabilities.
Provided the cues presented are linear (cues are linearily related to some
outcome in a situation, such as "every time I start to think about my

finances I start to drink"), learning is rather slow. If, however, a non-
linear relationship exists between outcome and cue, learning is even
slower. People are known to behave very inconsistently under such
circumstances, and what is worse, especially when one considers what
this means for decisional counseling, feedback does not help very much.
The situation is similar to one in a series of experiments done by
Brehmer (1974), where subjects were confronted with inferential tasks.
For some of these, the rules were entirely random. People tried to solve
the problems, generate rules, change them; they became very inconsis-
tent and would often revert (or relapse if you will) to previously unsuc-
cessful solutions. Even when tasks are not random, but nonlinearity
does persist, people will continue to have a hard time even when given
the proper rules. Real life feedback is similar to this situation in that
people in the client's social environment vary their behavior toward
them. Inconsistent behavior toward clients is also likely to occur in the
hospital setting and, last but not least, in the individual counselor's
behavior toward the client.
NOT MAKING A DECISION: CONTEMPLATION

VERSUS DETERMINATION
In the literature on cognition and decision making, not making a

decision has been given little attention. It is usually regarded as a conse-
quence of factors involved in the decisional process, and not as behavior
in its own right.
An aspect that is likely to exacerbate the problem of the alcoholic
client, as indeed of all people contemplating a major life decision, is the
fact that the decision to stay sober is usually tied into a whole sequence
of decisions (e.g., developing new skills for handling anxiety without
alcohol, changing one's leisure time activities or making new friends).
As noted before, in sequential decision making the final outcome is
affected by the probabilities in all steps in the sequence. Keeping the
inertial-phi effect in mind (Cohen et al., 1972), it is obvious that the more
steps involved, the less likely it is that the final outcome will be the one
predicted or hoped for.
There are, of course, many ways in which one could describe non-
decisions. I should like to address but three aspects of nondecision
closely related to what has been said about decision making in the pre-
vious paragraphs. Nondecision comes in at least three forms: refusals,
delays, and inattention.
REFUSALS
In many real life situations the status quo is associated with less
uncertainty than other alternatives. Many writers view it as the refer-
ence point against which all other alternatives are evaluated (Pitz, 1980).
Its appeal is that one knows more about it, whereas one knows less
about the alternatives. How much the lack of knowledge regarding alter-
natives may mean in the decisional process is nicely illustrated by a
recent encounter with a client of mine. My client strongly believed that
the discomfort experienced when reducing her alcohol and drug intake
would prevail for the rest of her life. Her refusal to change was under-
standable in the light of beliefs about what things would be like if she
decided to quit drinking.
Perceived risk is another related factor that may lead to both deci-
sions and nondecisions. Subjects will often have a maximum level of risk
they accept, and beyond that level an alternative will be rejected, no
matter how favorable the outcome may be (Pruitt, 1962). These levels are
usually quite low, as demonstrated in several studies (Irwin & Smith,
1957; Lanzetta & Driscroll, 1966). The research has led to the suggestion
that "risk tolerance" or the style of decision making could be viewed as a
personality trait, but the evidence for such a view has been scanty (Cox,
Chesnick, & Rieh, 1964; Goldsmith, 1968, Kogan & Wallach, 1967;
Lamm, 1967).
DELAY
Another form of nondecision is delay. The reasons for delay may

include (a) inspection of alternatives, (b) getting a second opinion, (c)
deliberation, and (d) waiting for the availability of the goal.
Inspection of Alternatives
A paradigm that has been frequently used to simulate real life
events has been the "secretary problem," proposed by Gilbert & Mostel-
ler (1966). It refers to a class of tasks where alternatives are presented
sequentially, for example, as when hiring a secretary. The subject may
stop the search for information at any time in the process or go on, at a
specified cost, to collect more information. Thus the number of alter-
natives can be controlled as well as their qualities. A major inference
from this type of study has been that the mere compilation of alter-
natives usually helps people to clarify their goal as well as the develop-
ment of means to achieve the goal in question. This is yet another piece
of evidence for the position that helping a client in the generation of
alternative ways of handling a problem is vital in the process of

counseling.
Getting a Second Opinion

Klahr (1967) has shown that people will intensify their search for
information when the alternatives at hand are similar. However, they
will go on searching for information about the available alternatives,
even when the outcome is unavoidable and the information is useless
(Lanzetta & Discroll, 1966). People have even been found willing to
invest a greater part of available resources (be it time or money) into
reducing this kind subjective uncertainty. Thus, the reduction of uncer-
tainty seems to be of major importance in the formative period of a
decision, and counseling will, to a large extent, always imply the reduc-
tion of such uncertainty.
Deliberation
The evaluation of available alternatives after they have been com-
piled has been another popular topic in research on decision making.
Among the prominent questions that have been asked are (a) Why do
some decisions take so long, and (b) What do people do while not
making a decision?
These questions have been studied in various ways, often by em-
ploying reaction-time measures. Irwin (1958) suggested, as did Pavlov
before him, that delay in choice behavior is intimately related to the
possibility (or capacity) to make discriminations. It would seem plausi-
ble that gathering information would play an important part in this
process. However, it has been frequently found that people gathering
little information will not arrive at a decision sooner than people gather-
ing much information. Those gathering little information merely seem to
use more time in reiterating the information available. Furthermore,
there is no evidence that the two strategies involved would lead to
differences in the quality of the decision made (Zajonc & Bernstein,
1961).
Svensson (1974) has more directly investigated what is done when
information is evaluated. He found that people usually attempt to ease
the cognitive burden involved in decision making through simplifying
the problem dealt with, and a decision is usually achieved by reducing
the number of aspects considered in the decision.
In a similar vein Slovic, Fischoff, and Lichtenstein (1977) have at-
tempted to take a closer look at decision making by having people
"think aloud" while in the process. From their compilation of verbal
protocols they concluded that subjects proceed by eliminating alter-

natives they feel are unacceptable. The major reason for a prolongation
of decision making, viewed from this angle, is the tendency to oscillate
with regard to the relative value of aspects.
Another major problem involved here is the revocability of a deci-
sion: it takes more than twice the time to make a decision that is viewed
as irrevocable than it takes to make one that is perceived as revocable
(Mann & Taylor, 1970), provided the decision is easy. If it is difficult, the
irrevocable decision may take still longer by comparison. As of yet, we
know only little about how revocability affects real life decisions, be-
cause the results reported come out of the laboratory. Yet there is little
reason to believe that people would behave more efficiently in real life,
when stakes usually are considerably higher and the amount of informa-
tion to be dealt with usually is more complicated than in the laboratory.
Waiting for the Availability of the Goal

Even though often observed in every day life, the topic of waiting
for the availability of a goal has attracted little attention in research,
apart from the finding by Mischel (1961) that people will opt for less
attractive alternatives when delay of gratification is required. In a recent
study by Vuchinich & Tucker (1983), subjects were confronted with a
choice between different amounts of alcohol and different amounts of
money, with varying degrees of delay. It was found that the preference
for alcohol, in all types of subjects, varied negatively with the amount of
money at stake and positively with the length of delay. Several other
aspects often considered of importance in the clinical context were in-
vestigated in this study, like the role of mood, but data did not show
that they modified choice behavior. The authors concluded that states
often Reported as immediately preceding the decision to have a drink
are probably given too much importance. Instead they suggest that the
availability of alternatives be given more attention when studying the
processes involved in deciding to have a drink.
Inattention
The preceding examples are based on the premise that subjects are
aware of alternatives. This might, however, not be the case. (Whether
this is so subjectively or objectively does not really come into play at this
level.) While inattention prevails it is unlikely that a decision is equiv-
alent to making a decision. Therefore it seems to be vital to ask why
situations fail to elicit a cognitive structure that will result in activities

normally associated with decision making.
There are mainly two lines of inquiry regarding this topic. One is
predominantly found in clinical literature and deals with the problem in
terms of avoidance whereas the other avenue of inquiry deals with it on
the level of task properties and is primarily found in the cognitive liter-
ature (Janis & Mann, 1977).
The propensity for the conscious avoidance of decision making
seems to be as old as mankind. It has been interpreted as shunning
responsibility, especially if blame, be it from others or from oneself, is
involved. Although this interpretation is quite frequently made (Kauf-
man, 1973; Mack, 1971), there is little evidence for such an interpretation
from research. The majority of support for such a view seems to ema-
nate from case studies and fiction. (This does not imply that such evi-
dence should be dismissed a priori.) Research on decision making that
might be interpreted to support the view that we are dealing with con-
scious avoidance comes from Kogan and Wallach (1967). In their work
they describe a phenomenon called "risky shift," which has been in-
terpreted as the avoidance of individual responsibility. The phe-
nomenon refers to the observation that a higher level of risk is accepted
when people make decisions collectively than when making decisions
individually.
The variable most frequently researched in the other line of inquiry
is the salience of cues (Yakimovich & Salz, 1971). Studies in which this
aspect has been varied systematically show that decisions often can be
reversed as a consequence of this variable. Salience is thus a step to-
wards discrimination, which in turn is a vital step towards reduction of
ambiguity. If low levels in any of these factors prevail at any of the
decisional stages, the likelihood of attention is reduced and with it the
probability that a decision will be made. Hence, raising the salience of
appropriate cues is likely to be a main objective in all kinds of counsel-
ing, including the counseling of the addicted client.
Even cognitively oriented researchers have speculated about the
possibility that decision making is aversive in character, but have
stressed intellectual rather than emotional strain (Festinger, 1964; Jams,
1959; Lewin, 1942). As of yet, there is little experimental corroboration of
this notion.
According to the view outlined here, nondecision may, to a large
extent, be understood in terms of cognitive factors, such as uncertainty,
level of risk tolerance or ambiguity, salience of cues, perceived conse-
quences, generation of alternatives, and search for information. Because
there is little reason to believe that alcoholic clients differ from other
people in the dimensions discussed, it seems important that more cur-
rent research be geared toward the development of techniques that

make for more proficient handling of these factors in the counseling
context.
DECISIONS: SOURCES OF CONFLICT AND STRESS
Treating the subject of decisions from a cognitive point of view by

no means excludes consideration of emotion. Not being able to make or
carry out a decision, changing one's value structure, finding out that one
has forgotten important aspects that should be incorporated into the
pending decision are the interface between cognition and emotion, and
therefore can not be overlooked in the present context. (For excellent
papers bearing on this topic see Folkman 1984; Simon 1967; and Zajonc,
1980.)
ANTICIPATORY REGRET
A feeling of regret may often creep in on us even before a decision is

made. This phenomenon is studied under the heading of anticipatory
regret (AR); it is probably the aspect of contemplation that has been
studied most intensively. Because it is reported that people during this
phase often display a maximum of anxiety (Epstein & Fenz, 1965; Janis &
Mann, 1977), it seems that contemplation is the phase of the decisional
process where we are most vulnerable. The level of anxiety displayed is
usually related to potential loss expected. Proficient counseling will thus
have to deal with the content of cognitions concomitent to this kind of
stress, which, as pointed out by Easterbrook (1959), is disruptive of
performance.
There are a number of conditions that can augment the severity of
AR. Among the more prominent ones is the anticipation that negative
consequences of a decision will materialize soon after the decision is
implemented. Under such circumstances it is likely that any decision
needs a good deal of support. The alcoholic who expects severe symp-
toms of abstinence is exactly in this predicament. The frequently ob-
served postponement of "getting on the wagon" is thus a normal behav-
ior. For this reason it is probably wise to alleviate medically the immedi-
ate negative physical consequences for the patient who tries to quit
drinking. Unfortunately, counselors can rarely offer such immediate
comfort for other types of more or less direct negative consequences of
stopping drinking, which, in the view of the patient, may result in a
pessimistic outlook regarding the benefit of the counseling relationship.
Another condition of importance is the expectancy that one has to
adhere to a decision, especially if this expectancy is expressed by signifi-

cant others. Although significant others generally are looked on as a
major source of support in the literature, it probably should be remem-
bered that cases exist where the opposite may be true with regard to
effects on the level of anticipatory regret. The youngster who has his
first performance in the little league watched by his parents, the para-
chutist who is about to execute her first jump in front of her compan-
ions, or, last but not least, the alcoholic client who is expected by spouse
and friends to succeed at staying sober, no matter what hassles might be
involved, all these are everyday examples where those close to us can
become a threat to success, in that they exacerbate anticipatory regret.
Thus the propensity towards inefficiency and premature decision mak-
ing might be heightened considerably. Some recent data from Appel &
Berglund (1985) may be interpreted along these lines of reasoning.
These data show that clients who have been in treatment before seem to
place less emphasis on the support of significant others than on medica-
tion (Antabuse) when entering treatment for the second time. The re-
verse is true for clients coming into treatment for the first time. (This
does not imply that significant others should, as a rule, be excluded
from participation in the counseling process.) Further evidence that the
concept of support, especially social support, may need some refine-
ment is found in a recently published article by Kessler, Price, and
Wortman (1985).
POSTDECISIONAL REGRET
While AR is a state of mind that is associated with the time before a

decision is made, post decisional regret (PDR) is the label linked to the
period immediately after one has made a decision. A look into the liter-
ature on this topic shows that PDR is a common phenomenon that may
be induced by events or credible communications that call attention to
potential losses, no matter what type of loss is at stake, be it financial,
social, or health related (Fe stinger, 1964). The stressful nature of PDR is
well documented. People complain about sleeplessness and psychoso-
matic problems. Defensive strategies, such as shifting responsibility,
procrastination, intensive work, sexual indulgence, or getting drunk,
may be employed to reduce the strain experienced. Apart from this,
some other issues dealt with in this line of research may be of some
bearing on the present topic of decision-making behavior. The research
demonstrates that a decision is not necessarily final just because it has
been implemented.
However, once a decision is made, it may be in need of a lot of
support to be maintained. For instance, people read ads about the prod-
uct they have bought and avoid ads of rival brands. This is tantamount
to looking for support for one's decision. Another form of support fre-
quently employed is bolstering (Mann & Abeless, 1970; Vroom, 1966). It
seems that once we have made a decision, we want to look good re-
gardless of the quality of the decision made. This is not, it seems, unlike
the situation where an alcoholic client decides to stick to drinking bud-
dies, or avoid them altogether depending on what the decision is. Even
though the outcome of both these behaviors is likely to differ consider-
ably, the quest for support for one's decision may very well be the same.
With this in mind, it is easy to understand why a client, having acted
contrary to advice given, would refrain from returning to the adviser.
Thus, a good counseling strategy would be to assure the client that he or
she is welcome back even if things do not turn out as anticipated.
Oscillations in PDR
One of the tricky qualities of PDR often overlooked in the clinical
literature is that PDR, although highest immediately after a decision,
seems to oscillate drastically even in the in context of normal decision
making. Walster (1964) was the first to report this phenomenon in a
study on draftees' job choices. She found that the rated attractiveness of
a choice diminished directly after a decision was made, was then up-
graded and finally reduced again. The opposite was observed for the
rejected alternative. To me this seems a plausible frame of reference for
understanding what happens to many alcoholic patients. Viewed this
way, their oscillation between bouts of inebriety and sobriety does not
necessarily imply a pathological trait. Unfortunately, few clinical studies
seem to have scrutinized the mechanisms that are in control of this
phenomenon.
In the context of general psychology it has been found that PDR is
controlled mainly by the same variables as AR, for example, the level of
anticipated negative consequences and expected losses following the
implementation of the decision made. Furthermore, it seems that bouts
of PDR often occur spontaneously without having to be triggered by
external stimuli or events. Both the immediacy and spontaneity of the
occurrence of PDR should probably be given attention in counseling
sessions, and in order to minimize the probability of relapse it would
seem wise not to dismiss the client immediately after a decision is
reached. Rather, the frequency of contact probably should be high im-
mediately after a decision is made.
In a sense, the emphasis on PDR is only half the story. In reality w~
might very well be dealing with what is called a double ap-

proach/avoidance conflict, in that both quitting and continuing are am-
bivalent alternatives. The aversive concomitants of such dilemmas have
been well documented for both animal and man since Pavlov. Change in
this context is impossible without change in valence of the alternatives at
hand (Lewin, 1942).
Loss OF FREEDOM
One aspect of choice that seems to produce a state of conflict is the

perception that by making a choice, freedom is reduced, especially
when one has to choose between two alternatives only. It seems reason-
able to expect that some clients feel that entering treatment may include
such a loss of freedom. My guess is that any aspect of treatment that
exacerbates that impression, as when the only alternative offered to
drinking is total abstinence, would run the risk of low compliance.
When only two alternatives are open to us the situation is usually han-
dled in either of two ways: one is to make the threatening alternative
more attractive and the other would be to derogate the valued alter-
native. From research it seems that people spontaneously follow the
latter course (Walster & Walster, 1973). With this in mind, it seems wise
for the counselor to assist a client in this process, not only because there
is a natural tendency to do so anyway, but because we know that the
loss of a valued alternative is very likely to result in that alternative being
rated as more attractive later on, especially when the choice was a forced
one (Brehm, Stires, Sensenig, & Shaben, 1966).
A forced choice may result in what Sherif and Hovland (1961) have
called the "Boomerang Effect," that is, a person coerced into a course of
action is likely to do exactly the opposite. Estimates of the occurrence of
this effect range between 40% and 80%; thus it seems to be a quite
frequent and natural reaction. (These figures remind me of figures con-
cerning dropout rates from treatment.) The lower the credibility of the
directive agent, the more frequent the boomerang effect. Worchel &
Brehm (1971) found that restrictions of freedom or choice evoked consid-
erable hostility in college students, and what is worse, the hostility was
elicted regardless of the quality of the alternative that was forced on the
subjects.
The above results may carry important implications for the preven-
tion of relapse and the design of treatment programs, suggesting treat-
ment programs, suggesting that patients should be offered a choice with
regard to treatment programs as well as freedom of choice regarding
drinking goals.
HELPING CLIENTS TO MAKE A DECISION: COGNITIVE AIDS
Apart from providing an alternative frame of reference for under-

standing client behaviors, it has been pointed out, that research results
from cognitive psychology, if more productively used, might be of great
benefit to counseling (Heesacker, Heppner, & Rogers, Janis, 1982; Rush,
1982; Smith, 1982; Strong & Claiborn, 1982). I would like to address this
issue in this final section of my presentation.
DECISIONAL AIDS IN COUNSELING
The transfer of results from one area to another always presents

problems of its own (Arnkoff, 1980), because the conditions between
areas may vary a great deal. For instance, in normal decision-making
experiments, problems usually are better defined than in the therapeutic
setting, and usually it is assumed that subjects have the necessary skills
to execute the steps required to carry out a decision. Even more impor-
tant, subjects in an experiment are usually willing to carry out these
steps, whereas clients may not be. Indeed, this may be the primary
problem a counselor may be dealing with. Thus, techniques geared to
promote better problem solving or better decision making are not meant
to replace traditional counseling techniques, but they could be valuable
additions to existing approaches to assist the client to make a satisfying
choice about some major aspects of his or her life (Heppner, Hibel, Neal,
Weinstein, & Rabinowitz, 1982; Kahneman, Slovic, & Tversky, 1982;
Kanfer & Busemeyer, 1982).
Most of the decisional aids advocated focus on the alleviation of the
commonly observed normal shortcomings discussed earlier. Their ad-
vantage lies in that they are nonspecific and thus can become tools in the
hand of any user, whether it be a counselor or a client.
The basic rationale of many of these aids is a self-regulatory process
similar to the one proposed by Miller, Galanter, and Pribram (1960) in a
different context. Usually it implies the execution of a series of steps: the
detection of a difference between a current state and a desired state, the
generation of means to reduce the difference, the execution of steps
implied by the means generated, and finally checking the result against
the originally desired state. In the context of counseling this type of self-
regulation is usually achieved by self-monitoring and self-evaluation
(Miller & Munoz, 1982).
A problem with the philosophy underlying this model is that the
client's predicament may very well be the result of incremental deci-
sions. Changes often proceed in steps below what in psychophysics is
called the difference threshold or "just noticable difference" OND) (Ste-
vens, 1951). As a result, recognition of change may be slow if it occurs at

all and problems are not often detected at a time when change is still
relatively easy to implement. (This is similar to the gradual deterioration
of the brakes of your car: it is difficult to detect in everyday usage but if
adjusted, the difference is apparent.) When a client enters a counseling
relationship, it seems that problem detection has been successful at least
at some level, but that does not mean that detection has been successful
at all levels necessary to handle the problem successfully. The objective
of counseling will, at least to some extent, imply that the patient
emerges out of the process more sophisticated in the detection of warn-
ing signals that might serve as indicators of a problem. Furthermore, it
seems equallv important that the patient develop a decisional rule about
when to react to these cues and, finally, how to react to these cues. This
may sound trivial, but a closer look at the problem will reveal that a
dilemma may be implicit in this line of reasoning. The way in which I
like to suggest that one look at this problem is to apply an elementary
signal detection view (Green & Swets, 1966), which has been used suc-
cessfully as a model in other areas of psychology (McNicol, 1972). In this
model detection is handled as illustrated in Table 1.
Table 1 illustrates four different types of outcomes with regard to
the detection of signs or cues that "something is the matter." These
different outcomes (hit, miss, correct rejection, and false alarm) are in-
terdependent. Thus it is impossible to maximize the number of hits
without substantially raising the frequency of false alarms, an outcome
that in the present context would characterize a hypochondriac. Con-
versly, it is impossible to maximize correct rejections without increasing
the number of misses, an outcome that would be characteristic- of what
often is referred to as denial. The detection of cues, or the decision to act
TABLE 1.
Cue Detection
R Patient's view
e Cue No cue
a
I Cue Hit Miss
i
t No False Correct
y cue alarm rejection
Table 1 shows the four possible types of outcome when a

therapist (symbolized by "Reality" in the table) or a pa-
tient make judgments concerning cues of impairment.
The outcomes are interdependent. (For further informa-
tion please consult the text.)
on cues, is a function of factors like the payoff for different courses of

action, the proportion of relevant to irrelevant cues, and the similarities
between relevant and irrelevant cues. Some of these factors are beyond
the control of both the counselor and the patient. Others, like the payoff
function, may be modified in the course of counseling, provided detec-
tion is followed by implementing appropriate coping skills for the situa-
tion at hand.
TREES, GOALS, AND DECISIONS
Actually the above reasoning implies more than the mere detection
of cues. It also implies that a cue is linked to a state that differs from a
goal state. It is only when this requirement is met that we can talk about
a "problem" (Wickelgren, 1981). From both attribution theory (Harvey
& Waery, 1984) and research on problem solving (Simon, 1979), it is
known that the way a problem is presented has a major impact on the
efficiency with which it is solved, or for that matter, on whether it will be
solved at all. In real life, problems are typically more complicated by the
involvement of emotion and thus it is not uncommon that people wish
to "maximize" two incompatible outcomes (as when one wants to avoid
hassles with a spouse and to go on drinking.) In decision research vari-
ous means have been developed to handle this and similar problems.
Two major approaches for handling this kind of problem are called
"goal trees" or "means-end analysis" (Edwards & Newman, 1982;
Keeny & Raiffa, 1976). They have been used successfully to structure the
utilities of outcomes. The procedure implies that a hierarchical model of
a decision is built, one with the final goal at the top. This goal is then
decomposed into lower-level objectives, which are broken down into
concrete events. The procedure is thus a major step toward detailed
description of the process leading to any goal a person may want to
reach and seems ideal for monitoring progress. The approach is close to
what is known as "backtracking" in problem solving (Lindsay & Nor-
man, 1977). Another advantage of this approach might be that a goal,
thus decomposed, will render a more vivid picture in one's memory
(Abelson, 1976), which may be important for keeping up motivation. A
major task for the counselor in this context would be to initiate this type
of process, which may be achieved by having the client project what
things would be like if he or she adopted different solutions.
What we have been discussing is actually a two-stage process.
Ideally, one would want to build a goal tree with the objective of arriving
at a hierarchy of goals. The second stage would require that one would
structure the sequence of possible or necessary actions to arrive at the
first goal. As a result, a decision tree composed of actions would be
produced. Such a tree would have branches, points of choice or deci-

sion, where a previous outcome would determine the next step. (This is
a little like fault-finding charts for electrical circuits, or troubleshooting
charts for ignition problems in cars.)
A question of some importance is how far into the future it is rea-
sonable to project such decision trees and how many branches one
should allow in the counseling context before matters become too com-
plex. One way to look at this is to postpone such decisions until several
representations of a given problem area have been produced. The visu-
alization in itself is likely to propel further work resulting in revisions or
"pruning." It is in this process of pruning that the counselor probably
has one of the most important tasks, that of advising on proper se-
quences (easy to difficult, now vs.later, generation of alternatives, etc.).
Another objective that might be achieved in this process is the pinpoint-
ing of a client's specific strength or weakness, which may specify the
therapeutic interventions necessary (e.g., the development of coping
skills, such as self-assertiveness training or negotiation skills).
As pointed out before, from a cognitive point of view, a major
objective of counseling is the generation of alternatives. The strategy of
clarifying subgoals inherent in this approach is a step in this direction.
Apart from the type of analyses suggested earlier, problem solving
has been shown to improve greatly from using analogy (Newell & Si-
mon, 1972). The approach involves the extraction of relevant features
from a situation with which the client is familiar and the demonstration
of the similarity between that situation and the target situation. It is
important that the client not only be familiar with the situation, but also
have been successful in handling the situation that is used as a refer-
ence. In my experience, a lot is to be gained by demonstrating to a client
that progress could be made, provided he or she made use of skills
already in her or his possession. Another way to use analogy is to move
from simple to more complex problems, with the essential features kept
unaltered.
Is THE LITERATURE RELEVANT IN THE PRESENT CONTEXT?
In modern society, decision making is increasingly transferred to

heuristics. Military commanders have been relieved of the burden of
integrating information; United States analysts make use of the Bayesian
model for the processing of intelligence. The design of man-machine
systems has been adapted to ease the burden of decision making (Slovic,
Fischoff, & Lichtenstein, 1977).
There are, of course, other real life contexts where decisional aids
might be welcome. One such context is science, if one remembers how
often researchers in psychology interpret a regression toward the mean

as an experimental effect (Furby, 1973), thereby differing little from the
subjects in Kahneman & Tversky's research (1973). A similar situation
prevails in the field of law, where judges have been shown to display a
number of the effects discussed, like anchoring, failure to consider base
rates, and insensitivity to unreliability of evidence. They seem to do little
better than do eyewitnesses, who show the same overconfidence in
court as they do in the laboratory (Brooks & Doob, 1975; Buckhout, 1974;
Fischoff, 1976; Shah, 1975; Sue, Smith, & Caldwell, 1973).
Simple heuristics are thus held by people in all walks of life. It is this
very simplicity that makes them attractive and useful tools in the man-
agement of everyday life. However, these very same qualities turn them
into vicious instruments if applied to situations where they no longer fit.
Knafl and Burkett (1975), in attempting to analyze the decisional rules of
surgeons, found one dominating rule: "Don't cut, unless you absolutely
have to." Imagine an addict abiding by the same heuristic: "Don't quit,
unless you absolutely have to .... "
REFERENCES
Abelson, R. P. (1976). Script processing in attitude formation and decision making. In J. S.

Carrol, & J. W. Payne (Eds.), Cognition and social behavior. Hillsdale, N. J.: Erlbaum.
Appel, C. P., & Berglund, M. (1985). What is important in treatment? The patient's view.
Unpublished manuscript, University of Lund, Sweden.
Arieti, S. (1955). Interpretation of schizophrenia. New York: Brunner/Maze\.
Arnkoff, D. B. (1980). Psychotherapy from the perspective of cognitive theory. In M. J.
Mahonney (Ed.), Psychotherapy process: Current issues and future directions. New York:
Plenum Press.
Baekelund, F., & Lundwall, L. (1975). Dropping out of treatment: A critical review. Psycho-
logical Bulletin, 5, 738-783.
Baekelund, F., & Lundwall, L. K. (1977). Engaging the alcoholic in treatment and keeping
him there. In B. Kissin & H. Begleiter (Eds.), The biology of alcoholism: Vo\. 5, Treatment
and rehabilitation of the chronic alcoholic (pp. 161-187). New York: Plenum Press.
Bartlett, F. C. (1958). Thinking. An experimental and social study. London: Allen & Unwin.
Birch, H. G., & Rabinowitz, H. S. (1951). The negative effect of previous experience on
productive thinking. Journal of Experimental Psychology, 41, 121-125.
Brehm, J. W., Stires, L. K., Sensenig, J., & Shaban, J. (1966). The attractiveness of an
eliminated choice alternative. Journal of Experimental & Social Psychology, 2, 301-313.
Brehmer, B. (1974). Hypotheses about relations between scaled variables in the learning of
probabilistic inference tasks. Organisational Behavior & Human Performance, 11, 1-27.
Brooks, W. N., & Doob, A. N. (1975). Justice and the jury. Journal of Social Issues, 31, 171-
182.
Bruner, J. S., Goodnow, J. J., & Austin, G. A. (1956). A study of thinking. New York: Wiley.
Buckhout, R. (1974). Eyewitness testimony. Scientific American, 231, 23-31.
Chapman, L. J., & Chapman, J. P. (1959). Atmosphere effect reexamined. Journal of Experi-
mental Psychology, 3, 220-226.
Cohen, J., Chesnick, E. I., & Haran, D. (1972). A confirmation of the inertial-phi effect in
sequential choice and decision. British Journal of Psychology, 1, 41-46.
Cox, D. F., Chesnick, E. I., & Rich, S. (1964). Perceived risk and consumer decision
making. Journal of Marketing Research, 1, 32-39.
Duncker, K. (1945). On problem-solving. Psychological Monographs, 270, 1-113.
Easterbrookk, J. A. (1959). The effect of emotion on cue utilization and the organization of
behavior. Psychological Review, 66, 183-201.
Edwards, W., & Newman, J. R. (1982). Multiattribute evaluation. Beverly Hills, CA: Sage.
Einstein, S. (1982). "The druguser": A semantic source for built in treatment failure. Drug
and Alcohol Dependence, 2-3, 203-209.
Epstein,S., & Fenz, N. P. (1965). Steepness of approach and avoidance gradients in
humans as a function of experience: Theory and experiment. Journal of Experimental
Etzioni, A. (1967). Mixed scanning: A third approach to decision making. Public Admin-
istration Review, 27, 385-392.
Etzioni, A. (1968). The active society. New York: Free Press.
Festinger, L. (Ed.), (1964). Conflict, decision, and dissonance. Stanford, CA: Stanford Univer-
sity Press.
Fischoff, B. (1975). Hindsight vs. foresight: The effect of outcome of knowledge on judge-
ment under uncertainty. Journal of Experimental Psychology, 1, 288-299.
Fischoff, B. (1976). Perceived informativeness of factual information. ORI Res. Bull. 16 (3),
Eugene, OR: Oregon Res. lnst.
Folkman, S. (1984). Personal control and stress and coping processes: A theoretical analy-
sis. Journal of Personality & Social Psychology, 4, 839-852.
Furby, L. (1973). Interpreting regression toward the mean in developmental research.
Developmental Psychology, 8, 172-179.
Gilbert, J. P., & Mosteller, F. (1966). Recognizing the maximum of a sequence. Journal of the
American Statistical Association, 61, 35-73.
Ginzberg, E., Ginsburg, S. W., Axelrad,S., & Herma, J. L., (1951). Occupational choice.
New York: Columbia University Press.
Goldsmith, R. (1968). Personlichkeitsspezifische Komponenten des Entscheidungsverhal-
tens. Psychol. Forsch., 32, 135-168.
Green, D. M., & Swets, J. A. (1966). Signal detection theory and psychophysics. New York:
Wiley.
Hansen, F. (1972). Consumer choice behavior. New York: Free Press.
Harvey, J. H., & Weary, G. (1984). Current issues in attribution theory and research.
Annual Review of Psychology, 35, 427-459.
Heather, N., & Robertson, I. (1983). Why is abstinence necessary for the recovery of some
problem drinkers? British Journal of Addiction, 78, 139-144.
Heesacker, M., Heppner, P. P., & Rogers, M. E. (1982). Classics and emerging classics in
counseling psychology. JournaL of Counseling Psychology, 29, 400-405.
Henle, M. (1962). On the relation between logic and thinking. Psychological Review, 69, 366-
378.
Heppner, P. P., Hibel, J., Neal, G. W., Weinstein, C. L., & Rabinowitz, F. E. (1982).
Personal problem solving: A descriptive study of individual differences. Journal of
Counseling Psychology, 29, 580-590.
Irwin, F. W. (1958). An analysis of the concept of discrimination. American Journal of
Irwin, F. W., & Smith, W. A. S. (1957). Value, cost and information as determiners of
decision. Journal of Experimental Psychology, 54, 229-232.
Ivey, A. E. (1980). Counseling and psychotherapy. Englewood Cliffs, NJ: Prentice Hall.
Jacobs, M. R. (1981). Problems presented by alcoholic clients. A handbook of counseling strategies.

Toronto: Addiction Research Foundation.
Janis, I. L. (1959). Decisional conflicts: A theoretical analysis. Journal of Conflict Resolution,
3,6-27.
Janis, I. L. (Ed.), (1982). Counseling on personal decisions. New Haven: Yale University Press.
Janis, I. L., & Mann, L. (1977). Decision making. New York: Free Press.
Johnes, E. J., & Tversky, A. (1983). Affect, generalization, and the perception of risk.
Journal of Personality & Social Psychology, 1, 20-3l.
Jones, B. M., Jones, M. K., & Hatcher, E. M. (1980). Cognitive deficits in woman alcoholics
as a function of gynecological status. Journal of Studies on Alcohol, 1, 140-146.
Kahneman, D., & Tversky, A. (1973). On the psychology of prediction. Psychological Re-
view, 80, 237-251.
Kahneman, D., S1ovic, P., & Tversky, A. (Eds.), (1982). Judgement under uncertainty:
Heuristics and biases. New York: Cambridge University Press.
Kanfer, F. H., & Busemeyer, J. R. (1982). The use of problem solving and decision making
in behavior therapy. Clinical Psychological Review, 2, 239-266.
Kaufman, W. (1973). Do you prefer a life without choice? Psychology Today, pp. 78-83.
Keeney, R. L., & Raiffa, H. (1976). Decisions with multiple objectives: Preferences and value trade
offs. New York: Wiley.
Kessler, R. c., Price, R. H., & Wortman, C. B. (1985). Social factors in psychopathology:
Stress, social support, and coping processes. Annual Review of Psychology, 36, 531-572.
Klahr, D. (1967). Decision-making in a complex environment. Unpublished Doctoral thesis,
Carnegie Institute of Technology.
Knafl, K., & Brukett, G. (1975). Professional socialisation in a surgical speciality: Acquiring
medical judgement. Social Science & Medicine, 9, 397-404.
Kogan, N., & Wallach, M. A. (1967). Risk taking as a function of the situation, the person,
and the group. In T. M. Newcomb (Ed.), New directions in psychology, (Vol. 3, pp. 223-
287). New York: Holt, Rinehart & Winston.
Kohler, W. (1947). Gestalt Psychology. New York: Liveright.
Kopp, S. (1960). Deductive reasoning in paranoid schizophrenics. Unpublished doctoral disser-
tation, New School for Social Research, New York.
Kuhn, T. S. (1962). The structure of scientific revolutions. Chicago, IL: University of Chicago
Press.
Lamm, H. (1967). Will an observer advise higher risk-taking after hearing a discussion of
the decision problem? /. Pers. & Soc. Psych., 6, 467-471.
Langer, E. J., & Roth, J. (1975). Heads I win, tails, it's chance: The illusion of control as a
function of the sequence of outcomes in a purely chance task. Journal of Personality &
Social Psychology, 32, 951-955.
Lanzetta, J. T., & Driscroll, J. N. (1966). Preference for information about an uncertain but
unavoidable outcome. Journal of Personality & Social Psychology, 6, 96-102.
Lee, W. (1971). Decision theory and human behavior. New York: Wiley.
Lewin, K. (1942). Field Theory and learning. In D. Cartwright, Field theory in social science
(pp. 76-82). New York: Harper & Row.
Lindsay, P. H., & Norman, D. A. (1977). Human information processing. New York: Academ-
ic Press.
Luborsky, L., Singer, B., & Luborsky, L. (1975). Comparative studies of psychotherapies:
Is it true that "everyone has won and all must get prizes"? Archives of General Psychia-
try, 32, 995-1008.
Mack, R. P. (1971). Planning on uncertainty. New York: Wiley.
Maier, N. R. F. (1931). Reasoning in humans. II: The solutibn of a problem and its ap-
pearance in consciousness. Journal of Comparative Psychology, 12, 181-194.
Mann, L., & Abeles, R. P. (1970). Evaluation of presidential candidates as a function of

time and stage of voting decision. Journal of Psychology, 74, 167-173.
Mann, L., & Taylor, V. (1970). The effects of commitment and choice difficulty on predeci-
sion processes. Journal of Social Psychology, 82, 221-230.
Matza, D. (1964). Delinquency and drift. New York: Wiley.
McNicol, D. (1972). A primer of signal detection theory. London: Allen & Unwin.
Mello, N. K. (1972). Behavioral studies of alcoholism. In B. Kissin, & H. Begleiter, (Eds.),
The biology of alcoholism, Vol. 2: Physiology and behavior (pp. 219-284). New York: Plen-
um Press.
Menaghen, E. F. (1983). Individual coping efforts: Moderators of the relationship between
life stress and mental health outcomes. In H. B. Kaplan, (Ed.), Psychosocial stress:
Trends in theory and research (pp. 157-191). New York: Academic Press.
Miller, D. W., & Starr, M. K. (1967). The structure of human decisions. Englewood Cliffs, NJ:
Prentice Hall.
Miller, G. A (1956). The magical number seven, plus or minus two: Some limits on our
capacity for processing information. Psychological Review, 2, 81-97.
Miller, G. A., Galanter, E., & Pribram, K. H. (1960). Plans and the structure of behavior. New
York: Holt, Rinehart & Winston.
Miller, W. R., & Munoz, R. F. (1982). How to control your drinking. Albuquerque, NM:
University of New Mexico Press.
Mischel, W. (1961). Delay of gratification, need for achievement and acquiescence in
another culture. Journal of Abnormal & Social Psychology, 62, 543-552.
Newell, A, & Simon, H. A. (1972). Human problem solving. Englewood Cliffs, NJ: Prentice
Hall.
Pearlin, L. I., & Schooler, C. (1978). The structure of coping. Journal of Health and Social
Behavior, 19, 2-21.
Pitz, G. F. (1980). The very gUide of life: The use of probablistic information for making
decisions. In T. S. Wallsten (Ed.), Cognitive processes in choice and decision behavior (pp.
77-94). Hillsdale, NJ: Erlbaum.
Popper, K. (1959). The logic of scientific discovery. New York: Oxford University Press.
Proschaska, J. 0., & DiClemente, C. C. (1982). Transtheoretical therapy: Toward a more
integrative model of change. Psychotherapy: Theory, Research & Practice, 3, 276-288.
Pruitt, D. G. (1962). Pattern and level of risk in gambling decisions. Psychological Review,
69, 187-201.
Rayner, E. H. (1958). A study of evaluative problem solving. Quarterly Journal of Experimen-
tal Psychology, 10, 155-165.
Rush, A. J. (Ed.). (1982). Short term psychotherapies for depression: Behavioral, interpersonal,
cognitive, and psychodynamic approaches. New York: Guilford.
Samsonowitz, V., & Sjoberg, L. (1981). Volitional problems in socially adjusted alcoholics.
Addictive Behaviors, 6, 385-398.
Schachter, S. (1981). Self-treatment of smoking and obesity American Psychology, 37, Am.
Psycho/" 37, 436-444.
Schwartz, S. (1970). Moral decision making and behavior. In J. MaCauly, & L. Berkowitz
(Eds.), Altruism and he/ping behavior. New York: Academic Press.
Shah, S. A. (1975). Dangerousness and civil commitment of the mentally ill: Some public
policy considerations. American Psychiatrist, 132, 501-505.
Sherif, M., & Hovland (1%1). Social judgement: Assimilation and contrast effects in communica-
tion and attitude change. New Haven, CT: Yale University Press.
Simon, H. (1967). Motivational and emotional controls of cognition. Psychological Review, 1,
29-39.
FROM CONTEMPLAnON TO DETERMINAnON 89
Simon, H. A. (1976). Administrative behavior: A study of decision-making processes in admin-

istrative organizations (3rd ed.). New York: Free Press.
Simon, H. A. (1979). Information processing models of cognition. Annual Review of Psychol-
ogy, 30, 363-396.
Slovic, P., & Lichtenstein, S. C. (1971). Comparison of Bayesian and regression ap-
proaches to the study of information processing in judgement. Organisational Behavior
& Human Performance, 6, 649-744.
Slovic, P., Fischoff, B., & Lichtenstein, S. C. (1976). The certainty illusion. ORI Research
Bulletin, 4. Eugene, OR: Ore. Res. Inst.
Slovic, P., Fischoff, B., & Lichtenstein, S. C. (1977). Behavioral decision theory. Annual
Review of Psychology, 28, 1-39.
Smith, D., (1982). Trends in counseling and psychotherapy American Psychologist, 37, 802-
809.
5
Critical Conditions for Change in
the Addictive Behaviors
JIM ORFORD
INTRODUCTION
What is so significant about this volume is that the disease model is

nowhere to be seen. We now have the collective confidence to develop a
genuinely alternative way of thinking about change in the addictive
behaviors, and it is as a contribution toward the development of this new
thinking that this chapter is directed. I will begin by outlining six conclu-
sions that I draw from what has been written about change. Not all of
these conclusions are comfortable to live with for those of us who see
ourselves as specialist or expert treaters of addictive behaviors, but each
has to be accommodated in any new model of change we care to develop.
I will then make some remarks about the directions in which I believe we
should look for ideas with which to build our new models of change. I will
attempt to make the point that we are in danger of reinventing the wheel;
a basic understanding of how those with addictive behaviors do change
has been with us for a long time. Finally, I will briefly speak of some of the
implications for practice of our new ways of understanding how those
troubled by addictive behaviors make changes. Whether we care to admit
it or not, most of us practice within a quasi-medical format, and I am not
sure whether we have yet faced up to the need for changes in our own
ways of working that new models suggest.
JIM ORFORD • Department of Psychology, University of Exeter and Exeter Health Au-
thority, Exeter, England.
91
92 JIM ORFORD
SOME UNPALATABLE FACTS
CHANGES ARE DIFFICULT TO PREDICT
As Frederick Kanfer points out in Chapter 2 of this volume, when it

comes to addictive behaviors, people do not always, or even often, act
rationally, or in ways we believe would be in their best interests. Mil-
lions of people continue to smoke despite the well-publicized evidence
that it is damaging to health, and the experience of hospital liver units is
that most patients continue to drink alcohol after being told that this has
already damaged a vital organ (Wodak, Saunders, Ewusi-Mensah,
Davis, & Williams, 1983). Nevertheless, although some people per-
severe with addictive behavior against all logic and reasoning, many
make dramatic changes and give up or curtail long-established and
deeply ingrained patterns of behavior. The point is that the outcome of
addictive behavior is very varied and, as far as I know, our ability to
predict who will make changes and who will not is poor. This is appar-
ent, for example, in the lives of the American playwright Eugene O'Neill
and novelist F. Scott Fitzgerald. These two cases of alcoholism have
been described by the psychiatrist Donald Goodwin (1970, 1971).
According to Goodwin, O'Neill already had a drinking problem in
his late teens, was a periodic binge drinker throughout his twenties and
most of his thirties, and was prone to mad and wild drunken outbursts
and monumental hangovers. At the age of 37, however, he underwent a
brief 6-week psychoanalysis and thereafter, with a few temporary
lapses, went on the wagon for the remaining 28 years of his life. Scott
Fitzgerald is described by Goodwin as "an alcoholic par excellence." He
and his friends recognized this in his twenties and although he made
regular attempts to try to control his drinking, unlike O'Neill he never
succeeded for more than a few months at a time. He refused any psychi-
atric treatment and died prematurely of a heart attack. I doubt we could
have predicted confidently that O'Neill would change his addictive be-
havior and Scott Fitzgerald not, nor that a brief psychoanalysis would
have done the trick in O'Neill's case. Samuel Pepys and James Boswell
both struggled to control their sexual appetites according to Stone's
(1979) The Family, Sex and Marriage in England 1500-1800. Pepys appears
to have mastered his urges in the end, whereas Boswell, who deteriorat-
ed and finally was having severe difficulty controlling his gambling and
alcohol intake as well, did not. Hypersexuality (Orford, 1978a), like
alcoholism, appears also to have a variable and unpredictable outcome.
We know from long-term follow-up studies of excessive drug takers
(e.g., Thorley, Oppenheimer, & Stimson, 1977) that the same is true of
drug addiction.
CRITICAL CONDITIONS FOR CHANGE IN THE ADDICTIVE BEHAVIORS 93
CHANGE OUTSIDE FORMAL TREATMENT
This uncomfortable fact has already been referred to by Jim Pro-

chaska in his chapter in this volume. Among the most interesting dem-
onstrations are the studies by Saunders and Kershaw (1979) of changes
by problem drinkers in Clydeside, Scotland, and Tuchfeld's (1981) sim-
ilar study in Texas. Saunders and Kershaw interviewed 60 people who
had once had drinking problems but no longer did. Nineteen were
"definitely alcoholics." The other 41 "problem drinkers" mentioned get-
ting married (most likely for those who had been under 30 at the time),
changing jobs, and having a physical illness (more common for those
over 30 at the time) as the three most common reasons for change.
Family advice, finances, general practitioner advice (three cases), and
ageing were other reasons stated by at least two respondents. This
group made no mention of specialized treatment at all. Even for the
smaller group of "definite alcoholics," getting married and changing
jobs still headed the list of stated causes of change. Treatment, including
Alcoholics Anonymous, was mentioned by seven members of this
group, but even here the majority claimed to have given up a severe
drinking problem without the aid of formal treatment. In our own study
of treatment versus advice for married problem drinkers referred to a
psychiatric clinic (Orford & Edwards, 1977), patients in the advice
group, who did just as well as those in the more intensive treatment
group, nominated changes in work and marriage as the two most com-
mon reasons for change.
Tuchfeld (1981) advertised for people who had given up drinking
problems without treatment. Many of the 51 people he subsequently
interviewed were resistant to being labeled alcoholics and were adamant
that they had helped themselves without the aid of others. For example:
The one thing I could never do is go into formal rehab.; For me to have to ask
somebody else to help with a self-made problem, I would rather drink myself
to death. (p. 631)
Like Premack (1970), who pad analyzed reasons people give for
giving up smoking, Tuchfeld found "humiliating events" to be the most
frequently cited reason for change. These included a pregnant woman
feeling her baby quiver and concluding she could be harming her un-
born child by drinking; a man who stopped drinking when his father
died, having concluded that his own drinking was one of the causes of
his father's death; and another lying in the hospital and coming to the
realization that drinking could be the main cause of his health and other
problems. Negative role models, such as Skid Row drinkers who
shocked people into considering change, and family members, particu-
larly when they were seen to have provided persistent support, were
94 JIM ORFORD
also influential. So was religion, although most cases showed a gradual

increase in commitment, rather than a classic and sudden religious
conversion.
Tuchfeld concluded that the change process was essentially social in
nature, but was rarely if ever "spontaneous" in the sense of developing
without apparent external influence. Schachter (1982) studied changes
in smoking and eating behavior without treatment and Robins, Davis,
and Wish's (1977) description of the apparently "spontaneous" remis-
sion of drug taking on the part of Vietnam veterans on returning home
to the United States is well-known. A number of years ago Winick (1962)
spoke of the "maturing out" of drug addiction and Drew (1968) wrote of
alcoholism as a "self-limiting disease." Both observed that the preva-
lence of addictive behaviors was less at certain ages than at others (fewer
drug addicts in their thirties than in their twenties, and fewer alcoholics
in their fifties and sixties than in their thirties and forties) and that these
differences were greater than could be accounted for by formal treat-
ment or mortality.
NONSPECIALIST PROFESSIONALS AS EFFECTIVE AGENTS OF CHANGE
It is some years since Chafetz, Blane, and Hill (1970) demonstrated

that physicians often failed to detect drinking problems, and that they
were particularly likely to do this if the patient appeared to be of higher
status (well dressed, employed, etc.) or if a straightforward physical
diagnosis was available. More recently in London, Cartwright and his
colleagues (Cartwright, 1980; Shaw, Spratley, Cartwright, & Harwin,
1978) have demonstrated that the therapeutic commitment to working
with problem drinkers is often low among such groups as general practi-
tioners, probation officers, and social workers. Two factors appear to be
necessary, from their studies, in order to overcome these negative at-
titudes. The first was practical experience of treating people with drink-
ing problems; it is so often the experience of helping agents that one or
two such cases come their way during training or shortly afterwards,
and that the outcome is apparently unsuccessful. As a result, such prob-
lems are avoided for the rest of the person's career. The second neces-
sary ingredient for encouraging positive attitudes among nonspecialists
was the availability of support from supervisors, colleagues, or others in
responding to drinking problems.
Some results from a study recently completed in Exeter illustrate the
positive role that such agents can play in the change process when their
attitudes are encouraging. The study was principally concerned with
abstinence versus controlled drinking as alternative outcomes (Orford &
Keddie, in press-a,b), but the one-year follow-up included questions

about the perceived helpfulness of agencies other than specialists.
The most frequently mentioned alternative help source was the
general practitioner (GP) (21 out of 34 clients). Although unfavorable
comments outnumbered favorable, six clients had only positive things
to say about their GP and a further six were mixed in their comments.
GPs appear to have been particularly helpful for those clients who had
finally adopted abstinence. The nature of the positive and negative com-
ments about GPs provide a fairly clear picture. The largest category of
positive comments suggested that the GP was available, particularly at
crucial times, and was able to address the issue of excessive drinking.
For example, "I have seen him once a week throughout the year, just
because he's there, he doesn't lose faith, he's always the same"; "I've
seen him five or six times, once at a crucial period, we always discuss
drink ... he talks commonsense"; "I saw myoid GP, a friend, he gave
me support"; "Saw him many times, often mention drink"; "I see him
regularly ... he tries to get me into hospital"; "I damaged my
ribs ... he said take the opportunity not to drink, excellent advice, he's
repeated it since"; "He gave me strong advice to abstain, he described it
as a chemical reaction like diabetes, told me I shouldn't think of drink-
ing." Other positive comments about GPs concerned the appropriate
and useful prescribing of medication. Two clients felt that feedback of
results of blood tests had been important, and one was grateful to her
GP for arranging admission at a crucial time.
Of the negative comments about GPs, three concerned medication.
The large majority of negative comments, however, referred not to spe-
cific procedures or methods, but rather to the absence of a positive
relationship with the GP for one reason or another. Some felt that their
GPs were too busy to do other than be concerned with aches and pains,
blood pressure, or the giving of prescriptions. Others felt their GPs were
unsympathetic, in one case stating that a year was not a long time to
have been virtually abstinent, and in another giving a patient what the
latter perceived as a "dressing down." One client felt that his GP was
inconsistent, on one occasion recommending reduced drinking and on
another abstinence, and another client complained that he saw a differ-
ent doctor every time he went to the Surgery. Several others felt that
their GPs were of little use, either because they could do nothing other
than prescribe or arrange hospital admission, or alternatively because
the client admitted she had been uncooperative.
Other agencies, such as social workers, were mentioned less fre-
quently, but our conclusions were the same. Hence a coherent picture of
the treatment process emerges. Although many wiIl understandably
96 JIM ORFORD
question the generalizability of this picture, it is one in which successful

treatment agents build upon their clients' beliefs and preferences by
offering a relationship that may be brief or prolonged, but that offers
understanding rather than condemnation, more often than not supports
a client's own goal, and helps bring objectivity to clients' perceptions of
the place of excessive drinking in their lives.
To anticipate a point I will make again later, it should be pointed out
that these conclusions are little different from those reached by Thomas
Trotter (1804) nearly two hundred years ago. His prescriptions were
mainly nonspecific and left little room for the paraphernalia of specific ex-
pert treatments. "The relationship is the essential tool," Trotter advised:
Within the setting of that relationship confrontation can be used-particular
opportunities are therefore to be taken to hold a mirror as it were, that he
may see the deformity of his conduct and represent the incurable maladies
which flow from perseverance in the course of intemperence. . .. confron-
tations should be joined with offer of hope-at the conclusion of every visit,
something consummatory must be left for amusement, and as food for his
recollection.
THE UNIFORM RESULTS OF SPECIALIST TREATMENTS FOR ADDICTIVE

BEHAVIORS
Goodwin (1971) attributed O'Neill's change to brief psychoanalysis,

although to my knowledge this particular form of treatment is not one
that has been highly commended for the treatment of alcohol problems
or other addictive behaviors. This illustrates the now familiar fact that
most forms of treatment for the addictive behaviors meet with some
success irrespective of the particular techniques employed. The conclu-
sion to be drawn from a great deal of research that has been carried out
on this subject is not that treatment does not work, but rather that when
it does work, it does not do so for the reasons supposed.
Bernstein's (1970) study of "unaided quitting" by smokers, and our
study of advice versus treatment for problem drinkers (Orford & Ed-
wards, 1977) both found that advice to quit (smoking and drinking re-
spectively) was as effective as more intensive treatment. It is important
to be clear, however, that our "advice" was given by a high-status
medical consultant, in the presence of the problem drinker's wife (all the
problem drinkers were men in our study) and two other members of
staff, at the end of a 4-hour-Iong period of assessment, during which
husband and wife had seen a number of members of the team separately
and jointly. These sessions took place in a high-prestige psychiatric
hospital to which the couple had been invited after referral from their
general practitioners. During the advice giving itself, the couple was
clearly told that the problem was one of drinking, that it could be over-
come if the client gave up drinking altogether-this was before the days
of flexible drinking goals-that no treatment as such was available to
help a person achieve this, and that the solution was clearly in their
hands. In other words, we gave each couple a lot of time and attention,
and there were a whole host of nonspecific factors encouraging commit-
ment to change. Bernstein, in his smoking study, also sent smokers
away with the clear message that the responsibility for stopping smok-
ing was theirs and that no further treatment would be forthcoming.
Other studies have even found countertheoretical treatments to be
as effective as theoretically appropriate ones. For example, Russell,
Armstrong, and Patel (1976) found noncontingent aversion theory (aver-
sive stimuli being delivered at the wrong time according to the learning
theory principles on which the effectiveness of aversion therapy is sup-
posed to be based) to be as effective as aversion therapy carried out in
the proper fashion. Similarly Ley, Bradshaw, Kincey, Couper-Smartt,
and Wilson (1974) found a "willpower" control group (overeaters were
advised to go into supermarkets when hungry, to leave tempting foods
around at home, etc.) to be as effective as theoretically appropriate ad-
vice in the context of a behavioral self-control treatment program. The
important point is that, although the specific procedures may have been
theoretically wrong in some of the treatment groups in these studies, all
treatments were similar in terms of nonspecific factors enhancing com-
mitment to change. DiClemente and Prochaska (1982) were surprised to
find that even clients who received aversion therapy rated "self-libera-
tion" as one of the most important factors in the change process.
Gardner (1964), on the other hand, perhaps because he was a cler-
gyman, was one who could appreciate the symbolic and self-liberating
elements in the aversion therapy he received for his drinking problem:
The great boon, for me anyway, was the feeling of freshness which accom-
panied the treatment: the body was livelier, eyes clearer, and a new alertness
took the place of the former mental lethargy. Even more than this were the
spiritual benefits. To one who had lived and thought for much of his life in
terms of sacramental symbolism, it was easy to see how the sudden, sharp
expulsion of alcohol, and its attendant poisons, from the body, could be
allied with the exorcism of that devil-desire to drink: and on each fresh
appointment with the trolley of drinks there was a decisive sense of waking
to a new life. (p. 215)
My argument, then, is that special treatments for addictive behav-
iors often work, but they rarely work for the reasons favored by our
cherished theories. The effective ingredients are more likely to lie, if my
reading of the literature is correct, in the direction of the consciousness-
raising and commitment elements. This leads me on to my final, and
perhaps the least palatable to modern professionals, fact.
98 JIM ORFORD
MORAL OR SPIRITUAL ELEMENTS OF CHANGE
There are a number of reasons for stating that change often contains
a large moral or spiritual element. Most obvious is the huge success of
Alcoholics Anonymous. Although A.A. publishes no figures about out-
come success rates that would match up to our high scientific standards,
its success as a confident, widespread, and ever-growing self-help
group is in no doubt. David Robinson's chapter in this volume makes
this fact quite clear. Furthermore, there can be little doubting its spiritual
aspects: God or a Higher Power is mentioned in no fewer than 6 of the
12 Steps. Glaser (1973) has traced A.A. and the origins of the drug-free
therapeutic communities, such as Daytop Village, Synanon, and Phoe-
nix House, to the Oxford Group Movement, a worldwide and still-
functioning organization, originally known as the First Century Chris-
tian Fellowship and later as Moral Re-armament. It was from this source
that A.A. received its ideas of self-examination, acknowledgment of
character defects, restitution for harm done to others, and working with
others. Among the key practices of the Oxford Group Movement was
"sharing," by which was meant the open confession of sins at large
public meetings or smaller "house parties."
Like many others, I have been intrigued by 19th-century attempts to
bring about change in excessive drinkers, and have wondered whether
the processes at work were the same as those that operate, under very
different circumstances, in our modern treatments. It may be impossible
to know now whether old-fashioned pledge taking was more or less
successful than our favored treatments, but it does appear that the for-
mer was sometimes on a rather larger scale. McPeek (1972), in his histo-
ry of the American temperance movement, tells us that claims were
being made that as many as a third of a million people had signed
abstinence pledges in 3 years following the foundation of the American
Temperence Union. The Washington Temperence Society claimed be-
tween 150 and 250 thousand pledged members in the few years of its
existence in the 1840s. Both McPeek (1972) and Longmate (1968), in his
history of the British temperance organizations, describe the influence of
Father Mathew of Cork, one of the most famous of all temperance re-
formers. By all accounts, this one man had such a sizable influence on
the volume of consumption of alcohol in Ireland that the drinks trade
did their best to break up his meetings when he attempted to carry his
message across the water, first to England and then the United States.
From the few sources available to him, McPeek estimated the follow-up
success rate after the Washingtonians and Father Mathew's campaigns
at around 25% to 30%. The similarity between these estimates of the
numbers who managed sustained change in drinking habits following
exhortation and pledge taking within a religious context and the per-
centages of success estimated by observers of the modern treatment
scene is striking.
Although the changed values of modern society would not allow us
to return wholeheartedly to the moral persuasion techniques of the last
century, I have argued elsewhere (Orford, 1985) that modern treat-
ments, when examined closely, are much more of a subtle blend of
directive and nondirective, the spiritual and the scientific, than we
would like to think. This is true of Alcoholics Anonymous (Tiebout,
1961), small therapeutic halfway houses (Otto & Orford, 1978), and
psychiatric clinics (Davies, 1979). Tiebout wrote of the need for "humili-
ty" and "surrender" to the A.A. program; the halfway house staff
whom we studied spoke of "putting in a lot," being "part of the house,"
being "realistic" and "thoughtful," and showing "an appropriate at-
titude"; and Davies' doctors spoke of "sincere determination to do
something," "realising that she must stop drinking," "a sensible and
constructive attitude," "motivation," and "insight." Self-control thera-
pists and others may argue that they have purged their treatments of all
such nonsense, but I doubt it.
TOWARD A MODEL OF CHANGE
Whatever our new understanding of the change process, it must,

therefore, account for dramatic but unpredictable changes in addictive
behavior, for the fact that most change occurs outside treatment, and
that those changes that occur in a treatment context often occur in the
nonspecialist setting and for nonspecific reasons, and that people have
been changing addictive behaviors for centuries and often in circum-
stances that seem to us very alien from our present perspectives. Models
of change must be able to account for Father Mathew converts, Synanon
graduates, A.A. successes, unaided quitters of smoking and reducers of
eating, Clydesiders and Texans who give up excessive drinking when
they get married, returning war veterans who give up their drugs, and
others who simply grow older and wiser.
Where should we look to gather the necessary ideas in order to
build a new model of change in the addictive behaviors? I have thought
for some time that we could profitably look outside the clinical sphere in
the direction of general social psychology (Orford, 1971, 1978b, 1985),
and in particular toward the work of Janis and his colleagues on the
processes of decision making (e.g., Janis & Mann, 1968, 1977). I am
delighted to see, from Claus Appel's chapter and others, that the impor-
tance of this work on how decisions are made is now being recognized.
100 JIM ORFORD
Janis and Mann's (1977) book, Decision Making: A Psychological Analysis of

Conflict, Choice, and Commitment, is concerned with all manner of impor-
tant life decisions, including making changes in addictive or potentially
addictive behaviors, such as smoking and eating, other decisions about
health, choices about jobs, housing, and marriage and divorce, what to
do when given warnings of impending natural and other disasters, as
well as decisions of a political and executive nature. Janis and Mann
would have appreciated Marcus Grant's assumption, in his chapter in
this volume, that governments go through a similar process in respond-
ing to national drinking problems as do individuals faced with a decision
about their own addictive behavior.
This breadth of thinking has great appeal. It immediately makes
available to us bodies of theory and knowledge about human experience
and action that would be closed to us if we confined attention to the
traditional clinical areas of alcoholism and drug addiction studies. I have
not the space here to do justice to the many insights that Janis and Mann
offer, and I would recommend that people read their book carefully.
There are, however, a number of features of their approach that I find
particularly applicable.
FEATURES OF JANIS AND MANN'S MODEL OF DECISION MAKING
One of the most attractive aspects, for our purposes, is the central
place that they give in their model to the ideas of loss and conflict. Many
decisions, particularly those concerning health, involve loss according to
Janis and Mann. The greater the loss involved, the greater the conflict
about taking health-promoting or illness-avoiding decisions. In the case
of addictive behaviors, the inclination to reduce or abstain from behavior
would be opposed by the positive incentives for carrying on with behav-
ior as before. This leads directly to the formulation of such conflicts in
terms of a payoff matrix or balance sheet of "pros" and "cons" for
different courses of action. Janis and Mann's model of general decision
making contains, then, the idea of dilemma or conflict, a vital ingredient
for an understanding of change in addictive behavior and one that has
been missing from disease and other previous models.
A second, and very significant, attraction of viewing addictive be-
havior change in terms of decision making is the light this throws on
what Prochaska calls the pre contemplation stage. In Janis and Mann's
terms, defensive avoidance of making a decision is particularly likely to
occur when such decisions are highly "ego-involving." Under the head-
ing of defensive avoidance, they list a number of tactics, including selec-
tive inattention to relevant informal or mass media communications,
distracting the self, buck-passing, bolstering by oversimplifying, distort-
ing, evading, omitting major considerations, exaggerating favorable

consequences, minimizing unfavorable consequences, exaggerating the
remoteness of any action required, as well as recourse to alcohol or
drugs! Previous models of addictive behavior change have no way of
handling such behavior, other than attributing it to " poor motivation"
or, what is worse, to "personality disorder." With the help of decision-
making theorists we now have a way of understanding such behaviors
in terms of the highly ego-involving personal conflict surrounding the
need to change addictive behavior.
Janis and Mann also speak of five stages of the decision making
process, although their correspondence with the stages considered in
this volume is far from exact. Their first two stages, reappraisal and
considering options, may correspond to contemplation, and their stages
three and four, selecting one option and acting upon this choice, to the
action stage. Their fifth, consolidation, looks like maintenance. They
acknowledge that reappraisal may take place over a long period of
time-the "slow burn" type of chronic reappraisal as they call it-and
that there may be many reversions to earlier stages of the process. Once
again, the idea of reversion to an earlier stage of the decision-making
process carries with it a greater understanding of what is happening
than does the term relapse.
Particularly valuable is the em'phasis that Janis and Mann place on
recommendations for new behavior. Once again, the term recommenda-
tion immediately broadens our appreciation of the influences that may
be brought to bear on someone embarking on addictive behavior
change. Recommendations may come from family, friends, or the mass
media. There is certainly no suggestion that the source of an influential
recommendation need be an expert or a professional, although it may
come from a prestigeful source, such as a government report on smok-
ing and lung cancer. What an effective source or recommendation for
addictive behavior change does require, perhaps, is some basis for social
power or influence over the person to whom the recommendation is
directed. Table 1 lists the types of social power outlined by French and
Raven (1959), with a rudimentary attempt to suggest the types of people
who may be in a position to exercise these forms of power for addictive
behavior change.
Finally, in this very partial list of attractive features of Janis and
Mann's decision making model, is their idea that vigilance is required in
order to make good, and stable, life decisions. This notion serves to
bring together a number of ideas discussed in this volume about the
personal actions required to consolidate addictive behavior change, as
well as similar ideas that have existed for decades or even centuries past.
Their central idea is that all factors for and against a particular course of
102 JIM ORFORD
TABLE 1.
Types of Social Influence and their Possible Uses for Addictive
Behavior Change
Reward power
Based on the perception that this person has the ability to give or withhold
rewards, e.g., partner, employer.
Coercive power
Based on the perception that this person has the ability to give or withhold
punishments, e.g., partner, the law.
Referent power
Based on identification, e.g., close friend, admired other, someone who has
previously made a similar change.
Expert power
Based on the perception that this person has some special knowledge, e.g., general
medical practitioner, addiction counselor, some mass media presentations ..
Legitimate power
Based on internalised norms and values that dictate acceptance of influence from
this person, e.g., a parent.
Note. Based on French and Raven's (1959) typology of the bases of social power.
action, including factors impinging on the self and those impinging on

others, those that involve material considerations and those that are
more subjective, should be laid out and carefully considered and recon-
sidered before corning to a decision. Anything, including defensive
avoidance, that detracts from this process will reduce the quality of the
decision-making process. In particular, they recommend the drawing up
of a balance sheet with pros and cons for different courses of action. To
make the point that there is nothing new about this procedure, they
provide the following quotation from Benjamin Franklin, who wrote to
the scientist Joseph Priestley in 1772 in the following terms:
When those difficult cases occur, they are difficult, chiefly because while we
have them under consideration, all the reasons pro and con are not present
to the mind at the same time; but sometimes one set present themselves, and
at other times another, the first being out of sight. Hence the various pur-
poses or inclinations that alternatively prevail, and the uncertainty that per-
plexes us. To get over this, my way is to divide half a sheet of paper by a line
into two columns; writing over the one Pro, and over the other Con. Then,
during three or four days consideration, I put down under the different
heads short hints of the different motives, that at different times occur to me,
for or against the measure. When I have thus got them all together in one
view, I endeavor to estimate their respective weights; and where I find two,
one on each side, that seem equal, I strike them both out. ... and thus
proceeding I find at length where the balance lies; and if, after a day or two of
further consideration, nothing new that is of importance occurs on either
side, I come to a determination accordingly. And, though the weight of
reasons cannot be taken with the precision of algebraic quantities, yet when
each is thus considered, separately and comparatively, and the whole lies
before me, I think I can judge better, and am less liable to make a rash step,
and in fact I have found great advantage from this kind of equation, in what
may be called moral or prudential algebra. (Janis & Mann, 1977, p. 149)
William James (1891), writing about habits in his Principles of Psychol-

ogy, already knew a lot about the need for action if permanent changes
were to be made in habitual behavior. He quoted with approval from
Professor Bain who wrote in his Moral Habits:
In the acquisition of a new habit, or the leaving off of an old one, we must
take care to launch ourselves with as strong and decided initiative as possible.
Accumulate all the possible circumstances which shall re-enforce the right
motives; put yourself assiduously in conditions that encourage the new way;
make engagements incompatible with the old; take a public pledge, if the
case allows; in short, envelope your resolution with every aid you know ....
Never suffer an exception to occur till the new habit is securely rooted in your
life . ... Keep the faculty of effort alive in you by a little gratuitous exercise every day.
That is, be systematically ascetic or heroic in little unnecessary points, do
every day or two something for no other reason than that you would rather
not do it, so that when the hour of dire need draws nigh, it may find you not
unnerved and untrained to stand the test. (James, 1981, pp. 122-126)
The need for action to consolidate decision is well known in Alco-

holics Anonymous. For example:
Don't for a split second allow yourself to think: "Isn't it a pity or a mean
injustice that I can't take a drink like so-called normal people" .... Don't
allow yourself to either think or talk about any real or imagined pleasure you
once did get from drinking.... Don't permit yourself to think a drink or two
would make some bad situation better, or at least easier to live with. Sub-
stitute the thought: "one drink will make it worse-one drink will mean
drunk". ... . Catalogue and re-catalogue the positive enjoyments of so-
briety .... Cultivate a helpful association of ideas: Associate a drink as being
the single cause of all the misery, shame and mortification you have ever
known." (A.A., undated)
Sjoberg and his colleagues in Sweden (Samsonowitz & Sjoberg,

1981; Sjoberg & Johnson, 1978) are almost alone among modern re-
searchers in daring to examine the notion of willpower in addictive
behavior change. They write of the need for "high-quality information
processing," and the occurrence of volitional breakdowns accompanied
by "low-quality information processing" and "twisted-reasoning." In a
study of a small number of excessive drinkers, Samsonowitz and
Sjoberg found an inverse correlation between the number of relapses
and the number of techniques used, such as bringing to mind the
positive consequences resulting from maintaining a decision, or the
negative consequences of going back on the decision, conscious plan-
104 JIM ORFORD
ning and preparation prior to the decision, avoiding difficult and tempt-
ing situations, and performing alternative activities.
THREE THINGS TO EXPECT OF A THEORY OF CHANGE IN ADDICTIVE

BEHAVIORS
As well as having a number of attractive features, including those

previously outlined, an understanding of change in terms of decision
making begins to do three general things that we should expect of a
theory of addictive behavior change, and which a disease model does
not do.
1. It should unite different addictive behaviors on an equal footing.
Disease models were always more comfortable with hard-drug addiction
and severe drinking problems. It was clear that they were stretched to
the breaking point when considering the full range of alcohol-related
problems, the full range of forms of drug misuse, and certain forms of
eating disorder. They were never serious contenders for embracing ex-
cessive gambling, most forms of excessive eating, excessive sexuality,
and even tobacco smoking. On the other hand, all can be embraced
within a model of decisional conflict.
2. Ideally, our theory should unite early and late choices or deci-
sions. We can now begin to talk in the same terms about early decisions
to take up a new form of potentially addictive behavior and late deci-
sions to give up a form of addictive behavior that has become trou-
blesome. In terms of practice, this is probably the single most important
aspect of the shift from disease to psychological models of addictive
behavior.
3. Our theory should unite the clinical and the social-epi-
demiological fields. It was always unhelpful to use one language to
describe a clinical change process (motivation to enter treatment, thera-
pist, patient, relapse, etc.) and another to describe the more numerous
changes that occur elsewhere (decision, spontaneous, unaided, etc.). A
theory of conflict, decision, and action, with the help of influence based
on social power of one kind or another, meets this criterion for a satisfac-
tory theory much more adequately.
THREE POINTS THAT MAY CAUSE TROUBLE
There are three aspects of change in addictive behaviors that, in my

view, receive less than adequate attention in this volume. The first of
these is probably easily accommodated in a decision-making model of
change, but the others may cause difficulties.
1. We have given little attention to the possibility that change is
most likely to occur at times of crisis. Is it the case that change is particu-
larly likely to be initiated at one of a limited number of occasional choice
points in a career of addictive behavior? There is a saying that problem
drinkers seek change only because of livers, lovers, livelihood, or the
law. Are changes confined to those times when one of these factors
plays up, when some humiliating event occurs, or perhaps when a
person enters a new role position (e.g., as father, mother, manager,
widow)?
2. The moral or spiritual aspects of the change process, discussed
above, are largely missing. Whether an understanding of change in
terms of a specific decision or action about an addictive behavior can do
justice to changes involving widespread modifications of attitudes and
values remains to be seen.
3. To those not brought up within a behavioral tradition, and not
familiar with a health-education approach, our deliberations must ap-
pear very addiction focused. The assumption has been that change oc-
curs because people contemplate or appraise their position regarding
the addictive behavior, that they consider the pros and cons regarding
it, that they monitor their behavior carefully, that they take action about
it. We have given little attention to the possibility that some change may
occur because the addictive behavior loses its meaning or its functional
significance, possibly without any direct contemplation or action, or
even without the person being aware that the addictive behavior was
changing at all. This is presumably one of the ways in which "maturing
out" works: as a person ages the formerly addictive behavior simply
ceases to perform the functions that the older person values.
IMPLICATIONS FOR PRACTICE
If we are going to abandon a disease model and replace it with one

based on human choice or decision making, and if this model, as I have
suggested, links a variety of addictive behaviors, equates early and late
choices, and views the therapeutic setting as simply a special case of
settings for change, then our practice will surely need to alter also. For
many of us this may present problems. Our training and our subsequent
careers have been spent in medical, or quasi-medical, institutions. Our
disdain for disease models notwithstanding, our skills lie in treating
people within a medical-like format of therapist and client.
As a teacher of clinical psychology, this concerns me not a little. My
perception is that graduates are increasingly applying for such training
in order to be therapists rather than to apply psychology in the most
effective way, whatever that turns out to be. On the other hand, I am
106 JIM ORFORD
much encouraged by the community psychology movement, which is

much better advanced in the United States than in the United Kingdom,
and in the development of community forms of alcohol service delivery
which I suspect are better developed in the United Kingdom where the
National Health Service has given strong backing (sceptics would say on
grounds of cost) to the development of community alcohol teams. My
experience of working in one such team in Exeter is that the techniques
and skills required are much broader than those implied by the term
therapist. The team has needed to develop knowledge of how to select
people to train as volunteers, to teach a multidisciplinary audience using
a range of visual aids, to provide a consultancy service to health, social
services, prison and probation personnel, to evaluate, to make decisions
in a team, and to use the local media to best advantage. Appropriately,
the team's institutional affiliation is a hybrid one. Its headquarters is a
health service building, part of which is leased by the voluntary Council
on Alcoholism, which is an integral part of the whole team. In addition,
there are members of the team from social services, probation, and
health education.
We have witnessed a revolution in thinking about alcohol problems
and other addictive behaviors in recent years (Sobell & Sobell, 1984). The
change has been rapid. Are our institutions for training and practice so
flexible?
REFERENCES
Alcoholics Anonymous. (Undated). 15 Points for an alcoholic to consider when confronted with
the urge to take a drink. London: Alcoholics Anonymous.
Bernstein, D. (1970). The modification of smoking behavior: An evaluation review. In W.
Hunt (Ed.), Learning mechanisms in smoking (pp. 3-41). Chicago, IL: Aldine.
Cartwright, A. (1980). The attitudes of helping agents towards the alcoholic client. British
Journal of Addiction, 75, 413-431.
Chafetz, M., Blane, H., & Hill, M. (Eds.). (1970). Frontiers of alcoholism. New York: Science
House.
Davies, P. (1979). Motivation, responsibility and sickness in the psychiatric treatment of
alcoholism. British Journal of Psychiatry, 134, 449-458.
DiClemente, c., & Prochaska, J. (1982). Self-change and therapy change of smoking be-
havior: A comparison of processes of change in cessation and maintenance. Addictive
Behaviors, 7, 133-142.
Drew, L. (1968). Alcoholism as a self-limiting disease. Quarterly Journal of Studies on Alcohol,
29, 956-966.
French, J. R., & Raven, B. H. (1959). The bases of social power. In D. Cartwright (Ed.),
Studies in Social Power. Ann Arbor, MI: University of Michigan Press.
Gardner, J. (1964). Spin the bottle: The autobiography of an alcoholic. London: Muller.
Glaser, F. (1973, May). Some historical aspects of the drug-free therapeutic community. Un-
published paper presented at a conference on Intervening in Drug Misuse: The

Therapeutic Community and other Self-Help Efforts, Saddlebrook, NJ.
Goodwin, D. (1970). The alcoholism of F. Scott Fitzgerald. Journal of American Medical
Association, 212, 86-90.
Goodwin, D. (1971). The alcoholism of Eugene O'Neill. Journal of American Medical Associa-
tion, 216, 99-104.
James, W. (1891). The principles of psychology. Vol. 1. London: MacMillan.
Janis, I., & Mann, L. (1968). A conflict theory approach to attitude change and decision
making. In A. Greenwald, T. Brock, & T. Ostrom (Eds.), Psychological foundations of
attitudes. New York: Academic Press.
Janis, I., & Mann, L. (1977). Decision-making: A psychological analysis of conflict, choice and
commitment. New York: Free Press.
Ley, P., Bradshaw, P., Kincey, J., Couper-Smartt, H., & Wilson, M. (1974). Psychological
variables in the control of obesity. In W. Burland, P. Samuel, & J. Yudkin (Eds.),
Obesity (pp. 316-337). London: Churchill Livingstone.
Longmate, N. (1968). The water drinkers: A history of temperance. London: Hamish Hamilton.
McPeek, F. (1972). The role of religiOUS bodies in the treatment of inebriety in the United
States. Alcohol, science and society: 29 lectures with discussions as given at the Yale Summer
School of Alcohol Studies. Westport, CT: Greenwood Press.
Orford, J. (1971). Aspects of the relationship between alcohol and drug abuse. In Proceed-
ings of the 29th International Congress on Alcoholism and Drug Dependence, Sydney, Aus-
tralia (pp. 130-144). Butterworth, Melbourne, Australia. Reprinted in: G. Edwards, M.
A. H. Russell, D. Hawks, & M. MacCafferty (Eds.), Drugs and Drug Dependence, 19,
199-213.
Orford, J. (1978a). Hypersexuality: Implications for a theory of dependence. British Journal
of Addiction, 73, 299-310.
Orford, J. (1978b). Alcoholism: what psychology offers. In G. Edwards & M. Grant (Eds.),
Alcoholism: New knowledge and new responses (pp. 88-99). London: Croom Helm.
Orford, J. (1985). Excessive appetites: A psychological view of addiction. Chichester, England:
Wiley.
Orford, J., & Edwards, G. (1977). Alcoholism: A comparison of treatment and advice, with a
study of the influence of marriage. Oxford: Oxford University Press.
Orford, J., & Keddie, A. (in press-a). Abstinence or controlled drinking in clinical practice:
Indications at initial assessment. Addictive Behaviors.
Orford, J., & Keddie, A. (in press-b). Abstinence or controlled drinking in clinical practice:
A test of the dependence and persuasion hypotheses. British Journal of Addiction.
Otto, S., & Orford, J. (1978). Not quite like home: Small hostels for alcoholics and others.
Chichester, England: Wiley.
Premack, D. (1970). Mechanisms of self-control. In W. Hunt (Ed.), Learning mechanisms in
smoking (pp. 107-123). Chicago, IL: Aldine.
Robins, L., Davis, D., & Wish, E. (1977). Detecting predictors of rare events: Demograph-
ic, family and personal deviance as predictors of stages in the progression toward
narcotic addiction. In S. Strauss, H. Babigian, & M. Roff (Eds.), The origins and course of
psychopathology: Methods of longitudinal research. New York: Plenum Press.
Russell, M., Amstrong, E., & Patel, U. (1976). Temporal contiguity in electric aversion
therapy for cigarette smoking. Behaviour Research and Therapy, 14, 103-123.
Samsonowitz, V., & Sjoberg, L. (1981). Volitional problems of socially adjusted alcoholics.
Saunders, W., & Kershaw, P. (1979). Spontaneous remission from alcoholism-A commu-
nity study. British Journal of Addiction, 74, 251-266.
108 JIM ORFORD
Schachter, S. (1982). Recidivism and self-cure of smoking and obesity. American Psychol-
ogist, 37, 436-444.
Shaw, S., Cartwright, A., Spratley, T., & Harwin, J. (1978). Responding to drinking problems.
London: Croom Helm.
Sjoberg, L., & Johnson, T. (1978). Trying to give up smoking: A study of volitional break-
downs. Addictive Behaviors, 4, 339-359.
Sobell, M. B., & Sobell, L. C. (1984). The aftermath of heresy: A response to Pendery et aI's
(1982) Critique of 'Individualized behavior therapy for alcoholics', Behavior Research
and Therapy, 22, 413-440.
Stone, L. (1979). The family, sex and marriage in England 1500-1800. Harmondsworth, Mid-
dlesex, England: Penguin. (Original unabridged version published by Weidenfeld &
Nicolson, 1977)
Thorley, A., Oppenheimer, E., & Stimson, G. (1977). Clinic attendance and opiate pre-
scription status of heroin addicts over a six-year period. British Journal of Psychiatry,
130, 565-569.
Tiebout, H. (1961). Alcoholics Anonymous-An experiment of nature. Quarterly Journal of
Studies on Alcohol, 22, 52-68.
Trotter, T. (1804). An essay, medical, philosophical and chemical, on drunkenness, and its effects on
the human body. London: Longman.
Tuchfeld, B. (1981). Spontaneous remission in alcoholics: Empirical observations and the-
oretical implications. Journal of Studies on Alcohol, 42, 626-641.
Winick, C. (1962). Maturing out of narcotic addiction. Bulletin of Narcotics, 14, 1.
Wodack, A. D., Saunders, J. B., Ewusi-Mensah, I., Davis, M., & Williams, R. (1983).
Severity of alcohol dependence in patients with alcoholic liver disease. British Medical
Journal, 287, 1420-1422.
6
Trying to Stop Smoking
A Decision-Making Perspective
STEPHEN SUTTON
The theoretical perspective that guides this chapter is that processes of

change in the domain of addictive behaviors can be understood in terms
of individuals' decisions based on evaluating the possible outcomes of the
courses of action available to them. The key decision is seen as one of
whether or not to try to change one's behavior (e.g., to try to stop
smoking, to try to reduce one's consumption of alcohol, to try to lose
weight). Once the person has embarked on such an attempt, he or she
will be faced repeatedly with another decision, namely whether to per-
severe with the attempt, often in spite of unpleasant withdrawal symp-
toms, or whether to abandon it. It is an unfortunate fact that the decision
to try to change can always be deferred and, when acted on, can be
revoked at any time.
Based on this perspective, this chapter presents a formal theoretical
model for explaining smokers' decisions to try to stop smoking. The
model is based on subjective expected utility (SEU) theory (Edwards,
1954). Expectancy-value models have been widely employed in psychol-
ogy, but there have been surprisingly few applications in the field of
addictive behaviors. The model presented here takes into account the
fact that an attempt to give up smoking mayor may not be successful
STEPHEN SUTTON • Addiction Research Unit, Institute of Psychiatry, 101 Denmark

Hill, London SE5 8AF, England. I would like to thank the Medical Research Council,
London, England, for their financial support.
109
110 STEPHEN SUITON
and that the consequences for the individual are conditional on the
success or failure of the attempt. After describing the model, I will
present data from three studies that were designed to evaluate this or
related models.
THE MODEL
The model used in this chapter is derived from subjective expected

utility or SEU theory. The subjective expected utility for a course of
action is the sum of the subjective values or utilities attached to the possi-
ble outcomes of that action, each weighted by the subjective probabilities
that the action will lead to those outcomes. According to the model, a
person faced with two or more alternative courses of action will choose
the one with the greatest subjective expected utility. Put simply, the
person will choose the course of action that they think, on balance, will
bring them more of what they want and less of what they dislike or fear.
It should be noted that the SEU model is an "as if" model. In other
words, it is not suggested that in making decisions people consciously
perform the multiplications and additions implied by the model, only
that they behave as if they do these calculations.
To see how this approach might be applied to smoking, consider a
smoker who has just watched a television program on the subject of
smoking and lung cancer. The choice facing the smoker can be seen as
one of continuing to smoke or trying to stop. Obviously other options
are available (e.g., cutting down, switching to a milder brand) but for
present purposes we are concerned only with the two options of con-
tinuing to smoke or trying to stop. Figure 1 shows the decision-tree
diagram for this situation where there is only one long-term outcome
(lung cancer). The second action alternative, trying to stop smoking,
may end in success or failure. The subjective probability of succeeding,
given that an attempt is made (Ps in the diagram), we call "confidence."
The subjective probabilities of success and failure are assumed to sum to
unity (Ps + P f = 1). There are two other subjective probabilities in the
model. Pc is the subjective probability of getting lung cancer given that
one continues to smoke (or tries and fails). Pr is the subjective proba-
bility of getting lung cancer given that one successfully stops smoking.
Finally, U is the utility of lung cancer, which will be negative.
Under the assumption that subjective probabilities and utilities are
combined multiplicatively, it is a simple matter to show that the decision
to try to stop smoking will depend on three factors:
1. The utility the person attaches to lung cancer (U)
TRYING TO STOP SMOKING 111
Pc
CONTINUE TO SMOKE-------..;~------LUNG CANCER U
Pr
S U C C E E D - - - - " ' - - - - - - L U N G CANCER U
TRY TO STOP
SMOKING
Pc
F A I L - - - - - - = - - - - - - L U N G CANCER U
FIGURE 1. Decision-tree diagram of the choice facing a smoker who is exposed to a fear-
arousing communication about lung cancer (from Sutton & Eiser, 1984).
2. The probability difference (Pc - Pc)' that is, the reduction in the
perceived risk of getting lung cancer that follows from success-
fully stopping smoking
3. The subjective probability of succeeding given that an attempt is
made (Ps)' or confidence
Thus, according to the model, a smoker will be more likely to try to stop
smoking to the extent that he or she believes: (a) that such an attempt is
very likely to end in success (high confidence); (b) that lung cancer is
extremely serious; and (c) that lung cancer will be much less likely if they
stop smoking than if they continue to smoke.
It is important to appreciate that the proposed model is entirely
compatible with the notion that smokers are physically or psychologi-
cally dependent on nicotine. For example, a smoker who is contemplat-
ing whether or not to try to stop smoking may anticipate unpleasant
withdrawal symptoms, and these may in fact be nicotine related. From
the standpoint of the model, however, it is the smoker's expectations
regarding such outcomes, regardless of their possible basis in nicotine
dependence, that will influence his or her decision.
Expectancy-value models are sometimes dismissed on the grounds
that they make the unrealistic assumption that people behave rationally.
This is true only in a limited sense, however. The model in fact permits
several kinds of what might be called irrational or suboptimal behavior.
112 STEPHEN SUTTON
For example, smokers may hold beliefs that conflict with the scientific
evidence: a two-pack-a-day inhaling smoker who has smoked for 20
years may believe that his or her chances of getting lung cancer are nil.
Again, people may fail to take into account all the relevant outcomes or
all the alternative courses of action open to them. On the other hand, the
model does assume that peoples' behavior is rational in the sense of
being future oriented and in the sense that they attempt to maximize
expected outcomes.
STUDY 1
The first study (Sutton & Eiser, 1984) was designed to investigate
the utility of this single-outcome SEU model in explaining the response
of smokers to a film about smoking and lung cancer. In an experimental
design, groups of smokers watched either a film about smoking and
lung cancer or a control film on a different health topic. The smoking
film we used was a television program called "Dying for a fag?" made
by Thames Television and originally broadcast in the United Kingdom in
1975. It consists of an extended interview with a man who is dying from
lung cancer and includes information about the health risks of smoking.
The effects of the film were assessed by means of questionnaires com-
pleted immediately and 3 months after exposure. Our subjects were 61
office workers from two companies based in London. They smoked on
average 17 cigarettes a day.
Figure 2 shows the causal model for this study. Film refers to the
between-film contrast (Le., the experimental film versus the control
film), and behavior refers to whether or not the smoker tried to stop
smoking or to cut down in the 3-month follow-up period. The other
variables in the diagram were all measured by means of single rating
scales on the questionnaire administered immediately after watching the
film. Intention is assumed to represent the person's decision at this time.
Probability difference, utility, and confidence were predicted to mediate any
effect of the film on intentions and subsequent behavior. Because fear
has traditionally been regarded as the central explanatory construct in
research on the effects of communications about health risks, we also
included a measure of the amount of fear aroused by the film.
Figure 2 shows the results of multiple regression analyses of the
data. The numbers on the paths are the standardized partial regression
coefficients, or betas, which can be interpreted as estimates of the direct
effects of one variable on another, given the causal ordering shown in
the diagram. Specifically, they estimate that portion of the observed
correlation between two variables that is due to the direct causal effect of
• p<.025
"p<.01 Behavior
FIGURE 2. The estimated path model for Study 1 showing the direct effects in terms of the
standardized partial regression coefficients (from Sutton & Eiser, 1984).
one variable on the other. Their size may be judged by reference to the
more familiar correlation coefficient (bearing in mind that, unlike cor-
relations, beta coefficients may in some circumstances exceed unity).
This approach to data analysis is useful in that it allows the total effect of
one variable on another to be decomposed into a single direct (unmedi-
ated) effect and one or more indirect (mediated) effects. Furthermore, it
enables each observed correlation to be decomposed into causal and
noncausal components. Sex, age, and cigarette consumption were con-
trolled for in the analysis, but are not shown in the diagram for the sake
of legibility.
As Figure 2 shows, there were a number of film effects. Relative to
the control film, "Dying for a fag?" aroused more fear. The direct effect
of the film on fear was .40 and the total effect (that is, the direct effect
plus the indirect effects mediated by other variables) was .45. The film
also strengthened intentions to try to stop smoking. Although the direct
effect of film on intention was only .07, the total effect was .41, signifi-
cant at the .025 level. The film also influenced the utility variable; it
apparently impressed on the subjects the importance of reducing their
chances of getting lung cancer. The largest effect of the film, however,
was on behavior. The total effect was .59, which was significant at
114 STEPHEN SUTTON
the .01 level: 86% of those who saw the smoking film tried to stop or to
cut down compared with 33% of those who saw the control film. Most of
this effect of the film on behavior was accounted for by the direct effect
of .43 (see Figure 2). Thus the five variables measured on the postex-
posure questionnaire did not mediate the effect of the film on behavior
to any great extent.
As predicted, confidence had a significant direct effect on intention.
Those smokers who thought they would be more likely to succeed if
they tried to stop smoking also expressed stronger intentions to try to
stop. The effects of probability difference and utility, however, were not
significant. Intention was also influenced by the amount of fear aroused
by the film-a direct effect of .39; those who were more frightened by
the film tended to have stronger intentions to try to stop. Finally, the
relationship between intention and behavior was, as would be expected,
positive and significant; those who expressed stronger intentions to try
to stop smoking tended to be more likely to try to stop or to cut down in
the 3-month follow-up period. Forty-six percent of the variance in inten-
tion and 57% of the variance in behavior was explained.
STUDY 2
The second study (Sutton, 1979) differed from the first in that a full
SEU model was used. Instead of focusing on one long-term outcome
(lung cancer), a number of different outcomes were assessed. It also
differed in employing a nonexperimental design.
The source of data for the study was a sample of 2,000 smokers
randomly selected from over half a million smokers who sent for a free
stop smoking kit offered by Granada Television's "Reports Action" pro-
gram. A questionnaire was included in the kits sent to these 2,000
people.
Ratings of confidence and intentions were obtained as in Study 1.
The questionnaire also listed 26 possible outcomes of continuing to
smoke or stopping smoking ("Be irritable with people," "Get lung can-
cer," "Put on weight," etc.). The subjects were asked to rate each of
these outcomes in terms of their subjective probabilities; for example, "If
you STOPPED smoking altogether would you be more likely or less
likely to be irritable with people than if you CONTINUED TO SMOKE?"
They indicated their response by ticking one of five boxes labeled from
"Much less likely" through to "Much more likely." Having rated the
subjective probabilities associated with the 26 outcomes in this way,
they then rated the utility of each outcome, which was operationalized
in terms of "importance." An SEU score was computed for each person
by summing the products of the probability and utility ratings over the
26 consequences. This score is assumed to represent the benefits of
stopping smoking relative to continuing to smoke, as perceived by the
person. A positive score means that stopping smoking is expected to
bring more benefits and fewer costs than continuing to smoke. A nega-
tive score means that continuing to smoke is expected to bring more
benefits and fewer costs than stopping smoking. A score of zero indi-
cates that the two alternatives are perceived as equally desirable/unde-
sirable.
Subjects who returned the first questionnaire were sent a short
follow-up questionnaire 3 months later to find out whether or not they
had tried to stop smoking or to cut down in the intervening period. Only
106 cigarette smokers provided adequate information on both question-
naires. The most likely reason for the poor response was that the "kit"
received by the subjects bore little resemblance to the one shown in the
program, consisting only of a broadsheet, a cardboard no-smoking sign,
a leaflet on how to stop smoking, and the questionnaire. The sample
consisted of fairly heavy smokers, smoking on average 25 cigarettes a
day: in fact, no subject smoked fewer than 10 cigarettes a day.
Figure 3 shows the results of regression analyses of the data. As in
Figure 2, intention refers to the intention to try to stop smoking and
behavior refers to whether or not the subject tried to stop smoking or to
cut down in the follow-up period. As before, the numbers on the paths
are the standardized partial regression coefficients.
Both SEU and confidence influenced intention. Those smokers who
had higher SEU scores (Le., who were more "motivated" to stop in the
specific SEU sense) tended to have stronger intentions to try to stop
BEH/<VIOR
FIGURE 3. The estimated path model for Study 2 (from Sutton, 1979).
116 STEPHEN SUTTON
smoking (p < .001). Furthermore, as in the previous study, those who

were more confident of succeeding had stronger intentions to try to stop
(p < .01). SEU and confidence correlated .35, that is, those who were
more strongly motivated to stop smoking were also more confident of
succeeding. These two variables accounted for 25% of the variance in
intention. Again, as expected, those who had stronger intentions to try
to stop were more likely to report at the 3-month follow-up that they had
tried to stop or to cut down (p < .02). Only 13% of the variance in
behavior could be accounted for.
STUDY 3
The data for Study 3 come from a survey of smokers' attitudes

conducted in the United Kingdom by the Office of Population Censuses
and Surveys (Marsh & Matheson, 1983). I am running some additional
analyses of the data in collaboration with the authors of the study. The
results reported below are preliminary findings from these analyses.
A large sample of smokers representative of the United Kingdom
population were interviewed and then completed a detailed question-
naire designed to assess, among other things, their confidence and in-
tentions, and their subjective probabilities and utilities with regard to 32
possible outcomes of continuing to smoke and stopping smoking. Six
months later they were sent a short questionnaire to find out whether or
not they had tried to stop smoking in the intervening period.
Figure 4 shows the results of regression analyses of the data on a
.21
FIGURE 4. The estimated path model for Study 3.

subsample of 966 cigarette smokers. The results show a pattern very

similar to that in Study 2. Again, those who had higher SEU scores, and
those who felt more confident, expressed stronger intentions to try to
stop; and those who expressed stronger intentions were more likely to
try to stop in the subsequent 6-month period. Thirty-five percent of the
variance in intentions and 9% of the variance in behavior was accounted
for.
DISCUSSION
This chapter has briefly reported three studies designed to evaluate

an expectancy-value model for explaining smokers' decisions to try to
stop smoking. The model is clearly inadequate as a full explanation of
smokers' decisions, whether these occur in response to a fear-arousing
film or as part of a "spontaneous" attempt to stop. This is shown by the
large unmediated effect of the film on behavior found in Study 1, and by
the relatively small amount of variance in intentions and behavior that
was explained in Studies 2 and 3. On the other hand, the model has
been shown to have heuristic value in enabling consistent and possibly
important relationships to be identified. Although sample charac-
teristics, such as average cigarette consumption, differed from study to
study, the results showed a surprising degree of consistency. For exam-
ple, a positive effect of confidence on intention was found in all three
studies, and in the two studies in which a full SEU model was em-
ployed, intention was also influenced by SEU score. These results sug-
gest that it is useful to regard the attempt to stop smoking as a decision
based in part on weighing up the pros and cons of stopping smoking
and continuing to smoke. Moreover, the results suggest that the
smoker's expectation of success or failure (confidence) also enters into
this decision. Thus a smoker who is motivated to stop smoking (in the
sense of having a high positive SEU score reflecting an expectation of net
personal benefits from stopping smoking) will be more likely to have
strong intentions to try to stop smoking if he or she also expects to
succeed in such an attempt. Confidence can be seen as a way of incorpo-
rating the addictive aspects of smoking into a social-psychological analy-
sis of smokers' decisions. From this viewpoint, what is important in
understanding smokers' decisions to stop is their expectations concern-
ing the possible outcomes of stopping smoking or trying to do so. To
what extent such expectations (for example, about withdrawal symp-
toms) have a basis in dependence is an important question that should
be addressed in future research.
118 STEPHEN SUTTON
ACKNOWLEDGMENTS
I am grateful to the members of the Smoking Section of the Addic-

tion Research Unit for their helpful comments on the first draft of this
chapter.
REFERENCES
Edwards, W. (1954). The theory of decision making. Psychological Bul/etin, 51, 380-417.
Marsh, A., & Matheson, J. (1983). Smoking attitudes and behaviour. London: Her Majesty's
Stationery Office.
Sutton, S. R. (1979). Can subjective expected utility (SEU) theory explain smokers' deci-
sions to try to stop smoking? In D. J. Oborne, Gruneberg, M. M., & Eiser, J. R. (Eds.),
Research in psychology and medicine (pp. 94-101). London: Academic Press.
Sutton, S. R., & Eiser, J. R. (1984). The effect of fear-arousing communications on cigarette
smoking: An expectancy-value approach. Journal of Behavioral Medicine, 7, 13-33.
III
Action
Aspects and Processes of Change
7
The Effectiveness of Alcoholism
Treatment
What Research Reveals
WILLIAM R. MILLER AND REID K. HESTER
INTRODUCTION
In 1979 we set out together on a journey. We decided to try to read every

study that had ever been published (in languages we could understand)
on the effectiveness· of different approaches to treating alcohol prob-
lems. We had no idea what lay in store for us.
In the course of our research we encountered four major surprises.
The first of these was the sheer volume of research. We had had no
conception of just how much research had been done. The chapter that
resulted from our search (W. R. Miller & Hester, 1980) comprised 130
printed pages and encompassed more than 600 references. It took us 6
months just to read the research.
Our second surprise was the encouraging number of alternative
treatments available for dealing with alcohol problems. We found em-
pirical information on more than 20 different treatment methods, a few
of which we had never encountered before. Some were old and familiar.
WILLIAM R. MILLER • Department of Psychology, University of New Mexico, Albuquer-

que, NM 87131. REID K. HESTER • Behavior Therapy Associates, Albuquerque, NM
87110.
121
122 WILLIAM R. MILLER AND REID K. HESTER
Others were novel applications of methods that had been used effective-
ly to treat other problems.
A third surprise was that we were pleased at how much clinically
relevant information is already available. Not only is the volume of
research large, but it is gratifyingly consistent. The results of well-con-
trolled studies in this area have seldom contradicted one another. Cer-
tain methods have a very good track record, working well across a wide
range of populations and settings. Others seem to have little therapeutic
value, and are rather consistently found to yield little impact on drinking
behavior when subjected to controlled evaluation. Although client char-
acteristics have not proved to be consistent prognostic indicators for
alcoholism treatment in general (Gibbs & Flanagan, 1977), certain meth-
ods do appear to be differentially beneficial for specific types of clients.
Encouraging gains have been made in deriving differential diagnostic
schemes that will allow the professional to help clients select an optimal
approach (Gottheil, McLellan, & Druley, 1981; d. Chapter 8, this vol-
ume).
But the fourth surprise was, for us, the most disturbing. As we
constructed a list of treatment approaches most clearly supported as
effective, based on current research, it was apparent that they all had
one thing in common as of 1979: they were very rarely used in American
treatment programs. The list of elements that are typically included in
alcoholism treatment in the United States likewise evidenced a com-
monality: virtually all of them lacked adequate scientific evidence of
effectiveness. We were shocked. The problem, it seemed, was not that
"we know not what we do," but rather that in the alcoholism field we
are not applying in treatment what is already known from research.
Since 1979 we have continued to read the emerging research on
treatment outcome and to search still further to accumulate whatever
knowledge might be available. Among the questions that have fasci-
nated us are: (a) Which treatment methods are most effective in treating
alcohol problems? (b) What types of individuals do best within each
alternative method? and (c) How is effectiveness influenced by the
length, intensity, or setting of treatment?
In the 6 years since our original review, more than 300 new treat-
ment reports have been published. A number of important controlled
investigations have appeared, and the information available on how to
select optimal interventions is still stronger than it was in 1979. Yet, as
far as we can see, this research has still had virtually no effect on treat-
ment practices in the United States, where alcoholism treatment has
become a major profit-making industry.
This chapter is an overview of the evidence on effectiveness of
alternative treatment approaches. Because of constraints on length, we
THE EFFECTIVENESS OF ALCOHOLISM TREATMENT 123
have had to impose several exclusionary criteria on what could be dis-

cussed here. First, we have restricted ourselves primarily to the evi-
dence from controlled research; that is, studies including either random
or matching assignment designs with control or comparison groups.
Given the wide variability in treatment outcome across populations, we
believe that controlled designs offer the best hope of providing reliable
and replicable results. Second, we have focused on studies evaluating
the impact of treatment on drinking behavior. Many studies have evalu-
ated treatment effects on alcoholics' mood, compliance, anxiety, insight,
or other variables, but we have emphasized here those studies including
measures of change in alcohol consumption. Third, we have discussed
only treatment interventions with problem drinking populations, and
have not included preventive interventions. Finally, we have concen-
trated on presenting the "bottom line," attempting to draw reasonable
and accurate conclusions from the data available. Space limitations con-
strain us from offering detailed methodological critiques of each study.
Our prior review (1980) offers methodological commentary on some
earlier studies.
SPECIFIC TREATMENT METHODS
We will give separate consideration to nine major classes of inter-

ventions, although we recognize that these methods frequently overlap
or are combined. First consideration will be given to four very common
elements of current treatment programs: pharmacotherapy, psycho-
therapy or counseling, Alcoholics Anonymous, and alcoholism educa-
tion. The review of specific approaches will then conclude with evalua-
tions of five less commonly employed approaches: family therapy,
aversion therapies, operant methods, controlled drinking, and broad
spectrum treatment. Finally, we will discuss the effects on outcome of
treatment length and setting, and briefly review the state of knowledge
on matching clients with interventions.
PHARMACOTHERAPY
The conception of alcoholism as a disease has fostered investigation

of a large number of medications as potential therapeutic agents. We
will summarize the evidence on three major alternative strategies of
pharmacotherapy for alcoholism: (a) antidipsotropic drugs, (b) psycho-
tropic medications, and (c) hallucinogens.
Although there has been a staggering number of studies of drug
therapies for alcoholics, there have been surprisingly few controlled
investigations that have included an adequate outcome measure of
drinking behavior. Follow-up periods have commonly been restricted to

those typical of short-term drug trials, brief periods inadequate for eval-
uating long-range impact on drinking. Dropouts from pharmacotherapy
studies (50% is not uncommon) as well as noncompliance with dosage
regimens have also posed major problems for interpretation of findings.
Antidipsotropics
Antidipsotropics represent a class of drugs that are prescribed with
the intention of creating an adverse physical reaction when the indi-
vidual consumes alcohol. Three agents of this type have been studied:
disulfiram, citrated calcium carbimide, and metronidazole.
Disulfiram. Disulfiram (trade name: Antabuse) is, by far, the most
popular American pharmacotherapy for alcoholism. A client taking an
adequate dose of disulfiram develops an extremely unpleasant physical
reaction upon ingesting alcohol. Although hundreds of articles and
commentaries have been published on this drug, we found fewer than a
dozen controlled studies.
The earliest of these was an extensive study by Wallerstein et al.
(1957), who reported 53% of alcoholics treated by disulfiram to be im-
proved, as compared with 24%,36%, and 26% in comparison groups. In
light of apparent deviations from random assignment, however, we
question the interpretability of these results (W. R. Miller & Hester,
1980). Reinert (1958) reported superiority for disulfiram over reserpine,
but the absence of placebo or unmedicated controls, combined with a
44% attrition rate at follow-up, renders these results difficult to in-
terpret. Gallant, Bishop, Faulkner et al. (1968) reported no advantage for
disulfiram over no treatment, but this study likewise was plagued by a
massive attrition rate.
An interesting experiment by Yalovoi (reported by Mottin, 1973)
compared two groups comprising 300 male alcoholics. The control
group received disulfiram, including a "challenge" procedure in which
the client was forced to consume alcohol in order to experience the
adverse reaction that would ensue. A comparison group for whom dis-
ulfiram was medically contraindicated (therefore not randomly as-
signed) received a parallel "challenge" experience of nausea induced by
the emetic drug, emetine. No significant differences were observed be-
tween the groups during 3 years of follow-up.
In an attempt to sort out the specific effects of disulfiram from those
of motivation and therapeutic attention, Gerrein, Rosenberg, & Man-
ohar (1973) randomly assigned outpatient alcoholics, who were willing
to take part in their experiment. The patients came to an outpatient clinic
either once or twice weekly. Some patients received disulfiram whilst
others did not receive the drug. Eight-week follow-up data pointed to
superiority of the group visiting twice weekly and receiving disulfiram.
Unfortunately, the follow-up was not extended beyond this brief period.
The latter problem was remedied in a study by Fuller and Roth (1979) in
which 128 alcoholic patients were randomly assigned to receive a
therapeutic dose of disulfiram, an inactive dose (1 mg.), or no medica-
tion. At one year, no significant differences were observed in absti-
nence, drinking days, appointments kept, family stability, or em-
ployment. The investigators did note, however, that both groups
receiving the drug (even if an inactive dose) showed a higher abstinence
rate (23%) than did those given no pill (12%), suggesting a placebo effect
in being told that one is taking the drug. Using a life-table method to
reanalyze their data, Fuller and Williford (1980) reported statistical sig-
nificance of this difference, whereas the previous analyses had yielded
only a nonsignificant trend. It must be recalled, however, that the placebo
group in this study (inactive dose) achieved the highest abstinence rate
at 12 months, indicating that the therapeutic effect is not attributable to a
specific pharmacological action of the drug. This study is a valuable
contribution to the literature, because it helps to untangle specific from
nonspecific effects that were confounded in much-cited earlier studies,
such as Hoff and McKeown (1953) where "experimental" patients who
were willing to take the drug were found to fare better than "controls"
(not randomly assigned) who refused the drug or for whom it was
contraindicated.
Another well-controlled study by Azrin, Sisson, Meyers, and God-
ley (1982) suggests that the effectiveness of disulfiram may be aug-
mented by an intervention to increase medication compliance. At 6-
month follow-up, two groups that received a behavioral compliance
program to encourage the taking of disulfiram showed superior out-
come (less drinking, intoxication, unemployment, institutionalization)
than a group receiving a typical alcoholism treatment regimen (dis-
ulfiram, education and films, individual counseling) but no compliance
intervention.
An alternative method for administering disulfiram is to implant it,
thus eliminating problems in compliance with daily oral dosage. Hus-
sain and Harinath (1972) reported, in a brief letter, that 91 % of implant
cases sustained abstinence whereas 52% of "controls" had relapsed
within 2 months. There was no indication of random assignment, how-
ever, and the placebo effects of surgery are potentially large. Whyte and
O'Brien (1974) reported a similar study in which post hoc matching was
used in an attempt to assemble a comparable control group, again re-
porting superior duration of abstinence for implant patients.
Wilson and his colleagues have conducted a well controlled series of
studies to evaluate the placebo component of disulfiram implantation.

In their initial investigation (Wilson, Davidson, & White, 1976), 20
chronic alcoholics received either disulfiram implant or sham surgery.
Although the number of abstainers at follow-up did not differ substan-
tially (5 versus 4), the disulfiram patients who drank had waited signifi-
cantly longer before relapsing, and were more likely to abstain again
following the initial episode. Both groups showed substantial increases
in abstinence relative to the 2 years prior to surgery. Reporting 2-year
follow-up data for these patients, Wilson, Davidson, Blanchard, and
White (1978) found a sustained advantage for the implant group over
the placebo surgery group. Nevertheless, the placebo effect (sham sur-
gery vs. no surgery) was substantially larger than the drug effect (im-
plant vs. sham surgery): Wilson (1979) reported a mean of 367 and 307
days of abstinence in disulfiram implant and placebo groups, respec-
tively, as compared with a mean of 27 abstinent days in an unoperated
comparison group.
In summary, the most striking differences in controlled evaluations
are those between alcoholics receiving no medication and those given a
"drug" they believe to be disulfiram (even if it is a placebo). Current
findings point to small differences between disulfiram and placebo, typ-
ically at marginal levels of statistical significance. Thus it appears that the
therapeutic effects of disulfiram (beyond those attributable to population
characteristics, such as motivation) derive from a substantial placebo
effect combined with a modest (at best) specific effect. Given the known
side-effects of disulfiram and current indications of potential deleterious
health effects of this drug (e.g., Burnett & Reading, 1970; Goyer & Major,
1979; Kwentus & Major, 1979; Lake, Major, Ziegler, & Kopin, 1977;
Lijinski, 1979; Van Thiel, Gavaler, Paul, & Smith, 1979) we question the
wisdom of its use as a routine therapeutic agent. We particularly question
the ethics and effectiveness of the rather common practice of mandating
disulfiram as a consequence of alcohol-related offenses.
Citra ted Calcium Carbimide. An alternative antidipsotropic agent is
citrated calcium carbimide (Ccq. Like disulfiram it produces an aver-
sive reaction in combination with alcohol. In comparison with the dis-
ulfiram-ethanol reaction, its effect apparently occurs sooner after inges-
tion, is less severe, and is shorter lived. Likewise, the unpleasant side
effects appear to be less severe and prolonged than those of disulfiram,
increasing the likelihood of compliance, and the health risks that ensue
if drinking occurs are less extreme (for review see W. R. Miller & Hester,
1980). Marketed under the trade names of Temposil and Abstem in Cana-
da and Britain, CCC is not currently approved for use in the United
States. Levy, Livingstone, and Collins (1967) reported a 37% abstinence
rate at 9 to 14 months in a group treated with CCC, contrasted with no
abstainers in a comparison group given disulfiram. No adequately con-

trolled evaluations of CCC have been published to date.
Metronidazole. The story of metronidazole (trade name: Flagyl) is an
instructive chapter in the history of alcoholism treatment. Metronida-
zole is a drug useful in the treatment of urinary and vaginal infections.
Early reports that it produced a taste aversion to alcohol (e.g., Taylor,
1964) generated substantial enthusiasm for and interest in this drug as a
pharmacotherapeutic agent for alcoholics.
The evidence from controlled evaluations, however, has been con-
sistently negative. Merry and Whitehead (1968) found that the addition
of metronidazole to a standard hospital program resulted in no signifi-
cant improvement in the maintenance of abstinence during the 30 days
of the trial. Egan and Goetz (1968) found identical outcomes for metroni-
dazole versus placebo treatments in a double-blind 6-month clinical trial.
Gallant, Bishop, Camp, and Tisdale (1968) found that neither metroni-
dazole nor chlordiazepoxide yielded a significant improvement over
routine group therapy. Tyndel, Fraser, and Hartleib (1969) found no
evidence of decreased desire for or consumption of alcohol in metroni-
dazole versus placebo groups in a double-blind design. Penick, Carrier,
and Sheldon (1969), in another double-blind study, found a higher rate
of improvement (abstinent with only an occasional drink) in placebo
(64%) than in drug-treated patients (42%) at 6 months. The same direc-
tion of findings persisted at 4 year follow-up (Penick, Sheldon, Templer,
& Carrier, 1971). The absence of any therapeutic effect from metronida-
zole has also been confirmed in other controlled evaluations (Lal, 1969:
Lowenstam, 1969; Lysloff, 1972: Platz, Panepinto, Kissin, & Charnoff,
1970). The single exception to this pattern is a positive double-blind
evaluation conducted by Swinson (1971), who found no differences be-
tween drug and placebo groups until the 12-month follow-up, at which
point more drug-treated (9/18) than placebo-treated (1/13) patients were
classified as improved.
The history of metronidazole treatment contains the important
lesson that one cannot rely on optimistic uncontrolled reports of effec-
tiveness, and that a consistent pattern of findings from adequately con-
trolled research may well differ from the findings of a single study and
from anecdotal reports and clinical impressions. The overwhelming
weight of findings here indicates that metronidazole produces no signif-
icant reduction in drinking behavior, and after two decades this drug
has fallen into justifiable disuse as a treatment for alcoholism.
Psychotropics
The rationale for using psychotropic medications is that by treating
underlying psychopathology that is presumably causing the excessive
drinking, the alcohol abuse will be eliminated. The following review is
organized according to the type of underlying pathology targeted by
these medications.
Antianxiety Drugs. An early study by Hoff (1961) reported a com-
bined improvement (abstinent, one slip, or better control) rate of 72% for
alcoholics receiving chlordiazepoxide (trade name: Librium) versus 52%
for matched controls over 3 to 12 months of follow-up. Subsequent
studies, however, have failed to find significant differences between
chlordiazepoxide and comparison groups on either drinking measures
or psychosocial functioning (Bartholomew & Guile, 1961; Charnoff,
Kissin, & Reed, 1963; Mooney, Ditman, & Cohen, 1961; C. M. Rosen-
berg, 1974; Shaffer, Freinek, Wolf, Foxwell, & Kurland, 1963). Overall,
controlled research provides no persuasive support for using antianxiety
agents with alcoholics, and many physicians caution against their use
because of the risks of multiple abuse of alcohol and medication.
Antipsychotics. Drugs intended for the treatment of psychoses have
also been tried with nonpsychotic alcoholics. Butterworth and Watts
(1974) evaluated the effectiveness of thiothixene, trifluoperazine, and
placebo by using global rating scales of adjustment in alcoholics. Over
the 3 weeks of the study no differential improvement was noted among
these groups. Turek, Ota, Brown, Massari, and Kurland (1973) Similarly
found no differential advantage among these same two drugs and
placebo.
Once again, although early anecdotal and uncontrolled reports
were quite optimistic (e.g., Fox & Smith, 1959), no evidence has
emerged from controlled research to indicate that these medications are
of value in treating alcoholism itself.
Antidepressants. Shaffer, Freinek, Wolf, Foxwell, and Kurland (1964)
reported a double-blind study of nialamide (an MAO inhibitor) versus a
placebo. During the 28 days of inpatient treatment no differences were
observed on "incidence of sobriety lapse." Butterworth (1971) reported
combined improvement rates of 79% versus 40% for clients receiving
imipramine versus a placebo, respectively, based on a global rating
scale, but follow-up was restricted to 3 weeks and the criteria used to
define "improvement" are unclear.
Kissin and Gross (1968) compared the combined effects of chlor-
diazepoxide and imipramine with those of either drug alone or placebo.
At 6-month follow-up they reported reduced-drinking rates of 28%,
19%,0%, and 13% with the combination, chlordiazepoxide, imipramine,
and placebo groups, respectively. In a subsequent study Kissin, Platz,

and Su (1970) compared this drug combination to controls receiving only
inpatient ward treatment, psychotherapy, or no treatment. Success rates
after an unspecified length of follow-up were 21 %, 15%, 36%, and 5%,
respectively.
A number of other studies have investigated the effects of anti-
depressant medications on depression among alcoholics, often finding
modestly superior symptom reduction in drug-treated groups (e.g.,
Baekeland & Lundwall, 1975; Shaw, Donley, Morgan, & Robinson,
1975). Unfortunately, such studies have typically been restricted to brief
periods of follow-up and often have included no measure of impact on
drinking behavior.
At the present time, findings are equivocal regarding the effect of
antidepressant medications on drinking behavior, although there may
be beneficial effects of certain medications on mood. Further research is
needed, and it would be particularly sensible to examine the differential
effectiveness of antidepressants with alcoholics who are clinically de-
pressed versus nondepressed. Various investigators are now exploring
whether antidepressants affect desire for or effects of alcohol in animals
and humans. At the present time it seems appropriate to consider the
use of antidepressant medications as one alternative for treating mood
disorders that persist with sobriety, but it would be ill-advised to rely on
these drugs as primary agents to bring about sobriety.
Lithium. Some of the best-controlled investigations of pharmaco-
therapy for alcoholism have been with lithium carbonate. Studies exam-
ining measures of depression in alcoholics have found no differences
between those given lithium or placebo (Kline, 1974; Merry, Reynolds,
Bailey, & Coppen, 1976; Pond et ai., 1981). When drinking behavior is
examined, however, a different picture has sometimes emerged. Kline et
al. (1974) reported that alcoholics receiving lithium had fewer drinking
episodes and relapses as compared with a placebo control group. Merry
et al. (1976) similarly found that depressed subjects receiving lithium
showed significantly fewer days of incapacitation from drinking than
did those given a placebo. Pond et al. (1981), by contrast, found no
significant differences in weekly alcohol consumption during lithium
versus placebo weeks using a within-subjects crossover design. Like
other pharmacotherapy studies, these have been compromised by high
drop-out rates.
McMillan (1981) has suggested that lithium may influence drinking
behavior by reducing alcohol's euphoric effects. With the controlled
studies numbering two positive and one negative, further research is
needed to clarify whether lithium has a specific or differential effect on
alcohol consumption of problem drinkers. As with antidepressants, re-
liance on lithium as a primary agent for modifying drinking behavior

cannot be recommended at present.
Summary. No psychotropic medication has yet been shown to pro-
duce reliable changes in drinking behavior. Where psychopathology
persists, particularly after initial sobriety has been achieved, a carefully
chosen medication may be appropriate for treating these concurrent
problems. There is very tentative evidence that certain antidepressants
and lithium may reduce desire for and consumption of alcohol, but
neither the magnitude of this effect nor the volume of research to date
can substantiate lithium as a primary therapeutic agent for alcoholism.
Hallucinogens
During the 1960s and early 1970s, the use of lysergic acid di-
ethylamide (LSD) as an alcoholism treatment enjoyed a rapid rise in
popularity, followed by an equally precipitous decline. The rationale
was that alcoholics would have a psychedelic experience or would un-
dergo an altered state of consciousness that would render them more
amenable to personality change. Early uncontrolled studies enthusi-
astically reported positive results, with abstinence rates ranging as high
as 94% (Chwelos, Blewett, Smith, & Hoffer, 1959).
As controlled studies began to appear, however, a different picture
emerged regarding the effectiveness of LSD therapy for alcoholics. With
two exceptions, controlled evaluations found no differential or additive
advantage for LSD in treating alcoholism (Bowen, Soskin, & Chotlos,
1970: Denson & Sydiaha, 1970; Hollister, Shelton, & Krieger, 1969; John-
son, 1970; Ludwig, Levine, Stark, & Lazar, 1969; Ludwig, Levine, &
Stark, 1970; Smart, Storm, Baker, & Solursh, 1966). Both of the studies
finding an advantage for LSD treatment Oensen & Ramsay, 1963; Tom-
sovic & Edwards, 1970) suffered high attrition rates at follow-up. Of the
two, the Tomsovic and Edwards study is better designed, having em-
ployed random assignment, a large sample, and follow-up to 12
months. Although the LSD-treated group in this study did show a high-
er abstinence rate (44%) than the control group volunteering for but not
receiving LSD (11 %), the experimental group did not differ in outcome
from another group not volunteering for LSD.
Here, at least, the weight of negative findings appears to have influ-
enced practice, and by the early 1970s the use of LSD in alcoholism
treatment had all but disappeared. Indeed, with the exception of a 6-
year follow-up of an uncontrolled study (Rydzynski & Gruszczynski,
1980), there have been no subsequent reports on this approach.
PSYCHOTHERAPY AND COUNSELING
Various types of counseling and psychotherapy have been pro-

posed as appropriate for alcoholics, and uncontrolled evaluations of
such interventions yield a wide ranging mixture of outcomes (W. R.
Miller & Hester, 1980). As in other sections of this chapter, we will
confine our discussion to controlled and comparative studies that yield
information about the absolute and relative effectiveness of these ap-
proaches in reducing drinking behavior and alcohol-related problems.
One type of design has assigned patients either to receive or not
receive counseling or psychotherapy in addition to a regular inpatient
ward regimen. Levinson and Sereny (1969) sequentially assigned alco-
holics to receive or not receive insight therapy (including individual and
group therapy and educational sessions). At one-year follow-up, con-
trary to predictions, no significant differences were found between
groups on measures of drinking, work status, psychological health, or
social adjustment. The control group, who received only occupational
and recreational therapy in an inpatient milieu, showed a somewhat
higher rate of improvement (33%) than the therapy-treated group (15%),
based on personal interviews. Working with "chronic character disor-
ders" (not restricted to alcoholics), Pattison, Brissenden, and Wohl
(1967) randomly assigned inpatients to receive or not receive psycho-
analytic group psychotherapy. The only significant difference found on
a broad range of outcome measures favored the control group, who
showed higher self-acceptance. Tomsovic (1970) studied the outcome for
alcoholic patients before and after the introduction of a new program
procedure, an intensive counseling conference session in which the pa-
tient met with staff to discuss problems. Contrary to expectations, those
receiving the counseling fared significantly worse. Finally, Bjornevoll
(1972) assigned 46 alcoholics to receive or not receive intensive encoun-
ter group therapy. Although nearly every patient in the encounter
groups rated them as helpful and indicated they would recommend
them to other patients, no differences were found (contrary to predic-
tions) at 6- to 7-month follow-up on measures of drinking status, with a
trend toward more abstainers in the control group. In sum, studies that
have examined the value of adding psychotherapy to an inpatient milieu
have found either no differences or, contrary to the expectations of the
investigators, trends favoring patients who did not receive additional
psychotherapy.
Another type of design has been used to study psychotherapy or
counseling in relation to more minimal interventions, usually on an
outpatient basis. Ogborne and Wilmot (1979) randomly assigned 40 Skid
Row alcoholics to receive or not receive outpatient counseling sessions
to discuss their drinking problems. Only 10 of the 20 assigned to coun-

seling remained in contact for the 6 months of intervention, and these
were matched with 10 controls. No significant differences were ob-
served during or 3 months after intervention, based on personal inter-
views and examination of records. Bruun (1963) assigned alcoholics ran-
domly to an intensive psychotherapy group plus disulfiram (32 visits) or
to disulfiram alone (10 clinic visits), and found no differences in drink-
ing, social, or psychological functioning between the two groups at fol-
low-ups ranging from 2 to 3.5 years. Zimberg (1974) contrasted a com-
prehensive outpatient treatment program including group and individ-
ual psychotherapy and medication (50 visits) with a minimal treatment
condition consisting of medication and brief supportive counseling by
an internist (24 visits). At 12 months, 40% of the "comprehensive"
group and 44% of the "minimal" group showed few or no symptoms of
alcoholism, and the latter group showed significantly more improve-
ment on 5 of 7 scales of emotional adjustment. Crumbaugh and Carr
(1979) randomly assigned inpatients to receive either or neither of two
forms of logotherapy (8 hr/week) and found no significant overall dif-
ferences at the end of treatment (measures restricted to subjective pur-
pose in life-no measures of drinking). Mindlin (1965) failed to find any
difference in sobriety and attitude change between alcoholics receiving
group therapy and those assigned to educational lectures. Thus in out-
patient settings, as in inpatient settings, controlled evaluations have
failed to demonstrate benefit from adding psychotherapy or counseling
to more minimal interventions.
Several studies have provided slightly more encouraging results.
Kissin et al. (1970) compared outpatient psychodynamic group psycho-
therapy or psychotropic medication versus inpatient rehabilitation on a
ward where group therapy was the primary intervention. No differences
in effectiveness were found among groups (thus failing to support psy-
chotherapy over medication), but all treated groups did fare better than a
randomly assigned group of untreated controls. Brandsma, Maultsby,
and Welsh (1980) randomly assigned alcohol-related offenders to insight
therapy, cognitive-behavior therapy, Alcoholics Anonymous, or a no-
treatment control. Although at early follow-ups all treated groups com-
bined showed more improvement than controls, by 12 months the only
remaining significant difference indicated fewer days of drinking among
insight and cognitive therapy groups relative to controls, with no other
differences on a wide range of measures. These two studies suggest that
psychotherapy may yield modest short-term gains relative to no treat-
ment at all, at least within certain alcoholic populations.
Other studies have compared different approaches to psycho-
therapy. Ends and Page (1957) contrasted client-centered and psycho-
analytic groups with a group involving discussion of learning principles

and with a control discussion group. At one to one and a half year blind
follow-up, the groups showed 53%, 40%, 13%, and 24% improvement,
respectively (abstinent or only 1 to 2 slips). Tomsovic (1976) found
slightly more improvement in self-concept among inpatients treated in a
closed encounter group as compared with those treated in an open-
membership group, but no drinking measures were reported. Waller-
stein et al. (1957) found less improvement among patients treated with
group milieu therapy than among those given disulfiram or hypno-
therapy. Pomerleau, Pertschuk, Adkins, and d' Aquili (1978) randomly
assigned problem drinkers to either insight-oriented or behaviorally ori-
ented therapy groups. Dropouts were significantly greater from the in-
sight group (43%) than from the behavioral group (11%), but among
completers the differences (which favored the behavioral group) fell
short of statistical significance. McCance and McCance (1969) and P. M.
Miller, Hersen, Eisler, and Hemphill (1973) failed to find differences in
effectiveness between group therapies and electrical aversive counter-
conditioning. The absence of untreated controls in these studies makes
it impossible to judge whether any of the treatments offered would have
exceeded the efficacy of no intervention.
Viewing the controlled and comparative studies as a whole, we are
struck by the absence of consistent and substantive support for the
efficacy of traditional psychotherapy and counseling approaches as eval-
uated to date. The majority of studies have found no differences be-
tween those receiving versus not receiving such therapy in spite of the
fact that in most of the studies the investigators expected to find an
advantage for the former. In several studies the existing differences have
favored those not receiving additional counseling or psychotherapy.
Studies reporting an advantage for therapy relative to controls have
violated random assignment (Kissin et al., 1970), lacked adequate out-
come measures of drinking (Ends & Page, 1959), or shown minimal
differences at best (Brandsma et al., 1980).
Of course specification of the precise procedures involved in coun-
seling or psychotherapy is often less than clear. This lack of clarity
would be a more serious problem were it the case that some studies
showed substantial success whereas others did not. In the absence of
persuasive evidence for efficacy, it seems futile to pursue differentiation
among relatively inert procedures. Suffice it to say that although group
and individual counseling and psychotherapy have become exceedingly
popular elements in the standard treatment of alcoholics, there is little or
no evidence to date that such interventions have a specific long-range
impact on drinking behavior.
CONFRONTATION
Within the popular lore on treating alcoholics, there is nearly uni-

versal agreement that confrontation is a valuable if not essential element
of counseling. The common belief is that alcoholics as a group tend to
deny or fail to recognize the reality of their problems, and that it is
therapeutic to confront them with reality.
Actual procedures for carrying out confrontation vary widely. The
literature on alcoholism treatment includes many dozens of descriptions
of how to carry out confrontation. The usual procedures include a force-
ful and factual presentation of evidence that the individual "has alco-
holism," refutation of the client's protestations to the contrary, and
application of any available leverage or contingencies to persuade or
coerce the individual into treatment.
In our research of the literature, we found not a single adequately
controlled evaluation of confrontational counseling with alcoholics. This
lack of evaluative research is disturbing, given the current ubiquity of
confrontational approaches in alcoholism treatment. In a comparative
study, MacDonough (1976) found, contrary to expectations, that two of
four alcoholics treated in a token economy ward showed improvement
in undesirable behavior, whereas of five alcoholics treated by an inten-
sive confrontation approach, none showed improvement. The general
literature on group therapy is no more encouraging regarding the effec-
tiveness of a confrontational approach. Lieberman, Yalom, and Miles
(1973), for example, reported that a hostile-confrontational style of
group leadership was associated with more negative outcomes than other
leader styles. Likewise, research on motivation for change does not
support an exhortatory or argumentative style as optimal for inspiring
behavior change (W. R. Miller, 1983).
Yet confrontation need not be equated with strategies of coercion
and extrinsic control. An alternative is a feedback model in which the
client is given information about his or her current health status (W. R.
Miller, 1985). Exemplary research has emerged from a Swedish team
headed by Hans Kristenson (Kristenson, 1983; Kristenson, Ohlin,
Hulten-Nosslin, Trell, & Hood, 1983). Individuals at risk for alcohol
problems were identified via elevated liver enzyme values on routine
physical examinations. Patients with at-risk values were then ran-
domized, and those assigned to an experimental condition were given
feedback of their test results, information about the meaning of such
enzyme elevations, and simple advice to reduce alcohol consumption.
Those given feedback and advice were found, at follow-ups ranging to 5
years, to have substantially lower mortality, illness, hospitalization, and
work-absence rates relative to control subjects.
One form of feedback that has received experimental attention is

videotape self-confrontation (VSC). The typical procedure has been to
videotape the client during a period of intoxication (either upon admis-
sion or following a period of supervised drinking), then to have the
client view the tape on a subsequent day while sober. Clinical reports
indicate that this procedure is quite stressful for alcoholics, yielding
depression, decreased self-esteem, and anxiety. The impact of this pro-
cedure on drinking behavior has been evaluated in several studies. One
common result is a high rate of relapse shortly after the treatment (Faia
& Shean, 1976; Feinstein & Tamerin, 1972), sometimes higher than that
of clinical controls (Schaefer, Sobell, & Mills, 1971). Schaefer et al. (1971)
likewise reported a differentially higher dropout rate among patients
exposed to VSC (56%) than among comparable patients assigned at
random to standard treatment without VSC (10%). Controlled com-
parisons of patients given versus not given VSC have indicated little or
no beneficial impact on drinking behavior. Baker, Udin, and Vogler
(1975) found a modest advantage at 6 weeks for clients receiving VSC
plus behavioral counseling (80% abstinent or controlled) versus those
receiving only counseling (73%), but this difference had disappeared by
6 months. Faia and Shean (1976) found no significant differences in
results for a hospital program with or without a VSC component.
Schaefer et al. (1971) and Shaefer, Sobell, and Sobell (1972) also observed
no significant differences in sobriety of inpatients receiving versus not
receiving VSc. Finally, Lanyon, Primo, Terrell, and Wener (1972) found
no benefit from a VSC intervention versus a comparison discussion
group, although the addition of systematic desensitization to VSC did
yield more encouraging results.
Taken together, these studies indicate that confrontational interven-
tions are not inert. Kristenson's findings in particular suggest that a
minimal feedback procedure can have a substantial impact on behavior
and health. At the same time, confrontational approaches must be un-
dertaken with care because of the potential for precipitating dropout,
negative emotional states, lowered self-esteem, and proximal relapse. It
seems likely that feedback interventions may be of value in prevention
and treatment programs, but the optimal approaches and safeguards
remain to be delineated.
ALCOHOLICS ANONYMOUS
In spite of the fact that it inspires nearly universal acclaim and

enthusiasm among alcoholism treatment personnel in the United States,
Alcoholics Anonymous (AA) wholly lacks experimental support for its
efficacy. A number of studies have reported positive correlations be-
tween A.A. attendance and abstinence (W. R. Miller & Hester, 1980),
but these studies have failed to control for multiple confounding vari-
ables and yield results that are virtually uninterpretable (Bebbington,
1976). Only two studies have employed random assignment and ade-
quate controls to compare the efficacy of A.A. versus no intervention or
alternative interventions. Brandsma et al. (1980) found no differences at
12-month follow-up between A.A. and no treatment, and at 3-month
follow-up those assigned to A.A. were found to be significantly more
likely to be binge drinking, relative to controls or those assigned to other
interventions (based on unverified self-reports). Oitman and Crawford
(1966) assigned court mandated "alcohol addicts" to A.A., clinic treat-
ment, or no treatment (probation only). Based on records of rearrest,
31 % of A.A. clients and 32% of clinic-treated clients were judged suc-
cessful, as compared with 44% successes in the untreated group (Oit-
man, Crawford, Forgy, Moskowitz, & MacAndrew, 1967).
Other studies have evaluated multidimensional programs in which
A.A. was one component. Edwards et al. (1977), for example, found that
a complex treatment program (including A.A., medication, outpatient,
and inpatient care) was no more effective in modifying alcohol con-
sumption and problems at 12-month follow-up than was a single session
of counseling consisting of feedback and advice.
To be sure, these studies (like most any research) can be criticized
for methodological weaknesses, and as always "further research is
needed." Given the absence of a single controlled evaluation supporting
the effectiveness of A.A. and the presence of these negative findings,
however, we must conclude that at the present time the alleged effec-
tiveness of A.A. remains unproved.
ALCOHOLISM EDUCATION
One additional element that has come to be common in American

alcoholism treatment centers is an educational component, usually con-
sisting of a series of lectures, films, readings, or discussions on the topic
of alcohol and alcoholism. Typical content includes the negative effects
of alcohol on health and behavior, combined with a disease model of the
etiology and treatment of alcohol problems. Such intervention proceeds
from the assumption that alcoholics are uninformed about alcohol and
their problems with it, and require education. Here there is a parallel
between "treatment" and "prevention" research, with the distinction
residing in the level of current problem development among those being
educated. We will focus on controlled evaluations of education pro-
grams for individuals already evidencing problem drinking.
Studies employing random assignment of subjects have been few,
and have failed to support the efficacy of education as a treatment inter-

vention. Scoles and Fine (1977), using a Solomon four-group design,
found that drinking drivers assigned at random to an education pro-
gram on the hazards of drinking and driving were no more improved on
measures of alcohol consumption and impairment than a control group
receiving assessment only. Swenson and Clay (1980) similarly found no
differential gains in a group of offenders assigned at random to 10 to 15
hours of education, as compared with controls given a minimal inter-
vention (home study course), with most subjects reporting more drink-
ing at posttest.
Studies not using random assignment have yielded mixed results.
At least two studies (Hagen, Williams, McConnell, & Fleming, 1978;
Salzberg & Klingberg, 1983) have found higher rates of repeat offenses
among offenders sent to education and treatment programs than among
those given standard legal sanctions. One possible mechanism for such
a paradoxical outcome is a desensitizing effect, such that a group educa-
tion program may diminish the deterrent impact of arrest and penalties.
Malfetti (1975), by contrast, reported significantly fewer subsequent of-
fenses among 500 drunken drivers attending an education class, versus
500 controls. McGuire (1978) reported that 1,000 participants in an edu-
cation program showed 78% fewer repeat offenses, 34% fewer acci-
dents, and 23% fewer moving violations compared to 1,000 control sub-
jects given only probation and fine. However, the assignment of
offenders to conditions in all of these studies was not only nonrandom,
but systematically biased so that groups were nonequivalent before in-
tervention. The findings may therefore be attributable to sample differ-
ences.
Least evaluated has been the value of alcohol education within the
context of an alcoholism treatment program. The only comparative eval-
uation to date contrasted three alternative modalities for education, but
lacked a no-education control group (Stalonas, Keane, & Foy, 1979). The
investigators found that patients watching presentations on videotape
showed significantly greater gain in knowledge than did patients at-
tending live lectures or reading written presentations of the same mate-
rial. All groups, however, showed return to baseline levels of knowl-
edge at follow-up. Furthermore, it is questionable whether changes in
knowledge or attitudes about alcohol can be expected to generalize to
behavior change (Uecker & Boutilier, 1976).
Although uncontrolled studies have sometimes found gains in
groups receiving alcohol education, the full body of such research sug-
gests that detrimental effects are at least as likely to occur (cf. review by
Kinder, Pape, & Walfish, 1980). Controlled studies employing random
assignment have failed to support the efficacy of alcohol education in
changing drinking behavior and problems. Future research should help

to clarify what kinds and content of education programs, if any, might
be effective in alleviating problem drinking. In the meantime, we are
inclined to agree with the assessment of Uecker and Solberg (1973),
made more than a decade ago:
Until it has been dearly demonstrated that alcohol education is effective and
essential in the ... treatment of alcoholics, this form of treatment should
probably be deemphasized in favor of methods that have a sounder basis in
research. (pp. 512-513)
MARITAL AND FAMILY THERAPY
Recognizing that alcohol problems both influence and are influ-

enced by the family, therapists and programs have increasingly in-
cluded the spouse and other family members in the treatment process.
A few controlled evaluations have been published to date, with mostly
encouraging results.
In an early quasi-experimental study, Corder, Corder, and Laidlaw
(1972) provided an intensive 4-day marital therapy workshop to 20 male
alcoholics and their wives following a 3-week inpatient program. A com-
parison group was constituted of 20 alcoholics treated in the same inpa-
tient program during the prior month. At 6-month follow-up, 85% of the
comparison group had relapsed, but in the group receiving marital ther-
apy in addition to inpatient treatment only 42% had relapsed.
In a more carefully controlled study, Cadogan (1973) assigned 40
alcoholics following inpatient treatment to either an outpatient marital
therapy group or a waiting list control group. At 6-month follow-up,
45% of those receiving marital therapy were abstinent, whereas only
10% had remained abstinent in the control group. '
Hedberg and Campbell (1974) permitted clients to choose either
abstinence or controlled drinking as a goal, then randomly assigned
them to one of four treatment conditions: behavioral family therapy,
electrical aversion, systematic desensitization, and covert sensitization.
Combining as "successful" all cases rated as abstinent, controlled, or
improved at 6 months, 87% of cases treated by family therapy showed
successful outcomes, as contrasted with 87% in desensitization, 67% in
covert sensitization, and 25% in aversion therapy. The family therapy
group yielded the highest rate (53%) of total abstainers.
A very thorough study with longer follow-up was contributed by
McCrady and her colleagues (1979), comparing joint inpatient admission
of alcoholic and spouse, outpatient involvement of spouse, and no
spouse involvement. At 6-month follow-up, patients in both groups
with spouse involvement showed significant decreases in drinking,
whereas those treated individually without spouse involvement showed

no change-clearly arguing for including the spouse in treatment. In-
terestingly, spouses who were admitted jointly with the alcoholics also
showed significant reduction in their own alcohol consumption at fol-
low-up, whereas spouses attending therapy sessions but not admitted
failed to show such change. Improvement in marital adjustment was
approximately equal across groups. By 4-year follow-up (McCrady, Mor-
eau, Paolino, Longabough, & Rossi, 1982), however, differences be-
tween the groups had disappeared, suggesting that marital therapy has
an important short-term impact but not necessarily an enduring advan-
tage over individual treatment.
In a similarly well-designed study, O'Farrell and his colleagues
(O'Farrell & Cutter, 1982; O'Farrell, Cutter & Floyd, 1984) contrasted
two styles of couples groups (behavioral versus interactional) with indi-
vidual outpatient alcoholism counseling. Assignment to groups was
random, and follow-up extended for 18 months. In immediate gains
(change from pretreatment to posttreatment), the behavioral marital
therapy group showed significantly more improvement in marital ad-
justment than the other two groups. Differences were less striking on
drinking measures, however, and by the 18-month follow-up most dif-
ferences between the groups had disappeared.
All of these studies indicate that marital or family therapy (at least
the types studied) when added to other treatment increases the level of
improvement observed at short-term follow-up (6 months). The absence
of enduring differences at later points appears to be due in part to
gradual improvement of the comparison groups and in part to erosion of
gains following marital therapy. The consistency of positive findings at
short follow-up certainly warrants further investigation, and indicates
that marital therapy is a worthwhile modality to consider for inclusion in
alcoholism treatment.
A VERSION THERAPIES
The aversion therapies have as their common goal the altering of an

individual's attraction for alcohol. Through counterconditioning pro-
cedures, alcohol is paired with any of a variety of unpleasant experi-
ences. If the conditioning is successful, the individual shows an auto-
matic negative response when later exposed to alcohol alone. Aversion
therapies must not be confused with antidipsotropic medication, in
which the intended effects rest not on conditioning by repeated aversive
pairings but rather on suppression of drinking by fear of immediate
aversive consequences.
The aversion therapies differ from each other in the kind of un pleas-
ant event with which alcohol is associated. We will consider four types:
nausea, apnea, electric shock, and imagery (including hypnosis).
Nausea. The oldest form of aversion therapy pairs alcohol with the
experience of nausea. In this type of treatment nausea is induced-
usually by chemical means-while the individual drinks favorite alcohol
beverages. Apomorphine, emetine hydrochloride, and lithium hydro-
chloride have all been used to induce nausea and vomiting in this type
of treatment. Although there are a number of uncontrolled reports with
long follow-ups reporting excellent results (averaging around 60% absti-
nent at 1 year), controlled evaluations have been few. The previously
mentioned study by Wallerstein et al. (1957) included an emetine-condi-
tioning modality that yielded a success rate comparable to that of milieu
therapy, but less favorable than hypnotherapy or disulfiram treatment.
Selection problems and high attrition at follow-up cloud these results,
however. Jackson and Smith (1978) reported abstinence rates of 57% and
55%, respectively, for emetine versus electrical aversion therapy, but
again assignment was not random and follow-up rate was low. Random
assignment was employed in a study by Cannon, Baker, and Wehl
(1981) comparing emetine conditioning, electrical aversion, and routine
inpatient treatment. At 12 months no differences were found between
the emetine group and control subjects (309 versus 304 days of absti-
nence), although both fared substantially better than those receiving
electrical aversion (188 days of abstinence).
Boland, Mellor, and Revusky (1978) have provided the only con-
trolled evaluation to date of chemical aversion therapy using lithium as
the aversive agent. At 6-month follow-up, 36% of patients in aversion
therapy reported total abstinence, as compared with 12% in a com-
parison group receiving citrated calcium carbimide (Ccq. The absence
of an unmedicated control group renders the results difficult to interpret
on an absolute scale.
Finally, Richard (1983) published the first controlled evaluation of
aversive counterconditioning based on nausea induced by motion sick-
ness. In a series of four well-designed studies, he found no support for
the superiority of this approach over control conditions at follow-ups as
long as 24 months.
Apnea. A terrifying type of aversion was practiced briefly on an
experimental basis during the 1960s. The aversive stimulus was an injec-
tion of succinylcholine, which induces total paralysis of movement and
breathing for an interval of about 60 seconds. During this interval, alco-
hol is placed on the lips of the paralyzed patient. Initial uncontrolled
studies provided glowing reports of effectiveness, but two controlled
evaluations yielded less optimistic findings. Clancy, Vanderhoof, and
Campbell (1967) compared apneic aversion with two controls: one re-
ceiving a saline injection, and another receiving standard hospital treat-
ment (given to all groups). At 12-month follow-up the apnea group

failed to show greater improvement than that observed in the saline
placebo condition. Laverty (1966) contrasted apneic aversion, a placebo
condition, and a group with apnea induced only after alcohol admin-
istration had been completed. Because of the absence of immediate pair-
ing, the latter group theoretically should not show conditioning or im-
provement. In fact both groups with induced apnea showed superior
outcome (defined by abstinence) relative to the placebo. The severe
nature of this treatment quickly discouraged its application in any but
experimental settings, and no new evaluations have appeared.
Electrical Aversion. The arduousness of chemical aversion and diffi-
culties in controlling its severity and onset led to the exploration, begin-
ning in the late 1960s, of electric shock as an aversive agent in treating
alcoholism. Electrical aversion has been used as a component in pro-
grams with goals of either abstinence or controlled drinking. Results
have been quite mixed. Vogler, Lunde, and Martin (1971) reported that
subjects receiving electrical aversion in addition to hospital treatment
fared significantly better (time to rehospitalization, number of readmis-
sions, time in hospital) than did ward controls, but noted that similar
improvement was shown by a group receiving noncontingent electrical
stimulation (which theoretically should not work). Lunde, Johnson, and
Martin (1970) similarly noted that patients receiving electrical aversion
took significantly longer to relapse, but by 8 months the number of
relapses was comparable to that in a control condition. Glover and Mc-
Cue (1977) reported superior improvement rates at 13 months for sub-
jects receiving versus not receiving electrical aversion (64% versus 33%).
Blake (1967) found that the 12-month improvement rate with electrical
aversion (50%) could be increased (59%) by the addition of relaxation
training. Success was also reported by Hallam, Rachman, and Falkowski
(1972), who rated 75% (6/8) of aversion subjects as successful outcomes,
as compared with 40% (4/10) of controls.
An important but unpublished study by Marlatt (1973) compared
alternative conditioning paradigms in electrical aversion with alcoholics.
Although no differences were found in abstinence rates at 3 months, the
two groups receiving paradigms that would be expected to produce
conditioning (escape or avoidance) showed substantial reductions in
consumption (69% and 65% reduction) as compared with modest reduc-
tions with noncontingent shock (23%) and a control group receiving
only the basic ward program (42%). These findings converge with other
studies of electrical aversion that suggest that this type of treatment may
be more successful in reducing consumption than in inducing total absti-
nence (d. W. R. Miller & Hester, 1980).
Other controlled studies have found either no differences or less
improvement among alcoholics treated by electrical aversion, as com-
pared with alternative or no treatment (Cannon et al., 1981; DevenyI &

Sereny, 1970; Hedberg & Campbell, 1974; McCance & McCance, 1969;
Regester, 1971).
Electrical aversion has also been included as a component in several
multimodal programs intended to produce controlled drinking out-
comes, with varying rates of success (Ewing & Rouse, 1976; Lovibond &
Caddy, 1970; W. R. Miller, 1978; Sobell & Sobell, 1973; Vogler,
Weissbach, & Compton, 1977; Vogler, Weissbach, Compton, & Martin,
1977). Only two studies of this type have evaluated the contribution of
electrical aversion to the larger package. Caddy and Lovibond (1976)
compared a group receiving electrical aversion alone, a group given self-
control training alone, and a combination group. The aversion-alone
group proved least effective and suffered the most dropouts. The com-
bination group was most effective overall, but only modestly more so
than the self-control training alone. W. R. Miller (1978) found compara-
ble outcomes at 12 months for electrical aversion, self-control training,
and a combination program, and noted that the aversion yielded the
lowest rates of successful outcome at early follow-ups. He concluded
that the beneficial effects to be derived from electrical aversion could be
achieved equally by alternative and less painful procedures.
Overall the research on electrical aversion presents a confusing pic-
ture. Some studies note strong and significant effects, whereas others
find none. The method of administration appears to be a crucial vari-
able, although it is puzzling that paradigms not expected to produce
conditioning nevertheless have been found to yield therapeutic effects
in some studies. It appears that electrical aversion procedures have po-
tential for supressing drinking behavior, at least over several months of
follow-up, but the mechanisms for such change are unknown at pre-
sent. There is reason, however, to question whether the painfulness and
risk of dropout associated with this approach are warranted, given the
availability of alternative methods with known efficacy in reducing alco-
hol consumption.
Covert Sensitization. Covert sensitization (discussed by Klinger in
this volume) is a relative newcomer among the aversion therapies. It
is conducted entirely in imagination, pairing aversive scenes with drink-
ing imagery. Sensitization bears close resemblance to some treatment
procedures that have been labelled "hypnosis" (W. R. Miller & Hes-
ter, 1980), and therefore these two approaches will be reviewed to-
gether.
In an early controlled study, Ashern and Donner (1968) reported
that 40% (6/15) of clients treated by sensitization were abstaining at 6
months, whereas none were abstinent among an untreated control
group. Maletzky (1974) found a modified sensitization procedure
yielded superior improvement when compared with outcome from a

halfway house program in a random assignment design. The sensitiza-
tion-treated clients reported fewer urges to drink, less drinking, and
fewer drinking-related problems. Hedberg and Campbell (1974) found a
74% improvement rate among clients assigned to sensitization, which
was higher than that for electrical aversion but lower than rates for
family counseling or desensitization. Olson, Ganley, Devine, and Dor-
sey (1981) assigned blocks of alcoholics in a milieu treatment to receive a
behavioral program including sensitization, or transactional analysis, or
both, or neither additional modality. Rates of total abstinence were not
different at any follow-up point over 4 years. The two groups receiving
behavioral treatment, however, did show significantly greater reduction
in drinking and better overall adjustment than either the control (neither
treatment) group or the group receiving only transactional analysis.
Two comparative evaluations have failed to find sensitization to be
superior to alternative interventions Fleiger and Zingle (1973) found no
differences in outcome for clients receiving either sensitization or prob-
lem-solving therapy. Piorkowsky and Mann (1975) found equivalent
outcomes for sensitization, desensitization, and insight therapy, but a
65% dropout rate renders these results uninterpretable.
A key to differences in outcome may lie in specification of pro-
cedures employed in administering sensitization. Strong support for
this view has been provided by the exemplary work of Elkins (Elkins,
1980; Elkins & Murdock, 1977). Working with inpatient alcoholics,
Elkins demonstrated that the success of covert sensitization could be
predicted from the degree of conditioning established during treatment.
Those showing conditioned nausea responses (verified by physiological
monitoring) remained abstinent significantly longer (14.9 versus 3.7
months) than those failing to show conditioning. Among those showing
conditioning, 31 % sustained abstinence for periods of 5 to 62 months,
whereas no subject without conditioning sustained abstinence for an
extended period. These findings, then, point to conditioning as the
effective mechanism in covert sensitization treatment of alcoholism, at
least within a taste aversion modality.
Following Elkins' work, W. R. Miller and Dougher (1984) compared
three alternative procedures for administering covert sensitization: Stan-
dard nausea scenes, nausea scenes augmented by aversive odors (Mal-
etzky, 1974), or individualized scenes focusing on plausible negative
consequences of drinking. At 2 weeks after treatment, all groups showed
comparable reductions in alcohol consumption. In both of the nausea
groups, reduction in drinking was associated with the establishment of a
conditioned aversive response to alcohol, confirming Elkins' prior report.
In the third (consequences) group, however, conditioning occurred less
frequently and was unrelated to outeome. At IS-month follow-up, the

mean drinking rate had reverted to baseline in the group given standard
nausea scenes, whereas the other two groups largely sustained the reduc-
tions observed immediately following treatment. These data support
Elkins' finding that covert sensitization using nausea scenes does operate
via the establishment of a conditioned aversion response (subject to
extinction). The therapeutic effect observed with "negative conse-
quences" scenes is more puzzling, in that the reduction in alcohol con-
sumption was substantial and sustained, yet unrelated to conditioning.
Hypnosis. The outcome literature on hypnosis in the treatment of
alcoholism is brief and readily reviewed. Two controlled studies found
hypnotherapy to be superior to milieu treatment alone (Smith-Moore-
house, 1969; Wallerstein et al., 1957), but both studies had substantial
methodological flaws. Edwards (1966) and Jacobson and Silfverskiold
(1973), by contrast, found no advantage in hypnotherapy over standard
inpatient treatment or psychotherapy. The two latter studies employed
random assignment and in general were more sound methodologically.
Variations of procedure in "hypnosis" are quite wide, ranging from
posthypnotic suggestion to induced aversion. Variations in procedure
combined with the inconsistency of findings from controlled research
render it impossible to judge the potential value of hypnosis in alco-
holism treatment at the present time.
Summary. Research on the aversion therapies constitutes one of the
largest literatures in the alcoholism treatment field. Few approaches
have been more thoroughly evaluated, and important gains have been
made in understanding the mechanisms of effectiveness, although there
is still a long way to go. Aversive conditioning strategies appear to be
effective in suppressing drinking behavior and urges to drink, at least
for a period of a few months. Nausea aversion is well founded in experi-
mental learning literature, and conditioning procedures are clearly capa-
ble of inducing a taste aversion to alcohol in both animals and humans.
The availability of covert sensitization makes aversion therapy applica-
ble to a wider range of clients, because it can be administered on an
outpatient basis and is considerably less dangerous and stressful than its
chemical predecessors. Reduction of consumption rather than total ab-
stinence is a common observation following aversion therapies, and
thus the success of these approaches would not be well reflected if
complete abstinence were used as the criterion. Better specification of
optimal procedures is needed, following the example of Elkins (1980).
With continued refinement in procedure, it seems likely that aversive
counterconditioning will remain a valuable modality for inclusion in
alcoholism treatment programs.
CONTROLLED DRINKING
Controlled drinking is not a treatment method, but rather an outcome

or a goal of treatment. For decades, studies of treatment outcome have
reported that some clients maintain patterns of moderate and non-
problem drinking over extended periods of follow-up (Heather &
Robertson, 1983). Even when the sole goal of treatment has been total
abstention, up to one third of clients have been reported to be nonabsti-
nent but improved (Emrick, 1974; W. R. Miller, 1983). Such "accidental"
controlled drinking outcomes are not to be confused, however, with the
results of studies to be reviewed here-evaluations of treatment pro-
grams intentionally designed to teach moderate and non problem
drinking.
The first published study of this kind was a report by Australian
psychologists Lovibond and Caddy (1970). Testing a complex multicom-
ponent treatment program, they reported that 21 of 28 problem drinkers
completing the program were controlled drinkers at 16 to 60-week fol-
low-up, and that 3 others were improved. Unfortunately a high dropout
rate in the control group prevented meaningful comparative analyses. A
subsequent study (Caddy & Lovibond, 1976) revealed an 80% success
rate (controlled or improved at 12 months) in the group receiving full
treatment, as compared with 60% in a group receiving the program
without its electrical aversion component, and 30% among those receiv-
ing a similar program without a self-control orientation. Assignment to
conditions was random, and follow-up was conducted by interviewers
blind to treatment condition.
A second series of studies was initiated at Patton State Hospital in
California. Schaefer (1972) compared 13 clients receiving standard hospi-
tal treatment with 13 others given an additional treatment consisting of
electrical aversion aimed at a moderation goal. He reported that at 12
months, two of the former versus seven of the latter were showing
favorable outcomes. Following this pilot research, Sobell and Sobell
(1973) conducted a major outcome study comparing behavioral ap-
proaches with hospital treatment alone in groups having either an absti-
nence or a moderation goal. Subjects selected for this study were all
diagnosed as chronic gamma alcoholics. Within the controlled-drinking
goal, patients receiving the behavioral treatment were reported to show
superior outcome as compared with control subjects at follow-ups rang-
ing to 3 years (Caddy, Addington, & Perkins, 1978). A major controver-
sy regarding this study emerged in 1982 with the publication of a to-year
follow-up by another group of researchers (Pendery, Maltzman, &
West, 1982). Based on an unspecified follow-up protocol, the Pendery
group reported that of the 20 patients receiving controlled drinking

training, only one had succeeded in maintaining it continuously over 10
years. No data were provided regarding the control group, except for
the comment that they had "fared badly." After a prolonged series of
allegations and investigations, the Sobells provided a detailed rebuttal of
the Pendery et al. report (Sobell & Sobell, 1984), and no further evidence
has been forthcoming from the Pendery team. A third study from the
Patton group (Baker et al., 1975) found modest advantages for groups
receiving moderation-oriented behavioral treatments, as compared with
a randomly assigned control group.
Further evaluation of the multicomponent approach developed at
Patton was pursued by Roger Vogler and his colleagues. In a study with
inpatient alcoholics (Vogler, Compton, & Weissbach, 1975), those re-
ceiving the full behavioral treatment were found to show somewhat
greater reduction in alcohol consumption and problems at 12 months, as
compared with a control group receiving standard hospital treatment
plus alcohol education. These differences fell short of statistical signifi-
cance, however. Similar results were obtained in two outpatient sampes
(Vogler, Weissbach, & Compton, 1977; Vogler, Weissbach, Compton, &
Martin, 1977) comparing various treatment programs differing in inten-
sity but sharing the goal of moderation. Overall percentages of suc-
cessful outcome were encouraging (averaging 60% to 70%), but more
extensive interventions showed only modest and nonsignificant advan-
tages over minimal educational interventions.
A series of studies by Miller and his colleagues have explored alter-
native methods for teaching controlled drinking to problem drinkers. In
all of these studies, individuals showing signs of severe dependence or
biomedical deterioration were screened out, leaving a population of
problem drinkers showing mild to moderate physical dependence
symptoms. Four controlled evaluations have yielded success rates con-
sistent with those found by Vogler's group for outpatients (W. R. Miller,
1978; W. R. Miller & Taylor, 1980; W. R. Miller, Taylor, & West, 1980; W.
R. Miller, Gribskov, & Mortell, 1981). All of these studies contrasted
more extensive therapist-directed interventions with less intensive self-
control approaches. All found equal effectiveness, and in the latter three
studies a self-directed bibliotherapy intervention based on a self-help
manual (W. R. Miller & Munoz, 1982) proved as effective as its therapist-
directed counterpart. All studies employed random assignment and
confirmation of self-report by collateral interviews.
Comparative evaluations from other clinics have yielded strikingly
similar findings. Problem drinkers receiving behavioral self-control
training show marked reduction in alcohol consumption and problems
at one-year follow-up, with approximately two thirds being rated as
successful outcomes (Alden, 1978; Carpenter, Lyons, & Miller, 1985:

Hedberg & Campbell, 1974; Lovibond, 1975; Pomerleau et al., 1978;
Sanchez-Craig, Annis, Bornet, & MacDonald, 1984). Alden found mod-
est differences between groups receiving more versus less intensive
treatment (Alden, 1978), and Sanchez-Craig et al. (1984) reported no
differential reductions among patients assigned at random to abstinence
versus moderation goals.
Another group of studies has examined the effectiveness of moder-
ation training methods with individuals convicted of driving while in-
toxicated. Lovibond (1975) compared outcomes of 27 offenders trained
in moderation with those of 16 matched controls. Of the former 27, 16
were rated as controlled drinkers and 7 as improved (at follow-up of 1 to
9 months). Among the 16 controls receiving only legal sanctions, just
one was rated as controlled and one as improved. Two other studies
have compared behavioral self-control training with traditional drink-
ing-driver education approaches, using random assignment designs
(Brown, 1980; Coghlan, 1979). Both studies found significantly greater
improvement in the behaviorally treated groups.
Although these studies have yielded consistent findings, none has
included an untreated control condition. In an unpublished study, Buck
and Miller (1981) compared therapist-directed and self-directed behav-
ioral training with two waiting-list control groups, one of which self-
monitored alcohol consumption whereas the other did not. Assignment
was random, and collaterals confirmed self-report of clients. Neither
control group evidenced significant reduction in drinking during the
treatment phase, whereas both treated groups showed gains consonant
with those observed in prior studies. This suggests that behavioral self-
control training has a specific effect on drinking behavior, but that it can
be equally effective in self- or therapist-administered formats.
Not every study has yielded positive findings, however. In a well-
controlled evaluation with inpatient alcoholics, Foy, Nunn, and Rych-
tarik (1984) found that patients given a broad-spectrum moderation-
oriented program in addition to hospital treatment showed significantly
less abstinence and more abusive drinking than controls receiving only
hospital regimen. The difference persisted for 6 months, but was non-
significant thereafter. Other studies have reported very low rates of
success in teaching moderation (Czypionka & Demel, 1976; Ewing &
Rouse, 1976; Maxwell, Baird, Wezl, & Ferguson, 1974). It is noteworthy
that all studies reporting negative findings with controlled drinking
training (including Pendery et al., 1982) have had one commonality: the
patients studied were inpatient chronic alcoholics, often with high levels
of physical dependence. By contrast, the positive studies reported above
have been predominantly evaluations of moderation training with out-
patient problem drinkers having little or no documented physical

dependence.
Behavioral self-control training methods (with a goal of moderation)
have been subjected to a large number of controlled and comparative
evaluations-more, in fact, than any other single method for treating
alcohol problems. The majority of these studies have employed random
assignment to treatment conditions and have verified self-reports with
collateral interviews or objective indexes. With outpatient problem
drinkers, consistent improvement rates of 60% to 70% have been found
at follow-ups as long as 2 years (W. R. Miller & Baca, 1983). With drunk
driving offenders, behavioral training has been found to be significantly
more effective than traditional approaches. Controlled drinking appears
to be an attainable and successful goal for problem drinkers who have
not established significant degrees of dependence (W. R. Miller, 1983;
W. R. Miller & Caddy, 1977; W. R. Miller & Hester, 1980). Chronic
inpatient alcoholics, by contrast, have presented a rather mixed picture,
and negative findings with controlled drinking training have been re-
stricted to this population. Current data (e.g., Foy et al., 1984; Pendery et
al., 1982) indicate that controlled drinking training is not an effective
method for chronic alcoholics who are severely dependent.
OPERANT METHODS
Operant conditioning techniques alter behavior through modifica-

tion of its consequences. With alcoholics, reinforcement and punish-
ment contingencies have been used to influence drinking and drinking-
related behaviors.
A very large literature attests to the effectiveness of reinforcement
contingencies in influencing drinking behavior within laboratory set-
tings (for reviews see Heather & Robertson, 1983; W. R. Miller & Hester,
1980). In vivo applications of operant methods have been fewer, and
published studies consist mostly of successful case reports. Taken to-
gether, the laboratory and case reports indicate that drinking behavior
can be influenced by contingencies of reinforcement and punishment.
Surprisingly few controlled reports have appeared, however. Lieb-
son, Bigelow, and Flamer (1973) were successful in increasing disulfiram
compliance among methadone patients by making methadone con-
tingent on taking disulfiram. Patients who were required to take dis-
ulfiram in order to obtain methadone showed significantly fewer drink-
ing days (1 %) than a randomly assigned control group receiving metha-
done noncontingently (17%). Rosenberg, Gerrein & Schnell (1978), how-
ever, were unsuccessful in decreasing alcoholics' drinking by offering
them marijuana cigarettes contingent on their attending sessions. This
failure may have been due to ineffectiveness of the reinforcer or to

ineffectiveness of the counseling sessions (or both).
Two populations that have commonly been coerced into treatment
(usually by threat of punishment for noncompliance) are drunk driving
offenders and problem drinking employees. It is quite clear that the
threat of imprisonment or job loss can be effective in increasing com-
pliance with treatment recommendations, although such compliance
typically ends as soon as the contingency is removed (e.g., c. M. Rosen-
berg & Liftik, 1976). In this sense, coercion is "effective." The long-term
impact of coerced treatment on drinking behavior is quite a different
matter, however, and one that cannot be answered without considering
the treatment into which the person is being coerced. Coerced compliance
with an ineffective treatment program will not produce long-term
change in drinking patterns unless the coercion itself has an effect inde-
pendent of treatment. A properly controlled evaluation would consist of
random assignment to (a) mandatory treatment, or (b) a nontreatment
or minimal treatment alternative. With regard to drinking drivers, three
studies (already reviewed) have found no effect of coerced treatment
(Brandsma et al., 1980; Ditman et al., 1967, Swenson & Clay, 1980). Three
other studies evaluating controlled drinking training found significantly
greater reductions in subsequent drinking behavior of offenders receiv-
ing mandated treatment, as compared with untreated offenders (Brown,
1980; Coghlan, 1979; Lovibond, 1975). Studies contrasting the improve-
ment of voluntary versus mandated clients within the same treatment
program have typically found no significant differences in outcome
(e.g., Freedberg & Johnston, 1980; W. R. Miller, 1978; Smart, 1974).
Overall, then, it is clear that reinforcement and punishment con-
tingencies can be used to enhance program compliance, but that ulti-
mate impact on drinking behavior depends on the effectiveness of the
program itself. Clients mandated to treatment respond similarly to vol-
untary clients undergoing the same treatment. It is meaningless, there-
fore, to discuss the "effectiveness" of coercion methods (e.g., employee
assistance programs or legal mandating of treatment) in general. The
treatment impact of coercion can be defined only in relation to the inter-
vention into which the individual is coerced. As currently practiced in
the United States, for example, coercion programs commonly refer indi-
viduals into costly disease-oriented treatment consisting of components
with little or no known effectiveness.
BROAD-SPECTRUM ApPROACHES
In the 1970s, the concept of broad-spectrum treatment began to be

applied in the alcoholism field. The premise of this approach is that
drinking behavior is functionally related to other problems in the per-

son's life, and that an approach addressing this broader spectrum of
problems is more effective than one that focuses on drinking alone.
Research to date has focused particularly on hypothesized skill deficits
of alcoholics. The typical prediction is that alcoholics who are taught
crucial coping skills will be more successful in maintaining sobriety than
those whose treatment focuses only on alcoholism.
Social Skills Training. Alcoholics are often deficient in social skills.
Several controlled studies have examined the value of adding social
skills training to an alcoholism treatment program, and the results to
date have been quite consistent. Freedberg and Johnston (1978b) found
that the addition of assertiveness training to a 3-week inpatient program
substantially improved treatment outcome at one-year follow-up (36%
versus 24% abstinent in the control group). Combining abstinent and
improved categories, the assertion and control groups yielded success
rates of 72% and 57%, respectively.
Chaney, O'Leary, and Marlatt (1978) also reported a one-year fol-
low-up of V.A. alcoholics given assertiveness training in addition to the
regular treatment program. The trained group practiced assertive re-
sponses to a range of problem situations, whereas a control group was
encouraged to express and discuss personal feelings about these same
situations without practicing new responses. A second control group
received regular hospital treatment. At one year, the group given asser-
tion training showed significantly greater reductions in total number of
days drunk, length of drinking period, and total alcohol consumption,
relative to both control groups.
Jackson and Oei (1978) compared social skills training with a cog-
nitive restructuring treatment intended to change clients' belief struc-
tures that inhibit assertive behavior. Alcoholics received either of these
treatments or traditional supportive therapy as a control condition. Both
treatments proved superior to supportive therapy. The social skills train-
ing group was significantly more improved than the cognitive therapy
group at posttreatment, but by 3 months the direction had reversed,
with the cognitive therapy group showing better maintenance of gains.
This study points to the importance of addressing cognitive patterns in
social skill training. In a subsequent study, Oei and Jackson (1980) ran-
domly assigned 32 alcoholics to receive either social skills training or
traditional supportive therapy, with each condition further divided into
individual or group format. The social skills training group showed
superior gains on measures of alcohol consumption, personality and
social functioning throughout one year of follow-up. Overall there were
no main effects of group versus individual therapy formats, although in
the social skills training condition there appeared to be a slight advan-

tage for group over individual training.
West (1979) compared alternative methods of teaching communica-
tion skills to alcoholics and found that a branching-programmed manual
yielded better acquisition than nonbranching material presented in ei-
ther videotape or written form. West further found that acquisition was
increased by providing the opportunity for practice of new social skills.
The group receiving both programmed instruction and practice reported
significantly fewer drinking episodes at a 4-week follow-up than a con-
trol group receiving neither.
Finally, Ferrell and Galassi (1981) assigned patients in a milieu pro-
gram to receive additional treatment consisting of either assertiveness
training or human relations training. Clients were selected for the pro-
gram on the basis of low scores on a scale of adult self-expression,
indicating a need for social skills training. The assertion training group
showed significantly higher rates of abstinence at both one year (38%
versus 11 %) and 2 years (25% versus 0%) after treatment. The assertion
group likewise reported significantly more months of abstinence, on the
average, over the 2 years of follow-up.
Controlled research to date clearly supports social skills training as a
helpful addition to alcoholism treatment. Comparative findings currently
available suggest that optimal elements include assertiveness training,
group training with practice, branching programmed instruction (if writ-
ten materials are used), and attention to cognitive inhibitions.
Stress Management. Stress has often been hypothesized as an ante-
cedent of drinking and relapse, and stress management has been evalu-
ated as a broad-spectrum adjunct to alcoholism treatment. Both relaxa-
tion training and systematic desensitization with alcoholics have been
subjected to controlled evaluation.
Blake (1967) reported that alcoholics who received relaxation train-
ing in combination with electrical aversion therapy showed superior
improvement (59%) at 12 months, relative to those assigned to receive
only aversion therapy (50%). Freedberg and Johnston (1978a) compared
the outcome of a group given relaxation training in addition to regular
inpatient treatment, with that of a control group receiving only the
latter. At 12 months, no differences were found on drinking measures,
although the relaxation group fared significantly better on measures of
employment, depression, and relaxation. In other controlled studies,
both Sisson (1981) and W. R. Miller and Taylor (1980) found no impact
on drinking measures of a relaxation intervention.
A study by S. D. Rosenberg (1979) suggests a possible reason for
these inconsistent findings. Rosenberg assigned 59 alcoholics to receive
either biofeedback relaxation training or alcoholism education (the con-

trol condition). He further differentiated clients as high versus low
scorers on the anxiety sub scale of the Alcohol Use Inventory. Among
the high scorers (indicative of high levels of anxiety), clients given relax-
ation training showed significantly greater reductions in alcohol con-
sumption than did controls. Among the low scorers, by contrast, no
benefit was found from the addition of relaxation training. This finding
suggests that broad-spectrum intervention may differentially benefit
those alcoholics who need it.
Systematic desensitization has also been evaluated in controlled
research. Lanyon et al. (1972) evaluated the effectiveness of a confronta-
tional "interpersonal aversion" procedure with or without the addition
of systematic desensitization. At 6- to 9-month follow-up, 71% (5/7) of
those taught desensitization were abstaining, as compared with 14%
(117) of those given confrontation alone, and 25% (114) in a randomly
assigned control group consisting of group discussion. Piorkowsky and
Mann (1975) set out to compare desensitization, covert sensitization,
and insight therapy, but a 65% attrition rate (26/40) precluded mean-
ingful interpretation of their findings (one abstinent in each group at 6
months). Storm and Cutler (1969) found no differences in treatment
outcome at 6-month follow-up between outpatients given desensitiza-
tion versus standard alcoholism clinic treatment, but 62% of patients
dropped out of each group, again clouding interpretation. Finally, in a
previously mentioned study, Hedberg and Campbell (1974) found an
87% (13/15) improvement rate in a group given desensitization, which
was equivalent to the improvement rate in family therapy condition, but
superior to rates reported for groups receiving (by random assignment)
either electrical aversion or covert sensitization.
Community Reinforcement Approach. If one were to judge the effec-
tiveness of alcoholism treatment methods based on the strength of sci-
entific support available for them, the community reinforcement ap-
proach (CRA) would surely be at the top of the list. A series of well-con-
trolled studies have provided strong evidence trhat this intervention has
a powerful impact on alcohol use and general adjustment. Yet the com-
munity reinforcement approach remains little known and seldom used.
The CRA is included with the broad-spectrum approaches because
it is designed to restructure family, social, and vocational reinforcers in a
manner that reinforces sobriety while discouraging further drinking
through operant extinction or "time-out." The original program (Hunt
& Azrin, 1973) included problem-solVing training, behavioral family
therapy, social counseling, and-for unemployed clients-job-finding
training. This program was put to a stringent test. Hunt and Azrin
(1973) tested its effectiveness when added to a full inpatient program for
chronic addicted inpatient alcoholics. Patients assigned (at random) to
CRA in this study showed such massively larger gains than the hospital
controls (alcohol education lectures and Alcoholics Anonymous) that by
6-month follow-up there was little overlap between the groups. CRA-
treated patients were drinking on 14% of days (versus 79%), unem-
ployed days were 12 times higher in the control group, and controls
spent 15 times more days in institutions. All marriages in the CRA group
remained intact, whereas 25% ended in separation or divorce in the
control group. Collaterals confirmed self-report measures.
In 1976, Azrin published an improved version of CRA incorporating
disulfiram, a behavioral program for disulfiram compliance, a "buddy"
system, and daily self-monitoring of moods as an early warning system
for impending relapse. He tested this intervention with a similar popula-
tion, again comparing it with standard hospital treatment alone. The
same counselors administered the hospital program (their accustomed
approach) and the CRA. In this study, CRA clients at 6 months were
drinking on 2% of days (versus 55%), spent 7% of days away from home
(versus 67%), 20% of days unemployed (versus 56%), and no days in-
stitutionalized (versus 45%). These gains maintained very well in the
long run, with CRA clients (all found at follow-up) showing more than
90% abstinent days at 12, 18, and 24 months.
Azrin et al. (1982) evaluated the contribution of disulfiram to their
program. They compared the full CRA (Azrin, 1976) with disulfiram
alone (but including the behavioral compliance program), both being
added to regular outpatient treatment. A randomly assigned control
group receiving the regular outpatient alcoholism treatment program
reported over 50% drinking days, and approximately one third of days
intoxicated and unemployed at 6-month follow-up. These rates were
roughly double those obtained in the disulfiram-compliance group. The
full CRA program, however, resulted in nearly total suppression of
drinking days (0.9/month), days intoxicated (O.4/month), and unem-
ployed days (2.2/month). By a 3-month follow-up, nearly all patients in
traditional outpatient treatment had relapsed, a rare occurrence in the
CRA group. It was noted that the CRA program was differentially bene-
ficial for unmarried clients, whereas for married clients comparable
gains were obtained from CRA and from disulfiram-compliance alone.
Mallams, Godley, Hall, and Meyers (1982) tested the value of one
component of the CRA, attendance at a nondrinkers' social club. Clients
were chosen at random to be encouraged or not encouraged to attend
the alcohol-free club. Those so encouraged showed higher rates of atten-
dance, greater reduction in drinking, less behavioral impairment, and
less time spent in environments associated with heavy drinking, as com-

pared with controls.
Summary. Current research provides sound support for at least
three broad-spectrum approaches: social skills training, stress manage-
ment training, and community reinforcement approach. All three in-
volve training clients in specific coping skills. Initial data on differential
improvement indicate that such training is of maximal benefit to clients
who are deficient in these coping skills. For this reason, not all popula-
tions will show increased benefit from broad-spectrum versus focused
approaches (e.g., W. R. Miller et al., 1980). Furthermore, there is some
wisdom in resolving the drinking problem first and then evaluating
remaining skill deficits, because many pretreatment problems show
marked improvement following treatment focusing on drinking alone
(W. R. Miller et al., 1984).
LENGTH AND SETTING OF TREATMENT
It is tempting to assume that more treatment is better treatment,

and that longer or more "intensive" interventions will yield superior
outcomes. Although there has been a modest shift from inpatient to
outpatient treatment for alcoholism in the United States (Knowles,
1983), expensive inpatient facilities continue to proliferate and to capture
the majority of treatment dollars. Is this justified?
Some studies have compared residential treatment with less inten-
sive and expensive nonresidential alternatives. Annis and Liban (1979),
for example, compared a group receiving detoxification and halfway
house treatment with a matched sample receiving detoxification only.
At 3 months, official records indicated no difference in total drunken-
ness episodes, although halfway house residents were more likely to
return for detoxification, whereas controls were more likely to be ar-
rested when intoxicated. Edwards and Guthrie (1966, 1967) have ran-
domly assigned alcoholics to inpatient (9 weeks) or outpatient (8 visits)
treatment, and have found no significant differences in improvement at
6 or 12 months, with trends favoring the outpatient group on drinking
and social adjustment measures. Edwards et al. (1977) and Orford, Op-
penheimer, and Edwards (1976) randomly assigned alcoholics to inten-
sive (inpatient plus outpatient options) treatment or a single session of
counseling and found no significant differences in outcome at any point
during 2 years of follow-up. Gallant et al. (1973) assigned offenders to
compulsory inpatient plus outpatient treatment or to compulsory outpa-
tient treatment alone, and reported no significant differences in out-
come, although a high attrition renders these findings uninterpretable.
Kissin et al. (1970) compared inpatient rehabilitation with two forms of

outpatient treatment and an untreated control. All treatments proved
better than no treatment, with no significant differences between inpa-
tient and outpatient settings. Random assignment in this study was
compromised, however, by an opt-out procedure that could be exer-
cised by clients displeased with their assignment. Pittman and Tate
(1972) randomly assigned alcoholics to 3 to 6 week inpatient treatment
followed by outpatient treatment and Alcoholics Anonymous, or to a 7-
to lO-day detoxification only, with no outpatient aftercare. At a 12-
month follow-up, four deaths were reported in the treated group as
compared with one in the control group, and no significant differences
between groups were reported on any measure. Of 19 abstinent cases in
the inpatient condition, 18 had made extensive use of outpatient after-
care. Smart, Finley, and Funston (1977) randomly assigned (with opt-
out) alcoholics to outpatient, halfway house, or inpatient treatment.
Defining success as at least a 50% reduction in (or no) detoxifications,
arrests, and convictions, the highest success rates (50%) were observed
among patients either receiving outpatient care or refusing their treat-
ment altogether. Inpatient treatment was associated with 25% suc-
cesses, and no successes were found in the halfway house group. Stein,
Newton, and Bowman (1975) randomly assigned alcoholics following
detoxification to 25-day inpatient treatment or to aftercare alone. No
significant differences in outcome were observed at 2, 4, 7, 10, or 13
months. Wilson, White, and Lange (1978) randomly referred alcoholic
patients into inpatient hospital or to unspecified "community pro-
grams," and found that at 5 months the community-treated controls
showed better self-concept, general adjustment, and reduction in alco-
holism symptoms. No differences were found at 10 or 15 months. Penk,
Charles, and Van-House (1978) used a matching design to contrast inpa-
tient treatment with day-hospital care. The day-treatment group
showed better outcome on employment and social activity, greater anx-
iety reduction, and equivalent reduction in drinking, as compared with
inpatients. McLachlan and Stein (1982) used random assignment to
place alcoholics in 4-week inpatient or day-clinic treatment. At 12-month
follow-up no significant differences were found in alcohol or drug use,
marital adjustment, or psychological measures. Day-clinic patients
showed a 79% reduction in days of hospitalization during the follow-up
year (compared with pretreatment year), whereas inpatients showed a
38% increase in days hospitalized. Finally, Longabaugh et al. (1983) em-
ployed a random assignment design with blocking to place alcoholics
after detoxification into inpatient treatment or day-hospital care. At 6
months, no differences were observed, with trends favoring the day-
hospital group on subjective adjustment measures.
Thus, among 12 controlled evaluations of inpatient treatment ver-

sus nonresidential alternatives, not a single study found superior out-
come for the former, and in several the existing differences favored
nonresidential settings. It is noteworthy that all of these studies em-
ployed either random assignment or careful matching, most included
extended follow-up, the patients studied were alcoholics who would
otherwise have been routinely admitted for inpatient care, and the find-
ings in most cases were contrary to the expectations of the investigators.
Is length or intensity of inpatient treatment a factor in outcome?
Mosher, Davis, Mulligan, and Iber (1975) assigned alcoholics at random
to receive either long (30 days) or short (9 days) inpatient stays, com-
bined with detoxification and outpatient aftercare. No differences were
found between groups on measures of drinking, drug use, work status,
or anxiety at either 3 or 6 months. Similarly, Page and Schaub (1979)
found no differences at 6-month follow-up between alcoholics assigned
at random to either 3 or 5 weeks of inpatient treatment. Willems, Lete-
mendia, and Arroyave (1973) likewise used random assignment to either
short (maximum 4 weeks) or long (8 to 26 weeks) inpatient treatment. At
2-year follow-up, five deaths were recorded in the long-treatment (LT)
group, none in the short-treatment (ST) group. No significant dif-
ferences were observed in rates of favorable outcome (abstinent or im-
proved) at either 12 months (ST: 71 %: LT: 55%) or 24 months (ST: 68%;
LT: 52%). Walker, Donovan, Kivlahan, and O'Leary (1983) randomly
assigned inpatient alcoholics to either 2 or 7 weeks of behaviorally ori-
ented treatment. No significant differences emerged between groups on
any measure at 3,6, or 9 months, with small existing differences favor-
ing the 2-week group. Finally, Stinson, Smith, and Kaplan (1979) ran-
domly assigned inpatients to either an "intensive incare" ward heavily
staffed with professionals or to a "peer-oriented incare" ward with few
staff and an emphasis on peer interactions among patients. Although
the investigators expected to find an advantage for the intensive ap-
proach, the opposite was found. Patients on the less intensive ward
were significantly more improved on a drinking outcome measure, with
no other significant differences appearing between groups at 3,6, 12, or
18 months. Thus, of five controlled studies evaluating more versus less
lengthy or intensive inpatient care, not one supported the more inten-
sive treatment, and differences tended to favor shorter, less intensive
methods.
Findings on intensity of outpatient care are less consistent. Pittman
and Tate (1972) found that success after inpatient treatment was highly
predictable from use of outpatient aftercare services. Robson, Paulus,
and Clarke (1965) used post hoc matching to equate outpatients seen for
one to four sessions versus five or more sessions. No differences in
abstinence rates was observed, but the longer-treated group showed

greater reduction in severity of drinking problems and more overall
improvement. Similar findings were reported by Smart and Gray (1978),
who used post hoc matching to assemble groups treated for one visit,
more than one visit but less than 6 months, or more than 6 months.
Differences at 12 months were modest, reflected primarily in increased
abstinence rates with longer treatment (3%, 11%, and 16%, respec-
tively). In a series of controlled clinical outcome studies, W. R. Miller
and his colleagues compared bibliotherapy (self-help manual plus mini-
mal therapist contact) with 6 to 18 weeks of outpatient sessions (Buck &
Miller, 1981; W. R. Miller & Baca, 1983; W. R. Miller & Taylor, 1980; W.
R. Miller et ai., 1980, W. R. Miller et al., 1981). In all studies, comparable
improvement was observed in self-directed versus therapist-directed
conditions at follow-ups as long as 2 years. This absence of difference is
consistent with findings of controlled studies reviewed earlier, evaluat-
ing more versus less intensive outpatient treatment (Bruun, 1963; Ed-
wards et ai., 1977; Ogborne & Wilmot, 1979; Vogler, Weissbach, &
Compton, 1977; Vogler, Weissbach, Compton, & Martin, 1977; Zimberg,
1974). Among studies of outpatient therapy, then, nonrandom (match-
ing) designs have yielded modest advantages for longer versus shorter
treatment, whereas studies employing random assignment controlled
designs have found no advantage in more intensive treatment.
Summary. Controlled research on length and intensity of treatment
provides a very clear and consistent message: More treatment is not
necessarily better treatment. If anything, differences that have emerged
in controlled research to date would favor shorter and less intensive
approaches, not only in cost-effectiveness but in absolute effectiveness.
This finding has remained consistent across a variety of theoretical ori-
entations and populations. The lack of an advantage for residential treat-
ment and for longer or more intensive treatment has been demonstrated
in precisely the population for which such treatment has been alleged to
be necessary: chronic alcoholics. Similar results have been obtained for
less severe problem drinkers, however. In the absence of any convincing
evidence in favor of intensive or residential treatment over less costly
alternatives, the burden of proof clearly lies with those who would
advocate more heroic interventions. With no significant differences in
effectiveness but massive differences in cost-effectiveness, it would ap-
pear that treatment should increasingly shift to an outpatient, communi-
ty-based approach.
Of course, it makes little sense to talk about length or intensity of
treatment without considering what kind of treatment is being offered.
Some modalities lack evidence for effectiveness at any length or level of
intensity. Future studies should determine whether length or intensity
are determinants of outcome in treatment having documented specific

effectiveness. Certain subpopulations of alcoholics may also be identi-
fied that will show differential benefits from longer or more intensive
levels of certain kinds of treatment. This brings us to our final considera-
tion: the matching hypothesis.
MATCHING CLIENTS WITH TREATMENTS
The matching hypothesis suggests that treatment will be more ef-

fective when clients are matched with optimal interventions (GottheiI et
ai., 1981). This idea is not only very sensible, but could provide substan-
tial savings if unnecessarily intensive and expensive treatment could be
averted through matching. Persuasive research evidence for the match-
ing hypothesis has begun to appear, although the field is clearly in its
infancy and sound empirical guidelines for matching must await further
research. We will point to a few variables that appear to be promising
predictors of differential treatment outcome.
CONCEPTUAL LEVEL
Early exemplary research on matching is that of McLachlan (1972,

1974), who has postulated that clients with a low conceptual level (CL;
e.g., preference for simpler constructs and rules) would show optimal
response to a directive approach, whereas clients with a high CL would
do better in nondirective therapy. McLachlan found that patients
matched to therapy (high CL in nondirective, low CL in directive)
showed a 70% recovery rate, whereas mismatched patients (low CL in
nondirective, high CL in directive) showed a 50% rate of recovery. This
is consistent with the common findings that successful responders to
Alcoholics Anonymous (a low CL approach) are characterized by high
authoritarianism, lower education, field dependence, religiosity, and
conformity (for reviews see W. R. Miller & Hester, 1980; Ogborne &
Glaser, 1981).
NEUROPSYCHOLOGICAL IMPAIRMENT
Alcoholic populations show rather consistent patterns of neuropsy-

chological impairment, and the degree of such deficits might be predic-
tive of differential response to treatment alternatives (W. R. Miller &
Saucedo, 1983). The first direct test of this hypothesis failed to confirm
the prediction that more impaired individuals would show more benefit
from longer and more intensive treatment (Walker et ai., 1983). More
severe neuropsychological impairment has been associated with poorer

overall outcomes within treatment, however (O'Leary, Donovan,
Chaney, & Walker, 1979), and this area deserves further exploration.
SEVERITY
A very promising differential predictor of outcome is severity of

alcoholism. One highly consistent finding is that individuals with less
severe problems are more likely to succeed in achieving controlled
drinking, whereas more severe alcoholics show better prognosis with
abstinence (W. R. Miller & Baca, 1983; W. R. Miller & Joyce, 1979; W. R.
Miller & Hester, 1980; Orford et al., 1976; Polich, Armor, & Braiker, 1981;
Popham & Schmidt, 1976; Smart, 1978). This has led to recommenda-
tions that treatment goals be negotiated in relation to problem severity
(W. R. Miller, 1983; W. R. Miller & Caddy, 1977). In a comparative study
of intensive versus minimal treatment, Orford et al. (1976) reported that
among gamma (severe) alcoholics, all successful cases had received in-
tensive treatment, whereas 80% of failures had received minimal treat-
ment-a pattern that was precisely reversed for less severe alcoholics.
Similarly, McLellan, O'Brian, Kron, Alterman, and Druly (1980), and
McLellan, Woody, Luborsky, O'Brian, and Druly (1983) found that
matched cases (severity of problem with intensity of treatment) showed
substantially better outcomes than mismatched cases (e.g., severe prob-
lem in less intensive outpatient treatment). These latter two studies
suggest that intensive treatment may be differentially beneficial for alco-
holics with more severe levels of problems and dependence.
Locus OF CONTROL
Rotter's (1966) Internal-External Locus of Control scale has been

employed in several evaluations of matching. Abramowitz, Abra-
mowitz, Roback, and Jackson (1974) found that alcoholics with an inter-
nallocus of control fared better in nondirective therapy, whereas those
with an external orientation showed better prognosis in directive treat-
ment. Obitz (1978) noted that patients volunteering to take disulfiram
showed a substantially more external orientation than those rejecting
disulfiram. O'Leary, Rohsenow, and Donovan (1976) found that atten-
dance at aftercare meetings could be predicted from locus of control
orientation, with externals being more likely to continue attending.
These are correlational findings, however, and the only controlled study
to date (Schmidt, 1978) failed to confirm the hypothesis that internals
would benefit differentially from a self-directed approach to treatment.
FAMIL Y HISTORY
For more than a decade, research has pointed to a "familial" type of

alcoholism characterized by a more rapid and severe progression (e.g.,
Winokur, Reich, Rimmer, & Pitts, 1970). Surprisingly, family history of
alcoholism has yet to be used as a matching variable in research on
selection of optimal treatment. W. R. Miller and Joyce (1979) found that
a history of paternal alcoholism predicted abstinence as an outcome,
whereas absence of paternal alcoholism was associated with successful
controlled drinking outcomes. If indeed familial alcoholism is qualita-
tively different from other types, family history is a good candidate for
inclusion in future research on matching.
LIFE PROBLEMS
The client's life problems (beyond alcoholism itself) may indicate

need for specific types of treatment. Psychotropic medications may be
appropriate in treating psychiatric syndromes that persist into sobriety.
Broad-spectrum interventions intended to treat a problem underlying
alcohol abuse are most likely to be effective with those individuals who
show evidence of the target problem (e.g., S. D. Rosenberg, 1979).
Training in social or job-finding skills is most appropriate for clients who
are currently deficient in these skills (Azrin et al., 1982). The addition of
such broad-spectrum components to routine treatment for all alcoholics
may have no impact or even a detrimental effect on overall outcome (W.
R. Miller et al., 1980). An individualized assessment of additional life
problems, however, can point to needed interventions beyond those
intended to deter drinking.
PERCEIVED CHOICE
In the absence of sound empirical grounds for matching clients with

treatments, an alternative approach is to offer clients a menu of alter-
natives and negotiate the treatment of choice. Several literature reviews
have concluded that clients offered such a choice will be more motivated
to participate in treatment and will show more favorable outcomes than
clients given no alternatives (Costello, 1975; W. R. Miller, 1985: Parker,
Winstead, & Willi, 1979). Cronkite and Moos (1978) found that a signifi-
cant proportion of explained variance in treatment outcome is accounted
for by the interaction of patient and therapy variables, supporting the
importance of matching patient and program characteristics. The only
direct experimental support for self-matching, however, comes from a
study by Kissin, Platz, and Su (1971). Alcoholic patients were assigned
at random to be offered one, two, or three treatment options. Patients

given an option frequently exercised it (one half to two thirds rejected
the original treatment offered to them, given a choice). Patient charac-
teristics were poor predictors of acceptance or outcome of particular
treatments. Choice, however, proved important. Individuals given op-
tions showed greater acceptance of their treatment and more favorable
outcomes, increasing in proportion to the number of choices offered. At
least until robust empirical matching schemes emerge, an optimal pro-
cedure may be to allow clients to make informed choices among a range
of plausible alternatives (see Chapter 8, this volume).
CONCLUSIONS
Several treatment approaches have been shown to be effective in

reducing drinking behavior, a necessary first step in treating alcohol-
related problems. Aversive counterconditioning methods have a long
history of positive outcomes in uncontrolled research, and controlled or
comparative studies have likewise supported the specific efficacy of
aversion therapies in fostering abstinence or reduced consumption. Al-
though both chemical and electrical aversion therapies have been associ-
ated with reduction of consumption, covert sensitization offers several
important advantages. Because it is based on imagery, covert sensitiza-
tion requires no physical aversive agents and is therefore less dangerous
and ethically problematic. It can be administered in either inpatient or
outpatient settings. Further research is needed to delineate optimal pro-
cedures for using this technique with alcoholics. Behavioral self-control
training is also well supported as an effective approach for reducing
consumption by problem drinkers.
The status of pharmacotherapeutic agents in reducing drinking be-
havior is currently in question. Few studies have demonstrated specific
effectiveness of antidipsotropic agents such as disulfiram, and the bulk
of their impact can be attributed to placebo effects. It is unclear whether
the minimal benefits that have been demonstrated outweigh the dan-
gers of side-effects and long-term health risks associated with these
drugs. There are some encouraging early data to indicate that certain
antidepressants as well as lithium may engender decreased desire for or
consumption of alcohol by some alcoholics. Findings are quite mixed,
however, and further data will be required to clarify the mechanisms of
such action and the specific subpopulations in which such benefit may
apply.
Once drinking behavior has been terminated or curtailed, the chal-
lenge is to maintain this pattern. Current research points to the value of
broad-spectrum interventions in increasing the probability of prolonged

sobriety. Specifically, social skills training, stress management training,
and marital and family therapy have been found to promote the mainte-
nance of sobriety. Azrin's community reinforcement approach, which
combines several of these elements, has been shown to increase absti-
nence and adjustment substantially when added to a program focusing
primarily on alcoholism.
The picture that emerges is that of a two-stage treatment process,
requiring different interventions. One set of interventions is optimal in
changing drinking behavior itself, in bringing about abstinence or mod-
eration. Here aversion therapies and behavioral self-control training are
well supported. Another set of interventions aims primarily at environ-
mental contingencies and other life problems, attempting to bring about
changes that will help to maintain sobriety. Neither set of interventions
may be sufficient in itself to bring about lasting change. A combination
of demonstrably effective procedures from each category would seem to
be optimal, and Azrin's data support this notion.
Table 1 contains a list of interventions that appear to have specific
effectiveness in the treatment of alcoholism, based on available em-
pirical evidence. Beside this list is another, specifying modalities com-
monly employed in current American alcoholism treatment programs.
The lack of overlap between these two lists is evident. American treat-
ment of alcoholism follows a standard formula that appears to be imper-
vious to emerging research evidence, and has not changed significantly
for at least two decades (W. R. Miller, in press; Moore, 1977). It is
noteworthy that the "standard practice" list in Table 1 contains no
modalities that have been soundly supported by research. Current em-
pirical evidence suggests that a combination of these ingredients would
TABLE 1.
Supported Versus Standard Alcoholism Treatment Methods
Treatment methods
Treatment methods currently currently employed as
supported by controlled outcome standard practice in
research alcoholism programs
Aversion therapies Alcoholics Anonymous

Behavioral self-control training Alcoholism education
Community reinforcement approach Confrontation
Marital and family therapy Disulfiram
Social skills training Group therapy
Stress management Individual counseling
not be expected to yield therapeutic gains substantially greater than the

spontaneous remission rate, and indeed this appears to have been the
overall result of American alcoholism treatment over the past few dec-
ades (W. R. Miller & Hester, 1980).
Another perplexing inertia is the persistence of expensive inpatient
treatment approaches despite clear evidence that they offer no advan-
tage in overall effectiveness. The substantially greater cost of such ap-
proaches increases the burden borne in taxes and insurance premiums;
yet every controlled evaluation to date reflects no increased benefits to
offset these costs. Future research may identify select subpopulations
for whom such intensive and expensive methods are differentially bene-
ficial. Overall, however, current data indicate that alcoholism can be
treated in outpatient settings with equal effectiveness but at substan-
tially lower cost.
In conclusion, we offer three basic principles as prudent guidelines
in designing future alcoholism treatment programs. First, treatment pro-
grams, both voluntary and involuntary, should be composed of modali-
ties supported by current research as having specific effectiveness, and
consideration should be given to preferential funding of programs so
constituted. Second, the first interventions offered should be the least
intensive and intrusive, with more heroic and expensive treatments em-
ployed only after others have failed. Third, as research warrants, clients
should be matched to optimal interventions based on predictors of dif-
ferential outcome. Clients should be informed participants in their own
treatment planning process, and should be offered a range of plausible
alternatives along with fair and accurate information on which to base a
choice.
REFERENCES
Abramowitz, C. V., Abramowitz, S. F., Roback, H. B., & Jackson, c. (1974). Differential
effectiveness of directive and nondirective group therapies as a function of client
internal-external control. Journal of Consulting and Clinical Psychology, 42, 849-853.
Alden, L. (1978). Evaluation of a preventive self-management programme for problem
drinkers. Canadian Journal of Behavioural Science, 10, 258-263.
Annis, H. M., & Liban, C. B. (1979). A follow-up study of male halfway-house residents
and matched nonresident controls. Journal of Studies on Alcohol, 40, 63-69.
Ashern, B., & Donner, L. (1968). Covert sensitization with alcoholics: A controlled replica-
tion. Behaviour Research and Therapy, 6, 7-12.
Azrin, N. H. (1976). Improvements in the community-reinforcement approach to alco-
holism. Behaviour Research and Therapy, 14, 339-348.
Azrin, N. H., Sisson, R W., Meyers, R, & Godley, M. (1982). Alcoholism treatment by
disulfiram and community reinforcement therapy. Journal of Behavior Therapy and Ex-
perimental Psychiatry, 13, 105-112.
164 WILLIAM R MILLER AND REID K. HESTER
Baekeland, F., & Lundwall, L. K. (1975). Effects of discontinuity of medication on the

results of a double-blind drug study in outpatient alcoholics. Journal of Studies on
Alcohol, 36, 1268-1272.
Baker, T. B., Udin, H., & Vogler, R E. (1975). The effects of videotaped modeling and self-
confrontation on the drinking behavior of alcoholics. International Journal of the Addic-
tions, 10, 779-793.
Bartholomew, A. A., & Guile, L. A. (1961). A controlled evaluation of "Librium in the
treatment of alcoholics." Medical Journal of Australia, 2 (15), 578-581.
Bebbington, P. E. (1976). The efficacy of Alcoholics Anonymous: The elusiveness of hard
data. British Journal of Psychiatry, 128, 572-580.
Bjernevoll, K. J. (1972). Alkoholikere i encountergruppe: Encountergruppe som tillegsbehandling
for en gruppe pasienter ved Victoria A-sentrum, A-klinikken for Vest/andet. Unpublished
psychology thesis, University of Bergen.
Blake, B. G. (1967). A follow-up of alcoholics treated by behaviour therapy. Behaviour
Research and Theory,S, 89-94.
Boland, F. J., Mellor, C. S., & Revusky, S. (1978). Chemical aversion treatment of alco-
holism: Lithium as the aversive agent. Behaviour Research and Therapy, 16, 401-409.
Bowen, W. T., Soskin, R. A., & Chotios, J. W. (1970). Lysergic acid diethylamide as a
variable in the hospital treatment of alcoholism: A follow-up study. Journal of Nervous
and Mental Disease, 150, 111-118.
Brandsma, J. M., Maultsby, M. c., & Welsh, R. J. (1980). The outpatient treatment of alco-
holism: A review and comparative study. Baltimore, MD: University Park Press.
Brown, R A. (1980). Conventional education and controlled drinking education courses
with convicted drunken drivers. Behavior Therapy, 11, 632-642.
Bruun, K. (1963). Outcome of different types of treatment of alcoholics. Quarterly Journal of
Buck, K. A., & Miller, W. R. (1981, November). Why does bibliotherapy work? A controlled
study. Paper presented at the annual meeting of the Association for Advancement of
Behavior Therapy, Toronto.
Burnett, G. B., & Reading, H. W. (1970). The pharmacology of disulfiram in the treatment
of alcoholism. British Journal of Addiction, 65, 281-288.
Butterworth, A. T. (1971). Depression associated with alcohol withdrawal: Imipramine
therapy combined with placebo. Quarterly Journal of Studies on Alcohol, 32, 343-348.
Butterworth, A. T., & Watts, R D. (1974). Double-blind comparison of thiothixene, tri-
fluoperazine, and placebo in chronic alcoholism. Psychosomatics, 15, 85-87.
Caddy, G. R, Addington, H. J., Jr., & Perkins, D. (1978). Individualized behavior therapy
for alcoholics: A third year independent double-blind follow-up. Behaviour Research
and Therapy, 16, 345-362.
Caddy, G. R., & Lovibond, S. H. (1976). Self-regulation and discriminated aversive condi-
tioning in the modification of alcoholics' drinking behavior. Behavior Therapy, 7, 223-
230.
Cadogan, D. A. (1973). Marital group therapy in the treatment of alcoholism. Quarterly
Journal of Studies on Alcohol, 34, 1187-1194.
Cannon, D. S., Baker, T. B., & Wehl, C. K. (1981). Emetic and electric shock alcohol
aversion therapy: Six- and twelve-month follow-up. Journal of Consulting and Clinical
Carpenter, R A., Lyons, C. A., & Miller, W. R (1985). Peer-managed self-control program
for prevention of alcohol abuse in American Indian high school students: A pilot
evaluation study. International Journal of the Addictions, 20, 299-310.
Chaney, E. F., O'Leary, M. R, & Marlatt, G. A. (1978). Skill training with alcoholics.
Journal of Consulting and Clinical Psychology, 46, 1092-1104.
Charnoff, S. M., Kissin, B., & Reed, J. I. (1963). An evaluation of various psycho-
therapeutic agents in the long term treatment of chronic alcoholism. American Journal
of Medical Sciences, 246, 172-179.
Chwelos, N., Blewett, D. B., Smith, C. M., & Hoffer, A. (1959). Use of d-Iysergic acid
diethylamide in the treatment of alcoholism. Quarterly Journal of Studies on Alcohol, 20,
577-590.
Clancy, J., Vanderhoof, E., & Campbell, P. (1967). Evaluation of an aversive technique as a
treatment for alcoholism. Quarterly Journal of Studies on Alcohol, 28, 476-485.
Coghlan, G. R. (1979). The investigation of behavioral self-control theory and techniques in a short-
term treatment of male alcohol abusers. Unpublished doctoral dissertation, State Univer-
sity of New York at Albany. (University Microfilms No. 7918818.)
Corder, B. F., Corder, R. F., & Laidlaw, N. D. (1972). An intensive treatment program for
alcoholics and their wives. Quarterly Journal of Studies on Alcohol, 33, 1144-1146.
Costello, R. M. (1975). Alcoholism treatment and evaluation: In search of methods. Interna-
tional Journal of the Addictions, 10, 251-275.
Cronkite, R. c., & Moos, R. H. (1978). Evaluating alcoholism treatment programs: An
integrated approach. Journal of Consulting and Clinical Psychology, 46, 1105-1119.
Crumbaugh, J. c., & Carr, G. L. (1979). Treatment of alcoholics with logotherapy. Interna-
Czypionka, A., & Demel, I. (1976). Kontrolliertes Trinken bei Alkoholkranken: Versuch
einer Behandlung mit verhaltenstherapeutischen Methoden. Zeitschrift fur klinische
Psychologie, 5, 92-108.
Denson, R., & Sydiaha, D. (1970). A controlled study of LSD treatment in alcoholism and
neurosis. British Journal of Psychiatry, 116, 443-445.
Devenyi, P., & Sereny, G. (1970). Aversive treatment with electro-conditioning for alco-
holism. British Journal of Addiction, 65, 289-292.
Ditman, K. S., & Crawford, G. G. (1966). The use of court probation in the management of
the alcohol addict. American Journal of Psychiatry, 122, 757-762.
Ditman, K. S., Crawford, G. G., Forgy, E. W., Moskowitz, H., & MacAndrew, C. (1967). A
controlled experiment on the use of court probation for drunk arrests. American Journal
of Psychiatry, 124, 160-163.
Edwards, G. (1966). Hypnosis in treatment of alcohol addiction. Quarterly Journal of Studies
on Alcohol, 27, 221-241.
Edwards, G., & Guthrie, S. (1966). A comparison of inpatient and outpatient treatment of
alcohol dependence. Lancet, 1, 467-468.
Edwards, G., & Guthrie, S. (1967). A controlled trial of inpatient and outpatient treatment
of alcohol dependency. Lancet, 1, 555-559.
Edwards, G., Orford, J., Egert, S., Guthrie, S., Hawker, A., Hensman, c., Mitcheson, M.,
Oppenheimer, E., & Taylor, C. (1977). Alcoholism: A controlled trial of "treatment"
and "advice." Journal of Studies on Alcohol, 38, 1004-1031.
Egan, W. P., & Goetz, R. (1968). Effect of metronidazole on drinking by alcoholics. Quar-
terly Journal of Studies on Alcohol, 29, 899-902.
Elkins, R. L. (1980). Covert sensitization treatment of alcoholism: Contributions of suc-
cessful conditioning to subsequent abstinence maintenance. Addictive Behaviors, 5, 67-
89.
Elkins, R. L., & Murdock, R. P. (1977). The contribution of successful conditioning to
abstinence maintenance following covert sensitization (verbal aversion) treatment of
alcoholism. IRCS Medical Science: Psychology & Psychiatry: Social & Occupational Medi-
cine, 5, 167.
Emrick, C. D. (1974). A review of psychologically oriented treatment of alcoholism. I. The
use and interrelationships of outcome criteria and drinking behavior following treat-
ment. Quarterly Journal of Studies on Alcohol, 35, 523-549.
Ends, E. J., & Page, C. W. (1957). A study of three types of group psychotherapy with
hospitalized inebriates. Quarterly Journal of Studies on Alcohol, 18, 263-277.
Ends, E. J., & Page, C. W. (1959). Group therapy and concomitant psychological change.
Psychological Monographs, 73 (Serial No. 480).
Ewing, J. A., & Rouse, B. A. (1976). Failure of an experimental treatment program to
inculcate controlled drinking in alcoholics. British Journal of Addiction, 71, 123-134.
Faia, c., & Shean, G. (1976). Using videotapes and group discussion in the treatment of
male chronic alcoholics. Hospital and Community Psychiatry, 27, 847, 851.
Feinstein, c., & Tamerin, J. S. (1972). Induced intoxication and videotape feedback in
alcoholism treatment. Quarterly Journal of Studies on Alcohol, 33, 408-416.
Ferrell, W. L., & Galassi, J. P. (1981). Assertion training and human relations training in
the treatment of chronic alcoholics. International Journal of the Addictions, 16, 959-968.
Fleiger, D. L., & Zingle, H. W. (1973). Covert sensitization treatment with alcoholics.
Canadian Counsellor, 7, 269-277.
Fox, V., & Smith, M. A. (1959). Evaluation of a chemopsychotherapeutic program for the
rehabilitation of alcoholics: Observations over a two-year period. Quarterly Journal of
Foy, D. W., Nunn, B. L., & Rychtarik, R. G. (1984). Broad-spectrum behavioral treatment
for chronic alcoholics: Effects of training controlled drinking skills. Journal of Consulting
and Clinical Psychology, 52, 218-230.
Freedberg, E. J., & Johnston, W. E. (1978a). The effects of relaxation training within the context
of a multi-modal alcoholism treatment program for employed alcoholics (Substudy No. 988).
Toronto, Ontario: Alcoholism and Drug Addiction Research Foundation.
Freedberg, E. J., & Johnston, W. E. (1978b). The effects of assertion training within the context
of a multi-modal alcoholism treatment program for employed alcoholics (Substudy No. 796).
Toronto, Ontario: Alcoholism and Drug Addiction Research Foundation.
Freedberg, E. J., & Johnston, W. E. (1980). Outcome with alcoholics seeking treatment
voluntarily or after confrontation by their employer. Journal of Occupational Medicine,
22,83-86.
Fuller, R. K., & Roth, H. P. (1979). Disulfiram for the treatment of alcoholism: An evalua-
tion in 128 men. Annals of Internal Medicine, 90, 901-904.
Fuller, R. K., & Williford, W. O. (1980). Life-table analysis of abstinence in a study evaluat-
ing the efficacy of disulfiram. Alcoholism: Clinical and Experimental Research, 4, 298-301.
Gallant, D. M., Bishop, M. P., Camp, E., & Tisdale, C. (1968). A six-month controlled
evaluation of metronidazole (Flagyl) in chronic alcoholic patients. Current Therapeutic
Research, 10, 82-87.
Gallant, D. M., Bishop, M. P., Faulkner, M. A., Simpson, L., Cooper, A., Lathrop, D.,
Brisolara, A. M., & Bossetta, J. R. (1968). A comparative evaluation of compulsory
(group therapy and/or Antabuse) and voluntary treatment of the chronic alcoholic
municipal court offender. Psychosomatics, 9, 306-310.
Gallant, D. M., Bishop, M. P., Mouledoux, A., Faulkner, M. A., Brisolara, A., & Swanson,
W. A. (1973). The revolving-door alcoholic: An impasse in the treatment of the chronic
alcoholic. Archives of General Psychiatry, 28, 633-635.
Gerrein, J. R., Rosenberg, C. M., & Manohar, V. (1973). Disulfiram maintenance in outpa-
tient treatment of alcoholism. Archives of General Psychiatry, 28, 798-802.
Gibbs, L., & Flanagan, J. (1977). Prognostic indicators of alcoholism treatment outcome.
International Journal of the Addictions, 12, 1097-1141.
Glover, J. H., & McCue, P. A. (1977). Electrical aversion therapy with alcoholics: A com-
parative follow-up study. British Journal of Psychiatry, 130, 279-286.
Gottheil, E., McLellan, A. T., & Druley, K. A. (Eds.), (1981). Matching patient needs and
treatment methods in alcoholism and drug abuse. Springfield, IL: Charles C Thomas.
Goyer, P. F., & Major, L. F. (1979). Hepatotoxicity in disulfiram-treated patients. Journal of
Hagen, R. E., Williams, R. L., McConnell, E. J., & Fleming, C. W. (1978). An evaluation of
alcohol abuse treatment as an alternative to drivers license suspension or revocation (Report
No. CAL-DMV-RSS-78-68 to Legislature of State of California).
Hallam, R., Rachman, S., & Falkowski, W. (1972). Subjective, attitudinal and physiological
effects of electrical aversion therapy. Behaviour Research and Therapy, 10, 1-13.
Heather, N., & Robertson, I. (1983). Controlled drinking (rev. ed.). London: Methuen.
Hedberg, A. G., & Campbell, L. M. (1974). A comparison of four behavioral treatment
approaches to alcoholism. Journal of Behavior Therapy and Experimental Psychiatry,S,
251-256.
Hoff, E. C. (1961). The use of pharmacological adjuncts in the psychotherapy of alcoholics.
Quarterly Journal of Studies on Alcohol, Supplement No.1, 138-150.
Hoff, E. c., & McKeown, C. E. (1953). An evaluation of the use of tetraethylthiuram
disulfide in the treatment of 560 cases of alcohol addiction. American Journal of Psychia-
try, 109, 670-673.
Hollister, L. E., Shelton, J., & Krieger, G. (1969). A controlled comparison of lysergic acid
diethylamide (LSD) and dextroamphetamine in alcoholics. American Journal of Psychia-
try, 125, 1352-1357.
Hunt, G. M., & Azrin, N. H. (1973). A community-reinforcement approach to alcoholism.
Behaviour Research and Therapy, 11, 91-104.
Hussain, M. Z., & Harinath, M. (1972). Helping alcoholics abstain: An implantable sub-
stance. American Journal of Psychiatry, 129, 363.
Jackson, P., & Oei, T. P. S. (1978). Social skills training and cognitive restructuring with
alcoholics. Drug and Alcohol Dellendence, 3, 369-374.
Jackson, T. R., & Smith, J. W. (1978). A comparison of two aversion treatment methods for
alcoholism. Journal of Studies on Alcohol, 39, 187-191.
Jacobson, N. 0., & Silfverskiold, N. P. (1973). A controlled study of a hypnotic method in
the treatment of alcoholism, with evaluation by objective criteria. British Journal of
Addiction, 68, 25-31.
Jensen, S. E., & Ramsay, R. (1963). Treatment of chronic alcoholism with lysergic acid
diethylamide. Canadian Psychiatric Association Journal, 8, 182-188.
Johnson, F. G. (1970). A comparison of short-term treatment effects of intravenous sodium
amytal-methedrine and LSD in the alcoholic. Canadian Pyschiatric Association Journal,
15, 493-497.
Kinder, B. N., Pape, N. E., & Walfish, S. (1980). Drug and alcohol education programs: A
review of outcome studies. International Journal of the Addictions, 15, 1035-1054.
Kissin, B., & Gross, M. M. (1968). Drug therapy in alcoholism. American Journal of Psychia-
try, 125, 31-41.
Kissin, B., Platz, A., & Su, W. H. (1970). Social and psychological factors in the treatment
of chronic alcoholism. Journal of Psychiatric Research, 8, 13-27.
Kissin, B., Platz, A., & Su, W. H. (1971). Selective factors in treatment choice and outcome
in alcoholics. In N. K. Mello & J. H. Mendleson (Eds.), Recent advances in studies of
alcoholism (pp. 781-802). Washington, DC: U.S. Government Printing Office.
Kline, N. S., Wren, J. c., Cooper, T. B., Varga, E., & Canal, O. (1974). Evaluation of
lithium therapy in chronic and periodic alcoholism. American Journal of Medical Science,
268, 15-22.
Knowles, P. L. (1983). Inpatient versus outpatient treatment of substance misuse in hospi-
tals, 1975-1980. Journal of Studies on Alcohol, 44, 384-387.
Kristenson, H. (1983). Studies on alcohol related disabilities in a medical intervention programme

in middle-aged males (2nd ed.). Skurup, Sweden: Lindbergs Blankett.
Kristenson, H., Ohlin, H., Hulten-Nosslin, M-B, TrelI, E., & Hood, B. (1983). Identifica-
tion and intervention of heavy drinking in middle-aged men: Results and follow-up of
24-60 months of long-term study with randomized controls. Alcoholism: Clinical and
Experimental Research, 7, 203-209.
Kwentus, J., & Major, L. F. (1979). Disulfiram in the treatment of alcoholism. Journal of
Lake, C. R., Major, L. F., Ziegler, M. G., & Kopin, I. J. (1977). Increased sympathetic
nervous system activity in alcoholic patients treated with disulfiram. American Journal
La!, S. (1969). Metronidazole in the treatment of alcoholism: A clinical trial and review of
the literature. Quarterly Journal of Studies on Alcohol, 30, 140-151.
Lanyon, R. I., Primo, R. V., Terrell, F., & Wener, A. (1972). An aversion-desensitization
treatment for alcoholism. Journal of Consulting and Clinical Psychology, 38, 394-
398.
Laverty, S. G. (1966). Aversion therapies in the treatment of alcoholism. Psychosomatic
Medicine, 28, 651-666.
Levinson, T., & Sereny, G. (1969). An experimental evaluation of "insight therapy" for the
chronic alcoholic. Canadian Psychiatric Association Journal, 14, 143-146.
Levy, M. S., Livingstone, B. L., & Collins, D. M. (1967). A clinical comparison of dis-
ulfiram and calcium carbimide. American Journal of Psychiatry, 123, 1018-1022.
Lieberman, M. A., Yalom, I. D., & Miles, M. B. (1973). Encounter groups: First facts. New
York: Basic Books.
Liebson, I., Bigelow, G., & Flamer, R. (1973). Alcoholism among methadone patients: A
specific treatment method. American Journal of Psychiatry, 130, 483-485.
Lijinski, W. (1979). N-nitrosamines as environmental carcinogens. In J-p Anselme (Ed.),
N-nitrosamines (American Chemical Society Symposium Series, No. 101). Washington,
DC: American Chemical Society.
Longabaugh, R., McCrady, B., Fink, E., Stout, R., McAuley, T., Doyle, c., & McNeill, D.
(1983). Cost-effectiveness of alcoholism treatment in partial vs. inpatient settings: Six-
month outcomes. Journal of Studies on Alcohol, 44, 1049-1071.
Lovibond, S. H. (1975). Use of behavior modification in the reduction of alcohol-related
road accidents. In T. Thompson & W. S. Dockens, 3rd (Eds.), Applications of behavior
modification. New York: Academic Press, pp. 399-406.
Lovibond, S. H., & Caddy, G. (1970). Discriminated aversive control in the moderation of
alcoholics' drinking behavior. Behavior Therapy, 1, 437-444.
Lowenstam, I. (1969). Metronidazole and placebo in the treatment of chronic alcoholism: A
comparative study. Psychosomatics, 10, 43-45.
Lowenstam, I. (1975). Drug treatment in a rehabilitation center for chronic alcoholics.
Journal of Chronic Diseases, 28, 431-434.
Ludwig, A., Levine, J., Stark, L., & Lazar, R. (1969). A clinical study of LSD treatment in
alcoholism. American Journal of Psychiatry, 126, 59-69.
Ludwig, A. M., Levine, J., & Stark, L. H. (1970). LSD and alcoholism. Springfield, IL:
Charles C Thomas.
Lunde, S. E., Johnson, G. R., & Martin, P. L. (1970). Electrical aversion conditioning with
chronic alcoholics. Journal of Consulting and Clinical Psychology, 34, 302-307.
Lysloff, G. O. (1972). Anti-addictive chemotherapy-metronidazole and alcohol aversion.
British Journal of Addiction, 67, 239-244.
MacDonough, T. S. (1976). Evaluation of the effectiveness of intensive confrontation in
changing the behavior of alcohol and drug abusers. Behavior Therapy, 7, 408-409.
Maletzky, B. M. (1974). Assisted covert sensitization for drug abuse. International Journal of
the Addictions, 9, 411-429.
Malfetti, J. L. (1975). Reeducation and rehabilitation of the drunken driver. Journal of Drug
Issues, 5, 255-269.
Mallams, J. H., Godley, M. D., Hall, G. M., & Meyers, R. J. (1982). A social-systems
approach to resocializing alcoholics in the community. Journal of Studies on Alcohol, 43,
1115-1123.
Marlatt, G. A. (1973, April). A comparison of aversive conditioning procedures in the treatment of
alcoholism. Paper presented at the annual meeting of the Western Psychological Asso-
ciation, Anaheim, CA.
Maxwell, W., Baird, R. L., Wezl, T., & Ferguson, L. (1974). Discriminated aversion condi-
tioning within an alcoholic treatment program in the training of controlled drinking.
Behavioral Engineering, 2, 17-19.
McCance, c., & McCance, P. F. (1969). Alcoholism in North-East Scotland: Its treatment
and outcome. British Journal of Psychiatry, 115, 189-198.
McCrady, B. S., Paolino, T. J., Jr., Longabough, R., & Rossi, J. (1979). Effects of joint
hospital admission and couples treatment for hospitalized alcoholics: A pilot study.
McCrady, B. S., Moreau, J., Paolino, T. J., Jr., & Longabaugh, R. (1982). Joint hospitaliza-
tion and couples therapy for alcoholism: A four-year follow-up. Journal of Studies of
Alcohol, 43, 1244-1250.
McGuire, F. L. (1978). The effectiveness of a treatment program for alcohol-involved
drivers in America. American Journal of Drug and Alcohol Abuse,S, 517-525.
McLachlan, J. F. C. (1972). Benefit from group therapy as a function of patient-therapist
match on conceptual level. Psychotherapy: Theory, Research and Practice, 9, 317-323.
McLachlan, J. F. C. (1974). Therapy strategies, personality orientation and recovery from
alcoholism. Canadian Psychiatric Association Journal, 19, 25-30.
McLachlan, J. F. c., & Stein, R. L. (1982). Evaluation of a day clinic for alcoholics. Journal of
Studies on Alcohol, 43, 26l-272.
McLellan, A. T., O'Brien, C. P., Kron, R., Alterman, A. I., & Druley, K. A. (1980).
Matching substance abuse patients to appropriate treatment: A conceptual and meth-
odological approach. Drug and Alcohol Dependence,S, 189-195.
McLellan, A. T., Woody, G. E., Luborsky, L., O'Brien, C. P., & Druley, K. A. (1983).
Increased effectiveness of substance abuse treatment: A prospective study of patient-
treatment "matching." Journal of Nervous and Mental Disease, 171, 597-605.
McMillan, T. M. (1981). Lithium and the treatment of alcoholism: A critical review. British
Merry, J., & Whitehead, A. (1968). Metronidazole and alcoholism. British Journal of Psychia-
try, 114, 859-861.
Merry, J., Reynolds, C. M., Bailey, J., & Coppen, A. (1976). Prophylactic treatment of
alcoholism by lithium carbonate: A controlled study. Lancet, 2, 481-482.
Miller, P. M., Hersen, M., Eisler, R. M., & Hemphill, D. P. (1973). Electrical aversion
therapy with alcoholics: An analogue study. Behaviour Research and Therapy, 11, 491-
497.
Miller, W. R. (1978). Behavioral treatment of problem drinkers: A comparative outcome
study of three controlled drinking therapies. Journal of Consulting and Clinical Psychol-
ogy, 46, 74-86.
Miller, W. R. (1983). Motivational interviewing with problem drinkers. Behavioural Psycho-
therapy, 11, 147-172.
Miller, W. R. (1985). Motivation for treatment: A review. Psychological Bul/etin, 98, 84-107.
Miller, W. R. (in press). Haunted by the Zeitgeist: 'Reflections on contrasting treatment
goals and conceptions of alcoholism in Europe and the United States. Annals of the New
York Academy of Sciences.
Miller, W. R., & Baca, L. M. (1983). Two-year follow-up of bibliotherapy and therapist-
directed controlled drinking training for problem drinkers. Behavior Therapy, 14,441-
448.
Miller, W. R, & Caddy, G. R. (1977). Abstinence and controlled drinking in the treatment
of problem drinkers. Journal of Studies on Alcohol, 38, 986-1003.
Miller, W. R, & Dougher, M. J. (1984, June). Covert sensitization; Alternative treatment
approaches for alcoholics. Paper presented at the Second Congress of the International
Society for Biomedical Research on Alcoholism, Santa Fe, NM.
Miller, W. R, & Hester, R K. (1980). Treating the problem drinker: Modern approaches.
In W. R Miller (Ed.), The addictive behaviors: Treatment of alcoholism, drug abuse, smoking,
and obesity (pp. 11-141). Oxford: Pergamon Press.
Miller, W. R, & Joyce, M. A. (1979). Prediction of abstinence, controlled drinking, and
heavy drinking outcomes following behavioral self-control training. Journal of Consult-
ing and Clinical Psychology, 47, 773-775.
Miller, W. R, & Munoz, R F. (1982). How to control your drinking (rev. ed.). Albuquerque,
NM: University of New Mexico Press.
Miller, W. R, & Saucedo, C. F. (1983). Assessment of neuropsychological impairment and
brain damage in problem drinkers. In C. J. Golden, J. A. Moses, Jr., J. A. Coffman, W.
R. Miller, & F. D. Strider (Eds.), Clinical neuropsychology: Interface with neurologic and
psychiatric disorders (pp. 141-195). New York: Grune & Stratton.
Miller, W. R, & Taylor, C. A. (1980). Relative effectiveness of bibliotherapy, individual
and group self-control training in the treatment of problem drinkers. Addictive Behav-
iors, 5, 13-24.
Miller, W. R, Taylor, C. A., & West, J. C. (1980). Focused versus broad-spectrum behavior
therapy for problem drinkers. Journal of Consulting and Clinical Psychology, 48, 590-601.
Miller, W. R., Gribskov, C. J., & Mortell, R. L. (1981). Effectiveness of a self-control
manual for problem drinkers with and without therapist contact. International Journal
of the Addictions, 16, 1247-1254.
Miller, W. R, Hedrick, K. A., & Taylor, C. A. (1983). Addictive behaviors and life prob-
403-412.
Mindlin, D. F. (1965). Group therapy for alcoholics: A study of the attitude and behavior
changes in relation to perceived group norms. Dissertation Abstracts, 26, 2323-2324.
Mooney, H. B., Ditman, K. S., & Cohen, S. (1961). Chlordiazepoxide in the treatment of
alcoholics. Diseases of the Nervous System, 22, (Supplement), 44-51.
Moore, R. A. (1977). Ten years of inpatient programs for alcoholic patients. American
Journal of Psychiatry, 134, 542-545.
Mosher, V., Davis, J., Mulligan, D., & Iber, F. L. (1975). Comparison of outcome in a 9-day
and 30-day alcoholism treatment program. Journal of Studies on Alcohol, 36, 1277-1281.
Mottin, J. L. (1973). Drug-induced attenuation of alcohol consumption: A review and
evaluation of claimed, potential or current therapies. Quarterly Journal of Studies on
Alcohol, 34, 444-472.
abitz, F. W. (1978). Control orientation and disulfiram. Journal of Studies on Alcohol, 39,
1297-1298.
Oei, T. P. S., & Jackson, P. (1980). Long-term effects of group and individual social skills
training with alcoholics. Addictive Behaviors, 5, 129-136.
O'Farrell, T., & Cutter, H. S. (in press). Evaluating behavioral therapy for alcoholics:
Procedures & preliminary results. In L. A. Hamerlynck (Ed.), Essentials of Behavior
Treatments for Families. New York: Brunner/Mazel.
O'Farrell, T. J., Cutter, H. S. G., & Floyd, F. J. (1984). Evaluating behavioral marital
therapy for male alcoholics: Effects on marital adjustment and communication before
to after treatment. Behavior Therapy, 15.
Ogborne, A. c., & Glaser, F. B. (1981). Characteristics of affiliates of Alcoholics Anony-
mous: A review of the literature. Journal of Studies on Alcohol, 42, 661-675.
Ogborne, A. c., & Wilmot, R. (1979). Evaluation of an experimental counseling service for
male Skid Row alcoholics. Journal of Studies on Alcohol, 40, 129-132.
O'Leary, M. R., Rohsenow, D. J., & Donovan, D. M. (1976). Locus of control and attrition
from an alcoholism treatment program. Journal of Consulting and Clinical Psychology, 44,
686-687.
O'Leary, M. R., Donovan, D. M., Chaney, E. F., & Walker, D. (1979). Cognitive impair-
ment and treatment outcome with alcoholics: Preliminary findings. Journal of Clinical
Psychiatry, 40, 397-398.
Olson, R. P., Ganley, R., Devine, V. T., & Dorsey, G. c., Jr. (1981). Long-term effects of
behavioral verses insight-oriented therapy with inpatient alcoholics. Journal of Consult-
Orford, J., Oppenheimer, E., & Edwards, G. (1976). Abstinence or control: The outcome
for excessive drinkers two years after consultation. Behaviour Research and Therapy, 14,
409-418.
Page, R. D., & Schaub, L. H. (1979). Efficacy of a three- versus a five-week alcohol
treatment program. International Journal of the Addictions, 14, 697-714.
Parker, M. W., Winstead, D. K., & Willi, F. J. P. (1979). Patient autonomy in alcohol
rehabilitation. I. Literature Review. International Journal of the Addictions, 14, 1015-
1022.
Pattison, E. M., Brissenden, A., & Wohl, T. (1967). Assessing specific effects of inpatient
group psychotherapy. International Journal of Group Psychotherapy, 17, 283-297.
Pendery, M. L., Maltzman, I. M., & West, L. J. (1982). Controlled drinking by alcoholics?
New findings and a reevaluation of a major affirmative study. Science, 217, 169-175.
Penick, S. B., Carrier, R. N., & Sheldon, J. B. (1969). Metronidazole in the treatment of
alcoholism. American Journal of Psychiatry, 125, 1063-1066.
Penick, S. B., Sheldon, J. B., Templer, D. I., & Carrier, R. N. (1971). Four year follow-up of
metronidazole treatment program for alcoholism. Industrial Medicine, 40, 30-32.
Penk, W. E., Charles, H. L., & Van-Hoose, T. A. (1978). Comparative effectiveness of day
hospital and inpatient psychiatric treatment. Journal of Consulting and Clinical Psychol-
ogy, 46, 94-101.
Piorkowski, G. K., & Mann, E. T. (1975). Issues in treatment efficacy research with alco-
holics. Perceptual and Motor Skills, 41, 695-700.
Pittman, D. J., & Tate, R. L. (1972). A comparison of two treatment programs for alco-
holics. International Journal of Social Psychiatry, 18, 183-193.
Platz, A., Panepinto, W. c., Kissin, 8., & Charnoff, S. M. (1970). Metronidazole and
alcoholism: An evaluation of specific and non specific factors in drug treatment.
Diseases of the Nervous System, 31, 631-636.
Polich, J. M., Armor, D. J., & Braiker, H. B. (1981). The course of alcoholism: Four years after
treatment. New York: Wiley.
Pomerleau, 0., Pertshuk, M., Adkins, D., & d' Aquili, E. (1978). Treatment for middle
income problem drinkers. In P. E. Nathan, G. A. Marlatt, & T. L",berg (Eds.), Alco-
holism: New directions in behavioral research and treatment (pp. 143-160). New York:
Plenum Press.
Pond, S. M., Becker, C. E., Vandervoort, R., Phillips, M., Bowler, R., & Peck, C. C. (1981).
An evaluation of the effects of lithium in the treatment of chronic alcoholism. I.
Clinical results. Alcoholism: Clinical and Experimental Research,S, 247-251.
Popham, R. E., & Schmidt, W. (1976). Some factors affecting the likelihood of moderate
drinking by treated alcoholics. Journal of Studies on Alcohol, 37, 868-882.
Regester, D. C. (1971). Changes in autonomic responsivity and drinking behavior of
alcoholics as a function of aversion therapy. Dissertation Abstracts International, 32,
1225-B.
Reinert, R E. (1958). A comparison of reserpine and disulfiram in the treatment of alco-
holism. Quarterly Journal of Studies on Alcohol, 19, 617-622.
Richard, G. P. (1983). Behavioural treatment of excessive drinking. Unpublished doctoral dis-
sertation, University of New South Wales.
Robson, R A. H., Paulus, I., & Clarke, G. G. (1965). An evaluation of the effects of a clinic
treatment program on the rehabilitation of chronic alcoholic patients. Quarterly Journal
of Studies on Alcohol, 26, 264-278.
Rosenberg, C. M. (1974). Drug maintenance in the outpatient treatment of chronic alco-
holism. Archives of General Psychiatry, 30, 373-377.
Rosenberg, C. M., & Liftik, J. (1976). Use of coercion in the outpatient treatment of
Rosenberg, C. M., Gerrein, J. R, & Schnell, C. (1978). Cannabis in the treatment of
Rosenberg, S. D. (1979). Relaxation training and a differential assessment of alcoholism. Un-
published doctoral dissertation, California School of Professional Psychology, San
Diego, CA. (University Microfilms No. 8004362).
Rotter, J. B. (1966). Generalized expectancies for internal versus external control of rein-
forcement. Psychological Monographs, 80, (Whole No. 609).
Rydzynski, Z., & Gruszczynski, W. (1980). Treatment of alcoholism with psychotomimetic
drugs: A follow-up study. Journal of Studies on Alcohol, 41, CL846. (Abstract).
Salzberg, P. M., & Klingberg, C. L. (1983). The effectiveness of deferred prosecution for
driving while intoxicated. Journal of Studies on Alcohol, 44, 299-306.
Sanchez-Craig, M., Annis, H. M., Bornet, A. R, & MacDonald, K. R. (1984). Random
assignment to abstinence and controlled drinking: Evaluation of a cognitive-be-
havioural program for problem drinkers. Journal of Consulting and Clinical Psychology,
52, 390-403.
Schaefer, H. H. (1972). Twelve-month follow-up of behaviorally trained ex-alcoholic social
drinkers. Behavior Therapy, 3, 286-289.
Schaefer, H. H., Sobell, M. B., & Mills, K. C. (1971). Some sobering data on the use of self-
confrontation with alcoholics. Behavior Therapy, 2, 28-39.
Schaefer, H. H., Sobell, M. B., & Sobell, L. C. (1972). Twelve month follow-up of hospi-
talized alcoholics given self-confrontation experiences by video-tape. Behavior Therapy,
3,283-285.
Schmidt, M. R. (1978). Structuring treatment programs on the basis of control orientation of
alcoholics. Unpublished doctoral dissertation, University of Nebraska-Lincoln. (Uni-
versity Microfilms No. 7814709).
Scoles, P., & Fine, E. W. (1977). Short-term effects of an educational program for drinking
drivers. Journal of Studies on Alcohol, 38, 633-637.
Shaffer, J. W., Freinek, W. R, Wolf, S., Foxwell, N. H., & Kurland, A. A. (1963). A
controlled evaluation of chlordiazepoxide (Librium) in the treatment of alcoholics.
Journal of Nervous and Mental Disease, 137, 494-507.
Shaffer, J. W., Freinek, W. R, Wolf, S., Foxwell, N. H., & Kurland, A. A. (1964). Replica-
tion of a study of nialamide in the treatment of convalescing alcoholics with emphasis
on prediction of response. Current Therapeutic Research, 6, 521-531.
Shaw, J. A., Donley, P., Morgan, D. W., & Robinson, J. A. (1975). Treatment of depression
in alcoholics. American Journal of Psychiatry, 132, 641-644.
Sisson, R. W. (1981). The effect of three relaxation procedures on tension reduction and subsequent
drinking of inpatient alcoholics. Unpublished doctoral dissertation, Southern Illinois Uni-
versity at Carbondale. (University Microfilms No. 8122668).
Smart, R. G. (1974). Employed alcoholics treated voluntarily and under constructive coer-
cion: A follow-up study. Quarterly Journal of Studies on Alcohol, 35, 196-209.
Smart, R. G. (1978). Characteristics of alcoholics who drink socially after treatment. Alco-
holism: Clinical and Experimental Research, 2, 49-52.
Smart, R. G., Finley, J., & Funston, R. (1977). The effectiveness of post-detoxification
referrals: Effects on later detoxification admissions, drunkenness and criminality.
Drug and Alcohol Dependence, 2, 149-155.
Smart, R. G., & Gray, G. (1978). Minimal, moderate and long-term treatment for alco-
holism. British Journal of Addiction, 73, 35-38.
Smart, R. G., Storm, T., Baker, E. F. W., & Solursh, L. (1966). A controlled study of
Iysergide in the treatment of alcoholism. I. The effects on drinking behavi~r. Quarterly
Smith-Moorhouse, P. M. (1969). Hypnosis in the treatment of alcoholism. British Journal of
Sobell, M. B., & Sobell, L. C. (1973). Individualized behavior therapy for alcoholics. Behav-
ior Therapy, 4, 49-72.
Stalonas, P. M., Keane, T. M., & Foy, D. W. (1979). Alcohol education for inpatient
alcoholics: A comparison of live, videotape and written presentation modalities. Ad-
dictive Behaviors, 4, 223-229.
Stein, L. I., Newton, J. R., & Bowman, R. S. (1975). Duration of hospitalization for alco-
holism. Archives of General Psychiatry, 32, 247-252.
Stinson, D. J., Smith, W. G., & Kaplan, J. M. (1979). Systems of care and treatment
outcomes for alcoholic patients. Archives of General Psychiatry, 36, 535-539.
Storm, T., & Cutler, R. E. (1969). Systematic desensitization in the treatment of alcoholics: An
experimental trial. Unpublished manuscript, Alcoholism Foundation of British
Columbia.
Swenson, P. R., & Clay, T. R. (1980). Effects of short-term rehabilitation on alcohol con-
sumption and drinking-related behaviours: An eight month follow-up study of drunk-
en drivers. International Journal of the Addictions, 15, 821-838.
Swinson, R. P. (1971). Long term trial of metronidazole in male alcoholics. British Journal of
Taylor, J. A. T. (1964). Metronidazole: A new agent for combined somatic and psychic
therapy of alcoholism. A case study and preliminary report. Bulletin of the Los Angeles
Neurological Society, 29, 158-162.
Tomsovic, M. (1970). A follow-up study of discharged alcoholics. Hospital and Community
Tomsovic, M. (1976). Group therapy and changes in the self-concept of alcoholics. Journal
of Studies on Alcohol, 37, 53-57.
Tomsovic, M., & Edwards, R. V. (1970). Lysergide treatment of schizophrenic and non-
schizophrenic alcoholics: A controlled evaluation. Quarterly Journal of Studies on Alco-
hol, 31, 932-949.
Turek, I. S., Ota, K., Brown, c., Massari, F., & Kurland, A. A. (1973). Tiotixene and
thioridazine in alcoholism treatment. Quarterly Journal of Studies on Alcohol, 34, 853-
859.
Tyndel, M., Fraser, J. G., & HartIeib, C. J. (1969). Metronidazole as an adjuvant in the
treatment of alcoholism. British Journal of Addiction, 64, 57-61.
Uecker, A. E., & Boutilier, L. R. (1976). Alcohol education for alcoholics: Relation to
attitude changes and posttreatment abstinence. Journal of Studies on Alcohol, 37, 965-
975.
Uecker, A. E., & Solberg, K. B. (1973). Alcoholics' knowledge about alcohol problems: Its
relationship to significant attitudes. Quarterly Journal of Studies on Alcohol, 34, 509-513.
Van Thiel, D. H., Gavaler, J. S., Paul, G. M., & Smith, W. I. (1979). Disulfiram induced
disturbances in hypothalamic-pituitary function. Alcoholism: Clinical and Experimental
Research, 3, 230-234.
Vogler, R E., Compton, J. V., & Weissbach, T. A. (1975). Integrated behavior change
techniques for alcoholism. Journal of Consulting and Clinical Psychology, 43, 233-243.
Vogler, R. E., Lunde, S. E., & Martin, P. L. (1971). Electrical aversion conditioning with
chronic alcoholics: Follow-up and suggestions for research. Journal of Consulting and
Clinical Psychology, 36, 450.
Vogler, R E., Weissbach, T. A, & Compton, J. V. (1977). Learning techniques for alcohol
abuse. Behaviour Research and Therapy, 15, 31-38.
Vogler, R E., Weissbach, T. A, Compton, J. V., & Martin, G. T. (1977). Integrated
behavior change techniques for problem drinkers in the community. Journal of Consult-
Walker, R D., Donovan, D. M., Kivlahan, D. R, & O'Leary, M. R (1983). Length of stay,
neuropsychological performance, and aftercare: Influences on alcohol treatment out-
come. Journal of Consulting and Clinical Psychology, 51, 900-911.
Wallerstein, R S., Chotios, J. W., Friend, M. B., Hammersley, D. W., Periswig, E. A., &
Winship, G. M. (1957). Hospital treatment of alcoholism: A comparative experimental study.
New York: Basic Books.
West, P. T. (1979). Three modes of training alcoholics in interpersonal communications skills: A
comparative study. Unpublished doctoral dissertation, University of Western Ontario.
Whyte, C. R, & O'Brien, P. M. (1974). Disulfiram implant: A controlled trial. British Journal
Willems, P. J. A., Letemendia, F. J. J., & Arroyave, F. (1973). A two-year follow-up study
comparing short with long stay in-patient treatment of alcoholics. British Journal of
Wilson, A (1979). Patient management in disulfiram implant therapy. Canadian Journal of
Wilson, A., Davidson, W. J., & White, J. (1976). Disulfiram implantation: Placebo, psycho-
logical deterrent and pharmacological deterrent effects. British Journal of Psychiatry,
129, 277-280.
Wilson, A, White, J., & Lange, D. E. (1978). Outcome evaluation of a hospital-based
alcoholism treatment programme. British Journal of Addiction, 73, 39-45.
Wilson, A, Davidson, W. J., Blanchard, R., & White, J. (1978). Disulfiram implantation: A
placebo-controlled trial with two-year follow-up. Journal of Studies on Alcohol, 39, 809-
819.
Winokur, G., Reich, T., Rimmer, J., & Pitts, F. N., Jr. (1970). Alcoholism. III. Diagnosis
and familial psychiatric illness in 259 alcoholic probands. Archives of General Psychia-
try, 23, 104-111.
Zimberg, S. (1974). Evaluation of alcoholism treatment in Harlem. Quarterly Journal of
8
Matching Problem Drinkers With
Optimal Treatments
WILLIAM R. MILLER AND REID K. HESTER
It seems the most obvious of commonsense assertions: that individuals

with varying needs and characteristics will respond optimally to differ-
ent kinds of intervention, and therefore that clients should be matched
with optimal approaches rather than all being treated in the same way.
Indeed many if not most alcoholism treatment programs give lip service
to the need for individually tailoring intervention programs. Yet this
apparently simple and uncontentious assertion is, in fact, a highly com-
plicated and interesting issue which-if taken seriously-has important
research and clinical implications that are not only controversial, but
potentially revolutionary, at least for alcoholism treatment practices as
they currently exist in the United States.
UNDIFFERENTIATED TREATMENT: THE STATUS QUO
The typical alcoholism program, although claiming to individualize

intervention, in fact affords few choices and offers a rather standard set
of treatment procedures in which all individuals participate (Costello,
1975; Orford & Hawker, 1974). Where alternatives exist, assignment
WILLIAM R. MILLER • Department of Psychology, University of New Mexico, Albuquer-

que, NM 87131. REID K. HESTER • Behavior Therapy Associates, Albuquerque, NM
87110. Portions of this chapter were presented at the Third International Conference on
Treatment of Addictive Behaviors, North Berwick, Scotland, August, 1984.
175
appears to be unrelated to patient characteristics (Bromet, Moos,

Wuthmann, & Bliss, 1977; Finney & Moos, 1979; Gibbs, 1980, 1981;
Hague, Donovan, & O'Leary, 1976; Hansen & Emrick, 1983; Martin,
1979; Smart, 1978), and even when differential diagnosis is made at
intake it is often ignored during treatment (Schmidt, Smart, & Moss,
1968). The best predictor of the treatment offered to an individual ap-
pears to be the type of place to which the person goes for evaluation:
each program tends to recommend its own services (Hansen & Emrick,
1983).
Still more perplexing is the fact that the standard elements of a
typical United States alcoholism treatment program (detoxification, alco-
holism education, A.A. meetings, group confrontation therapy, and dis-
ulfiram) individually and collectively lack adequate experimental sup-
port of effectiveness, whereas other approaches with well-documented
efficacy remain largely unused (see Chapter 7, this volume). There is little
convincing evidence that the effectiveness of this "standard formula"
alcoholism treatment significantly exceeds spontaneous remission
or minimal intervention (Edwards et al., 1977). The status quo, then,
appears to be undifferentiated and at best modestly effective treat-
ment.
One plausible explanation for mediocre outcomes is the absence of
appropriate client-treatment matching. Imagine a physician whose min-
istrations were limited to prescribing a single broad-spectrum antibiotic
and offering reassurance. For some patients (those with conditions ap-
propriately treated with this antibiotic, or who respond well to placebo)
the treatment would be wonderfully effective, but for those with other
serious maladies (or even infections requiring a different antibiotic) the
physician's efforts would be to no avail. The failure here is not in provid-
ing a totally ineffective treatment, but rather in the absence of differen-
tial diagnosis and alternative efficacious interventions.
THE MATCHING HYPOTHESIS
The underlying premise of a matching strategy is a hypothesis: that

clients who are appropriately matched to treatment will show superior
outcome relative to those who are unmatched or mismatched. In experi-
mental-design terms, this hypothesis predicts an interaction effect even
in the absence of main effects of treatment or predictor variables. This
relationship is illustrated in Figure I, which displays the efficacy of two
hypothetical alcoholism treatment approaches: A and B. The horizontal
axis represents a hypothetical client-predictor variable, ranging from
"low" to "high" (imagine, for example, that it represents severity of
MATCHING PROBLEM DRINKERS WITH OPTIMAL TREATMENTS 177
POSITIVE Study 1: A<8

Study 2: A=8
OUTCOME Study 3: A>8
A
NEGATIVE *1 2 3
OUTCOME
lOW HIGH
level of Client Predictor Variable
FIGURE 1. A hypothetical client-by-treatment interaction.
dependence or socioeconomic status). The relative effectiveness of treat-

ments A and B found in a given study will depend on the population
being treated. In Study I, a "low" population is treated, and B is found
to be superior to A. In Study 2, with a "medium" population, no dif-
ferences are found, and in Study 3 ("high") A exceeds B in effectiveness.
Note that if a full range of clients were included in the population stud-
ied and were all averaged together, no substantial differences between
A and B would be found, leading to the mistaken conclusion that treat-
ments A"a~d B are equivalent.
The failure to take matching into account, then, may explain the
typical findings across studies in the alcoholism treatment field: either
no significant differences, or inconsistent differences. If the matching
hypothesis is correct, then knowledge and clinical efficacy will be signifi-
cantly advanced by determining which treatments are optimal for which
types of clients. The alternative (null) hypothesis is that client charac-
teristics do not differentially affect the outcome of various alcoholism
treatment methods.
A variety of rational schemes for matching have been proposed
(Brown & Lyons, 1981; Ewing, 1977; Gibbs, 1980; Glaser, 1980; Gottheil,
McLellan, & Druley, 1981; O'Leary, Donovan, Chaney, & O'Leary,
1980). For the matching hypothesis to be testable, however, its elements
must be operationalized. What are the "treatment" elements to be com-
pared? These might be alternative methods (e.g., aversion therapy vs.

disulfiram), goals (e.g., moderation vs. abstinence), settings (e.g., inpa-
tient vs. outpatient), or therapists (e.g., peer vs. professional). The crite-
ria for success also must be specified. Some studies have relied upon
patient ratings of helpfulness as the criterion of efficacy (Obitz, 1975;
Price & Curlee-Salisbury, 1975), but more direct measures of interven-
tion impact are desirable. One could predict superior motivation for treat-
ment with matching (e.g., more likely to accept, continue in, and com-
ply with treatment; Miller, 1985b). Matching might also improve the
effectiveness of treatment (e.g., reduction in drinking and symptoms, or
increased duration of sobriety), or the efficiency of treatment (less costly
interventions, fewer relapses and less need for renewed treatment).
Adequate periods of follow-up and measures of effectiveness must be
employed. Price and Curlee-Salisbury (1975), for example, reported dif-
ferential predictors of therapeutic success with alcoholics, but used as
their outcome measure the patients' immediate postdischarge percep-
tions of how helpful various elements of a standard program had been.
The design therefore lacked differential treatment, follow-up, and objec-
tive measures of therapeutic impact-all of which are important condi-
tions of an adequate matching research strategy.
RESEARCH STRATEGIES
Before proceeding to a review of current knowledge on matching,

we wish to distinguish three research strategies for studying matching
effects. We will term these the predictor, differential, and modeling
approaches.
PREDICTOR ApPROACH
By far the largest number of studies reporting data relevant to

matching have employed a predictor strategy. In this method a single
treatment approach is studied, and individual difference variables are
examined for their ability to predict outcome following this particular
intervention. The design is correlational, and appropriate statistical
methods include multiple regression, discriminant function analysis,
and canonical correlation-for continuous, nominal, and multiple con-
tinuous outcome variables, respectively (Harris, 1985).
Knowledge contributed by predictor designs is valuable as it accu-
mulates over time. Each study suggests predictors of effectiveness of a
particular kind of treatment. To the extent that such findings are repli-
cated and yield consistency across studies, one can begin to derive
promising schemas for matching clients with interventions.
DIFFERENTIAL ApPROACH
A weakness of the predictor approach is its failure to demonstrate

directly that alternative treatments differ with regard to predictors of
outcome. The differential approach remedies this fault by comparing two
or more types of treatment within the same population and study. Ide-
ally clients are assigned at random to alternative interventions, out-
comes are determined, and then statistical analyses are performed to
compare the characteristics of "successes" within each treatment. Cli-
ents may be blocked before treatment on one or more relevant variables,
or post hoc analyses may be performed to examine the predictive valid-
ity of multiple client variables. Success profiles can be developed for
each individual treatment by using statistical methods described above,
or treatment assignment can be regarded as one more predictor variable
and entered into multivariate analyses along with client characteristics
to yield differential equations.
MODELING ApPROACH
A third research strategy uses treatment assignment as the depen-

dent variable, as the subject of study. The underlying question is: what
criteria are being used to assign clients to treatments within an existing
system? This modeling strategy has been introduced by Goldberg (1968,
1971) for studying the judgment processes of clinicians. Quantitative
client variables are used as predictors, and treatment assignment
(chosen by a therapist) is the criterion to be predicted. Through discrimi-
nant function analysis it is possible to derive a multivariate equation that
models the judgment processes of the clinician in making treatment
assignments.
This strategy, though seldom used, has several possible applica-
tions. One is to determine the implicit criteria being used to decide who
gets what type of treatment. Studies of this type to date have found
assignment to be at best modestly related to client characteristics stud-
ied, suggesting an arbitrary or undifferential matching process (e.g.,
Bromet et al., 1977). This strategy would be appropriate in examining for
arbitrary discrimination (on the basis of race, socioeconomic status, sex,
age) in assignment to different treatment approaches (e.g., Hollings-
head & Redlich, 1958).
A fascinating and as yet untried application of this approach would
be to model the judgment processes of the most successful diagnosti-
cians (e.g., Goldberg, 1968, 1971). Individual clinicians could be permit-

ted to use their own criteria to match clients with treatments. Patterns of
client outcome could then be studied (how many accepted treatment,
completed it, showed positive outcomes), identifying the diagnosticians
with higher success rates at matching. The judgment processes of these
"good matchers" could then be modeled and compared with the judg-
ment processes of less successful matchers. Likewise the efficacy of
matching by these clinicians could be compared with alternative match-
ing procedures based on actuarial methods (Goldberg, 1965, 1972) or
choices by the clients themselves (Parker, Winstead, & Willi, 1979; Par-
ker, Winstead, Willi, & Fisher, 1979).
Because the modeling approach has not yet been adequately ap-
plied in alcoholism treatment research, our review will be restricted to a
summary of findings from predictor and differential approaches. We
begin with predictor studies, based on single treatment conditions.
PREDICTOR STUDIES
Are there client characteristics that predict favorable outcome re-

gardless of the type of treatment given? Gibbs and Flanagan (1977)
posed this question in an extensive review of the predictor literature in
alcoholism treatment. They defined a general predictor as "one which
has been investigated for its predictive power in six or more treatment
groups ... has been found of predictive value in all of these treatment
groups" (p. 1101). Using this stringent criterion, they found no general
predictors. By loosening their criterion, however, they were able to
point to a set of "somewhat less than perfect predictors." These in-
cluded employment status and work history, marital or cohabitation
status, occupational or social status, arrest record, diagnosis of neurosis,
IQ variables, and history of Alcoholics Anonymous contact prior to
treatment. Such variables have been found to be predictive of outcome
in some studies, but both the presence and direction of prognostic pre-
dictiveness are variable. Data published since the Gibbs and Flanagan
review have yielded no greater hope for the utility of general predictors.
It appears that prognosis for recovery from alcoholism cannot be ade-
quately predicted from pretreatment variables without reference to the
type of treatment received.
If there are no stable predictors for all treatments, are there client
characteristics that predict successful outcome within specific treatment
modalities? To examine this question we have grouped predictor studies
into 11 categories, according to the type of treatment examined.
PSYCHOTROPIC MEDICATIONS
Because psychotropic medications are used in alcoholism treatment

with the intent of alleviating a related form of psychopathology, an
obvious candidate for prediction studies would be the pretreatment
presence and level of the targeted pathology. Thus it could be hypoth-
esized that alcoholics treated with antidepressant medication would
show improvement (in alcoholism) to the extent that they evidenced
pretreatment depression. Merry, Reynold, Bailey, & Coppen (1976) re-
ported that depressed alcoholics receiving lithium carbonate, relative to
controls receiving placebo, spent fewer days drinking and had fewer
days of incapacitation by alcohol. Surprisingly, pretreatment level of
pathology is yet to be adequately explored as a predictor of change in
alcoholism following psychotropic medication.
Certain forms of pharmacotherapy may produce negative iatrogenic
effects with specific diagnostic groups. An illustrative example is a study
by Tomsovic and Edwards (1970) comparing alcoholics receiving or not
receiving psychedelic medication (lysergide) as part of their treatment.
Not surprisingly, patients with diagnosed schizophrenia (in addition to
alcoholism) showed detrimental effects from the hallucinogen, whereas
transient benefits were reported for nonschizophrenic alcoholics.
A difficulty in using pretreatment pathology is the fact that many
types of problems concurrent to alcohol abuse will remit spontaneously
once the drinking problem has been resolved (Miller, Hedrick, & Taylor,
1983). For this reason, it has been recommended that the use of psycho-
tropic medication might be delayed until alcohol abuse has been re-
solved, and that the persistence of psychopathology during aftercare
may be a better predictor of benefit from appropriate medications (Dit-
man & Crawford, 1966; Miller & Hester, 1980).
DISULFIRAM
Numerous studies have explored pretreatment client characteristics

as predictors of favorable outcome with disulfiram (d., Lundwall &
Baekeland, 1971). Reported favorable prognostic factors include being
older (Baekeland, Lundwall, Kissin, & Shanahan, 1971; Hoff &
McKeown, 1953), male (Hoff & McKeown, 1953), socially stable (Shaw,
1951), married (Azrin, Sisson, Meyers, & Godley, 1982), less depressed
(Baekeland et al., 1971), more motivated (Baekeland et al., 1971; Fuller &
Roth, 1979; Wexberg, 1953; d. Miller, 1985b), more compulsive (Waller-
stein, 1958), and having less psychopathology (Bowman et al., 1951), a
more external locus of control (Obitz, 1978), and a longer history of
alcohol abuse (Baekeland et al., 1971; Hoff & McKeown, 1953).
We have discussed elsewhere the absence of adequate outcome

studies of A.A. and the methodological problems inherent in the avail-
able research (Miller & Hester, 1980). We found only one study (David-
son, 1976) relating prognostic factors to outcome in A.A.: Males have
more favorable outcomes than females. Otherwise the literature has
focused on predictors of affiliation or attendance, which are at best
indirect indicators of outcome. A.A. attendance has been reported to be
related to authoritarianism and lower educational levels (Canter, 1966;
Ditman, Crawford, Forgy, Moskowitz, & MacAndrew, 1967), less psy-
chopathology (Gerard, Saenger, & Wile, 1962), affiliative and dependen-
cy needs (Trice & Roman, 1970), field dependence (Reilly & Sugarman,
1967), greater severity of alcohol-related problems, and higher overall
use of external sources of aid to stop drinking (O'Leary et al., 1980). In a
review of this literature, Ogborne and Glaser (1981) concluded that "af-
filiates of A.A. are more likely to be men, over 40 years of age, white,
middle or upper class and socially stable" (p. 666), and they speculated
on a range of other predictive factors.
PSYCHOTHERAPY
Controlled research points to no demonstrable benefit of individual

or group psychotherapy in alcoholism treatment (Miller & Hester, 1980).
Do certain types of clients benefit differentially? Wallerstein (1958) noted
that clients with borderline depression or psychosis seemed to do better
with psychotherapy, whereas compulsiveness was predictive of favor-
able outcome regardless of the type of treatment received. Pomerleau
and Adkins (1980) found that higher baseline levels of alcohol consump-
tion were associated with less favorable outcome, whereas longer dura-
tion of drinking problems predicted greater persistence in treatment and
greater reduction in consumption. These factors, however, were not
differential predictors, but were prognostic of outcome both in behavior
therapy with a moderation goal and in traditional group psychotherapy.
Kissin, Platz, and Su (1970) reported that patients who accepted psycho-
therapy had higher verbal intelligence and occupational stability, and
that successful responders to psychotherapy were more intelligent and
field independent. Finally McLachlan (1972), in research to be reviewed
later, found that client conceptual level is predictive of outcome in psy-
chotherapy, with the direction of prediction depending on the orienta-
tion of the therapist.
CHEMICAL AVERSION THERAPY
One of the oldest forms of alcoholism treatment employs the nau-

sea-inducing drug emetine as an unconditioned stimulus within an
aversive counterconditioning paradigm (Miller & Hester, 1980). Favor-
able prognostic factors in aversion therapy were well described by
Voegtlin and Broz (1949): being married, older, having a longer drinking
history (regardless of age), prior successful periods of abstinence, better
occupation, and higher financial status. More than 30 years later, Neu-
berger et ai. (1982) reported a similar pattern: Married, employed alco-
holics had the highest abstinence rates at follow-up, and success rates
increased in older groups.
Because aversion therapy relies (theoretically) on the establishment
of a conditioned aversion to alcohol, it is sensible that outcome might be
related to the establishment of such an aversive response during treat-
ment. Consistent with this prediction, Boland, Mellor, & Revusky (1978)
found that individuals who developed an illness reaction during taste
aversion (with lithium) showed higher rates of abstinence at 6-month
follow-up. Cannon, Baker, and Wehl (1981) likewise found an inverse
relationship between heart rate response during conditioning and post-
treatment (12-month follow-up) drinking among subjects receiving
chemical aversion therapy. Unfortunately there are as yet no known
pretreatment predictors of the establishment of conditioned aversion.
Vogel (1960, 1961a,b), in an experimental study, reported that steady-
drinking introverted alcoholics conditioned faster and extinguished
more slowly than their extraverted counterparts, and she speculated
that introverts might therefore be more appropriate for aversive counter-
conditioning. To date, however, this hypothesis has not been tested in a
clinical trial.
COVERT SENSITIZAnON
An alternative approach to aversion therapy employs negative im-

agery to establish conditioned avoidance of alcohol. Elkins (1980) re-
ported that patients who developed conditioned nausea during covert
sensitization treatment remained abstinent significantly longer than
than those who did not. (Interestingly, this difference disappeared
when he expanded his remission criteria to include moderate consump-
tion.) Miller and Dougher (1985) were able to replicate this finding,
reporting that drinking status at 18-month follow-up was predictable
from the establishment of conditioned aversion during nausea sensitiza-
tion. A comparison group received covert sensitization based on imag-
ery of aversive consequences of drinking (other than nausea). Within
this group, no relationship was observed between conditioning and

outcome, although overall success rate was comparable to that in the
nausea aversion groups. Again, it would be helpful to know pretreat-
ment predictors of responsiveness to this promising form of treatment.
RELAXATION TRAINING
Although relaxation training has not been found to add substan-

tially to the effectiveness of alcoholism treatment (Miller & Hester,
1980), it is plausible that certain clients-particularly those with greater
anxiety-might benefit differentially from learning relaxation skills. The
only study to examine this hypothesis supports it. Rosenberg (1979)
found that alcoholics with higher scores on the anxiety subscale of the
Alcohol Use Inventory showed significantly better outcome at 12
months if given relaxation training, relative to those not receiving such
training. Among clients low in anxiety, however, no differential benefit
was found. This finding, though in need of replication, demonstrates
the danger in prematurely dismissing a treatment as ineffective because
of a lack of impact within an undifferentiated population. Failure to
differentiate on critical predictor dimensions may mask the value of a
treatment technique for a particular subgroup.
SOCIAL SKILLS TRAINING
Although social skills training is well supported as an effective ad-

junct to other alcoholism treatment modalities, little information is avail-
able regarding the characteristics of optimal candidates. Some studies
have preselected alcoholics deficient in social skills, thereby removing
variance of potential predictive value (e.g., Adinolfi, McCourt, & Geog-
hegan, 1976; Ferrell & Galassi, 1981). Others have not explored differen-
tial benefits in relation to pretreatment social skill level, perhaps because
of the large main effect of treatment (e.g., Chaney, O'Leary, & Marlatt,
1978; Freedberg & Johnston, 1978). Although it is logical that social skills
training would have greater impact (on alcoholism) for clients with
larger skill deficits in this area, this hypothesis remains to be tested.
FAMILY THERAPY
Studies of family therapy necessarily preselect clients on the basis of

having a family willing to participate in treatment. That this is a signifi-
cant selection factor is supported by Smith's (1969) finding of better
outcome for alcoholics with wives willing to attend a treatment group.
Cadogan (1973) reported more favorable outcome following family ther-
apy when the alcoholic was employed, showed little or no organic brain
damage, was not psychotic, sought treatment early, and had a spouse
who showed trust and acceptance at the beginning of treatment. Marital
cohesion, however, has been found to be a predictor of outcome in other
modalities as well (e.g., Orford, Oppenheimer, Egert, Hensman &
Guthrie, 1976).
COMMUNITY REINFORCEMENT ApPROACH (CRA)

One of the best supported and most complex treatment programs
for alcoholism is the "community reinforcement approach" (CRA)
(Azrin et ai., 1982; Miller, 1985c). The CRA seeks to make substantive
changes in the client's life-style in order to increase social stability and
reinforce nondrinking. Azrin et al. (1982) compared disulfiram com-
pliance alone with disulfiram plus CRA. Among married clients these
approaches were equally effective, suggesting no additive effect of CRA.
Among single clients, however, disulfiram alone was significantly less
effective than the combination. If one hypothesizes less social stability
and greater life-style reliance on drinking within a single population, it
is understandable that CRA might be of differential benefit.
BEHAVIORAL SELF-CONTROL TRAINING (BSCT)

BSCT is a set of self-management approaches designed to teach
moderate and problem-free ("controlled") drinking. Several studies
have examined pretreatment predictors of the successful establishment
of controlled drinking following BSCT. Although findings have been
mixed (Elal-Lawrence, 1984), the most consistent predictors of favorable
prognosis for controlled drinking have been lower duration and severity
of drinking symptoms and problems (Edwards, Duckitt, Oppenheimer,
Sheehan, & Taylor, 1983; Finney & Moos, 1981; Miller & Baca, 1983;
Miller & Joyce, 1979; Orford, Oppenheimer, & Edwards, 1976; Polich,
Armor, & Braiker, 1981; Popham & Schmidt, 1976; Smart, 1978; Vogler,
Compton, & Weissbach, 1975; Vogler, Weissbach, Compton, & Martin,
1977). The implication is that BSCT may be a more effective approach
with less severe problem drinkers. Orford and Keddie (in press), by
contrast, found no relationship between severity of alcohol dependence
and abstinent versus controlled drinking outcomes. Instead, they found
that outcome was predictable from client beliefs about alcohol and alco-
holism. Clients endorsing traditional disease conceptions of their alco-
hol problems tended to become abstinent, whereas clients rejecting ten-
ets of a disease conception were more likely to attain moderation.
SUMMARY
No general predictors of alcoholism treatment outcome have

emerged, suggesting that prognosis involves an interaction of client and
treatment variables. There seem to be neither "good prognosis" clients
(without reference to type of treatment) nor "effective" treatments
(without considering type of client).
Current data, though far from conclusive, suggest that adjunctive
treatment techniques aimed at problems presumed to be related to or
underlying alcoholism (e.g., depression, social skill deficit, anxiety) are
most effective when the client manifests a significant level of the target
problem. Offering such broad-spectrum approaches to an undifferenti-
ated population is unlikely to result in a substantial increment in pro-
gram effectiveness (e.g., Miller, Taylor, & West, 1980). With a selected
population, however, these strategies may significantly prolong
sobriety.
Predictor studies likewise point to certain predictor variables that
may be of substantive value in selecting optimal treatment approaches.
Problem severity and duration, for example, have been found to be
positively correlated with AA attendance and success with disulfiram,
but inversely related to the effectiveness of behavioral self-control train-
ing. This suggests that whereas one approach may be optimal for severe
alcoholics, a quite different approach may be maximally effective with
less severe problem drinkers (d. Miller & Caddy, 1977).
This leads us to the final major section of this chapter: a survey of
research on differential predictors of outcome.
DIFFERENTIAL STUDIES
The differential research strategy in matching research examines the

utility of particular client characteristics in predicting the relative proba-
bility of success in alternative approaches. Two or more different inter-
vention methods are compared in the search for discriminative predic-
tors of outcome. Whereas for predictor studies we summarized findings
by treatment method, here we will present current knowledge according
to specific matching variables that have appeared promising.
PROBLEM SEVERITY
Severity of problem can be defined in a variety of ways including,

(a) severity of current or cumulative consequences of drinking, (b) level
of alcohol consumption, (c) severity of current or cumulative signs of
alcohol dependence, or (d) problem duration (d. Horn, Wanberg, &

Foster, 1974; Miller & Marlatt, 1984; Polich et al., 1981). Gross measures
of problem severity (e.g., Selzer, 1971) often confound these dimen-
sions, which are only modestly intercorrelated and cannot be considered
to be interchangeable (Horn, 1978). Pomerleau and Adkins (1980), for
example, found favorable outcome to be related to lower baseline alco-
hol consumption but longer problem duration.
As indicated earlier, controlled drinking outcomes (as contrasted
with abstinent outcomes) have been associated in uncontrolled studies
with lower pretreatment alcohol consumption, shorter problem dura-
tion, fewer consequences and symptoms of pathological drinking. The
only two studies to assign clients at random to controlled drinking ver-
sus strict abstinence goals, however, have failed to yield strong differen-
tial predictors of outcome (Foy, Nunn, & Rychtarik, 1984; Sanchez-
Craig, Annis, Bornet & MacDonald, 1984), and data from uncontrolled
studies are not wholly consistent (Elal-Lawrence, 1984). Although in our
opinion the data point to controlled drinking as an optimal goal for less
severe problem drinkers, it would be premature to define strict indica-
tions or contraindications for treatment goals at this time (Marlatt, et al.,
1985).
Two research teams have provided data regarding the differential
benefit of an intensive treatment approach with more severely impaired
alcoholics. Orford, Oppenheimer, and Edwards (1976) reported 2-year
follow~up data for clients assigned at random to either an intensive
(inpatient hospital program, outpatient therapy, and medication) or a
minimal (evaluation plus one session) intervention. Clients diagnosed at
intake as gamma alcoholics (severe dependence with loss of control)
fared better with intensive treatment and tended to become abstainers,
whereas those diagnosed as alpha alcoholics (problem drinkers without
dependence or loss of control) fared better if not given intensive treat-
ment and tended to become moderate drinkers.
McLellan, Luborsky, Woody, O'Brien, & Druley (1983) employed a
measure of "psychiatric severity" (McLellan, O'Brien, Kron, Alterman,
& Druley, 1980) to predict outcome retrospectively from a variety of
inpatient and outpatient programs. Relying on self-report measures at 6-
month follow-up, they found that patients with high levels of problem
severity fared equally poorly in inpatient and outpatient approaches,
whereas at low levels of severity, patients did equally well regardless of
treatment setting. Within the intermediate severity range, however
(60% of patients), levels of other life problems (family, employment, and
legal) showed complex relationships to outcome.
Pursuing these findings in a prospective study, McLellan, Woody,
Luborsky, O'Brien, and Druley (1983) attempted to match patients to

treatments based on the data from their retrospective study. They suc-
ceeded in matching 53% of cases, with the rest mismatched because of
refusal to accept treatment, assignment errors, clinical overriding of the
match, or unavailability of the desired treatment slot. Because high se-
verity patients had fared poorly in the earlier study, all patients with
severe problems were classed as mismatched regardless of the treatment
they received. This procedure, of course, created an artifactual bias
favoring matched cases. With this bias removed (by excluding severe
cases from analyses) significant differences remained between matched
and mismatched cases on a multivariate analysis of covariance and 8 of
19 outcome measures. It must be recognized, however, that assignment
to matching versus mismatching was nonrandom, and that outcome
was judged from unverified self-report. Nevertheless the work of the
McLellan team represents a methodological advance in research on
client-treatment matching.
Taken together, these studies present an inconsistent picture. Un-
controlled predictor studies point to more favorable outcomes of low
severity clients in moderation-oriented treatment, but two studies em-
ploying random assignment to moderation versus abstinence goals have
failed to confirm this. Retrospective data from the Orford, Oppenhei-
mer, and Edwards (1976) investigation suggest a more favorable prog-
nosis for severely dependent persons in intensive treatment, whereas
retrospective data from McLellan, Luborsky, et al. (1983) point to poor
outcome in this group regardless of treatment locus. It may be that the
Orford, Oppenheimer, and Edwards cohort, which was required to
have an intact family, may resemble the middle-severity group of the
McLellan, Luborsky, et al. study, in which case these findings could be
seen as more similar. Unfortunately the severity measures employed in
these two studies do not overlap. At most, then, there is a suggestion of
a matching interaction between problem severity and treatment inten-
sity.
COGNITIVE STYLE
Another type of predictor variable that appears promising has to do

with a client's "cognitive style," by which we mean relatively enduring
patterns of perception and information processing that the person evi-
dences in a broad range of situations. This is, we note, very similar to
the definition of "personality" adopted by Hall & Lindzey (1970).
One such client characteristic that has been widely studied is Rot-
ter's (1966) construct of internal versus external locus of control. Exter-

nals on this dimension tend to perceive their lives as being largely con-
trolled by forces beyond their own influence (luck, fate, powerful
others), whereas Internals view themselves as efficacious and responsi-
ble in determining what happens to them. Most studies to date linking
locus of control to differential outcome have been correlational in de-
sign, and have suggested that Internals fare better with nondirective
than with directive approaches (Abramowitz, Abramowitz, Roback, &
Jackson, 1974), are less likely to accept disulfiram (abitz, 1978), and
participate less in aftercare (O'Leary, Rohsenow, & Donovan, 1976). All
of these findings are consistent with a more self-directed approach for
Internals. The only experimental study to date, however, failed to find
an interaction between locus of control and directiveness of treatment
(Schmidt, 1978).
Some of the strongest matching effects in the literature have been
found with regard to a cognitive style dimension often referred to as
"conceptual level" (CL). Clients with a low CL (e.g., preference for
simpler rules and fewer constructs, dependence on authority) are hy-
pothesized to be optimal for highly structured directive approaches that
stress adherence to rules and minimal self-direction. By contrast, those
with a high CL (e.g., independent, complex thinkers) are predicted to be
optimal for less structured nondirective approaches emphasizing per-
sonal control. McLachlan (1972) studied the effectiveness of directive
versus nondirective therapeutic approaches with high versus low CL
clients. Matched cases (high CL with nondirective and low CL with
directive therapist) reported greater perceived benefit, change, and satis-
faction with their treatment than did mismatched cases. (Matching was
determined post hoc rather than by intentional assignment.) More im-
portantly, when recovery rates were derived from collateral reports at
12- to 16-month follow-up for these same 92 clients, 70% of matched
cases versus 50% of unmatched cases were rated as recovered (McLach-
lan, 1974). McLachlan also examined CL in relation to style of aftercare
provided: The city dwellers were offered weekly aftercare meetings
(structured), whereas out-of-town patients received only a letter encour-
aging them to write to other patients (unstructured). (The confound
with place of residence must be noted.) Matched cases (high CL in
unstructured and low CL in structured) showed 71% recovery, versus
49% recovery in unmatched cases. When both treatment and aftercare
style were considered, the separation was even greater. Patients appro-
priately matched to treatment and aftercare style showed a 77% recovery
rate. Rates for patients matched on treatment alone (65%) or aftercare
alone (61 %) were intermediate and well above those for patients mis-
matched to both (38%), suggesting that either appropriate treatment or

appropriate aftercare may improve long-term outcome. Although treat-
ment assignment was nonrandom, there were no apparent biases in
assignment of cases to treatments. An unfortunate weakness of the
McLachlan study is the relatively crude outcome measure, a 4-point
rating scale completed by the aftercare physicians and counselors, other
patients, and clinical secretaries. Finally, it is noteworthy that although
there were strong interaction effects, no main effects of treatment or
aftercare style were observed. Had McLachlan not separated patients
according to CL, he would have been forced to conclude that directive
and nondirective styles of treatment were equivalent in effectiveness.
This underlines the potential importance of matching in clarifying the
therapeutic impact of specific interventions.
Thornton, Gottheil, Gellens, and Alterman (1977, 1981), in retro-
spective prediction of outcome, studied the relationship of posttreat-
ment drinking pattern (assessed by unverified patient questionnaires) to
"developmental level" (DL), a construct measured from Rorschach re-
sponses and conceptually similar to CL. They reported that although
high and low DL patients did not differ with regard to the percentage
achieving abstinence at 6-month follow-up, high DL patients were more
likely to achieve moderate drinking, whereas low DL drinkers drank
more frequently and more heavily following treatment.
Karp, Kissin, and Hustmyer (1970) studied a related dimension of
cognitive style: field dependence. They found that alcoholics selecting
and selected for psychotherapy were highly field independent, and fur-
ther that this predictor discriminated dropouts within treatment modes.
Clients who dropped out of psychotherapy were significantly more field
dependent than those who remained. For drug therapy, by contrast,
dropouts were slightly (not significantly) more field independent. When
the treatment modes were combined, field dependence failed to discrim-
inate dropouts from those continuing. Once again, outcome prediction
could not be divorced from the particular nature of the intervention.
Similar findings have been reported by Kissin et al. (1970).
Taken together, these findings suggest that certain clients may evi-
dence a cognitive style (external control, low CL, low DL) which renders
them optimal candidates for more directive and structured treatments,
whereas others (internal control, high CL, high DL) may respond more
favorably to less structured and more self-directed approaches. This is
consistent with correlational data reviewed earlier, indicating that affili-
ates of Alcoholics Anonymous show higher authoritarianism, depen-
dency, affiliation, field dependence, and reliance on external sources of
aid (all plausible correlates of low CL).
NEUROPSYCHOLOGICAL STATUS
Prolonged heavy consumption of alcohol is known to produce a

characteristic pattern of brain impairment (Miller & Saucedo, 1983). An-
other logical candidate as a matching variable, then, is the degree of
neuropsychological deficits. Greater neuropsychological impairment
has been associated with less favorable overall outcomes (O'Leary, Don-
ovan, Chaney, & Walker, 1979), but to date no treatment has been
shown to be differentially effective for more impaired individuals. Walk-
er, Donovan, Kivlahan, and O'Leary (1983) failed to confirm their hy-
pothesis that patients with greater cognitive impairment would respond
more favorably to longer and more intensive treatment. Rather, their
findings resemble those of McLellan, Luborsky, et ai. (1983), that higher
levels of severity are associated with equally poor outcome regardless of
treatment approach. The previously cited finding of Karp et ai. (1970)-
that alcoholism treatment dropout patterns are related to field depen-
dence-is noteworthy here because field dependence has been shown
to be a correlate of neuropsychological impairment in alcoholism (Miller
& Saucedo, 1983). Thus even though intensity and treatment may not
interact with impairment among treatment completers, it may be that
clients with greater cognitive deficits may find certain types of treatment
more appealing or comprehensible, and thereby attrition rates may be
affected. Clearer conclusions must await further research (Wilkinson &
Sanchez-Craig, 1981).
SELF-ESTEEM
Research on motivation for treatment points to self-esteem as a

potentially important factor in determining treatment acceptance and
perseverence (Miller, 1985b). To our knowledge, however, only one
study has examined differential efficacy of treatments based on client
self-esteem level. Annis & Chan (1983) randomly assigned alcohol-relat-
ed offenders undergoing institutional care either to receive or not to
receive a highly confrontational group therapy intervention. Neither
treatment nor client type produced a main effect: high and low self-
esteem clients fared equally well overall, and the group therapy made
no significant difference. Consistent with the matching hypothesis,
however, a significant interaction effect appeared, such that high self-
esteem clients fared better if they received the group treatment (outcome
assessed by reconvictions), whereas clients with low self-esteem
showed a detrimental effect of the group and fared better without it.
SOCIAL STABILITY
Clients with stable family, residence, and employment are often

reported to have more successful outcomes overall (Adinolfi, DiDario, &
Kelso, 1981; Armor, Polich, & Stambul, 1978; Gerard & Saenger, 1966;
Gibbs & Flanagan, 1977; Orford, Oppenheimer, Egert, et al., 1976), but
are there treatments that are differentially beneficial to clients with low
(or high) social stability? Married and employed clients have been re-
ported to be more likely to establish controlled drinking outcomes than
less socially stable individuals (e.g., Levinson, 1977; Smart, 1978), but
controlled studies have failed to confirm this finding. Azrin et aI. (1982)
found that a broad-spectrum community reinforcement approach in-
creased the effectiveness of treatment for unmarried but not for married
clients, suggesting that a broader life-style intervention may differen-
tially benefit clients without a stable family situation.
OTHER LIFE PROBLEMS
Therapists have observed that treatment focused exclusively on al-

coholism may be less effective with clients who show broader problems
and pathology (McLellan, Luborsky, et aI., 1983; Miller, Pechacek, &
Hamburg, 1981; Orford, Oppenheimer, Egert, et aI., 1976). The implica-
tion (which might be termed the "broad-spectrum matching hypoth-
esis") is that clients with broader problems could benefit differentially
from broad-spectrum treatments that address not only alcohol con-
sumption but other specific problem areas as well. This hypothesis may
explain why comparative studies with undifferentiated populations of
problem drinkers have found little or no advantage in broad-spectrum
approaches over alcohol-focused treatment (Alden, 1978; Miller et al.,
1980; Vogler, Weissbach, Compton, & Martin, et aI., 1975; 1977).
Support for the broad-spectrum matching hypothesis is found in
studies pointing to the differential effectiveness of relaxation training for
anxious alcoholics (Rosenberg, 1979) and of a community reinforcement
approach for unmarried alcoholics (Azrin et al., 1982). Other investiga-
tors have reported favorable (though not differential) response of de-
pressed alcoholics to lithium (Merry et aI., 1976), of unassertive alco-
holics to assertion training (Ferrell & Galassi, 1981), and of socially un-
skilled alcoholics to social skills training (Adinolfi et aI., 1976). Other
correlational data indicate a less favorable response to very alcohol-
focused interventions, such as Alcoholics Anonymous (Gerard et al.,
1962) and disulfiram (Baekeland et al., 1971; Bowman et al., 1951) when
other major psychopathology is present.
These data point toward a very sensible though necessarily tenta-
tive conclusion: that alcoholics will benefit from additional treatment to
the extent that they manifest the problem that the treatment effectively
alleviates. Future research on broad-spectrum matching would best
focus on problem-specific interventions for documented pretreatment
deficits, rather than seeking differential benefit of all-purpose "shot-
gun" additions to alcohol-focused treatment.
CLIENT CHOICE
Numerous writers have posited beneficial effects of participation by

clients in the selection of their own treatment approaches (Costello,
1975; Ewing, 1977; Parker, Winstead, & Willi, 1979; Parker, Winstead,
Willi, & Fisher, 1979). Predictive data indicate that clients who have (or
at least perceive that they have) a voluntary choice about the goal and
nature of their treatment show more favorable satisfaction (Vannicelli,
1978, 1979), compliance (Sanchez-Craig, 1980), and outcome (Thornton
et al., 1977) during and following the treatment process (d. Miller,
1985b). In an experimental study, Kissin, Platz, & Su (1971) assigned
patients at random to be offered three, two, one, or no alternative treat-
ments for alcoholism. Individual patient characteristics proved to be
poor predictors of either acceptance of or success in specific types of
treatment. One half to two thirds of patients who were given a choice
rejected the first treatment offered to them and opted for an alternative.
Findings indicated that patients given a choice of treatment options
showed greater acceptance of treatment and superior rates of recovery at
12-month follow-up. Success rates increased with the number of choices
available to the client.
These·findings suggest at least an interim practice until more sub-
stantial data on client-treatment matching become available: to involve
clients directly in the choice of their own treatment. Through mecha-
nisms of perceived control and intrinsic motivation, clients may show
increased acceptance of, continuation in, and compliance with a treat-
ment that they select themselves (Deci, 1975; Miller, 1983). Indeed,
given adequate information about the alternatives, clients may be better
than their therapists in selecting an optimal treatment approach. As
more reliable information becomes available regarding differential prob-
abilities of success in alternative treatment goals and strategies, such
information can be shared with the client as part of the decision-making
process.
OTHER PREDICTOR VARIABLES
Finally, we would point to a few other potential predictor variables

that seem to us to be promising, although we know of no adequate data
to support their usefulness in differential treatment choice. This is not
meant to be an exhaustive list, but rather a set of suggestions for further

exploration.
Family history of alcoholism and other types of psychopathology
would seem a logical candidate. Family history is often an important
clue to differential diagnosis and treatment in other domains (e.g., the
affective disorders), and numerous researchers have pointed to the
plausibility of different types of alcoholism Oacobson, 1976; Winokur,
Reich, Rimmer, & Pitts, 1970; Winokur, Rimmer, & Reich, 1971). Miller
and Joyce (1979) reported that problem drinkers with alcoholic fathers
were more likely to abstain and less likely to sustain controlled drinking
following behavioral self-control training. If indeed there are familial
and nonfamilial types of alcoholism, differential treatment goals and
strategies may be optimal for them.
Alcohol dependence is another clear candidate for exploration as a
differential predictor variable. Although we have speculated that level of
dependence may predict success in controlled drinking versus absti-
nence goals (Miller & Caddy, 1977; Miller & Hester, 1980), predictive
studies to date have focused instead on problem severity, a dimension
modestly correlated with severity of pharmacologic addiction. Orford
and Keddie (in press) found no relationship between severity of depen-
dence and outcomes of moderation versus abstinence. We hope that
future clinical trials will assess differential treatment outcome against
valid measures of physical dependence.
Various predictors may be of particular utility in evaluating the
probable effectiveness of conditioning therapies (e.g., covert sensitiza-
tion). Because the establishment of a conditioned aversive response is
predictive of success (Elkins, 1980; Miller & Dougher, 1985), pretreat-
ment predictors of conditioning may be helpful. These might include
conditionability measures, hypnotic susceptibility, or imagery vivid-
ness. Likewise, because a goal of aversion therapies is to reduce desire
for alcohol, these procedures may be particularly helpful for clients who
report strong cravings or urges.
Finally, new theoretical and psychometric developments in the ad-
dictive behaviors may point to robust predictors of differential outcome.
The well-constructed Alcohol Use Inventory (Horn et al., 1974) was de-
rived by extensive factor analytic research into crucial dimensions of
alcohol abuse, and its scales provide low-cost and promising tools for
differential treatment choice. The "stages of change" model recently
introduced by Prochaska and DiClemente (1983; d. Chap. 1 in this vol-
ume) proposes measurable steps through which clients pass in the pro-
cess of change, and the model poses specific predictions regarding
which interventions would be optimal at each stage. Apter's provocative
theory of psychological reversals (Apter, 1982; Miller, 1985a) may yield
diagnostic and psychometric methods useful in matching patients to

treatments.
SUMMARY
Although alcoholism treatment approaches do show evidence of

substantial differences in characteristic success rates (Miller & Hester,
1980), a more appropriate topic for future research may be the interac-
tion of client and treatment characteristics in facilitating recovery. Multi-
variate studies indicate that there is indeed substantial shared variance
between client and treatment factors (Cronkite & Moos, 1978). It must
also be recognized that posttreatment experiences of clients are likely to
account for at least as much variance as pretreatment and treatment
characteristics (Finney, Moos, & Mewborn, 1980).
The matching hypothesis proposes that clients who are matched to
appropriate treatments will show greater improvement than will those
who are unmatched or mismatched. The criteria for such optimal match-
ing, however, are far from clear at present. A few conclusions that can
be drawn from the data presently available are the following:
1. The degree of differential benefit from a broad-spectrum inter-
vention depends upon the degree to which the problem drinker man-
ifests the life problem or deficit for which the additional intervention is
an effective treatment.
2. Clients show greater improvement when matched with a treat-
ment that is congruent with their cognitive style, relative to clients who
are unmatched or mismatched.
3. Clients with more severe alcohol-related problems benefit differ-
entially from more intensive (though not necessarily inpatient) treat-
ment, whereas clients with less severe problems benefit at least as much
if not more from a minimal intervention.
4. Clients who choose their treatment approach from among alter-
natives show greater acceptance of, compliance in, and improvement
following treatment, relative to clients offered only a single program or
approach.
Given the limited research available at present, however, these are
best regarded as tentatively supported hypotheses in need of further
verification. Future research will likely confirm some degree of truth in
these assertions, but also reveal them to be overly simplistic.
A few points of methodology are warranted here as advice to pre-
vent repetition of past errors in future studies. All have to do with clear
specification and operationalization of terms. First, the predictor vari-
ables in a matching study should be small in number (due to usual
limitations of multivariate analyses with smaller samples), carefully se-

lected on empirical or theoretical grounds, and measured in a manner
that facilitates replication in research and clinical settings. Gross mea-
sures of "alcoholism" or "severity" that confound different types of
impairment should be avoided in favor of more specific dimensions that
will clarify the interactive processes involved in matching. An exem-
plary instrument is the Alcohol Use Inventory (Horn et al., 1974), which
was developed by factor analysis to represent orthogonal domains with-
in diverse problem drinking populations. The subscales of this instru-
ment appear particularly promising for purposes of individualized treat-
ment planning.
Secondly, treatment procedures should be specified and differenti-
ated as clearly as possible. A comparative study of two global multicom-
ponent programs may, for example, be less informative than a com-
parison of the same program with and without one clearly described
additional component which is offered to a random sample of program
participants. Matching data will be of little use if the nature of matched
treatments is vague.
Finally, a matching study with well specified predictor variables
and clearly differentiated treatments can be rendered uninterpretable if
the criteria for improvement are inadequate. Client self-ratings of satis-
faction, perceived helpfulness, or change are insufficient bases for judg-
ing treatment impact. Alcohol consumption during follow-up should be
carefully quantified, rather than recorded as merely present or absent.
Alcohol-related problems and signs of dependence should also be
monitored (Polich et al., 1981). Verification of self-report by collaterals is
desirable (e.g., Miller, Crawford, & Taylor, 1979) although outcome
should not be judged by collateral report alone. Follow-up interviews
and interviewer ratings should be completed, whenever possible, by
staff who are blind to client treatment assignment. Immediate posttreat-
ment status is not a reliable indicator of long-term impact of an interven-
tion, and a minimum of 6 to 12 months of posttreatment follow-up
should be completed with at least 80% of treated cases (or with a ran-
dom, representative sample in larger studies).
Although most treatment programs now acknowledge the impor-
tance of matching and individually tailoring treatment to client charac-
teristics and needs, we still have a very long way to go toward achieving
this goal. The standard formula approach to treatment must be aban-
doned in favor of offering a range of real and accessible alternative
approaches. Private financial interests of treatment providers favor the
assignment of a maximal number of patients to intensive and expensive
programs, and in the United States this is likely to be a substantial
barrier to optimal matching (Hansen & Emrick, 1983). Much more pre-
dictive, differential, and modeling research is needed to provide an

adequate empirical data base from which to make competent treatment
recommenda tions.
Meanwhile it would seem that the most ethical (and perhaps also
the most effective) approach to alcoholism treatment matching is one
of consumer advocacy: to provide individuals with full and accurate
information about the nature and effectiveness of the alternatives avail-
able to them. Neither the bureaucracies of public treatment systems nor
the competition of for-profit private providers has yet even begun to en-
able clients to make such informed choices about their own treatment.
Caveat emptor!
REFERENCES
Abramowitz, C. V., Abramowitz, S. F., Roback, H. B., & Jackson, c. (1974). Differential
effectiveness of directive and nondirective group therapies as a function of client
internal-external control. Journal of Consulting and Clinical Psychology, 42, 849-853.
Adinolfi, A A, DiDario, B., & Kelso, F. W. (1981). The relationship between drinking
patterns at therapy termination and intake and termination status on social variables:
A replication study. International Journal of the Addictions, 16, 555-565.
Adinolfi, A. A., McCourt, W. F., & Geoghegan, S. (1976). Group assertiveness training for
alcoholics. Journal of Studies on Alcohol, 37, 311-320.
Alden, L. (1978). Evaluation of a preventive self-management programme for problem
drinkers. Canadian Journal of Behavioural Science, 10, 258-263.
Annis, H. M., & Chan, D. (1983). The differential treatment model: Empirical evidence
from a personality typology of adult offenders. Criminal Justice and Behavior, 10, 159-
173.
Apter, M. J. (1982). The experience of motivation. London: Academic Press.
Armor, D. J., Polich, J. M., & Stambul, H. B. (1978). Alcoholism and treatment. New York:
Wiley, 1978.
Azrin, N. H., Sisson, R W., Meyers, R, & Godley, M. (1982). Alcoholism treatment by
disulfiram and community reinforcement therapy. Journal of Behavior Therapy and Ex-
perimental Psychiatry, 13, 105-112.
Baekeland, F., Lundwall, L., Kissin, B., & Shanahan, T. (1971). Correlates of outcome in
disulfiram treatment of alcoholism. Journal of Nervous and Mental Disease, 153, 1-9.
Boland, F. J., Mellor, C. S., & Revusky, S. (1978). Chemical aversion treatment of alco-
holism: Lithium as the aversive agent. Behaviour Research and Therapy, 16, 401-409.
Bowman, K. M., Simon, A., Hine, C. H., Macklin, E. A., Crook, G. H., Burbridge, N., &
Hanson, K. (1951). A clinical evaluation of tetraethylthiuramdisulphide (Antabuse) in
the treatment of problem drinkers. American lournal of Psychiatry, 107, 832-838.
Bromet, E., Moos, R, Wuthmann, c., & Bliss, F. (1977). Treatment experiences of alco-
holic patients: An analysis of five residential alcoholism programs. International Journal
of the Addictions, 12, 953-958.
Brown, J., & Lyons, J. P. (1981). A progressive diagnostic schema for alcoholism with
evidence of clinical efficacy. Alcoholism: Clinical and Experimental Research,S, 17-25.
Cadogan, D. A (1973). Marital group therapy in the treatment of alcoholism. Quarterly
198 WILLIAM R. MILLER AND REID K HESTER
Cannon, D. 5., Baker, T. B., & Wehl, C. K (1981). Emetic and electric shock aversion
therapy: Six- and twelve-month follow-up. Journal of Consulting and Clinical Psychology,
49, 360-368.
Canter, F. M. (1966). Personality factors related to participation in treatment by hospi-
talized male alcoholics. Journal of Clinical Psychology, 22, 114-116.
Chaney, E. F., O'Leary, M. R., & Marlatt, G. A. (1978). Skill training with alcoholics.
Costello, R. M. (1975). Alcoholism treatment and evaluation: In search of methods. Interna-
Cronkite, R. c., & Moos, R. H. (1978). Evaluating alcoholism treatment programs: An
integrated approach. Journal of Consulting and Clinical Psychology, 46, 1105-1119.
Davidson, A. F. (1976). An evaluation of the treatment and after-care of a hundred alco-
holics. British Journal of Addiction, 71, 217-224.
Deci, E. L. (1975). Intrinsic motivation. New York: Plenum Press.
Ditman, K 5., & Crawford, G. G. (1966). The use of court probation in the management of
the alcohol addict. American Journal of Psychiatry, 122, 757-762.
OHman, K 5., Crawford, G. G., Forgy, E. W., Moskowitz, H., & MacAndrew, C. (1967). A
controlled experiment on the use of court probation for drunk arrests. American Journal
Edwards, G., Orford, J., Egert,S., Guthrie,S., Hawker, A., Hensman, c., Mitcheson, M.,
Oppenheimer, E., & Taylor, C. (1977). Alcoholism: A controlled trial of "treatment"
and "advice." Journal of Studies on Alcohol, 38, loo4-103l.
Edwards, G., Duckitt, A., Oppenheimer, E., Sheehan, M., & Taylor, C. (1983). What
happens to alcoholics? Lancet, 2, 269-271.
Elkins, R. L. (1980). Covert sensitization treatment of alcoholism: Contributions of suc-
cessful conditioning to subsequent abstinence maintenance. Addictive Behaviors, 5,67-
89.
Elal-Lawrence, G. (1984, August). Predictors of alcoholism treatment outcome: With special
reference to controlled drinking. Paper presented at the Third International Conference
on Treatment of Addictive Behaviors, North Berwick, Scotland.
Ewing, J. A. (1977). Matching therapy and patients: The cafeteria plan. British Journal of
Ferrell, W. L., & Galassi, J. P. (1981). Assertion training and human relations training in
the treatment of chronic alcoholics. International Journal of the Addictions, 16, 959-
968.
Finney, J. W., & Moos, R. H. (1979). Treatment and outcome for empirical subtypes of
alcoholic patients. Journal of Consulting and Clinical Psychology, 47, 25-38.
Finney, J. W., & Moos, R. H. (1981). Characteristics and prognoses of alcoholics who
become moderate drinkers and abstainers after treatment. Journal of Studies on Alcohol,
42,94-105.
Finney, J. W., Moos, R. H., & Mewborn, C. R. (1980). Posttreatment experiences and
treatment outcome of alcoholic patients six months and two years after hospitaliza-
tion. Journal of Consulting and Clinical Psychology, 48, 17-29.
Foy, D. W., Nunn, B. L., & Rychtarik, R. G. (1984). Broad-spectrum behavioral treatment
for chronic alcoholics: Effects of training controlled drinking skills. Journal of Consulting
and Clinical Psychology, 52, 218-230.
Freedberg, E. J., & Johnston, W. E. (1978). The effects of assertion training within the context of
a multi-modal alcoholism treatment program for employed alcoholics (Substudy No. 796).
Toronto: Alcoholism and Drug Addiction Research Foundation.
Fuller, R. K, & Roth, H. P. (1979). Disulfiram for the treatment of alcoholism: An evalua-
tion in 128 men. Annals of Internal Medicine, 90, 901-904.
Gerard, D. L., & Saenger, G. (1966). Outpatient treatment of alcoholism. Toronto: University
of Toronto Press.
Gerard, D. L., Saenger, G., & Wile, R. (1962). The abstinent alcoholic. Archives of General
Gibbs, L. E. (1980). A classification of alcoholics relevant to type-specific treatment. Inter-
national Journal of the Addictions, 15, 461-488.
Gibbs, L. E. (1981). The need for a new design for evaluating alcoholism treatment pro-
grams. Drug and Alcohol Dependence, 8, 287-299.
International Journal of the Addictions, 12, 1097-1141.
Glaser, F. B. (1980). Anybody got a match? Treatment research and the matching hypoth-
esis. In G. Edwards & M. Grant (Eds.), Alcoholism treatment in transition (pp. 178-196).
Baltimore, MD: University Park Press.
Goldberg, L. R. (1965). Diagnosticians vs. diagnostic signs: The diagnosis of psychosis vs.
neurosis from the MMPI. Psychological Monographs: General and Applied, 1965, 79, 1-28
(Whole No. 602).
Goldberg, L. R. (1968). Simple models or simple processes? Some research on clinical
judgments. American Psychologist, 23, 483-496.
Goldberg, L. R. (1971). Five models of clinical judgment: An empirical comparison be-
tween linear and nonlinear representations of the human inference process. Organiza-
tional Behavior and Human Performance, 6, 458-479.
Goldberg, L. R. (1972). Man versus mean: The exploitation of group profiles for the
construction of diagnostic classification systems. Journal of Abnormal Psychology, 79,
121-131.
Gottheil, E., McLellan, A. T., & Druley, K. A. (Eds.) (1981). Matching patient needs and
treatment methods in alcoholism and drug abuse. Springfield, IL: Charles C Thomas.
Hague, W. H., Donovan, D. M., & O'Leary, M. R. (1976). Personality characteristics
related to treatment decisions among inpatient alcoholics: A nonrelationship. Journal
of Clinical Psychology, 32, 476-479.
Hall, C. S., & Lindzey, G. (1970). Theories of personality (2nd ed.). New York: Wiley.
Hansen, J., & Emrick, C. D. (1983). Whom are we calling "alcoholic"? Bulletin of the Society
of Psychologists in Addictive Behaviors, 2, 164-178.
Harris, R. J. (1985). A primer of multivariate statistics (2nd ed.). San Francisco: Academic
Press.
Hoff, E. c., & McKeown, C. E. (1953). An evaluation of the use of tetraethylthiuram
disulfide in the treatment of 560 cases of alcohol addiction. American Journal of Psychia-
try. 109, 670-673.
Hollingshead, A. B., & Redlich, F. C. (1958). Social class and mental illness: A community
study. New York: Wiley.
Horn, J. L. (1978). Comments on the many faces of alcoholism. In P. E. Nathan, G. A.
Marlatt, & T. L0berg (Eds.), Alcoholism: New directions in behavioral research and treatment
(pp. 1-40). New York: Plenum Press.
Horn, J. L., Wanberg, K., & Foster, F. M. (1974). The Alcohol Use Inventory. Odessa, FL:
Psychological Assessment Resources.
Jacobson, G. R. (1976). The alcoholisms: Detection, diagnosis and assessment. New York:
Human Sciences Press.
Karp, S. A., Kissin, B., & Hustmyer, F. E., Jr. (1970). Field dependence as a predictor of
alcoholic therapy dropouts. Journal of Nervous and Mental Disease, 150, 77-83.
Kissin, B., Platz, A., & Su, W. H. (1970). Social and psychological factors in the treatment
of chronic alcoholism. Journal of Psychiatric Research, 8, 13-27.
Kissin, B., Platz, A., & Su, W. H. (1971). Selective factors in treatment choice and outcome
in alcoholics. In N. K. Mello & J. H. Mendelson (Eds.), Recent advances in studies of

alcoholism (pp. 781-802). Washington, DC: U.S. Government Printing Office.
Levinson, T. (1977). Controlled drinking in the alcoholic: A search for common features. In
J. S. Madden, R. Walker, & W. H. Kenyon (Eds.), Alcoholism and drug dependence: A
multidisciplinary approach (pp. 297-308). New York: Plenum Press.
Lundwall, L., & Baekeland, F. (1971). Disulfiram treatment of alcoholism. Journal of Ner-
vous and Mental Disease, 153, 381-394.
Marlatt, G. A, Miller, W. R., Duckert, F., G0testam, G., Heather, N., Peele,S., Sanchez-
Craig, M., Sobell, L. c., & Sobell, M. B. (1985). Abstinence and controlled drinking:
Alternative treatment goals for alcoholism and problem drinking? Bulletin of the Society
of Psychologists in Addictive Behaviors, 4, 123-150.
Martin, P. Y. (1979). Clients' characteristics and the expectations of staff in halfway houses
for alcoholics. Journal of Studies on Alcohol, 40, 211-221.
McLachlan, J. F. C. (1972). Benefit from group therapy as a function of patient-therapist
match on conceptual level. Psychotherapy: Theory, Research and Practice, 9, 317-323.
McLellan, A. T., Luborsky, L., Woody, G. E., O'Brien, C. P., & Druley, K. A. (1983).
Predicting response to alcohol and drug abuse treatments: Role of psychiatric severity.
Archives of General Psychiatry, 40, 620-625.
McLellan, A. T., O'Brien, C. P., Kron, R., Alterman, A. I., & Druley, K. A. (1980).
Matching substance abuse patients to appropriate treatment: A conceptual and meth-
odological approach. Drug and Alcohol Dependence,S, 189-195.
McLellan, A. T., Woody, G. E., Luborsky, L., O'Brien, C. P., & Druley, K. A. (1983).
Increased effectiveness of substance abuse treatment: A prospective study of patient-
treatment "matching." Journal of Nervous and Mental Disease, 171, 597-605.
Merry, J., Reynolds, C. M., Bailey, J., & Coppen, A. (1976). Prophylactic treatment of
alcoholism by lithium carbonate: A controlled study. Lancet, 2, 481-482.
Miller, W. R (1983). Motivational interviewing with problem drinkers. Behavioural Psycho-
therapy, 11. 147-172.
Miller, W. R. (1985a). Addictive behavior and the theory of psychological reversals. Addic-
tive Behaviors, 10, 177-180.
Miller, W. R. (1985b). Motivation for treatment: A review. Psychological Bul/etin, 98, 84-107.
Miller, W. R (1985c). Community reinforcement approach. In A S. Bellack & M. Hersen
(Eds.), Dictionary of behavior therapy techniques (pp. 64-66). New York: Plenum Press.
directed controlled drinking training for problem drinkers. Behavior Therapy, 14, 441-448.
Miller, W. R., & Caddy, G. R. (1977). Abstinence and controlled drinking in the treatment
of problem drinkers. Journal of Studies on Alcohol, 38, 986-1003.
Miller, W. R., & Dougher, M. J. (1985). Covert sensitization: Alternative treatment procedures
for alcoholics. Manuscript submitted for publication.
Miller, W. R., & Hester, R. K. (1980). Treating the problem drinker: Modern approaches.
In W. R. Miller (Ed.), The addictive behaviors: Treatment of alcoholism, drug abuse, smoking,
and obesity (pp. 11-141). Oxford: Pergamon Press.
Miller, W. R., & Joyce, M. A (1979). Prediction of abstinence, controlled drinking, and
heavy drinking outcomes following behavioral self-control training. Journal of Consult-
Miller, W. R, & Mariatt, G. A. (1984). Manual for the Comprehensive Drinker Profile. Odessa,
FL: Psychological Assessment Resources.
Miller, W. R., & Saucedo, C. F. (1983). Assessment of neuropsychological impairment and
brain damage in problem drinkers. In C J. Golden, J. A. Moses, Jr., J. A. Coffman, W.

R. Miller, & F. D. Strider (Eds.), Clinical neuropsychology: Interface with neurologic and
psychiatric disorders (pp. 141-195). New York: Grune & Stratton.
Miller, W. R., Crawford, V. L., & Taylor, C A. (1979). Significant others as corroborative
sources for problem drinkers. Addictive Behaviors, 4, 67-70.
Miller, W. R., Taylor, CA., & West, J. B. (1980). Focused versus broad-spectrum behav-
ior therapy for problem drinkers. Journal of Consulting and Clinical Psychology, 48, 590-
60l.
Miller, W. R., Pechacek, T. F., & Hamburg, S. (1981). Group behavior therapy for problem
drinkers. International Journal of the Addictions, 16, 827-837.
Miller, W. R., Hedrick, K E., & Taylor, C A. (1983). Addictive behaviors and life prob-
403-412.
Neuberger, O. W., Miller, S. I., Schmitz, R. E., Matarazzo, J. D., Pratt, H., & Hasha, W.
(1982). Replicable abstinence rates in an alcoholism treatment program. Journal of the
American Medical Association, 248, 960-963.
Obitz, F. W. (1975). Alcoholics' perceptions of selected counseling techniques. British
Journal of Addiction, 70, 187-19l.
Obitz, F. W. (1978). Control orientation and disulfiram. Journal of Studies on Alcohol, 39,
1297-1298.
Ogborne, A. C, & Glaser, F. B. (1981). Characteristics of affiliates of Alcoholics Ano-
nynous: A review of the literature. Journal of Studies on Alcohol, 42, 661-675.
O'Leary, M. R., Rohsenow, D. J., & Donovan, D. M. (1976). Locus of control and attrition
from an alcoholism treatment program. Journal of Consulting and Clinical Psychology, 44,
686-687.
O'Leary, M. R., Donovan, D. M., Chaney, E. F., & Walker, D. (1979). Cognitive impair-
ment and treatment outcome with alcoholics: Preliminary findings. Journal of Clinical
O'Leary, M. R., Donovan, D. M., Chaney, E. F., & O'Leary, D. E. (1980). Relationship of
alcoholic personality subtypes to treatment follow-up measures. Journal of Nervous and
Mental Disease, 168, 475-480.
Orford, J., & Hawker, A. (1974). An investigation of an alcoholism rehabilitation halfway
house: II. The complex question of client motivation. British Journal of Addiction, 69,
315-323.
Orford, J., & Keddie, A. (in press). Abstinence or controlled drinking in clinical practice: A
test of the dependence and persuasion hypotheses. British Journal of Addiction.
409-418.
Orford, J., Oppenheimer, E., Egert, S., Hensman, C, & Guthrie, S. (1976). The co-
hesiveness of alcoholism-complicated marriages and its influence on treatment out-
come. British Journal of Psychiatry, 128, 318-339.
Parker, M. W., Winstead, D. K, & Willi, F. J. P. (1979). Patient autonomy in alcohol
rehabilitation. I. Literature review. International Journal of the Addictions, 14, 1015-
1022.
Parker, M. W., Winstead, D. K, Willi, F. J. P., & Fisher, P. (1979). Patient autonomy in
alcohol rehabilitation. II. Program evaluation. International Journal of the Addictions, 14,
1177-1184.
Pomerleau, 0., & Adkins, D. (1980). Evaluating behavioral and traditional treatment for
problem drinkers. In L. C. Sobell, M. B. Sobell, & E. Ward (Eds.), Evaluating alcohol and
drug abuse treatment effectiveness: Recent advances (pp. 93-108). New York: Pergamon
Press.
Popham, R. E., & Schmidt, W. (1976). Some factors affecting the likelihood of moderate
drinking by treated alcoholics. Journal of Studies on Alcohol, 37, 868-882.
Price, R. H., & Curlee-Salisbury, J. (1975). Patient-treatment interactions among alco-
holics. Journal of Studies on Alcohol, 36, 659-669.
Prochaska, J. 0., & DiClemente, C. C. (1983). Stages and processes of self-change of
smoking: Toward an integrative model of change. Journal of Consulting and Clinical
Reilly, D. H., & Sugarman, A. A. (1967). Conceptual complexity and psychological differ-
entiation in alcoholics. Journal of Nervous and Mental Disease, 144, 14-17.
Rosenberg, S. D. (1979). Relaxation training and a differential assessment of alcoholism. Un-
published doctoral dissertation, California School of Professional Psychology, San
Diego. (University Microfilms No. 8004362)
Rotter, J. B. (1966). Generalized expectancies for internal versus external control of rein-
forcement. Psychological Monographs, 80, (Whole No. 609).
Sanchez-Craig, M. (1980). Random assignment to abstinence or controlled drinking in a
cognitive-behavioral program: Short-term effects on drinking behavior. Addictive Be-
haviors, 5, 35-39.
Sanchez-Craig, M., Annis, H. M., Bornet, A. R., & MacDonald, K. R. (1984). Random
assignment to abstinence and controlled drinking: Evaluation of a cognitive-be-
havioural program for problem drinkers. Journal of Consulting and Clinical Psychology,
52, 390-403.
Schmidt, M. R. (1978). Structuring treatment programs on the basis of control orientation of
alcoholics. Unpublished doctoral dissertation, University of Nebraska, Lincoln, NE.
(University Microfilms No. 7814709)
Schmidt, W., Smart, R. G., & Moss, M. K. (1968). Social class and the treatment of alcoholism.
Toronto: University of Toronto Press.
Selzer, M. L. (1971). The Michigan Alcoholism Screening Test: The quest for a new diag-
nostic instrument. American Journal of Psychiatry, 127, 1653-1658.
Shaw, I. A. (1951). The treatment of alcoholism with tetraethylthiuram disulfide in a state
mental hospital. Quarterly Journal of Studies on Alcohol, 12, 576-586.
Smart, R. G. (1978). Characteristics of alcoholics who drink socially after treatment. Alco-
holism: Clinical and Experimental Research, 2, 49-52.
Smith, C. J. (1969). Alcoholics: Their treatment and their wives. British Journal of Addiction,
115, 1039-1042.
Thornton, C. c., Gottheil, E., Gellens, H. K., & Alterman, A. I. (1977). Voluntary versus
involuntary abstinence in the treatment of alcoholics. Journal of Studies on Alcohol, 38,
1740-1748.
Thornton, C. c., Gottheil, E., Gellens, H. K., & Alterman, A. I. (1981). Developmental
level and treatment response in male alcoholics. In E. Gottheil, A. T. McLellan, & K.
A. Druley (Eds.), Matching patient needs and treatment methods in alcoholism and drug
abuse. Springfield, IL: Charles C Thomas.
Tomsovic, M., & Edwards, R. V. (1970). Lysergide treatment of schizophrenic and non-
schizophrenic alcoholics. Quarterly Journal of Studies on Alcohol, 31, 932-949.
Trice, H. M., & Roman, P. M. (1970). Sociopsychological predictors of affiliation with
Alcoholics Anonymous: A longitudinal study of "treatment success." Social Psychia-
try, 5, 51-59.
Vannicelli, M. (1978). Impact of aftercare in the treatment of alcoholics: A cross-lagged

panel analysis. Journal of Studies on Alcohol, 39, 1875-1886.
Vannicelli, M. (1979). Treatment contracts in an inpatient alcoholism treatment setting.
Voegtlin, W. L., & Broz, W. R. (1949). The conditioned reflex treatment of chronic alco-
holism. X. An analysis of 3125 admissions over a period of ten and a half years. Annals
of Internal Medicine, 30, 580-597.
Vogel, M. D. (1960). The relation of personality factors to GSR conditioning of alcoholics:
An exploratory study. Canadian Journal of Psychology, 14, 275-280.
Vogel, M. D. (1961a). The relationship of personality factors to drinking patterns of alco-
holics: An exploratory study. Quarterly Journal of Studies on Alcohol, 22, 394-400.
Vogel, M. D. (1961b). The relationship of GSR conditioning to drinking patterns of alco-
holics. Quarterly Journal of Studies on Alcohol, 22, 401-410.
Vogler, R. E., Compton, J. V., & Weissbach, T. A. (1975). Integrated behavior change
techniques for alcoholism. Journal of Consulting and Clinical Psychology, 43, 233-243.
Vogler, R. E., Weissbach, T. A., Compton, J. V., & Martin, G. T. (1977). Integrated
Walker, R. D., Donovan, D. M., Kivlaha~, D. R., & O'Leary, M. R. (1983). Length of stay,
neuropsychological performance, and aftercare: Influences on alcohol treatment out-
Wallerstein, R. S. (1958). Psychologic factors in chronic alcoholism. Annals of Internal
Medicine, 48, 114-122.
Wexberg, L. E. (1953). The outpatient treatment of alcoholism in the District of Columbia.
Quarterly Journal of Studies on Alcohol, 14, 514-524.
Wilkinson, D. A., & Sanchez-Craig, M. (1981). Relevance of brain dysfunction to treatment
objectives: Should alcohol-related cognitive deficits influence the way we think about
treatment? Addictive Behaviors, 6, 253-260.
Winokur, G., Reich, T., Rimmer, J., & Pitts, F. N., Jr. (1970). Alcoholism. III. Diagnosis
and familial psychiatric illness in 259 alcoholic probands. Archives of General Psychiatry,
23, 104-111.
Winokur, G., Rimmer, J., & Reich, T. (1971). Alcoholism. IV. Is there more than one type
of alcoholism? British Journal of Psychiatry, 118, 525-531.
9
Early Intervention with Problem
Drinkers
GEIR BERG AND ARVID SKUTLE
Early-stage problem drinkers have received relatively little attention

from either researchers or clinicians even though they are in the majority
compared to those fitting diagnostic criteria for alcoholism or alcohol
dependence. The reasons for their being ignored are many. Among
them are the following:
1. There are no standard screening instruments or procedures for
identifying early-stage problem drinkers.
2. "Either/or" thinking still dominates among professionals: "You
are either an alcoholic or not." The alternative, defining different de-
grees of alcohol dependency, is missing.
3. The general attitude towards alcoholics and problem drinkers is
negative and stigmatizing. Many professionals are frustrated by these
groups and pessimistic about their prognosis.
4. Most of the treatment centers for alcoholics are not attractive
enough to the early-stage problem drinker because of their dependence
on total abstinence as sole treatment goal.
5. Once identified and recruited, the practitioners' competence to
treat the problem drinker are, in many cases, not adequate. There is a
GEIR BERG AND ARVID SKUTLE • The Hjellestad Clinic, Hjellestad, Norway. The ma-
terial in this chapter is taken from a paper presented at the Third International Conference
on Treatment of Addictive Behaviors, North Berwick, Scotland, August 12 to 16,1984.
205
206 GEIR BERG AND ARVID SKUTLE
need for special training programs for the professionals who would treat
problem drinkers.
The present study is an evaluation of four behavioral treatment
methods with 48 self-referred early-stage problem drinkers. The treat-
ment took place at out outpatient unit at Hjellestad-Klinikken. The
study was conducted in Bergen in 1983 to 1984. The following questions
were raised:
1. Is it possible to recruit and motivate problem drinkers for par-
ticipation in an early intervention program? Until now this has not been
done systematically in Norway.
2. If so, which of four intervention methods, varying in content and
cost, is the most effective in attaining the intervention goal?
In terms of the Prochaska and DiClemente model (1982), the ques-
tion is how to motivate contemplators to make a decision and take
action, to participate in the treatment program and reduce their alcohol
consumption.
In the study the following treatment programs were used.
• Group 1: bibliotherapy based on behavioral self-control training.
The basic elements of this program are described in the book How
to Control Your Drinking (Miller & Munoz, 1982). (2 group sessions
x 2 hours = 4 hours.)
• Group 2: behavioral self-control training with therapist-adminis-
trated group sessions. The content is based on the same reference
as mentioned for Group 1, but the setting is different. (6 group
sessions x 2 hours = 12 hours.)
• Group 3: training in coping skills. This method is a modified ver-
sion of the Relapse Prevention Program (Chaney, O'Leary & Mar-
latt, 1978; Marlatt, 1980). (6 group sessions x 2 hours = 12 hours.)
• Group 4: a combination of behavioral self-control training and
training in coping skills. (8 group sessions x 2 hours = 16 hours.)
The four intervention methods were given to four groups of clients.
Because of ethical concerns and because the question was which of the
four treatment methods would be most effective, no untreated control
group was included.
METHOD
A MOTIVATIONAL ApPROACH
During all contact with the clients the therapists (the authors) tried
to avoid a moralistic attitude, which is characterized by blaming the
EARLY INTERVENTION WITH PROBLEM DRINKERS 207
client for his or her drinking behavior and by creating guilt and feelings
of personal insufficiency. The clients were not treated as helpless victims
suffering from an "alcoholic disease," without individual responsibility
and positive resources. As an alternative we adopted an empathic
therapeutic style that we presumed had a more motivational effect on
the clients (Miller, 1983). The motivational approach is characterized by
the following features:
1. A de-emphasis on labeling. Instead of the terms alcoholic and
alcoholism, "different degrees of dependency" and "problems related to
drinking" were used;
2. Individual responsibility. The clients were provided with assess-
ment information, but it was up to the clients to decide if there were any
problems and what to do with them. The therapist clarifies options, but
the client makes the decisions.
3. Internal attribution. The client is not assumed to be a helpless
victim of external events. He or she is in control of the situation and is
able to reduce alcohol consumption with adequate counseling. Progress
is attributed to the client's own efforts.
4. Cognitive dissonance. By informational feedback of the assess-
ment results to the client, an inconsistency between emotions-attitudes
and behavior is produced. A motivational condition is created and the
next stage is to restore consistency through behavior change.
RECRUITMENT OF CLIENTS
Clients were recruited through a local newspaper announcement

and through a presentation of the project plans in the same newspaper
and on the local radio. Newspaper announcement has been found to be
an effective way of recruiting subjects for early treatment programs
(Duckert, 1982; Miller, Taylor, & West, 1980; Pomerleau, Pertscuck,
Adkins, & Brady, 1978; Sanchez-Craig, Wilkinson, & Walker, 1984;
Vogler, Weissbach, & Compton, 1977). The response was very positive.
Within one week the four groups were filled. In 83% of the cases the
contact was established on the basis of the newspaper announcement.
All the clients were self-referred.
SCREENING AND GROUP ASSIGNMENT
After a short telephone orientation with clients about the program,

a 45-minute screening interview was conducted. Because the target
group was early-stage problem drinkers, all severely dependent subjects
were excluded from the study. This was the case for two subjects, who
were referred to the outpatient unit at Hjellestad-Klinikken. Other crite-
ria for exclusion from our study were, (a) pregnancy, (b) previous treat-
ment for alcoholism, (c) diagnosis (DSM-III) of alcohol idiosyncratic in-
toxication, (d) history of withdrawal delirium, (e) dependency on other
drugs, (f) evidence of liver damage, (g) a self-reported duration of prob-
lem drinking in excess of 10 years, or (h) medical illness. Severe depen-
dence was defined as increased tolerance for alcohol with severe with-
drawal symptoms.
During the screening interview, clients were informed about the
program and signed a statement of informed consent to participate. All
clients with collaterals gave us permission to interview the collaterals,
for which they signed another statement of informed consent. The as-
signment of the final 48 clients to the four groups was random, based on
a table of random numbers. Each group had 12 clients.
ASSESSMENT
The assessment instruments were the following:

1. The Comprehensive Drinker Profile, developed by Marlatt and Miller
(1984). This is a standardized interview including questions on demog-
raphy, the Michigan Alcoholism Screening Test and other alcohol-relat-
ed questions. In the follow-up interviews we used a shorter version.
2. The Severity of Alcohol Dependence Questionnaire, developed by
Stockwell, Murphy, and Hodgson (1979), measured the degree to which
clients were experiencing the syndrome of alcohol dependence.
3. Symptom Check List 90 (SCL-90), Derogatis, Lipman, & Covi (1973),
a well-known personality questionnaire.
4. High Risk Situation Questionnaire, measuring the client's perceived
ability to cope with high-risk situations. The questionnaire is based on
the Situation Difficulty Questionnaire by Chaney, O'Leary, & Marlatt
(1978). Forty-eight more typically Norwegian situations are selected, for
example: how difficult is it to pass the spirits- and wine-monopoly-shop
[that we have in Norway] on the way from the job Friday afternoon
without entering the shop and buying a bottle?
TREATMENT
A group treatment format was chosen because of the mutual sup-

port among group members it provides and the possibilities of discus-
sion and role playing. The authors served as the group leaders. In
addition, group format was preferred because of low cost and relatively
high effectiveness compared to an individual format, as documented by
Miller and Taylor (1980).
Group 1: Bibliotherapy (BSCT)

The bibliotherapy group received two group sessions: one introduc-
tory session and one final assessment session. In between, they received
weekly materials from us, including a self-help manual and a supply of
self-monitoring cards. The content of the manual was as follows:
Session 1. Goal Setting and Self-Monitoring. The clients set a concrete
and realistic goal toward which progress could be measured. To monitor
the strength of ethanol in different alcohol beverages, a Standard Eth-
anol Content (SEC) or standard "one-drink" unit was used. One such
unit contains 0.5 oz or 15 ml pure ethanol,-for example the ethanol
content of either 10 oz of 5% beer, one glass or 4 oz of 12% table wine,
one glass or 2.5 oz of sherry (20%), or one glass or 1 oz of 50% whisky.
Clients were provided with self-monitoring cards which could increase
their awareness of their actual alcohol consumption and provided a
record of progress.
Session 2. Controlling Drinking Rate. Clients were encouraged to re-
duce their drinking rate, for example by switching to a less preferred
beverage, slowing the pace of drinking by increasing number of sips per
drink and by avoiding gulping, and refusing unwanted drinks.
Session 3. Self-Reinforcement. In this session clients learned to rein-
force progress with a material reward or by self-reinforcement, for exam-
ple by saying, "so far I have done well."
Session 4. Functional Analysis. The purpose of the functional analysis
was to identify antecedents of overdrinking and high-risk situations and
to find coping strategies, either by avoidance of the problem situations
or by encountering and handling the situations better.
Session 5. The Meaning of Drinking. Clients were encouraged to deter-
mine the meaning of drinking for them and "new roads" from anteced-
ents of drinking to the pleasant effects that alcohol usually gives them,
but this time without alcohol. For example: "I shall not have a drink or
two in order to be accepted by my friends. Instead 1 shall be more
assertive and say 'no thanks' to unwanted drinks." In this way psycho-
logical dependence on alcohol can be reduced.
Session 6. Final Assessment. The main purpose of this session was to
evaluate a client's progress through the program and to consider relapse
strategies. Clients were provided with a list of DO's (to use) and
DON'Ts (to avoid).
Group 2: BSCT-Therapist Directed

This group received BSCT in groups during 6 sessions. The content
of the treatment was the same as that described above.
Group 3: Training in Coping Skills

The main purpose of this group was to identify high-risk situations,
that is, situations that involve overdrinking, and to teach clients to cope
with them, for example by role playing. The method is based on a
modified version of Marlatt's relapse prevention model (Cummings,
Gordon, & Marlatt, 1980). It was emphasized, to increase their aware-
ness of the decision stage, that overdrinking to a large extent was a
result of their own decisions and behavioral responses. Another compo-
nent of this method was relaxation training. After a short presentation of
the techniques in the group, clients received an audio casset with a
relaxation program for rehearsals at home.
Group 4: A Combination of BSCT and Coping Skill Training

The purpose of this group was to see if treatment beyond BSCT
would have any effect. After the 6 BSCT-sessions, clients had 2 addi-
tional sessions with special training in coping skills.
Groups I, 2, and 4 received a very didactic, educational, and stan-
dardized treatment format whereas Group 3 was more open to indi-
vidual initiatives and benefitted more from "group process," that is, the
expression of emotions and personal attitudes, interpersonal feedback,
and a higher activity level among the clients during the sessions.
FOLLOW-UP ASSESSMENT
Interviews with the clients were conducted at 3, 6, and 12 months

following treatment termination, and with collaterals at 6 and 12 months.
RESULTS
CLIENT ATTRITION
The number of excluded clients, no-shows and dropouts were rela-

tively low. After the screening interview 2 clients were excluded. Of the
remaining 48 clients, one did not show up at the start of treatment.
During treatment 4 clients attended less than 50% of the sessions. They
were categorized as "not treated" and were excluded from the follow-up
interviews. During the follow-up one person was lost.
PRETREATMENT MEASURES
The majority of the clients were married, middle-aged men (see

Table 1). There were no significant differences among the four groups
on income or years of education. Because of a 72-year-old man in Group
2, there was a significant difference in age between Group 2 and the
three others.
ALCOHOL CONSUMPTION
Figure 1 shows the changes in alcohol consumption from intake to

3-, 6-, and 12-month follow-up. Consumption was measured in SECs
(Standard Ethanol Content or Standard Units) per week. There were no
significant differences among the four groups at either intake or follow-
up points. The differences in consumption from intake to follow-up at 3,
6, and 12 months however, were significant within all four groups (p
values range from .0001 to .035). Results at 3-month follow-up are pre-
dictive of status at later points.
All clients had done self-monitoring by using special cards every
TABLE 1.
Demographic Data
Groups: 2 3 4 All
Sex Male 10 8 7 9 34
Female 1 2 4 2 9
Age Mean 38 49 44 40 43
Range 29-72
Marital status Single, never married 1 0 1 1 3
Married 7 7 7 6 27
Separated 1 1 0 2 4
Divorced 2 2 3 2 9
Years of educa- Mean 12.5 13.3 12.7 13.6 13
tion Range 7-20
Employment Worker 5 4 3 3 15
Supervisor 0 0 2 3 5
Officer 5 3 4 4 16
Employer 1 2 2 1 6
Retired 0 1 0 0 1
Family income 6,250-12,500 1 2 2 1 6
in US dollars 12,500-18,750 3 2 4 5 14
18,750-25,000 2 2 1 1 6
25,000-31,250 3 1 1 2 7
31,250-37,500 2 3 2 1 8
>37,500 0 0 1 1 2
•
SEC. PEA WEEK
45
G,..oup 1
40
315 ~ Group 2
30
II Group 3
nm
2 !!I
Group 4
20
15
10
!!I
0
IHTAk£ 3 IIOHTH 8 IIOHTH 12 IIOHTH
TIME
FIGURE 1. Weekly alcohol consumption.
week during treatment. The self-monitoring data (Figure 2) show that

the drop in consumption had already taken place by the week after start
of treatment. At this point subjects had done self-monitoring for one
week. The changes are significant within all groups, and there are no
significant differences among groups. This pattern is strikingly similar to
American data reported by Miller (1978; Miller & Taylor, 1980; Miller,
SECe/WEEK
40
30
2!5
20
1!5
10
oL-----------------------------------------------
00 ~ ~ • m • ~ • • •
TIME (WEEKS/MONTHS)
FIGURE 2. Weekly alcohol consumption (with self-monitoring during the treatment
period).
Taylor & West, 1980). After this initial drop, weekly consumption re-
mains stable through treatment and the three follow-up periods (Figure
2). The decrease in alcohol consumption took place before the treatment
programs started. This could mean that a decision was made to reduce
drinking in this early and important phase.
At intake (baseline) a majority of the clients, that is, 51 % (23 clients),
had a mixed drinking pattern (Figure 3). A mixed drinking pattern or
combination pattern drinking means a pattern whereby a person drinks
at least once per week with a regular weekly pattern, but also has heav-
ier episodes deviating from the typical pattern by at least 5 SEes within
one day. It was especially on the weekends that these subjects had
heavy drinking episodes. This is a typical Norwegian way of drinking.
Fifteen clients (36%) had a regular drinking pattern (defined as drinking
at least once per week and about the same amount every week without
periodic episodes of heavier drinking). Five clients (13%) had a periodic
drinking pattern (defined as a client drinking less often than once a
week, and being abstinent between drinking episodes). Drinking pat-
terns changed from intake to follow-up, and there are only small dif-
ferences from 3 to 12 months (Figure 3). At one-year follow-up the
largest group of subjects were no longer regular pattern drinkers with a
concentrated consumption on the weekends, but were periodic drinkers
using alcohol less often than once a week, and were abstinent between
drinking episodes.
Another way to analyze drinking behavior is to divide weekly alco-
hol consumption into categories (see Figure 4). At intake more than 53%
" of cHent. Drinking-pettern

!So
P"r10d1c
40 Aegular
M1x"d
30
Abst1nsnt
20
10
0'-----
B••• llne
TIME
FIGURE 3. Drinking patterns.

" OF CLIENTS
SEC. PER NEEK:
so 0-10 SEC a
40 11-20 SEC.
21-30 SECa
30
> 30 sec.
20
10
0'------
TIME
FIGURE 4. Distribution of alcohol consumption (SEes per week).
(23 clients) of the clients drank over 30 SECs per week. (That is about 1.5
bottles, or 110 cl, of 80 proof whisky or vodka). By contrast, at one-year
follow-up exactly half (21 clients) were drinking 10 SECs (0.5 bottle) or
less per week. If 20 SECs per week or less, that is, a consumption within
Categories 1 and 2 (one bottle or 70 cl of 80 proof vodka), is accepted as
safe or acceptable drinking, 78.5% (33 clients) of the clients reached this
goal at the one-year follow-up. At intake only 23.2% (10) of them drank
20 SECs or less per week.
All nine clients in Group 3 who drank above 20 SECs at intake
reduced their consumption below that level one year later. The figures
for the four groups are presented in Table 2. The total mean reduction
for all clients (from intake to one-year follow-up) was 64%. The reduc-
tion in consumption within each group are presented in Table 3.
TABLE 2.
Number of Clients Drinking More Than
20 Sees Per Week
At 12-month
Group At intake follow-up
1 7 4
2 7 2
3 9 0
4 10 3
TABLE 3.
Percent of Group
Reduction in Alcohol
Consumption
Group % Reduction
1 57
2 63
3 67
4 70
PROBLEMS RELATED TO DRINKING
The Michigan Alcoholism Screening Test is a part of the Compre-

hensive Drinker Profile and was administered at intake and at follow-
up. The questionnaire measures two variables, problems related to
drinking (Mast) and physical dependence (Ph). Figure 5 shows the per-
cent of clients in each of the four MAST score categories. Scores at intake
were based on the total life span of the clients, but the one-year follow-
up scores related to the 3 months prior to interview. At intake 56% (24
clients) had scores within Category 3 (indicating significant life problems
related to alcohol). At one-year follow-up the percentage within the
same category was 12 (5 clients). At one-year follow-up 69% (29 clients)
:I OF CLIENTS
SECa PER WEEK:
so 0-10 SEC"
40 ~ tt-::!O SEC.
30
~ 21-30 SECs
lIll > 30 SEC_

::!O
10
oL-----
INTAKE 12
IUlNlHS
TIME
FIGURE 5. Distribution of MAST scores.

had scores within Category 1 (indicating no or mild problems with

drinking).
Scores on physical dependence showed a similar pattern (Figure 6).
Intake scores on PH reflected experience from the total past, whereas
one-year follow-up scores were based on the last 3 months. At intake we
can see that 70% (30 clients) had scores in Category 2 (significant symp-
toms of physical dependence). 21 % (9) fell into Category 1 (mild symp-
toms of alcohol dependence). At one-year follow-up 93% (39) had scores
in the lowest category. Seven percent (2) had scores in Category 2, and
no one was in Category 3 (more serious dependence on alcohol). One
person did not show a reduction in Ph values from intake to the one-
year follow-up.
There have been some differences in views on the question of
whether or not a reduction in alcohol consumption leads to a reduction
in life problems. By using the intake and parallel follow-up interviews,
in which questions about significant life problems were asked, it was
possible to analyze this issue (Figure 7). There were no significant dif-
ferences among groups, but within each group the differences were
significant from intake to the 3- and the 12-month follow-up (the results
at 6 months were almost identical to those at the one-year follow-up).
Only three clients (from different groups) did not report a reduction in
life problems. The mean number of life problems for all clients decreased
from 6.3 (at intake) to 2.2 (at one-year follow-up). At the 12-month
follow-up, 64% (27 clients) reported no problems related to drinking,
whereas all had reported such problems at intake.
I OF CLIENTS
PH-ecorea:
90
0-4
eo
70 5-10
60
11-20
50
40
20
10
oL-----
TINE
FIGURE 6. Distribution of PH scores.

•
PROBl.EMS
7
GFIOUP
6
FZ1 GFIOUP
II
5
.. GFIOUP 3
mJ GROUP ..
I
~II
o ~
INTAKE
~12
MONTHS
TIME
FIGURE 7. Life problems.
High-risk situations, factors that could increase the probability of

heavy drinking in a particular situation, were assessed too (Table 4). At
intake most subjects had an "internal" or intra personal high-risk situa-
tion, but at one-year follow-up most of them reported an "external" or
interpersonal factor as their high-risk situation. Thirteen clients did not
report any high-risk situation at the 12-month follow-up.
In the statement of informed consent that all subjects completed
there was an agreement with the clients that, if necessary (e.g., in a
crisis), they could make contact if they needed help. Four clients in
TABLE 4.
High-Risk Situations (H-R-S)
I. At intake
• 30% (13 clients) reported no high-risk situation. (There is a small change compared to
one-year follow-up).
• Most of them, that is, 42% (18), reported that a negative emotional state (depression,
anxiety, boredom, etc.) could lead to problems.
• 16% (7) reported craving for alcohol as a high-risk factor.
II. At 12 months follow-up

• 45% (19) now reported that a positive emotional state together with other people is
the most important H-R-S.
• The decrease in the number of clients reporting a negative emotional state as a H-R-S,
12% (5), is also obvious.
Group 1, one client in Group 2, one in Group 3, and two clients in Group
4 had contact with us, either by telephone or personal contact. One
client in Group 1 and one in Group 3 received treatment elsewhere (our
own outpatient unit and local detoxification center). In total, we had
contact with 8 clients or 19% of them, and Group 1 (the bibliotherapy)
received the most help.
In the follow-up interviews the clients did self-evaluation with re-
gard to alcohol consumption. Most (60%) said they were drinking much
less, one third (33%) said they were drinking less, 5% reported drinking
the same, and one client indicated an increase in consumption. In spite
of the fact that 93% reported drinking less, 65% still wanted to reduce
their drinking even more.
COLLATERAL DATA
Collaterals were interviewed by telephone, and were asked ques-

tions about the clients' alcohol consumption. In 19 cases collaterals re-
ported the same alcohol consumption as the clients, in 7 cases collaterals
reported more drinking than the clients, and in 7 cases collaterals re-
ported less drinking than the clients. Collateral data were unavailable
for 9 cases. There were no significant differences on drinking variables
or other variables between clients with collaterals and those without,
either at intake or at any follow-up point.
DISCUSSION
There were no significant differences among the four groups on any

variable at follow-up. Mean values in each group showed significant
reductions in alcohol consumption and in life problems. In all groups
weekly consumption decreased significantly from intake to the first
week after self-monitoring of drinking behavior. This indicates that a
motivational approach during the assessment period and informational
feedback to the clients about issues concerning drinking behavior might
have been sufficient to help this target group of early-stage problem
drinkers to change behavior. Another hypothesis is that these early-
stage problem drinkers were, so to speak, "ready for treatment," and
were inspired because they were accepted to participate in the study.
Therefore they made decisions to activate self-healing or self-help pro-
cesses before the treatment actually started. The treatment programs
may have functioned as maintaining factors with regard to the behavior
change. Both these hypotheses remain to be verified.
In any case, the minimal treatment program (bibliotherapy) can be
EARL Y INTERVENTION WITH PROBLEM DRINKERS 219
recommended to early problem drinkers who have a relatively low con-

sumption at intake. Clients with a high consumption at intake may
profit less from minimal treatment. For example four of the eight clients
who received additional help between 3 and 12 month follow-ups be-
longed to Group 1 (bibliotherapy), and all four showed a higher than
mean alcohol consumption at intake, and at one year follow-up nine
clients still had a weekly consumption above 20 SECs. Four of these
clients came from Group 1, and they were the same clients who had
received help.
The results indicate the necessity for a control group receiving no
treatment, as that could have made it possible to answer the question of
whether participation in assessment and informational feedback from
this assessment would lead to a behavior change at a one-year follow-
up. Nevertheless, in spite of these reservations, there is no doubt that
the majority of the clients in this study really did something positive
with their problem behavior. They reduced their problem drinking, and
their life problems decreased.
The study also shows that a newspaper announcement is an effec-
tive way of recruiting early-stage problem drinkers, at least in Norway.
A motivational approach emphasizing individual responsibility, infor-
mational feed-back creating cognitive dissonance, de-emphasis on label-
ing, and internal attribution of positive behavior change can be a helpful
approach in clinical work with early-stage problem drinkers. This ap-
proach may be particularly helpful for clients in the contemplation stage
as a way of encouraging them toward action.
REFERENCES
Cummings, c., Gordon, J. R., & Marlatt, G. A. (1980). Relapse: Prevention and prediction.
In W. R. Miller (Ed.), The Addictive Behaviors. Treatment of alcoholism, drug abuse, smok-
ing, and obesity. (pp. 291-321). Oxford: Pergamon Press.
Derogatis, L. R., Lipman, R. S., & Covi, L. (1973). SLL-90: An outpatient psychiatric rating
scale-preliminary report. Psychopharmacology Bulletin, 9, 13-28.
Duckert, F. (1982). Control training in the treatment of alcohol abusers (Mimeograph No. 62).
Oslo: National Institute of Alcohol Research.
Marlatt, G. A. (1980, August). Relapse Prevention: A Self-Control Program for the Treat-
ment of Addictive Behaviors. Unpublished manuscript, University of Washington.
Marlatt, G. A., & Miller, W. R. (1984). The Comprehensive Drinker Profile. Odessa, FL:
Psychological Assessment Resources.
Miller, W. R. (1978). Behavioral treatment of problem drinkers: A comparative outcome
study of three controlled drinking therapies. Journal of Consulting and Clinical Psychol-
ogy, 46, 74-86.
Miller, W. R. (1983). Motivational interviewing with problem drinkers. Behavioral Psycho-

therapy, 11, 147-172.
Miller, W. R., & Munoz, R. F. (1982). How to Control Your Drinking. Albuquerque, NM:
University of New Mexico Press.
Miller, W. R., & Taylor, C. A. (1980). Relative effectiveness of bibliotherapy, individual
and group self-control. Addictive Behaviors,S, 13-24.
Miller, W. R., Taylor, C. A, & West, J. (1980). Focused versus broad-spectrum behavior
Pomerleau, O. F., Pertscuck, M., Adkins, D., & Brady, J. P. (1978). A comparison of
behavioral and traditional treatment methods for middle-income problem drinkers.
Journal of Behavioral Medicine, I, 187-200.
Sanchez-Craig, M., Wilkinson, D. A, & Walker, K. (in press). Theory and methods for
secondary prevention of alcohol problems: A cognitively-based approach. In M. Cox
(Ed.), Treatment and Prevention of Alcohol Problems: A Resource Manual, Academic Press.
New York.
Stockwell, T. R., Murphy, D., & Hodgson, R. J. (1983). The Severity of Alcohol Depen-
dence Questionnaire: Its use, reliability and validity. British Journal of Addiction, 78,
145-155.
Vogler, R. c., Weissbach, T. A, & Compton, J. C. (1977). Learning techniques for alcohol
abuse. Behavioral Research and Therapy, IS, 31-38.
10
Strategies of Change in Eating
Disorders
CHRISTOPHER P. FREEMAN
INTRODUCTION
I have been asked to provide a review of strategies of change in eating

disorders. This chapter may be somewhat out of place in this volume,
firstly because this is the only paper on eating disorders and, secondly,
because the theme of the book is that of addictive behavior and I have
considerable reservations about whether two of the syndromes I am
going to discuss have, in fact, any element of addictive behavior at all.
It concerns me a great deal that, in a number of papers in this
volume, eating disorders appear to have been grouped with other sorts
of behaviors, such as drinking and smoking. The implicit assumption is
that obesity is a disorder of overconsumption of food in the same way
that alcohol dependence is related to excessive alcohol intake. As I hope
to show, the evidence for this is minimal.
I intend to discuss three groups of disorders: Obesity, anorexia
nervosa, and bulimia. I will devote more space to a discussion of obesity
and bulimia and relatively little to the more uncommon syndrome of
anorexia nervosa.
CHRISTOPHER P. FREEMAN • Department of Psychiatry, University of Edinburgh,

Royal Edinburgh Hospital, Morningside Park, Edinburgh EHlO 5HF Scotland.
221
222 CHRISTOPHER P. FREEMAN
OBESITY
STRATEGIES OF CHANGE IN OBESITY
For the last 20 or 30 years, treatments for obesity have been based
on the essential assumption that fatness is due to some abnormality of
behavior, and that this abnormality is either the intake of an excessive
amount of food on a regular basis or, less frequently, subnormal phys-
ical activity. Treatments have therefore concentrated on efforts to correct
the psychological causes of overeating, to educate people about diet, or
to use learning-theory techniques to modify eating behavior in a system-
atic way.
The results of all this effort have been very disappointing. Few
people change and even fewer maintain their change. If one looks more
closely at recent research it is perhaps clear why this is so: the majority
of studies have shown no difference in the food intake of obese infants,
obese children, obese adolescents, or obese adults, when compared
with their lean counterparts.
Whether obesity is truly caused by overeating, or by periods of
overeating, is not yet certain, but what is clear is that obesity can be
maintained without overeating. In fact, it can often be maintained even
with undereating. Physicians will describe patients admitted to meta-
bolic wards who are put on a strictly controlled 750 to 1000 K calorie a
day diet, yet do not lose weight. Such findings are incompatible with the
notion of curing obesity by normalizing eating behavior.
One of the most powerful ways that body weight is maintained
appears to be regulation around a predetermined set point. This theory
has recently been reviewed by Keesey and Corbut (1981). They have
shown that rats who are overfed until obese develop physiological con-
trols that act to sustain their weight at their new obese level. In other
words, once a set-point weight becomes established, be it overweight or
underweight, the body seems to defend this new set point against
changes in calorie intake, As yet, relatively little is known about how
such shifts in set point occur.
A study by Kromhout (1983), the Zutphen Study from the Nether-
lands, showed that middle-aged men in the highest quartile of body fat
range consumed on average 300 to 400 kilocalories less than those in the
lowest quartile. The Department of Agriculture figures from the United
States (Friend, 1974) show that per capita calorie intake has fallen by
about 5% during this century; yet the incidence of obesity is higher than
it was 70 to 80 years ago. A study in Glasgow on adolescent boys be-
tween 1964 and 1971 (Durnin, Lonergan, Good, & Ewan, 1974) showed
an average reduction in daily calorie intake from 2795 to 2610 kilocalo-
STRATEGIES OF CHANGE IN EATING DISORDERS 223
ries; over the same period body fat increased from 16.3% to 18.4%.
These findings point to obesity being, if anything, a state caused by
inactivity, rather than a state caused by abnormal eating. The decrease
in activity may be small but may, over many months or years, help
establish and maintain obesity.
Stern (1984) quotes a simple example:
The Illinois Bell Telephone Company has estimated that in the course of one
year, an extension phone saves approximately 70 miles of walking. For some
people, this could be the calorific equivalent of 2-31b of fat or 7,000-10,000
Kcal. (p. 133)
As Stern has concluded, obesity is not a single disorder. For many

people obesity appears to be a consequence of inactivity, low calorific
requirements, or both.
Two other important factors should be discussed in relation to strat-
egies of change in obesity. The first is the widely held belief that to
become thinner is to become healthier. Most people are not thin. In fact
by widely accepted, but entirely arbitrary standards, 40% of American
woman are fat. As Wooley and Wooley (1984) point out:
It would be very cruel of nature if it were to persistently thwart the best
efforts of a substantial proportion of the population in their efforts to reach
and stay in a weight range which caused healthy survival. (p. 186)
Is, therefore, thinner-healthier? Keys (1980) reviewed 13 prospec-

tive studies on obesity and mortality and concluded that the risk of early
death increases only in extremes of underweight and overweight, with
no impact on the middle 80%. This finding applied only to women. A
similar conclusion can be drawn from data from the Framlington Study
in 1980. In a paper entitled "Body Build and Mortality," Sorlie, Gordon,
and Kennel (1980) showed that being underweight is more dangerous
than being overweight. There was no relationship between being over-
weight and increased mortality for the middle 60% of the weight range.
In a stud~T carried out in Arizona by Pettitt, Lisse, Knowles, and
Bennett (1982), the safest weight range was 167% to 190% for women
and 145% to 176% for men. (These percentages are of the Society of
Actuaries standard "desirable" weights.) Similarly, Noppa, Bengtsson,
Wedel and Wilhelmsen (1980) found an inverse relationship between
death from all causes of obesity.
These studies, then, raise questions about one of the main rational
bases used to justify the treatment of the mild to moderately obese.
Perhaps the massively obese should be considered separately. In this
group there seems clear evidence of increased mortality and morbidity,
let alone simple discomfort and lack of mobility.
The second widely held belief is that dieting makes you feel better.
Although this may be true in the early stages of a diet, the price that
many obese individuals have to pay to try to achieve a socially accept-
able body is considerable. Studies of both successful and unsuccessful
dieters indicate that there is a considerable psychological morbidity as-
sociated with dieting. A recent review by McReynolds (1982) shows that
obese people undergoing help for their weight problems show evidence
of psychological disturbance in terms of depression, anxiety, and gener-
al distress, whereas obese individuals in the general population who are
not dieting show comparable or better psychological adjustment than
the nonobese. Woolley and Woolley (1984) note that there are no good
studies of the attitudes, life-styles, and coping strategies of well adjusted
obese people.
Finally, we have to consider the impact that medicine's demand for
universal slenderness has on society. Although it seems unlikely that
this demand has caused the weight obsession of our current society,
medicine could go a long way to defuse the situation by refusing to
define fatness per se as a disease and refusing to treat it. The current
epidemic of eating disorders, such as bulimia, seem closely related to
such attitudes, namely the universal desire for slimness and the anti-fat
prejudices that our society has-such as that being fat is ugly, that being
fat is sexually undesirable, and that being fat indicates weakness. The
conclusions of this argument from Woolley and Woolley (1984) are sum-
marised in Table 1.
It seems to me that Woolley and Wolley's arguments are very
cogent. The burden of proof clearly rests with those who claim that mild
to moderate obesity is either physically or psychologically unhealthy
and with those who claim that there are successful strategies for change,
the benefits of which outweigh the harm they may do (Table 2).
What I have said so far applies to mild and moderate obesity. What
about those classified as severely obese? Stunkard (1984a) estimates that
there are about 40 million mildly obese people in the United States, 2
million moderately obese, and 200,000 severely obese individuals. The
only strategies that reliably produce enduring change in the treatment of
the severely obese are surgical interventions. Such treatments were in-
troduced about 20 years ago. The first generation of operations was
mainly the technique of jejunoileal by-pass. The aim of this type of
operation was to reduce dramatically the area of the small intestine so
that only about 18 inches was active. The surgical complications of this
operation were often serious; mortality was around 5% and postopera-
tive complications, such as severe flatulence and recurrent vomiting,
were common. Thus, although the operations were successful in pro-
moting weight loss, the risk/benefit ratio was not clearly in favor of
surgical intervention, even in the most severe and intractable cases of
obesity.
TABLE l.
Should We Treat Obesity At All?
Findings
• Obesity treatment with the exception of surgical techniques carrying high physical
risks are generally ineffective.
• Individual differences in body size appear to have a strong basis in biology, helping
to account for the extreme measures required to maintain successful weight loss and
high number of therapeutic failures.
• Mild to moderate obesity does not appear to constitute a significant health risk for
women, and possibly not for men.
• An increasingly stringent cultural standard of thinness for women largely supported
by the medical and psychological professions has been accompanied by a steadily
increasing incidence of serious eating disorders in women.
Conclusions
It is hard to construct a rational case for treating any obesity other than massive life
endangering obesity.
We must vigorously treat weight obsession and its manifestations, which are:
(a) Poor self and body image
(b) Disordered eating patterns created by dieting
(c) Metabolic depression produced by dieting
(d) Inadequate nutrition due to constricted eating behavior
(e) Disordered life-styles, often marked by excessive or inadequate exercise.
Note. Reproduced with permission from Woolley and Woolley (1984).
TABLE 2.
Classification of Obesity
Label Mild Moderate Severe
Percent 20 to 40% 41 to 100% > 100%

overweight
Prevalence 90.5% 9% 0.5%
(among
obese women)
Pathology hypertropic hypertropic, hypertropic,

hyperplastic hyperplastic
Complications uncertain conditional severe
Treatment behavior diet and surgical

therapy behavior
(lay) therapy
(medical)
The second generation of operations consists of a variety of gastric

restriction procedures. The aim of such operations is to reduce the vol-
ume of the stomach to as little as 50 ml. The commonest such operation
currently in use is gastric stapling. This consists of restricting the lumen
of the stomach with a row of staples so that the individual postopera-
tively feels full after just one or two mouthfuls of food-the "one bite"
stomach. Gastric stapling is a much safer operation with fewer side
effects and a markedly lower mortality (Mason, 1981). It is as successful
at promoting weight loss as earlier operations.
From a psychological point of view, one of the most interesting
aspects of such surgical interventions is the behavioral and cognitive
changes that occur postoperatively. Halmi, Stunkard, and Mason (1980)
have shown that reducing diets are associated with a high degree of
distress, whereas the emotional responses of postsurgical cases are
much more benign, despite the subjects losing far more weight (Table
3). Favorable consequences of such operations are also more commonly
reported. Seventy-five percent report increased well-being and 53% in-
creased self-confidence. Changes in body image appear to occur, even
before significant weight loss is achieved. Food likes and dislikes
change, and there are increased feelings of satiety after food and de-
creased binge eating. It would seem therefore that gastric stapling sur-
gery does far more than simply alter the functioning of the gastroin-
testinal tract; major changes in both biolugy and in cognitive functioning
occur. Such individuals no longer have to struggle with the biological
TABLE 3.
Emotional Changes and Dieting
Emotional response to dieting
Mild Moderate Severe Total
Depression 20% 25% 15% 60%

Anxiety 19% 30% 23% 72%
Irritability 38% 27% 14% 79%
Preoccupation with food 6% 21% 55% 82%
Reduction in emotional responses when dieting compared with postgastric

bypass state
Less Much less Total
Depression 10% 45% 55%

Anxiety 14% 46% 60%
Irritability 21% 49% 70%
Preoccupation with food 17% 48% 65%
pressures to support a higher weight. They can limit their food intake
with relative ease until a new lower set point is achieved.
MILD AND MODERATE OBESITY
A review of all the strategies of change for individuals who fall

within the group of the mildly and moderately obese is outwith the
scope of this article. There are literally thousands of diet programs,
hardly any of which have been subjected to any sort of evaluation.
When viewed critically, it would appear that the outcome of such pro-
grams is universally dismal and that for every kilogram lost, a kilogram
or more is eventually gained. The saying "dieting makes you fat" has
much truth in it. Weight loss by carbohydrate restriction is achieved by
loss of both lean muscle and body fat. Weight gain that occurs when
diets fail is largely adipose tissue. The individual who has lost weight by
dieting and then regains weight to his or her original level probably has
a higher percentage of body fat. .
- ~/""--1'" ""'M'"
2
........
0
-2
-4
01 COMBINED
.z TREATMENT
.,
01
-6 MEDICATION PHARMACO-
c: (~OCTOR'S
THERAPY
0 OFFICE I
6 -8
BEHAVIOR
~ THERAPY
•• -10
01
-12
-14
-16
o 2 4 6 e
.~----~~~--~',~--------~--------------~,
10 12 14 16 18
Treatment FOIIO~w'Up
MONTHS
FIGURE 1. Weight changes during and after treatment for obesity Reproduced with permis-
sion from Craigshead, Stunkard, & O'Brian (1981).
Two developments are worthy of mention here. First, an increasing

number of diet programs are promoting regular, vigorous exercise. It
does seem clear that 20 to 30 minutes of such exercise at least three times
a week can raise basal metabolic rate, lower body set-point weight, and
produce weight loss with relatively limited restriction of food intake.
The term aerobic exercise, or "aerobics," is misleading and wrong. If any
such term is appropriate, it is anaerobics, namely exercise that exceeds
the body oxygen supply and produces an oxygen debt.
Second, the approach of using pharmacological agents has until
recently been frowned upon by most experts. Stunkard has always been
a strong advocate of behavioral intervention. When he published the
preliminary results of a trial in 1981 (Craigshead, Stunkard, & O'Brien,
1981), he concluded that behavior therapy was the best available treat-
ment and that drugs were contraindicated. The trial whose results are
summarised in Figure 1 compared the effects of pharmacotherapy alone,
behavior therapy alone, and a combination of the two in 98 obese wom-
en over a 6-month period with a one-year follow-up. It is worth noting
that tolerance did not develop to the drug used (Fenfluramine) over the
6-month period. When reviewing the results again at a conference (1983)
(see Stunkard, 1984b), the author concluded that it was possible to lower
a set point on a long-term basis using such drugs and that there was
strong evidence to believe that tolerance does not develop. His conclu-
sions were that appetite supressant drugs should either not be used at
all or used on a chronic, long-term basis.
ANOREXIA NERVOSA
I will mention relatively little about the syndrome of Anorexia Ner-

vosa, which is defined as follows by DSM-III (American Psychiatric As-
sociation, 1980, p. 67):
a) Intense fear of becoming obese, which does not diminish as weight
loss progresses.
b) Disturbance of body image, e.g., claiming to "feel fat" even when
emaciated.
c) Weight loss of at least 25% of original body weight or, if under 18 years
of age, weight loss from original body weight plus projected weight
expected from growth charts may be combined to make the 25%.
d) Refusal to maintain body weight over minimal normal weight for age
and height.
e) No known physical illness that would account for weight loss.
Although it is probably increasing in prevalence, anorexia nervosa

is still relatively uncommon. There seems little doubt that it is a multi-
determined condition. Hsu (1983) reviewed six different groups of theo-

ries for its etiology. Ploog (1983) has recently suggested a seventh,
which is that anorexia nervosa is an addictive behavior. Ploog compares
anorexia nervosa to obligatory running, which has been shown to in-
duce high cerebrospinal fluid endorphin levels, and suggests that diet-
ing may do the same, producing addictive behavior and dependence.
This theory has not gained wide acceptance and generally does not
appear to fit the clinical picture as described, in that women with
anorexia do not usually report cravings or withdrawal symptoms but
much more frequently a highly controlled constant vigilance around
food and eating.
The main reason for discussing anorexia nervosa only very briefly is
that, as far as strategies for change are concerned, there is little or noth-
ing to report that is new. Most treatment approaches are well described
elsewhere and most have been in clinical use over the past 10 to 15
years. It is perhaps worth pointing out that, despite anorexia nervosa
being a relatively circumscribed syndrome, which is easy to identify,
worthy of treatment, and the main research interest of a number of
professors of psychiatry in the United Kingdom, there have been no
systematic attempts to evaluate treatment approaches. Lucid, up-to-
date, and eclectic reviews are provided in two recent books by Garfinkel
and Garner (1982) and Garner and Garfinkel (1984).
BULIMIA OR BINGE EATING
Bulimia is a relatively recently described syndrome in its circum-

scribed form, though as part of or as a late development of the anorexia
nervosa syndrome, it has been recognised for many years. Intervention
in a therapeutic sense is probably only warranted when the syndrome is
severe. Unlike anorexia nervosa, the initial results of treatment pro-
grams have been promising and systematic attempts have been made to
evaluate different forms of treatment.
The syndrome of bulimia is defined as follows by DSM-II1 (Ameri-
can Psychiatric Association, 1980, p. 69):
a) Recurrent episodes of binge-eating (rapid consumption of a large amount of
food in a discrete period of time, usually less than 2 hours).
b) Awareness that eating pattern is abnormal and fear of not being able to stop
eating voluntarily.
c) Depressed mood and self-depreciating thoughts following eating binges.
d) Bulimic episodes are not due to anorexia or any known physical disorder.
e) At least three of the following:
1. Consumption of high calorie, easily digested food during a binge.
2. Inconspicuous eating during a binge.

3. Termination of such eating episodes by abdominal pain, sleep, social inter-
ruption or self-induced vomiting.
4. Repeated attempts to lose weight by severely restrictive diets, self-induced
vomiting or use of laxatives and/or diuretics.
5. Frequent weight fluctuations greater than 10 lbs due to alternating binges and
fasts.
Bulimia nervosa has been defined by Russell (1979) as follows:

a) The patients suffer from powerful and intractable urges to overeat.
b) They seek to avoid the "fattening" effect of food by inducing vomiting or abus-
ing purgatives or both.
c) They have a morbid fear of becoming fat. (p. 429)
The major difference between the two definitions is that Russell's is

more restrictive by virtue of including the requirement of "a morbid fear
of becoming fat," and that binging alone is not sufficient. The syndrome
has many other names, including dietary chaos syndrome, bulimarexia,
stuffing syndrome, purging/vomiting syndrome, thin-fat people, and
binge eating syndrome.
The development of the syndrome can be divided into three areas:
these include vulnerability, triggering, and maintaining factors. Vul-
nerability factors to bulimia include a biological predisposition to obesity
and probably to depressive illness. There is a family history of obesity
and many sufferers are slightly overweight during adolescence and up
to the onset of the syndrome. There is an excess family history of prima-
ry depressive illness and many individuals have a mixture of symptoms
of bulimia and depression. Perhaps the most important vulnerability
factor is the tremendous social pressure on women to be slim, athletic,
and attractive. In Eastern and African societies where such pressures do
not exist, there are no reported cases of bulimia. When African and
Middle Eastern women move to the United Kingdom or the United
States and become westernized, bulimia and anorexia nervosa do then
occur. As yet, no clear-cut individual or family psychopathology has
been described for the syndrome. In general, sufferers are more extra-
vert, outgoing, and sexually experienced than women who have devel-
oped anorexia nervosa. A final factor that may predispose some women
to bulimia is that they have very marked carbohydrate craving in the
premenstrual period of their cycle. We have treated a number of cases
who only binge and vomit during the week before their menses.
The triggering factors for bulimia are often quite trivial. It nearly
always starts in the context of a period of intense dietary restraint. Most
sufferers have been somewhat overweight and go on a strict diet in this
setting. Carbohydrate craving increases and they break their diet by
binging. Some women use food to cope with feelings of depression and
dysphoria, and comfort eating is very common in adolescent females.

The syndrome may start after a relatively trivial personal remark about
appearance or a minor life event, such as the break-up of a relationship.
Many of the students that we see have put on weight at a time of
stress-for example, while studying before exams-and resolve after
this to diet even more intensively.
The maintaining factors for bulimia are more complex and a number
of feedback circuits seem to operate to continue the behavior. The binges
themselves initially produce relief of dysphoria and are therefore re-
warding. As the binge continues, guilt and shame about the behavior
increase, as does the general level of distress. The starve-binge-starve-
binge cycle appears to become self-perpetuating. This is driven partly by
emotional factors, such as relief of depression and anxiety by bingeing,
which produces further depression and anxiety, leading to increased
carbohydrate restraint with further binges following. It is also driven by
biological factors. Carbohydrate restraint itself, as in severe dieting, pro-
duces marked carbohydrate craving and increases the likelihood of bing-
ing. The discovery of vomiting is initially intensely rewarding. Weight
loss is usually quite dramatic and for the first time a woman may get
down to a weight that is close to the one she desires. Unfortunately,
vomitig allows relaxation of dietary control and it may also encourage
overeating. Many women, once they know that they can vomit at the
end of a binge, will continue to eat vast amounts in an uncontrolled
way. Another maintaining factor may be that certain physical symp-
toms, such as intermittent fluid retention or parotid gland swelling, may
cause panics about weight gain.
Clearly, there are many similarities with the syndrome of anorexia
nervosa but there are also some subtle differences, which I think justify
bulimia being seen as a syndrome in its own right. Less than half of
sufferers have a previous history of anorexia nervosa. The majority start
with bulimia as a new behavior. Perhaps prolonged dietary restraint,
which is required for anorexia nervosa, is so difficult that only a small
proportion of women can achieve and maintain this. About half of wom-
en who begin with restricting anorexia nervosa graduate to bulimia. The
term bulimia nervosa is probably best used for this group of anorexic
graduates. There are clear personality differences as mentioned earlier,
bulimics being more extravert, more socially skilled, and showing less
impulse control. The age of onset would also appear to be different;
bulimia tends to start in late adolescence, or early adulthood, whereas
anorexia starts in early adolescence. Perhaps the two syndromes repre-
sent responses to different maturational tasks occuring at different
stages of adolescence.
In my view, bulimia (whether it be a syndrome in its own right,
associated with obesity, or chronic anorexia nervosa) is the eating disor-

der that fits the addiction/dependence model best. In fact, it is the only
eating disorder that clearly does so.
STAGES OF CHANGE IN BULIMIA
In terms of the stages of change that form the central theme of this
volume, bulimics are quite different to women with anorexia nervosa.
They are acutely aware that they have a problem, and rarely if ever use
denial. They may be secretive in the extreme about their behavior but
this secretiveness is usually deliberate and highly motivated.
Precontemplation
Very few women that we see are in the precontemplation stage.
Very occasionally we have had a women referred, say by her general
practitioner, because of family pressure after having been discovered
vomiting and/or binging; or a woman may be referred by her dentist,
because her tooth enamel is dissolving. Such individuals form a very
small proportion of referrals.
Our view of such individuals has been to accept their state and not
offer or coerce them into treatment. If a woman chooses to maintain her
weight at 15% to 20% below her biologically determined weight, and if
she chooses to do this by constant calorie restriction and uses vomiting
and/or laxative abuse to cope with the binge eating the restriction pre-
cipitates, then that is her choice. Providing she is not distressed by her
behavior and providing the behavior is not extreme, it is relatively
harmless. It is quite an effective way of weight control and probably not
an uncommon practice among late adolescent and young adult females.
Contemplation
Most who seek help are in the contemplation stage and they have
been so for many years. The average length of time from onset of dis-
tressing symptoms to seeking help for bulimics is about 4 years. This
delay in seeking help is usually not because of obsessional ruminations,
or obsessional indecisiveness. It is usually because of guilt about the
behavior and shame that will ensue when the behavior is made public.
The majority of women that we see have told no one about their prob-
lem, or at most only one or two close women friends.
The confessional process in itself seems to be highly therapeutic. In
our research, we have had a number of problems at this stage. Firstly,
having confessed, subjects are highly eager to talk and pour out all their
distress, problems, and abnormal behavior. They want feedback on how
they compare with other subjects: Are there other people as distressed
or as bad as they, is there any hope for a cure? They have often read
much about the syndrome in lay articles and in women's magazines. In
our clinic, this catharsis usually occurs with the interviewer/rater who is
doing the initial assessment interview and who is not going to be the
subject's continuing therapist. This obviously creates problems.
Secondly, the confession is often so therapeutic in itself that the
behavior stops for a few days, or even a few weeks, and occasionally
stops completely, so it is not possible to get reasonable pretreatment
baseline measures without waiting for the behavior to return. Distress at
this stage is often marked and suicidal ideation and attempts are
common.
Action
At the point, about half of our subjects appear ready to move on to
the stage of action. They find the other assessment procedures irritating
and slow; they are reluctant to take time over making careful baseline
measures before treatment starts; and they want rapid if not instant
action.
The other 50% remain stuck in the stage of contemplation, con-
cerned with what they will have to give up if their behavior is to change.
They are terrified that they will put on weight, concerned at how they
will cope with their dysthmic feelings without the use of food, and about
how they will cope with their carbohydrate craving. However, apart
from group treatments, the dropout rate in our treatment has not been
high.
Maintenance
As far as the maintenance stage is concerned, there is really not
enough evidence of how individuals cope with this, nor do we have
enough information of self-change or even whether this occurs in
bulimia.
MANAGEMENT OF BULIMIA
Despite the fact that the syndrome has only relatively recently been
described, there have been a large number of suggested regimes pub-
lished. There are also several controlled trials of treatment recently pub-
lished or in progress. It would appear then that the management of
bulimia is being much more systematically evaluated than the manage-
ment of anorexia nervosa.
TABLE 4.
Management of Bulimia
Treatment Authors Outcome
Drugs
Phenytoin Wermuth, Davis, Hollister, & negative study
Stunkard (1977)
Greenway, Dahms, & Bray positive study
(1977)
Imipramine Pope, Hudson, Jonas, & positive study
Yurgelan-Todd (1983)
Johnson & Larsen (1982) positive study
Phenelzine Walsh et al. (1982) positive study
Mianserin Sabine, Yonacre, Farrington, negative study
Barratt, & Wakeling (1983)
Cabamazine Kaplan, Garfinkel, Darby, & negative study
Garner (1983)
Group psychotherapy
Once weekly eclectic groups Lacey (1983) positive study
with some individual counsel-
ing
Short-term group, 9 weeks, 12 Johnson, Connors, & Stuckey positive study

sessions self-monitoring, di- (1983)
dactic information, alternative
coping strategies
Eclectic group combining psy- Roy-Byrne, Lee-Benner, & Yager positive study
chodynamic interpretations (1983)
Once weekly 90 mins 15 ses- Freeman, Sinclair, Annandale, positive, but

sions eclectic group & Turnbull (1985) not as good
as individual
Individual psychotherapy
Cognitive behavioral Fairburn (1981) uncontrolled
Cognitive vs. behavioral Freeman, Sinclair, Annandale, both effective

& Turnbull (1985)
Experimental/behavioral Boskind-Lodahl & White (1978) uncontrolled
Exposure and response pre- Rosen & Leitenberg (1982) uncontrolled

vention
There are a number of important issues concerning treatment and

there is no general agreement about any of them. Firstly, when to treat?
The boundaries of the syndrome are so blurred that it is very difficult to
set any definite cutoff point, as far as severity is concerned, beyond
which treatment intervention is justifiable. For example, should people
who only binge and vomit or purge very occasionally be offered treat-
ment? Their behavior may cause them great distress, but happen only a
few times a year. Should individuals who binge eat at times of stress,
but who do not have an accompanying fear of fatness, be offered the
same kind of treatment, or should they primarily have treatment for
their anxiety? Is bulimia best treated on an outpatient or an inpatient
basis? Some suggested regimes have closely modeled themselves on
treatment for anorexia nervosa, which usually involves prolonged, in-
tensive inpatient care. Another issue that has been relatively little dis-
cussed is the sex of the therapist undertaking the treatment (Van-
dereycken & Meermann, 1984). There is also the issue of how much the
management of this syndrome should be medicalized. Over recent years
self-help groups have sprung up and it may be that for many individuals
professional help is unnecessary. Finally, there is the debate about the
use of drugs, particularly anticonvulsant and antidepressant drugs, in
the management of the syndrome.
Some of the suggested regimes of treatment are listed in Table 4. In
this chapter I will highlight only a few of the treatments listed there but
references are given to all the others.
The Use of Drugs

There have been two main groups of drugs suggested, the anticon-
vulsants and the antidepressants. Rau, Struve, and Breen (1979) have
used diphenylhydantoin; Wermuth, Davis, Hollister, and Stunkard
(1977) phenytoin; and Kaplan, Garfinkel, Darby, and Garner (1983) car-
bamazepine. The justification for anticonvulsants is a rather weak one,
but the hypothesis is that, in those who show episodic abnormal eating
and display an abnormal EEG, bulimia may represent a form of epilep-
toid behavior. The results of studies on anticonvulsants have been dis-
appointing and, at present, I think that there is no justification for their
use in the treatment of bulimia.
Three main groups of antidepressants have been used. Sabine,
Yonace, Farrington, Barratt, and Wakeling (1983) used mianserin and
found no difference between mianserin and a placebo in bulimic pa-
tients. Hudson, Pope, and Jonas (1982) and Johnson and Larson (1982)
used imipramine. The most recent placebo controlled study (Pope, Hud-
son, Jonas, & Yurgelun-Todd, 1983) found that imipramine was associ-
ated with a significantly reduced frequency of binge eating. Walsh et al.

(1982) have claimed efficacy for monoamine oxidase inhibitors and have
produced some very promising results using phenelzine. At this stage it
is not clear whether antidepressant drugs are acting as anxiolytics, anti-
depressants, or specific antibulimic drugs. Nor is it clear whether drugs
successfully treat only those bulimics who have depressive symptoms.
Although these drug studies report significant levels of improvement,
careful inspection shows that for most subjects binging does not stop
entirely, and that patients are left with significant residual symp-
tomatology. In our experience, there is a small group of perhaps 10% to
15% of the total number of cases referred who have clear symptoms of
biological depression, such as early morning wakening, retardation,
poor concentration, etc. They respond dramatically to antidepressant
medication and bulimic symptoms stop when the depression is relieved.
In the much larger group of more typical bulimics, with some dysphoria
but not a true biological depression, antidepressants have a very limited
effect and probably work by reducing anxiety and stress.
Psychotherapeutic Treatments
Psychotherapeutic treatments are summarised in Table 4. Many dif-
ferent psychotherapeutic approaches have been suggested. Many are a
combination of straightforward behavioral techniques and other types of
psychotherapy. The most promising results so far have been group
treatments (Lacey, 1983) and a type of cognitive therapy (Fairburn,
1982). The preliminary results of our own study (Freeman, Sinclair,
Annandale, & Turnbull, 1985) comparing three different types of psy-
chotherapy-namely, cognitive, behavioral, and group-indicate that
all three types are successful in reducing the level of symptomatology,
but that cognitive therapy has a greater effect on depressive symptoms
and may have a more powerful effect on prevention of relapse, although
relapse rates are high. So far there has been relatively little published on
either family-based or psychodynamic approaches. Schwartz (1982) de-
scribes a single case of a 17-year-old girl treated with a family therapy
approach and Linden (1980) describes the psychodynamic treatment of a
patient with a ravenous appetite.
The next 18 months should see the publication of several large
controlled trials of psychotherapy taking place in different parts of the
world. This will give clinicians a much clearer idea of which treatments
are effective and, importantly, which are cost-effective.
REFERENCES
American Psychiatric Association. (1980). Diagnostic and statistical manual of mental disorders
(3rd ed.). Washington, DC: Author.
Boskind-LodahI, M., & White, W. (1978). The definition and treatment of bulimarexia in
college women: A pilot study. Journal of the American Health College Association, 27, 84-
87.
Craigshead, L. W., Stunkard, A. J., & O'Brien, R. (1981). Behaviour therapy and phar-
macotherapy of obesity. Archives of General Psyclliatry, 38, 763-768.
Durnin, J. V. G. A., Lonergan, M. E., Good, J., & Ewan, A. (1974). A cross sectional
nutritional and anthropometric study, with an interval of 7 years, on 611 young
adolescent school children. British Journal of Nutrition, 32, 169-179.
Fairburn, C. (1981). A cognitive behavioural approach to the treatment of bulimia. Psycho-
logical Medicine, 11, 707-711.
Freeman, C. P. L., Sinclair, F., Annandale, A., & Turnbull, J. (in press). A controlled trial
of three different treatments for bulimia. Journal of Psychiatric Research, 19, 473-478.
Friend, B. (1974, May). Changes in nutrients in the US diet caused by alterations in food patterns.
Paper given at Conference on Changing Food Supply in America, Washington, DC.
Garfinkel, P. E., & Garner, D. M. (1982). Anorexia nervosa: A multidimensional perspective.
New York: Brunner/Maze!.
Garner, D. M., & Garfinkel, P. E. (1984). Handbook of psychotherapy for anorexia nervosa and
bulimia. New York: Guildford Press.
Greenway, F. L., Dahms, W. T., & Bray, G. A. (1977). Phenytoin as a treatment of obf'sity
associated with compulsive eating. Current Therapeutic Research, 21, 338-342.
Halmi, K. A., Stunkard, A. J., & Mason, E. E. (1980). Emotional responses to weight
reduction by three methods: Diet, jejunoileal bypass and gastric binges. American
Journal of Clinical Nutrition, 33, 446-451.
Hsu, L. K. E. (1983). The aetiology of anorexia nervosa. Psychological Medicine, 13, 231-238.
Hudson, J. I., Pope, H. G., & Jonas, J. M. (1984). Treatment of bulimia with antidepres-
sants: Theoretical considerations and clinical findings. In A. J. Stunkard & E. Stellar
(Eds.), Eating and its disorders (pp. 259-273). New York: Raven Press.
Johnson, c., Connors, M., & Stuckey, M. (1983). Short term group treatment of bulimia.
International Journal of Eating Disorders, 2, 199-208.
Johnson, c., & Larsen, R. (1982). Bulimia: An analysis of moods and behavior. Psychoso-
matic Medicine, 44, 341-351.
Kaplan, A. S., Garfinkel, P. E., Darby, P. L., & Garner, D. M. (1983). Carbamazepine in
the treatment of bulimia. American Journal of Psychiatry, 140, 1225-1226.
Keesey, R. F., & Corbut, S. W. (1984). Metabolic defense of the body weight set-point. In
A. S. Stunkard & E. Stellar (Eds.), Eating and its disorders (pp. 87-96). New York:
Raven Press.
Keys, A. (1980). Overweight, obesity, coronary heart disease and mortality. Nutritional
Review, 38, 297-307.
Kromhout, J. (1983). Changes in energy and macronutrients in 871 middle aged men
during 10 years of follow-up (the Zutphen Study). American Journal of Clinical Nutri-
tion, 37, 287-294.
Lacey, J. H. (1983). Bulimia nervosa, binge eating and pychogenic vomiting: a controlled
treatment study and long-term outcome. British Medical Journal, 286, 1609-1613.
Linden, W. (1980). Multicomponent behavior therapy in a case of compulsive binge-eating
followed by vomiting. Journal of Behavior Therapy and Experimental Psychiatry, 11, 297-
300.
Mason, E. E. (1981). Surgical treatment of obesity. Philadelphia, PA: Saunders.
McReynolds, W. T. (1982). Toward a psychology of obesity: Review of research on the role
of personality and level of adjustment. International Journal of Eating Disorders, 2, 37-
57.
Noppa, H., Bengtsson, c., Wedel, H., & Wilhelmsen, L. (1980). Obesity in relation to
morbidity and mortality from cardiovascular disease. American Journal of Epidemiology,
111, 682-692.
Pettitt, D. L Lisse, J. R, Knowles, W. c., & Bennett, P. M. (1982). Mortality as a function

of obesity and diabetes mellitus. American Journal of Epidemiology, 115, 359-366.
Ploog, D. W. (1983, July). Psychobiology of anorexia nervosa. Paper given at VII World Con-
ference of Psychiatry, Vienna.
Pope, H. G., Hudson, J. I., Jonas, J. M., & Yurgelun-Todd, D. (1983). Bulimia treated with
imipramine: A placebo-controlled double-blind study. American Journal of Psychiatry,
140, 554-558.
Rau, J. H., Struve, F. A., & Green, R J. (1979). Electroencephalographic correlates of
compulsive eating. Clinical Electroencophalography, 10, 180-189.
Rosen, J. c., & Leitenberg, H. (1982). Bulimia Nervosa: Treatment with exposure and
response prevention. Behavior Therapy, 13, 117-124.
Roy-Byrne, P., Lee-Benner, K., & Yager, J. (1984). Group therapy for bulimia. International
Journal of Eating Disorders, 33, 97-116.
Russell, G. (1979). Bulimia nervosa: An ominous variant of anorexia nervosa. Psychological
Medicine, 9, 429-448.
Sabine, E. J., Yonace, A., Farrington, A. J., Barratt, K. H., & Wakeling, A. (1983). Bulimia
nervosa: A placebo-controlled double-blind therapeutic trial of mianserin. British Jour-
nal of Pharmacology,S (Suppl. 2), 1958-2025.
Schwartz, R. C. (1982). Bulimia and family therapy: A case study. International Journal of
Eating Disorders, 2, 75-82.
Sorlic, P., Gordon, P., & Kennel, W. B. (1980). Body build and mortality. Journal of the
American Medical Association, 243, 1828-1831.
Stern, J. (1984). Is obesity a disease of inactivity? In A. J. Stunkard & E. Stellar (Eds.),
Eating and its disorders (pp. 131-140). New York: Raven Press.
Stunkard, A. J. (1984a). New developments in the drug treatment of obesity. Clinical
Neuropharmacology, 7 (Suppl. No.1), 704-707.
Stunkard, A. J. (1984b). The current status of treatment for obesity in adults. In A. J.
Stunkard & E. Stellar (Eds.), Eating and its disorders. New York: Raven Press.
Vandereycken, W., & Meermann, R (1984). Anorexia nervosa: A clinician's guide to treatment.
New York: Walter de Gruyter.
Walsh, B. T., Stewart, J. W., Wright, L., Harrison, W., Roose, S. P., & Glassman, A. H.
(1982). Treatment of bulimia with monoamine oxydase inhibitors. American Journal of
Psychiatry, 139, 1629-1630.
Wermuth, B. M., Davis, K. L., Hollister, L. E., & Stunkard, A. J. (1977). Phenytoin
treatment of the binge-eating syndrome. American Journal of Psychiatry, 140, 1225-
1226.
Woolley, S. c., & Woolley, O. W. (1984). Should obesity be treated at all? In A. J. Stunkard
& E. Stellar (Eds.), Eating and its disorders (pp. 185-192). New York: Raven Press.
11
Early Indications of Treatment
Outcome in Multiple Drug Users
D. ADRIAN WILKINSON AND SIMONNE LEBRETON
INTRODUCTION
The past two decades have witnessed a dramatic increase in the use of a
variety of psychoactive substances in the developed countries, and a
predictable increase in the need for treatment services directed towards
this problem. When presenting to treatment, most drug users report
recent use of a wide variety of psychoactive substances (Farley, Santo, &
Speck, 1979; Sadava, 1984; Wilkinson & Martin, 1983). Nonetheless,
there is scant literature on the effectiveness of treatments for multiple
substance abuse, also referred to as polydrug abuse (Sobell, Sobell,
Ersner-Hirshfield, & Nirenberg, 1982).
In the field of substance-abuse treatment generally, there has re-
cently been a trend towards the use of brief interventions. It can be
argued that several factors have accounted for the trend. One factor is a
body of evidence that brief treatments can be as effective as much more
intensive and costly interventions. This finding seems to hold true in
particular for clients whose substance-abuse problem is at a relatively
early stage of development (Miller, Taylor, & West, 1980; Sanchez-
Craig, Wilkinson, & Walker, 1986). Brief treatments have the significant
D. ADRIAN WILKINSON AND SIMONNE LEBRETON· Addiction Research Founda-

tion, Toronto, Ontario, Canada, M5S 2S1. The views expressed in this document are those
of the authors and do not necessarily reflect those of the Addiction Research Foundation.
239
240 D. ADRIAN WILKINSON AND SIMONNE LEBRETON
advantage that they are minimally disruptive of other activities in which

clients need to participate (e.g., professional and family respon-
sibilities), and hence they are able to attract to treatment clients who
would otherwise demur. Again this appears to be particularly true of
clients whose substance-abuse problem is at an early stage of develop-
ment, and who have not yet experienced the kinds of physical and social
deterioration that are common in more chronic cases (Sanchez-Craig et
al., 1986).
Despite the finding that brief interventions can be effective for
many persons with substance-abuse problems, it is plausible to believe
that more intensive treatments would be more suitable for some clients
(Orford, Oppenheimer, & Edwards, 1976). Hence, it behooves those
providing brief treatments to attempt to identify the characteristics of
clients who succeed in such interventions. The present chapter de-
scribes an initial evaluation of factors associated with treatment outcome
in a group of multiple drug users who received a brief outpatient pro-
gram of self-control training, and were followed-up one year later.
Treatment outcome was defined exclusively on the basis of change in
the frequency and nature of drug use from assessment to one-year fol-
low-up (Wilkinson & Martin, 1983).
In planning this evaluation, we identified five clusters of variables
that appeared to be predictors of treatment outcome. These were (a) the
level of drug use in the year preceding treatment; (b) the multiplicity of
drug problems and other life problems reported by the client; (c) clients'
perceptions of need for professional assistance in resolving these prob-
lems; (d) measures of client motivation; and (e) the initial progress to-
wards treatment object:ves during treatment.
PRETREATMENT DRUG USE
There is abundant evidence that the intensity and chronicity of

substance-abuse problems, before treatment begins, are significant pre-
dictors of treatment outcome (Moos & Finney, 1983; Ogborne, 1978;
Polich, Armor, & Braiker, 1981). In the present study this dimension
was assessed by measuring the self-reported levels of drug consumption
during the intial stages of treatment, and retrospectively for the year
preceding treatment.
PROBLEM MULTIPLICITY
Perceived multiplicity of problems was indexed by the number of

goal areas clients indicated on a self-completed questionnaire during
assessment. Persons with substance-abuse problems frequently have
TREATMENT OUTCOME IN MULTIPLE DRUG USERS 241
problems in other important life areas, and it is accepted that evaluation

of such problems and their resolution is an important aspect of sub-
stance-abuse treatments (Lorei, 1982, Pattison, 1978; Sobell & Sobell,
1980; Wilkinson & Martin, 1984).
PERCEIVED NEED FOR PROFESSIONAL HELP (SELF-RELIANCE)
We assessed the clients' perceived need for professional assistance

(self-reliance) by asking them to indicate those goals for which such help
was required. The larger the number of goals a client selected for which
professional assistance was needed, the lower that person scored on the
measure of self-reliance. This measure may reflect the clients' self-
efficacy in regard to those goals. However, because the procedure for
assessing this dimension does not conform to Bandura's (1977, 1982)
methods of assessing self-efficacy (in lacking behavioral specificity), we
have restricted ourselves to the use of the term self-reliance.
CLIENT MonvA nON
A frequent, and understandable, attribution that is made for treat-

ment failure is low client motivation (Miller, 1985). This is particularly
true in the field of substance-abuse treatment, where the concept of
denial (a hypothetical motivational variable) holds considerable sway.
Miller (1985) has comprehensively reviewed the issues of client moti-
vation in alcoholism treatment and, largely on the basis of his work, a
set of four variables were selected to reflect initial client motivation for
treatment. It seemed particularly important to assess motivation because
the treatment itself (Wilkinson & Martin, 1983) essentially comprised
what Miller has termed a motivational intervention.
The first measures of motivation were the frequency of reported
cravings for, and refusals of, drugs during the week preceding and the
week following the start of treatment. Cravings were defined as the
experience of a strong urge to use drugs, to which the person did not
succumb. We reasoned that cravings thus defined would reflect restraint
of drug use, and that such restraint indicates motivation. Similarly, the
refusal of offers of drugs would constitute an indication of motivation to
refrain from drug use. In addition, motivation was assessed on the basis
of the degree of compliance to the therapist's request that drug use,
cravings, and refusals be self-monitored during the first week of treat-
ment. Finally, the goals that clients set for drug use during the first week
of treatment were taken as a measure of client motivation. The more
conservative the level of anticipated drug use set by the client during the
first treatment session, the higher the level of motivation was judged to
be. We recognize the inherent circularity of these operational definitions

of motivation. Nonetheless they represent the behavioral signs on
which attributions of client motivation are frequently based (Miller,
1985).
PROGRESS IN TREATMENT
The last cluster of predictor variables involved the clients' initial

progress in treatment. Specifically, relaxation of drug use goals during
treatment and attrition from treatment would seem to indicate change in
client motivation during treatment. These measures may reflect change
in client motivation during treatment, or the failure of the motivational
treatment.
METHOD
SUBJECTS
The subjects of the study were 49 (40 males and 9 females) multiple
drug users who were randomly assigned to the treatment, as part of a
larger study, and successfully followed-up one year later. Criteria for
admission into the study included a client age from 16 to 30 years;
clients' presenting for treatment of a substance abuse problem; not being
psychotic or requiring hospitalization or psychotropic medication; will-
ingness to accept either outpatient or residential counseling; normal
cognitive ability; and clients' consenting to participate in the research
project. Characteristics of the group are presented in Table 1.
INSTRUMENTS
A variety of information was collected from clients throughout the

project by means of self-administered forms and standardized inter-
views. In the present chapter we describe only those procedures evalu-
ated in this part of the study.
The Psychoactive Drug Use History form is a standardized question-
naire on drug use for the previous year (Wilkinson & Martin, 1983).
Information is collected on use of 10 classes of drugs: alcohol; cannabis;
hallucinogens; narcotic analgesics; sedative hypnotics; solvents and
aerosols; stimulants; tranquillizers; volatile nitrites; and miscellaneous
others. For all drug classes involving some reported use in the past year
information is collected on recency; months with any use; days of use in
months of use; frequency of use in days of use; typical dosage; mode of
TABLE 1.
Characteristics of 49 Clients Who Started Treatment and were Followed-Up
One Year Later
Age (Mdn, Range) 23 16-30

Sex M 81.6% F 18.4%
Marital status %
Single 79.6
Married/common law 10.2
Separated/divorced 10.2
Employment status Unemployed 63.3
Employed 28.6
Student 6.1
Other 2.0
Education (Mdn highest grade) 10
Social stability' Low (0-4) 18.4
Medium (5-10) 53.1
High (11-14) 28.6
Legal status On probation, parole, 45.8
or awaiting trial
No legal problems 54.2
'Skinner (1979).
administration; and the client's view of whether use of the drug was
ever a problem, and if so whether it remains a problem. This history
takes 30 to 90 minutes to administer.
The Treatment Goals form is a self-administered questionnaire in
which the client indicates current goal areas, and, for each indicated
goal, whether professional assistance is required. Listed goal areas in-
clude alcohol use; use of other drugs; anxiety; assertiveness; family
problems; social skills; employment; leisure; accommodation; legal prob-
lems; sex education; and a write-in option. This form takes about 5
minutes to complete.
Self-Monitoring Cards were adapted from those published by Miller
and Munoz (1976). For purposes of the present study the format of the
cards was retained, but the form was adapted to permit recording on
several drug classes, for the identification of cravings and refusals of
drugs, and for consumption. As with the Miller and Munoz cards, cli-
ents are asked to record the time of events, dose, and context.
Information from self-monitoring cards was transcribed on to the
Drug Avoidance Inventory during treatment sessions. This form is orga-
nized to permit day-by-day-recording of uses, cravings, and refusals of
drugs in the 10 drug classes, for the week preceding completion of the
form. The client's goals with respect to drug use in the interval between
appointments was recorded on the Drug Use Goals form. If use of any
drug class was anticipated, the maximum frequency, quantity, and dose
were indicated, as well as identification of situations in which use would
be consistent with the goal, and situations in which drug use would be
inappropriate.
INITIAL ASSESSMENT
After intake and orientation to procedures of the treatment center,

clients were given an appointment for one day of assessment. Assess-
ments were conducted by persons independent of the study.
Cognitive abilities were assessed by means of the Wide Range
Achievement Test (WRAT, reading; Jastak & Jastak, 1978), the Clarke-
WAIS Vocabulary Test (an earlier version of the Clarke Vocabulary
Scale; Paitich, 1979), the Digit Symbol Substitution test of the Wechsler
Adult Intelligence Scale (WAIS) (Matarazzo, 1972) and the Benton Visual
Retention Test (Benton, 1963). Potential clients were included in the
study if the following criteria were met: either WRAT scaled scored >45
or Clarke-WAIS scaled score >6; and either Digit Symbol scaled score
>6 or Benton scaled score >-2. All clients received a medical examina-
tion, during which they were screened for medical or psychiatric prob-
lems (e.g., psychosis) that would exclude them from the study. A urine
sample for drug screening was collected at that time. Clients meeting all
of the admission criteria (see Subjects) were randomly assigned to one of
the conditions of the larger study, and presented with a standardized
description of the treatment condition to which they had been assigned,
and details of the study requirements. Those consenting to participate
then completed various questionnaires specific to the study, including
the Psychoactive Drug Use History and the Treatment Goals form, and
the first treatment appointment was scheduled for approximately one
week later.
TREATMENT PROCEDURES
The treatment consisted of three training sessions and six follow-up

sessions, spread over 70 weeks. Clients could elect to contact the thera-
pist for additional sessions. The central features of the treatment pro-
cedures were cognitive and behavioral self-control measures aimed at
reducing drug and alcohol consumption to levels the client considered
appropriate. Essential components of this process were self-monitoring
of drug use, setting specific goals for reduced use, and identifying cog-
nitive and behavioral strategies for avoiding drug use in situations of
high risk. In addition, problems in other life areas were identified. Per-
sonal or community resources were selected to help in resolving these

problems. At the conclusion of each session the client received a copy of
the form indicating the current goals with respect to drug use. From the
second session on, he or she would also receive copies of forms com-
pleted in the session, indicating strategies selected for the avoidance of
drug use, and plans to resolve other life problems. On the average,
sessions lasted between 60 and 90 minutes.
All clients in the study received treatment from one of the authors
(S. LeB.), a registered nurse, who received additional specific training
for the study and had had 8 years of experience working with substance
abusers. Training consisted of planning sessions about the treatment
procedures to be used in the study, and participation in the develop-
ment of a self-help manual. Literature on self-help manuals was re-
viewed, and a large selection of such manuals was purchased for subse-
quent use by clients. Potentially useful community resources were
identified, and a filing system describing their characteristics was estab-
lished. As part of the pilot study, the first author (D. Adrian Wilkinson)
modeled the procedures, and then observed the therapist (Simonne
LeBreton) in sessions with a total of 20 clients.
INDEPENDENT FOLLOW-UP
An extensive follow-up interview and assessment was conducted,

independently of project staff, at one and two years after the initial
assessment. As part of this procedure the Psychoactive Drug Use Histo-
ry was readministered, and urine drug screen was repeated. A payment
of $25.00 was provided to the client for attending the follow-up inter-
view, which lasted about 2Y2 hours.
RESULTS
CATEGORIZATION OF TREATMENT OUTCOME
The first stage of the analysis was to assign the subjects to one of
three outcome categories on the basis of information collected on the
Psychoactive Drug Use History. This information was used to categorize
the outcomes of clients as Successful (S), Significantly Improved (I), and
Unimproved (U), Four raters (blind to client identity) independently
compared each subject'S data at pretreatment and one year and rated
each subject (Wilkinson & Martin, 1983). All four raters had several
years experience in research on substance abuse, and three had exten-
sive clinical experience with multiple substance users. Subjects rated
TABLE 2_
Psychoactive Drug Use History
Typical
Months Typical # Years
Number of drug of use in frequency times since Example of
types used in Time since last past in months used Typical dosage Usual mode of Typical Use problem Use problem frrst rating of
Drug class drug class use year of use per day &. comments administration source ever? still? problem consumption
Circle types 1 Past 24 hrs Range Range Enter the units 1 Oral 1 Retail Has your use Ooes your Range
used 224-48 hrs 1-12 1-30 in which the use 2 Nasal (snort- 2 Prescription of . .. ever current use 0-4
3 48 hrs-7 days of particular ing) 31llegaJ caused of. .. still
47 days-1 mo drugs is 3 Inhalation 4 Gift problems cause
51-3 months specified 4 Injection IV 5 Self-pro- with your problems for
6 More than 3 Also note any 5 Injection- duced work, family, you in any of
months complex other friends, these areas?
patterns of use 6 Other health or the 1 No
law? 2 Yes
1 No
2 Yes
1. Alcohol Beer; wine; 12 20 6 Usually drinks 1 2
fortified wine; an average of 6
liquor; other pints of beer
2. Cannabis Marijuana; 4 12 25 8 Smokes 10-12 3 4 2 2 3 4
hashish; hash joints of mainly
oil; other hash every 2-3
hours
throughout day
3. Hallucinogens LSD; MDA; 4 12 8 Uses mainly 3 3
Mescaline; LSD
phencyclidine; 4 hits a day
psilocybin; All at one time
other
4. Narcotics Codeine; 5 4 3 3 Takes 5-6 4 2
heroin; percodan a day
hydrocodone; 2 at a time
hydromorphone
meperidene;
methadone;
oxycodone;
pentazocine;
propoxyphene;
other
5. Sedative Hyp- Barbiturates; 5 4 3 3 5-6 4 2
notics chloral Quaaludes/day
hydrates; 2 at a time
diphenhydr-
amine;
ethchlorvynol;
flurazepam;
glotethimide;
methaqualone;
other
6. Solvents &< Aerosols; 0
Aerosols cleaning
solvents;
gasoline; glue;
other
7. Stimulants Amphetamines; 6 4 30 S Taking 8 Ritalin 2
chlorphenter- for hyperactivity
mine; cocaine; Siday Bennies-
diethyl propion; only a few times
methamphet-
amine; methyl-
phenidate;
phenmetrazine;
phentermine;
other
8. Tranquilizers Chlordiazep- 4 14 Taken while in 4
oxide; jail for 2
diazepam; weeks-2
meprobamate; valium to go to
other minor sleep
tranquilizers;
major
tranqUilizers
9. Volatile Amyl nitrate; OveraU low
Nitrates isobutyl nitrate; incidence
other
10. Miscellaneous Anticholin- Gasses not
ergics; rated
antiemetics;
antihistamines;
other
unimproved by any rater were termed "unimproved" (U), subjects rated

as treatment successes by at least 3 raters were termed "treatment suc-
cesses" (S) and subjects falling between those ratings were termed "sig-
nificantly improved" (I). Thus, three groups of subjects (11 S, 20 I, and
18 U) were formed. The categorization of outcome was based exclusively
on change in overall reported frequency and quantity of drug use for 10
drug classes. Self-reported drug use was validated by means of drug
screening of urine samples (Martin, Wilkinson, & Kapur, 1984), and the
employment of drug use as the sole dependent measure was con-
vergently validated by examining concurrent changes in social stability,
criminality, employment, incarceration, and social relationships (Wil-
kinson, & Martin, 1984).
DRUG USE AT ASSESSMENT AND ONE YEAR
Categorization of treatment outcome was based on a comparison of

pairs of Drug Use History forms for all subjects; that is, it was a measure
of change in drug use. It was clearly also desirable to assess the levels of
drug use reported by subjects at both assessment and follow-up, for
each of the three outcome groups. To permit this comparison an ordinal
scale of frequency of drug use was constructed for each of eight of the
drug classes: Alcohol, cannabis, hallucinogens, narcotics, sedative hyp-
notics, solvents, stimulants, and tranquillizers. Volatile nitrites and mis-
cellaneous others were omitted as irrelevant after review of the assess-
ment data.
Drug consumption in the previous year was measured in frequency
for all drug classes except alcohol, for which annual consumption was
estimated in standard drinks. In addition, pertinent information on typ-
ical dose was recorded (see Table 2). This dose information requires
coding before full-scale analysis of the drug consumption data can be
conducted. Hence, in the initial analysis, consumption for the 8 classes
was rated on a five point scale by 3 of the 4 raters described above.
Typical data and the ratings assigned are presented in Table 2. The data
in Table 2 were collected from an actual subject at assessment. Guide-
lines for the ratings were issued, but could be overruled by the raters on
the basis of additional information on the form. The guidelines are pre-
sented in Table 3. Satisfactory interrater reliability was obtained by this
procedure (for all drug classes the range of correlation coefficients
was .83 to .99).
For each client an index of drug use severity was constructed as
follows: The three ratings for each of eight classes were added. Because
the ratings were on a 5-point scale (0-4) this yielded a possible range of
summed scores of 96 (3 raters x 8 drug classes x score of 4). The
summed score was then divided by (3 x 8) to convert it to a mean rating
TABLE 3.
Guidelines for Ratings of Drug Use in Past Year
Other drug
Alcohol Cannabis classes
o= Abstinent No use of drugs from this class during past year.

1 = Low :520 drinks/week :54 joints/week :51 use/month
and problem and :52 joints/day of
still? = No use and problem
still? = No
2 = Intermediate :510 drinks/day <10 joints/week > 1 use/month
and :542 but <1 use/week
drinks/week
3 = High > 10 drinks/day ~1O joints/week ~1 use/week
or >42 drinks/week
4 = "Outrageous" Very high level of consumption, even within this sample. At
rater's discretion.
per drug class for each subject, on a 5-point scale similar to that presented
in Table 3. The averages of these mean ratings for the three outcome
groups are presented in Figure 1. This shows the mean of mean ratings
at both assessment and one year. Such means represent complex pat-
terns of drug use. The two clients with the most representative reported
s---
I ...-----...
U ..............
U) 2.0
10
"0
Cl
2 1.5
"0
*!.,.~
..............................................•
............
~ ............
Cl
c:
~
1.0 ......
.------~
c:
m
:::?i 0.5
Assessment 1·Year
FIGURE 1. Mean ratings per class for clients in the three outcome categories at assessment
and I-year follow-up. The S group reported significantly less drug use in the year preced-
ing treatment than did the other two groups. To clarify the meaning of these values the
data from the two most representative clients from each outcome category are presented in
Table 4.
TABLE 4.
Drug Use by Two Subjects in Each Outcome Group Whose Data Were Closest
to Points on Graph
Assessment I-Year follow-up
Successful 51 Reported consumption of an Reported 3 months abstinence

average of 12 "pints" of beer from alcohol, then 9 months
and one bottle of liquor al- of 3-4 beers/day 3 times per
most every day. He bought week. He had continued to
"lots" of valium or Iibrium on use prescribed analgesics oc-
the street in units of 50-100 casionally and reported no
tablets, of which he would use of benzodiazepines.
use about 25 tablets a day till
finished. This would occupy
about 16 days in the year.
Once he received a 14-day
prescription for a narcotic an-
algesic which he took as pre-
scribed.
Successful 52 Reported using 3 joints of can- There were 3 uses of cannabis

nabis 6 days/week through in past year, all of V4-Y2 joint,
the past year. For seven 10 days use of narcotic anal-
months he had used about Y2 gesic as prescribed, and con-
gm cocaine/day 2-3 sumption of 10 drinks/day on
days/week, sometimes com- 3 occasions per month.
bined with 3-4 diet pills.
About three times in the year
he had used a headache rem-
edy containing codeine. Alco-
hol consumption was 3
beers/day on weekends.
Improved 11 Reported smoking 2-4 joints of Reported consuming one joint

cannabis daily throughout the cannabis and 6 beers/day on
past year. LSD in doses of 2 weekends. On fewer than 10
"hits" /day was used on occasions one "bennie" had
weekends. About three been taken, and 10 mg valium
days/week for eight months on three occasions.
2-4 assorted "uppers" would
be used once per day in a
single dose. About four
days/week 8 drinks of liquor
would be consumed. One
three day episode of use of
narcotic analgesic, illicitly ob-
tained, was reported.
(continued)
TABLE 4. (Continued)
Assessment I-Year follow-up
Improved 12 Reported smoking an average of Reported use of 1 joint of can-

7-8 joints cannabis/day daily nabis almost every other day
throughout the year. About 6 for 10 months. There had
days/week he would drink 8 been 4 uses of hallucinogens,
beers. Every other day he use of one tablet of barbitu-
would use 6 "yellow jacket" rate, and no narcotic anal-
diet pills in two doses. He gesics. There were 4 uses of
had used narcotic analgesics amyl nitrite, and valium (5
about 8 times 4 capsules/time mg) once. During one month
"for a different high." Seda- there had been 14 days of use
tive hypnotics were taken or- of 1 "yellow jacket" three
ally daily 3 times/day times/day. For 6 months he
(methaqualone, 2 tabs.) for 3 used 2 beers/day 5
months. About 30 mg valium times/month.
had been used on three occa-
sions and amyl nitrite once.
Unimproved U1 Reported smoking an average of Cannabis use was 2 joints/day

3 joints of cannabis/day 20 every other day. For 5 months
days per month. For 10 6 beers/day 20 days/month
months 6 pints of beer/day had been used, but client had
had been used 20 days per been abstinent from all drugs
month. Psilocybin had been save cannabis for more than 3
used 6 times in the past year. months. For 6 months 2 tab-
One "bennie" or cocaine had lets of percodan had been
been used about 5 used 3x /day, 10 days/month.
times/month through the There were 4 episodes with
year. Occasionally 15 mg di- psilocybin and 2 with barbitu-
azepam was used for sleep. rates. For 6 months cocaine
was used 15 days/month, 3
times/day, and valium (5 mg)
once per day 10 days/month.
Unimproved U2 Reported smoking 10-12 joints Reported 1h gm speed/week

cannabis daily throughout the (IV), 3 times/day, 15
year. On 12 days/month 15 days/month. Cannabis daily 7
beers would be drunk. Use of joints, and 12 beer/day, 20
"speed" (IV) or "bennies" 3 days/month. Eight months of
times/day, 20 days/month. use of PCP 3 times/day.
One use each of LSD and val-
ium.
patterns of drug use, for each of the various points in Figure I, are
described in Table 4.
To summarize the data in Figure I, at follow-up the three outcome
groups differed predictably on the index of mean drug use. This result
did not arise simply out of differences in drug use in the year before
treatment. At that time the U and the I groups reported essentially the
same high levels of drug consumption; however, the 5 group reported
significantly lower levels of consumption before treatment than the
other two outcome groups.
TREATMENT GOALS ASSESSMENT
A further analysis of the pretreatment status of the three outcome

groups involved examination of the Treatment Goals form. The results
of these analyses for each of the two categories of goal area are present-
ed in Figure 2. The figure indicates that the U group had more drug
DRUG PROBLEMS OTHER PROBLEMS

1.8 7
1.0
<II
E
Q)
:0 T
0
a.
'0 SELF-RELIANCE
ci
c
.,
c
Q)
::?; 3.5
1.2
3.0
1.0
•
2.5
.8 2.0
T T
s 5
value significantly different from other groups
FIGURE 2. Mean scores on measures of problem multiplicity and self-reliance for clients in
the three outcome categories.
goals than the other two groups, which were similar on this measure.
With regard to other problems the U group identified the most prob-
lems, the S group fewest, and the I group were intermediate. Analyses
of variance confirmed these impressions (for Drug Goals, F2 ,46 = 4.00,
p<.05, U group significantly separated from I and from S, p<.Ol by
Tukey's HSD; for Other Problems, F2 ,46 = 5.76, p<.Ol, all groups signifi-
cantly separated at p<.Ol by Tukey's HSD). Analyses of variance for
self-reliance scores revealed that the U group was significantly removed
from the other two groups for drug goals (F 2 ,46 = 5.25, p<.Ol). There
was a similar trend for other problems, but this was not statistically
significant (F 2 ,46 = 1.74). In summary, the U group consistently identi-
fied more goals than the I and S groups, and indicated need for profes-
sional assistance with a larger number of goals. The I group did not
differ from the S group in number of goals for which they needed pro-
fessional help, or in number of drug use goals. However they did report
more problems in other life areas.
WITHIN TREATMENT MEASURES
The next phase of the analysis was the examination of performance

of the three outcome groups on three clusters of variables measured
within treatment. As previously indicated, the three clusters were as-
sumed to represent (a) initial client motivation, (b) initial problem sever-
ity, and (c) changes in motivation within treatment. Some of the mea-
sures were collected during Session 1 and others during Session 2. As a
check for the appropriateness of this procedure an analysis was made of
subject attrition from Session 1 to Session 2. A total of 11 subjects did not
appear for Session 2, one from the S group and five from each of the I
and U groups. Fisher's Exact Probability test was applied to check for
differential attrition from group S, but the hypothesis was not confirmed
(p = .220).
Initial Client Motivation

Self-Monitoring. The degree to which clients complied with the re-
quest to self-monitor drug use behaviors was categorized into three
levels: Daily self-monitoring; self-monitoring, but not each day; and not
self-monitoring. The percentages of subjects in each category, from each
of the outcome groups, are presented in Table 5. The three groups did
not differ in the degree of compliance in Session 2. This was assessed by
Chi-square analysis of Daily versus Not Daily or Not Self-Monitoring (X 2
TABLE 5.
Measures Obtained During Session 1 and Session 2
Groups
S u
Session 1
N 11 20 18
Drug use:
Mean no. drug uses (past week)a 13.4 26.6 40.6
% of clients abstinent 18 10 6
Goals set:
Mean max. drug uses b 14 25 26
% of clients planning abstinence 36 20 22
Session 2
N 10 15 13
Self-monitoring: b
Daily (%) 70 67 77
Not every day (% ) 10 20 15
Not self-monitoring (%) 20 13 8
Cravings: b
% Reporting 40 53 47
Refusals:
% Reporting 20 60 c 22
-Significant difference between groups.

bNo significant between-group differences.
<Trend to between-group difference.
= .36, df = 2). Overall, 71 % of clients attending this session self-

monitored daily, and an additional 16% self-monitored, though not
daily.
Craving. The frequency of cravings for drugs, recorded on the self-
monitoring cards, was assessed in Session 2. The percentage of subjects
reporting any cravings are presented in Table 5. The groups did not
differ on this measure (X2 = .44, df = 2), which ranged between 40% and
53%.
Refusals. The percentage of subjects reporting any refusals of drugs
in the previous week was also assessed in Session 2 (Table 5). Overall,
37% of subjects reported some refusal. The I group were more likely
than the other groups to report refusal (60% vs. 20% and 23%). A Chi-
square analysis indicated a trend to significance of this result (X 2 = 5.73,
df = 2, p< .10), however this finding should be interpreted with great
caution because the expected values in two cells fell below 5 (Siegel,
1956).
Initial Drug Use Goals. The fourth index of motivation was assessed
in Session I, in the form of goals for drug use in the coming week. Two
kinds of assessment were made on these data: (a) The percentage of
subjects in each group aiming for total abstinence are presented in Table
5. The frequencies were too low to permit statistical analysis, but the
data do not indicate a systematic difference between groups on this
measure. (b) The mean maximum frequency of drug uses was computed
for each group. Maximum frequency of drug uses was defined as the
sum of the product of maximum days of use by maximum frequency in
days of use, summed across all drug classes. (Almost all anticipated
drug uses were for alcohol or cannabis.) The obtained values are pre-
sented in Table 5. One-way analysis of variance indicated no overall
main effect of groups (F2 ,46 = 1.55).
Initial Problem Severity

The measure selected to indicate initial problem severity was the
frequency of drug uses reported in the retrospective history of the week
preceding Session 1 (Table 5). One-way analysis of variance indicated a
significant effect of groups (F2 ,46 = 4.55, p<.05). Between group com-
parisons indicated that the S group reported significantly lower frequen-
cy than the I group, who in turn reported lower frequency than the U
group (Tukey's HSD a=.05, 12.00). The proportions of clients reporting
abstinence for that week (Table 5) reveal a similar pattern of results.
CHANGE IN CLIENT MonvATION

Change in Drug Use Goals. The first analysis of changes in client
motivation involved examination of the numbers of clients who raised
any of their drug use goals in Session 2, by increasing the maximum
drug uses for any drug class from the level set in Session 1. One client in
Group S, two in Group I, and six in Group U raised their goals, indicat-
ing that such change was more likely in group U (Fisher's Exact Proba-
bility test, p= .028).
Client Attrition from Treatment. Although attrition from treatment did
not differ significantly between Sessions 1 and 2, there was an indication
of greater attrition of U subjects between Sessions 1 and 3. This hypoth-
esis was evaluated by contrasting the proportions of clients remaining in
treatment for the three groups at Session 3 (X 2 = 10.67, df = 2, p<.OI).
Most of this effect appeared to derive from attrition of U clients between
Sessions 2 and 3; the proportion of subjects failing to return to Session 3
after attending Session 2 was significantly higher in the U group (Fish-
er's p<.OOl). At Session 3, 90% of the S group, 73% of the I group, and
30% of the U group attended.
DISCUSSION
We stress at the outset of the discussion that our purpose has been
to describe variables that are predictive of outcome, and that can be
measured during the early stages of treatment. Such information may be
of use to clinicians in selecting a treatment strategy. This chapter does
not constitute an attempt to evaluate the relative importance of the
variables assessed. Furthermore, the findings that have been presented
represent group data, and their usefulness in making individual predic-
tions has yet to be evaluated. Nonetheless, in the absence of any identi-
fied individualized predictors of outcome, the present findings can serve
a useful heuristic function in developing future research and in manag-
ing individual cases.
PRETREATMENT DRUG USE
The initial finding of the study was that pretreatment drug use was
predictive of outcome status. The S group, who had the lowest drug use
at follow-up, also reported lower levels of overall drug use at assess-
ment. However, the I and the U groups were not discriminated on this
measure of drug use at assessment, though they differed at follow-up.
This finding suggests two interpretations: either variables other than
initial drug use accounted for the superior performance of the I group
over the U group, or the measure of overall drug use was too insensitive
to discriminate pretreatment differences between these groups. Because
the measure separated the S group from the other groups at assessment,
we favor the former hypothesis. The reader should bear in mind that
outcome status (successful, improved, unimproved) was based on the
clients' status at follow-up, and, as indicated by Figure 1, the I group
may have achieved as great a change in their drug use as the S group.
Thus, though their outcomes were different, these two groups may none-
theless have progressed equally, and differences at follow-up may simply
reflect pretreatment differences.
Measures of drug use in the week following assessment but preced-
ing Session 1 revealed differences between the three groups, which
mirrored their status one year later. That is, before the intervention
stage had begun the three groups had separated themselves. This sug-
gests either that measures of recent drug use (past week) are better
predictors of outcome status than measures representing a longer period
(one year), or that self-initiated change in behaviour (by the I and S
groups) is a good predictor of follow-up status. Whatever the explana-
tion of this finding, it is clear that information about recent drug use may
be useful in determining whether clients are likely to fare well after a

brief intervention, such as the one described.
The information collected on the Treatment Goals Assessment form

was used to estimate the multiplicity of client problems. The unim-
proved group was significantly more likely that the successful and im-
proved groups to identify problems with their use of alcohol and other
drugs. The multiplicity of other problems mirrored the outcome status
of the three groups, the S group identifying fewest and the U group
identifying most. Thus the multiplicity of client-identified presenting
problems is related to level of drug use at follow-up. This information
can be very rapidly assessed before assignment to treatment and may
prove useful in determining the intensity and complexity of treatment
that clients should receive.
SELF-RELIANCE
The measure of self-reliance can be viewed as an indirect assess-

ment of the clients' perceived self-efficacy (Bandura, 1977, 1982) for
achievement of their various goals. As was indicated in the results sec-
tion, this measure separated the U group from the S and I groups. In
contrast the I group, though they indicated a significantly greater multi-
plicity of problems than the S group, were no more likely to request help
for problems. Again, though by objective measures their drug problems
seemed as severe as those of the U group, by subjective assessment they
were not. A compelling suggestion from this finding is that the clients'
subjective estimates of the severity of their problems should be consid-
ered, as well as more objective measures, in formulating a treatment
plan. This assertion is consistent with a variety of studies by Bandura
and others (Bandura, 1982; Locke, Frederick, Lee, & Bobko, 1984) indi-
cating that subjects' assessments of their future performance are among
the best available predictors of their subsequent behavior.
In summary, the pretreatment severity of drug use and the multi-
plicity of problems were related to the outcome of treatment. Further-
more, clients' subjective judgments about the need for professional as-
sistance in achieving their goals were predictors of treatment outcome.
Clients low on self-reliance were more likely to have an unsuccessful
outcome. Information concerning levels of drug use, problem multi-
plicity, and self-reliance can easily be collected in the early stages of
treatment and should serve as a basis for planning the nature of the
intervention to which clients are assigned. This is probably common

practice in regard to intensity of drug use and problem multiplicity.
What is more novel about the present findings is the indication of the
usefulness of the clients' subjective judgments. Clinicians could use
such judgments for more extensive assessment of their origins, and the
nature of the treatment the client considers necessary. In short, it may
be possible to exploit clients' ability to make personal prognostications
so as to improve the effectiveness of treatment services.
CLIENT MOTIVATION
In the present study we employed five measures as indirect indica-

tors of client motivation: Compliance with the request to self-monitor
behavior; frequency of drug refusals; frequency of cravings; relaxation of
drug goals; and attrition from treatment. The first three of these mea-
sures we interpreted as indicators of the clients' initial level of moti-
vation for treatment. The latter two measures, on the other hand, are
interpreted as reflecting change in client motivation during the course of
treatment. In his extensive review of the literature on motivation for
treatment, Miller (1985) indicates that a wide variety of indexes have
been used to operationalize this variable. The conceptual bases for this
multiplicity of measures are obscure, and appear to lie in the idiosyn-
cratic subjective judgments of investigators or clinicians. In formulating
the measures of initial motivation we conceptualized them as behaviors
that have at least the appearance of aversive properties: Effort in record-
ing one's own behavior, restraint from drug use leading to "cravings"
that have to be resisted, and refusals of the presumably friendly offers of
drugs from friends and acquaintances. Change in motivation was ex-
pressed by behaviors contrary to the explicit objectives of treatment
(reduction of drug use, and completion of the program).
The measures selected to represent initial levels of motivation for
treatment (self-monitoring, craving, refusals) yielded no evidence that
the unsuccessful group exhibited lower levels of motivation than the
other two groups, (though the I group indicated higher levels than the U
and the S in frequency of refusals). In short, there was no indication that
treatment failure was associated with lower levels of initial motivation.
Parenthetically, it may be noted that Miller (1985) indicates that
expressed need for help is taken by some as evidence of client moti-
vation. By this index one encounters the paradoxical finding that the
group with the highest levels of initial motivation was the unsuccessful
group.
PROGRESS IN TREATMENT
Change in motivation, as measured here by relaxation of goals and

attrition from treatment, was associated with treatment failure. It is
interesting to note that on the average the S and the I groups achieved at
Session 2 the goals they had set in Session 1. In contrast, the U group
failed on the average to achieve the goals they had set. It may be the case
that this failure experience contributed to the change in motivation ob-
served in this group, as expressed in attrition from treatment. Sandura
and Cervone (1983) have suggested that goal setting and performance
feedback are important elements of effective treatments. The present
finding suggests the importance of attempting, in clinical practice, to
ensure that the goals clients set for themselves are goals that they be-
lieve they can achieve. Performance feedback about failure to attain the
initial goals of treatment may lead to client attrition.
In considering motivation for treatment it may be useful to delineate
three aspects that, at the phenomenonological level, are importantly
distinct. Clients are likely to enter psychological treatment having re-
solved to change behavior or affective states-this we term motivation
for change. In terms of the theme of this volume this would be after
contemplation and into decision making and action. Some persons with
motivation for change initiate the changes themselves, whereas others
seek professional assistance, which may be said to indicate motivation
for treatment. In other words the action phase may be autonomous or
assisted. Various treatment options are usually available. The extent to
which a client views the particular treatment that is offered as appropri-
ate to his or her needs could be described as motivation for the treat-
ment. The present findings suggest that failing to meet goals of treat-
ment may undermine such motivation. In such circumstances the
treatment options are to set more achievable goals; to select alternative
treatments that the client may view as more promising; or to make
motivation attributions about the client to account for the lack of suc-
cess. We suggest that either of the first two options have merit, and the
person best qualified to choose between them is probably the client. In
our study we found no justification for the third recourse.
In summary, three major conclusions seem to be justified by the
present study. First, a majority of the clients were significantly im-
proved at one year follow-up, after receiving the brief outpatient inter-
vention. Hence, for selected multiple substance users, such treatment is
apparently helpful. Second, problem multiplicity was, as expected, pre-
dictive of outcome status. Furthermore, the higher the clients' self-re-
liance for coping with problems, the more likely they were to have
improved significantly at follow-up. Clearly, measures of problem mul-
tiplicity and client self-reliance may be useful in making initial decisions

about treatment disposition. Finally, the outcome groups did not differ
systematically on a variety of measures of initial motivation. However,
clients who were subsequently unsuccessful showed evidence of change
in motivation for the treatment within one or two sessions. Such infor-
mation may be used clinically as the basis for raising with the client the
advisability of seeking a more intensive form of treatment.
ACKNOWLEDGMENTS
The authors wish to acknowledge the valuable contribution of our

colleagues Joanne Cordingley, Garth Martin, and Kim Ankers who as-
sisted with data analysis. We also thank the staff of the Assessment and
Follow-Up Units who collected clinical data. For their valuable com-
ments on the original draft of the manuscript, we thank Howard Cap-
pell, Doug Chaudron, William Miller, Wayne Skinner, Linda Sobell, and
Mark Sobell. Finally, we thank Loretta Pavan for her assistance in pre-
paring this manuscript.
REFERENCES
Bandura, A. (1977). Self-efficacy: Toward a unifying theory of behavioral change. Psycho-

logical Review, 84, 191-215.
Bandura, A. (1982). Self-efficacy mechanism in human agency. American Psychologist, 37,
122-147.
Bandura, A., & Cervone, D. (1983). Self-evaluative and self-efficacy mechanisms govern-
ing the motivational effects of goal systems. Journal of Personality and Social Psychology,
45, 1017-1028.
Benton, A. L. (1963). The revised visual retention test: Clinical and experimental applications (3rd
ed.). Iowa City, IA: The State University of Iowa.
Farley, E. c., Santo, Y., & Speck, D. W. (1979). Multiple drug-abuse patterns of youths in
treatment. In G. M. Beschner & A. S. Friedman (Eds.), Youth drug abuse: Problems,
issues and treatment (pp. 149-168). Lexington, MA: Lexington Books.
Jastak, J. F., & Jastak, S. (1978). Wide range achievement test: Manual of instructions. Wilming-
ton, DE: Jastak Associates.
Locke, E. A., Frederick, E., Lee, c., & Bobko, P. (1984). Effects of self-efficacy, goals, and
task strategies on task performance. Journal of Applied Psychology, 69, 241-251.
Lorei, T. W. (1982). Goal-oriented outcome evaluation in drug dependence treatment
programs. In R. J. Craig & S. L. Baker (Eds.), Drug dependent patients: Treatment and
research (pp. 263-307). Springfield, IL: Charles C Thomas.
Martin, G. W., Wilkinson, D. A., & Kapur, B. (1984, August). Validation of self-reported
cannabis use by urine analysis. Paper presented at the Third International Conference on
the Treatment of Addictive Behaviors, North Berwick, Scotland.
Matarazzo, J. D. (1972). Wechsler's measurement and appraisal of adult intelligence (5th edition).
Baltimore, MD: Williams & Wilkins.
Miller, W. R (1985). Motivation for treatment. A review with special emphasis on alco-
holism. Psychological Bul/etin, 98, 84-107.
Miller, W. R, & Munoz, R F. (1976). How to control your drinking. Englewood Cliffs, NJ:
Prentice-Hall.
Miller, W. R., Taylor, C. A., & West, J. (1980). Focused versus broad-spectrum behavior
therapy for problem drinkers. Journal of Clinical and Consulting Psychology, 48, 590-601.
Moos, R. H., & Finney, J. W. (1983). The expanding scope of alcoholism treatment evalua-
tion. American Psychologist, 38, 1036-1044.
Ogborne, A. (1978). Patient characteristics as predictors of treatment outcomes for alcohol
and drug abusers. In Y. Israel, F. B. Glaser, H. Kalant, R E. Popham, W. Schmidt, &
R G. Smart (Eds.), Research advances in alcohol and drug problems (Vol. 4, pp. 177-224).
New York: Plenum Press.
for excessive drinkers two years after consultation. Behavior Research and Therapy, 14,
409-418.
Paitich, D. (1979). The Clarke Vocabulary Scale: A multiple-choice estimate of WAIS vocabulary
level. Unpublished manuscript. Toronto: Clarke Institute of Psychiatry.
Pattison, E. M. (1978). Differential approaches to multiple problems associated with alco-
holism. Contemporary Drug Problems, 7, 265-309.
Sadava, S. W. (1984). Concurrent multiple drug use: Review and implications. Journal of
Drug Issues, 14, 623-636.
Sanchez-Craig, M., Wilkinson, D. A., & Walker, K. (1986). Theory and methods for
secondary prevention of alcohol problems: A cognitively-based approach. In M. Cox
(Ed.), Treatment and prevention of alcohol problems: A resource manual. New York: Aca-
demic Press.
Siegel, S. (1956). Nonparametric statistics for the behavioral sciences. New York: McGraw-Hill.
Skinner, H. A. (1979). Assessment of clients with alcohol problems: Basic principles, critical issues
and future trends (ARF Substudy #1061). Toronto: Addiction Research Foundation.
Sobell, L. c., & Sobell, M. B. (1980). Convergent validity: An approach to increasing
confidence in treatment outcome conclusions with alcohol and drug abusers. In L C.
Sobell, M. B. Sobell, & E. Ward (Eds.), Evaluating alcohol and drug abuse treatment
effectiveness: Recent advances (pp. 177-183). New York: Pergamon Press.
Sobell, M. B., Sobell, L. c., Ersner-Hirshfield, S., & Nirenberg, T. D. (1982). Alcohol and
drug problems. In A. S. Bellack, M. Hersen, & A. E. Kazdin (Eds.), International
handbook of behavior modification and therapy (pp. 501-533). New York: Plenum Press.
Wilkinson, D. A., & Martin, G. (1983, December). Experimental comparison of behavioral
treatments for multiple drug abuse: Brief outpatient self-control training and two broad spec-
trum residential treatments. Paper presented at the 17th Annual Convention of the
AABT, the World Congress of Behavior Therapy, Washington, DC.
Wilkinson, D. A., & Martin, G. (1984, August). Concordant findings with multiple outcome
measures in evaluation of behavioral treatment of multiple drug abuse. Paper presented at the
Third International Conference on the Treatment of Addictive Behaviors, North Ber-
wick, Scotland.
12
Description and First Results of an
Outpatient Drug-Free Treatment
Program for Opiate Dependents
SABINE DEHMEL, FRANZ KLETT, AND

GERHARD BUHRINGER
INTRODUCTION
At the end of the 1960s an attempt was made in the Federal Republic of
Germany to establish treatment programs for the increasing number of
people dependent on illicit drugs. The major goal was the development
of residential drug-free programs. In addition to a variety of other ap-
proaches, a research group at the Max-Planck Institute for Psychiatry
developed and implemented a treatment program based on behavior
therapy (Biihringer & De Jong, 1980; Biihringer et al., 1978). Information
is available concerning program implementation, effectiveness (De Jong
& Henrich, 1978) and follow-up results to 8 years after treatment (De
Jong & Henrich, 1978, 1980; Klett, Hanel, & Biihringer, 1984). Until the
beginning of the 1980s, residential long-term therapy was judged by the
majority of those working in the field as the only treatment that prom-
ised success. Only recently has outpatient treatment received some ac-
ceptance in Germany.
SABINE DEHMEL, FRANZ KLETT, AND GERHARD BDHRINGER • Addiction Research

Group, Department of Psychology, Max-Planck Institute for Psychiatry, 8000 Munich 40,
Federal Republic of Germany. This study was funded by a grant from the Federal Depart-
ment of Youth, Family, and Health.
263
264 SABINE DEHMEL ET AL.
The experiences gained in the residential treatment program, and in

two pilot studies by the same research group of outpatient treatment of
drug dependents (Feldhege, Krauthan, Schneider, Schulze, & Vollmer,
1977) and young alcohol dependents (Vollmer & Kraemer, 1982; Vollmer
et al., 1982), were the basis for an ongoing, 4-year research project on the
outpatient treatment of drug dependents. The overall goals of this pro-
ject are (a) the development and evaluation of a comprehensive drug-
free treatment program based on behavior therapy, including follow-up
data; (b) analysis of the organizational aspects of the implementation of
such a program in German outpatient treatment centers (e.g., coopera-
tion with physicians' private practices); and (c) analysis of the target
group for such a program (Le., indications for outpatient treatment).
This chapter includes a description of the program and some first results
from the subgroup of those clients whose treatment has now been
finished.
PROGRAM PHILOSOPHY AND TREATMENT GOALS
The outpatient program is based on a behavioral approach. The

principle of changing behavior and cognitive structures by replacing
inadequate behavioral patterns with more appropriate responses and
coping skills is applied to the treatment of drug dependence. The
change model integrates aspects such as therapist-client interaction
(Kanfer & Grimm, 1980), the analysis of individual problems that cause
or are related to the dependent behavior, the client's resources and
abilities, and the issue of the correct timing of the next step within the
change process (DiClemente & Prochaska, 1985).
An outpatient program has the advantage of being able to change
behavior within the context of the client's actual, day-to-day life situa-
tion. Because change is not a linear process, but takes place in a system
of interactions and response patterns, integration of family members or
significant others is necessary at certain points during treatment. In
order to describe this therapeutic approach by its objectives, structure,
and interventions, the image of a circle, as shown in Figure I, may be
helpful.
There are two sections within the circle shown in Figure I, corre-
sponding to two basic assumptions for effective drug-free treatment.
The first section focuses on the extinction of overt drug-using behavior
and the second on the maintenance of drug abstinence in a long-term,
drug-free life. There are four different phases in the cycle: drug abuse,
detoxification, stabilization of drug abstinence, and relapse. The thera-
peutic approach is analogous to the model of processes and stages of
DRUG-FREE TREATMENT PROGRAM FOR OPIATE DEPENDENTS 265
~"""h~---'" Termination of
treatment
G Extinction
of drug
consumption
~ Maintenance
.
of abstinence
FIGURE 1. Model of treatment stages.
change described by Prochaska and DiClemente; clients who enter the

program can be at different phases of the cycle and it is likely that they
will move through this cycle more than once. Each phase can be seen as
an important treatment objective and movement from one stage to the
next involves different processes. Therapy has to be orientated to the
client's deficits, needs, resources, and abilities to enable the client to
achieve long-standing drug abstinence.
The outpatient treatment combines a structured framework, de-
fined by outcome goals (such as abstinence, social integration, solution
of individual problems) and rules, with an individual procedure related
to the client's present situation and strains. Regardless of the phase in
which the client enters treatment, there is a set procedure at intake. The
first contact has the following components: obtaining demographic data
from clients concerning drug abuse and present situation (e.g., court
referral); giving a brief description of therapy by pointing out goals and
rules (Le., controlled urinanalysis, regular attendance, etc.); and estab-
lishing motivation for therapy (e.g., reinforcing the reasons for drug
abstinence). It is important that the therapist shows evidence of such
interactional variables as acceptance, interest, understanding, confiden-
tiality, and role clarity (Kanfer & Grimm, 1980), which are part of the
whole therapeutic process. Ineligible for admission are those clients who
need only aftercare services, who are not dependent on hard drugs, and
who disagree with the rules and basic goals of the program. The treat-
ment duration for planned discharge ranges from 8 to 12 months, with

an average of two individual therapy sessions per week.
The first part of the treatment cycle may be the phase of consuming
drugs. The focus is on overt behavior, with the goal being to establish a
motivation for detoxification. This phase resembles that described by
Kanfer and Grimm (1980) as the development of a commitment for
change and can be compared with Prochaska and DiClemente's stage of
contemplation, in which clients have considered giving up drugs but
have not yet changed their drug-consuming behavior. The therapy has
to build up and strengthen motivation for abstinence by the following
interventions: Working out goals incompatible with drug abuse; analyz-
ing the value of these goals; examining the pros and cons of drug absti-
nence and their short-term versus long-term consequences; building up
self-confidence by pointing out prior successful changes and positive
experiences with abstinence; using drug-free models; and, finally, end-
ing the process by making a contract regarding detoxification. The possi-
ble methods of withdrawal are the reduction of drugs, an abrupt absti-
nence with the support of a physician or significant others, and going
into a hospital. Through these processes, the client acquires self-respon-
sibility and moves from a more passive cognitive state to the active stage
of detoxification.
A special problem in drug abuse treatment programs is related to
the percentage of clients who enter therapy because of court pressure.
Quite often, the external motivation makes it difficult to move a client
into a state of self-responsibility. These clients appear to commit them-
selves to change and attainable goals, but constantly try to find ways to
continue their old behaviors.
To complete detoxification, clients learn to anticipate withdrawal
symptoms in order to lower anxiety, to analyze situations that are cues
for drug consumption, to develop alternative behaviors in tempting sit-
uations, to rehearse goals and values concerning abstinence, and to
change their usual daily pattern. The focus of the interventions is always
on a client's potential, given his personal limitations.
This last phase lasts until the client has abstained from drug use for
at least a week. Changing into the stage of maintenance is a very impor-
tant and critical step. Abstinence is necessary because of its links with
achieving goals in other areas of daily life, such as jobs, housing, and
relationships that are measures of personal success and satisfaction. If
clients change in positive ways and are still consuming drugs, the dan-
ger of the positive conditioning of drug abuse is very likely.
The other part of the cycle can be called the stabilization of drug
abstinence. It is important to emphasize that processes in this part of
treatment have, to a certain extent, also been issues in the previous part.
The positive value of a drug-free life has to be supported by reducing

personal inadequacies and attaining new resources. Social integration,
problem solving, self-management, restructuring of self-concept, expec-
tations, and vulnerabilities are the goals of this phase. Clients learn to
set attainable goals, to be aware of the original function of drugs and
replace it by adequate coping skills, and to observe and control them-
selves. The treatment of alternative behavior also includes therapeutic
techniques like covert conditioning, contingency contracting, and as-
sertiveness training (Gotestam & Melin, 1980). Timing is an essential
issue because change can have an overwhelming effect that may lead to
a relapse, and this starts the cycle all over again. As change is not a
linear process, we expect relapses to occur and there are many reasons
for them. There can be external factors like an extremely stressful life
situation, temptation events with which the client cannot yet cope, or
therapy-related factors that have to be attended to in the treatment
process.
It is always important, in terms of relapse prevention, to address
motivation for change, the modification of goals, the extent of alter-
native behavior, the connection between personal problems and the
original drug use, and self-responsibility. The anticipation of a possible
relapse and working it through by sensitizing clients for signals within
their behavior are very useful intervention techniques (Cummings, Gor-
don, & Marlatt, 1980). If a relapse occurs, therapy has to start again at
the first point of the cycle. The time spent moving towards the point of
detoxification might decrease, but it is essential to address every step
within this phase (pro/cons, decision, negotiation). In addition, the
function and meaning of the relapse should be analyzed and addressed
in terms of new information for self-control.
In the stage of maintenance, there is a move from acute problem
solving towards a generalization of learning. Clients should be able to
manage their lives without professional help and cope with future prob-
lems in an adequate way.
METHOD
The study was designed using a pre-post treatment assessment and

further assessments are made during the course of treatment. Follow-up
interviews are conducted at 6, 12, and 24 months after the end of treat-
ment. At enrollment into treatment, information about frequency and
type of drug use, employment, personal relationships, and housing
situation is gathered. These data refer to the client's situation in the
periods of 6 or 12 months before entering treatment. No control group is
included for both ethical and practical reasons. It is very difficult to

obtain a group of untreated drug dependents because, if they are in-
terested in treatment and if entry is not possible within a certain time,
they choose to enroll in a different program or disappear.
In order to evaluate the results of the program, we plan to compare
clients finishing therapy to dropouts. In addition, comparisons with
clients of residential treatment facilities are to be made.
ASSESSMENT
Standardized questionnaires, rating scales, and clinical evaluation

are used to document the status of the client upon enrollment in therapy
and at the time of completion. This assessment includes personality
tests, socioeconomic data, and the life situation before regular drug use
and in the 6 months prior to therapy, as well as detailed information
about the client's use of legal and illegal drugs and other addictive
substances. Information about the course of therapy is collected from
the notes of therapy sessions and ratings, which focus on treatment
objectives and individual goals. In order to assess the client's drug-
abstaining behavior, unscheduled, controlled urinanalyses are carried
out.
Follow-up assessment, which includes all clients, is carried out by
staff unconnected with the therapy team. The interview includes stan-
dardized questionnaires and rating scales, adjusted and comparable to
the pre- and postmeasurements, as well as urinanalysis. This assess-
ment corresponds largely with the follow-up standards of the German
Society for Addiction, Research & Treatment (Deutsche Gesellschaft fur
Suchtforschung und. Suchttherapie, 1985).
SUBJECTS
The outpatient treatment program is planned for people dependent

on hard drugs, which include predominantly heroin and some other
addictive substances, like amphetamines and barbiturates. Enrollment
in treatment is independent of the type of drug use (oral, i.v.) and
detoxification is not a required condition for enrollment. Those free of
drugs for more than 6 months or those who have been imprisoned for
more than 2 years are ineligible for admission to the program. These
requirements prevent the outpatient treatment program from being
used as an aftercare facility. It is also required that clients live within a
given distance from Munich in order to ensure that they are able to
attend frequent therapy sessions. Clients are referred by different
sources, such as the courts, probation officers, drug counseling centers,
physicians, and other clients or friends.
The data that follow are based on 46 clients who had completed
treatment (planned discharge and dropouts) at the end of 1984. 21 cli-
ents are still enrolled in treatment. The distribution of male and female
clients matches the figures from other therapeutic institutions, with 70%
(32) males and 30% (14) females. With four exceptions, all are German
nationals. In the entire sample of 46, only three clients are married and
one is divorced. Average age is 25 years on enrollment, with a range
from 18 to 44.
Table 1 shows that 30% of the clients had failed to finish secondary
education and nearly one fifth never completed elementary school. In
terms of diagnosis, all clients were dependent on hard drugs. (The defi-
nition of hard drugs includes all sorts of opiates, cocaine, ampheta-
mines, such as speed, and all medically used narcotics, as recommended
by the Federal Criminal Investigation Department.) Only 9% used nar-
cotics or opiates (morphine, codeine) other than heroin as their major
drug.
Nearly all our clients have injected drugs (91 %). The average age of
first regular use of hard drugs is approximately 18 years and the average
duration of dependence on hard drugs is 7 years before enrolling in
treatment. Figure 2 shows the drug consumption pattern during the 4
weeks prior to treatment. This reflects the efforts of some clients to
reduce consumption before entering treatment.
TABLE 1.
Education Characteristics at the Beginning
of Treatment
N %
Elementary school, no degree 8 17

Elementary school graduate 7 15
High school, no degree 14 30
High school graduate 10 22
College education, no degree 1 2
Unknown/other 6 13
Total 46 100
Percent
50
40
30
20
10
abstinent rare consump· frequent frequent

tion of soft consumption consumption
or hard drugs of hard drugs of hard drugs
(oral) (injection)
FIGURE 2. Drug consumption during the 4 weeks prior to treatment.
RESULTS
PROBLEMA TIC BEHAVIOR
Drug Behavior and Previous Treatment

At the beginning of treatment, 26% of the 46 clients were abstinent,
57% used hard drugs, and the remainder were dependent on medica-
tions, soft drugs, or abused alcohol. Nearly 22% failed to detoxify. 12
clients were detoxified at home, 11 stopped their drug use by abrupt
withdrawal, and one reduced his dose gradually. Three clients under-
went detoxification in a psychiatric ward of a hospital.
Reports of previous treatment experiences show that half the clients
had never attended a residential treatment program, 26% had once been
involved in residential treatment, and 24% had participated in two to
five different treatment programs. As only four clients reported com-
pleting previous treatment, the percentage of dropouts (83%) is very
high.
Education and Employment

Eight clients (17% ) had dropped out of primary school and 15 clients
(33%) failed to complete secondary education. An even higher percent-
age (63%) dropped out of vocational training, with six failing to com-
plete two or more training experiences. Only 22% passed the final exam-
inations of their vocational training. These data reflect the problems
surrounding vocational integration: clients are too old to enroll in voca-

tional training program and therefore can achieve only an unskilled
working position.
In the 6 months before beginning therapy, 4% were enrolled in
educational activities, 11 % were regularly engaged in unskilled em-
ployment, and approximately 11 % were regularly employed as skilled
workers. About 13% were without steady employment and more than
39% were unemployed. This percentage of unemployment is four times
higher than the average rate in Germany.
The major source of income for eight clients (17%) in the 6 months
preceding therapy was their own jobs. Seventeen (37%) received sick
pay, social support, or were supported by their relatives, and eight
(17%) lived on illegal income.
Legal Status
Approximately one fifth (22%) of the clients had neither appeared
or been sentenced in court before enrolling in the program. Sixty-one
percent had been found guilty of buying, dealing, and/or possessing
illicit drugs. Seventeen percent had been convicted of non-drug-related
criminal activities. Twelve (26%) of the 46 clients had been sentenced to
jail without probation.
Compared to residential treatment facilities, the percentage of cli-
ents who had a court order for therapy was low and more than half
(54%) entered voluntarily. Forty-six percent were required to undergo
drug treatment, with five cases being forced to choose between entering
treatment or receiving a jail sentence. This is in accordance with Para-
graph 35 of the German Narcotic Act.
OUTCOME
Total Sample
Thirty percent (14) of the participants in the program completed
treatment and 52% (24) dropped out. Therapy was terminated in 11
cases by the client, in 8 cases by the therapist, and 5 clients left for other
reasons. Six (13%) clients were placed in a long-term inpatient treatment
program because outpatient treatment was too difficult for them, and
two were imprisoned.
Clients attended an average of 25 sessions of therapy, with a range
from 4 to 89 sessions. Thirty-nine meetings was the average duration of
therapy for those clients finishing treatment. Those who quit partici-
pated in 17 sessions and those sent to residential programs attended 24
sessions. The average length of therapy was 9 months when regularly

finished and 4 months when participants were sent to other treatment
programs. Those clients not finishing therapy quit after an average peri-
od of 3 months.
Comparisons Between Clients with and without Court-Ordered Therapy

As consuming, possessing, and dealing with drugs is illegal, a large
number of drug dependents enter therapy programs as the result of
court orders. In our program, 46% of the clients had to prove to the
court, in different ways, that they were involved in therapy. Comparing
this figure with recent data from an inpatient treatment program, where
85% of the clients are court ordered, there is a suggestion that an outpa-
tient, drug-free program is more attractive to clients who enroll
voluntarily.
Our sample includes a small group who were ordered to a treat-
ment program in accordance with Paragraph 35 of the German Narcotic
Act. These clients had been in jail prior to entering court-ordered thera-
py. Personal demographic data suggest some differences betweeen
these clients and others with court orders. The former had generally
been addicted for longer periods (ranging from 6 to 11 years) and had
previously participated in treatment programs more often than the refer-
ence group of court-ordered clients.
More information can be gathered from the comparison between
clients who started therapy as a result of a court order and those who
came voluntarily. The outpatient treatment program was the first thera-
py experience for more than two thirds of the clients (70%) who had
voluntarily entered treatment. In comparison, 24% of those with court
orders had no previous treatment experiences.
Upon enrolling in therapy, 80% of those who had voluntarily en-
tered the treatment program were employed or in school and 71 % of
these lived in their own apartments. By contrast, 25% of those required
to enter therapy by the court were employed and 53% were living on
their own. Forty-three percent of the voluntary clients had never been
convicted of criminal or drug-related activities. This shows the higher
social integration of clients who enter therapy on a voluntary basis. No
group differences were found with respect to duration of dependence.
The program was also voluntarily chosen by clients with a long history
of drug dependence.
At the beginning of therapy, 22% of the voluntary group were ab-
staining and the remaining 78% were still consuming hard drugs. Of
those clients required to enter therapy, 50% were abstaining, whereas
the remainder had changed to alcohol, medication, and/or soft drugs.
Of those required to undergo treatment, 38% completed treatment,

compared to 20% among the voluntary clients.
Comparisons Between Dropouts and Planned Discharges

Completion of the program was defined as abstinence from drug
use for at least 3 months, social integration in accordance with the cli-
ent's competence, and a consensus between client and therapist con-
cerning the achievement of individual goals, the stabilization of the
social situation, and an awareness that the client was able to cope with
problems adequately.
All clients who terminated treatment before achieving all these
goals, and were not referred to other programs or imprisoned, are in-
cluded in the dropout group, irrespective of the reasons for termination.
The following examples point to some reasons for dropout: obvious
violation of rules, such as ignoring a negotiated contract concerning
abstinence; unsuccessful detoxification attempts; the client's satisfaction
with his present life situation.
Of the 24 clients who dropped out, a high percentage (54%) quit
treatment before the 15th session. The group who were referred to resi-
dential treatment stayed an average of seven sessions longer in treat-
ment. This is related to the fact that, after the decision for residential
treatment has been made, there is a preparation period before referral.
Table 2 shows that the critical period for treatment termination is within
the first 3 months. This finding corresponds to results from American
outpatient treatment programs (Craddock, Hubbard, Bray, Cavanaugh,
& Rachal, 1984).
At the beginning of treatment, 46% of the clients who became
planned discharges had been drug free, compared to 29% of those who
TABLE 2.
Treatment Duration According to Termination, Drop Out and Referral
Planned Referral!
discharge Dropout arrest
Mean number of therapy 39 17 23

sessions
Mean treatment duration 8.7 2.9 4.4
(months)
Mean number of relapses dur- 2.8 1.7 3.3
ing treatment
dropped out. This shows that abstinence favors planned termination.

However, there are some problems with this interpretation because the
data do not yet differentiate between clients who were detoxified for
only a week, those who had been abstinent for up to 5 months, and
those who had come directly from hospital, jail, or inpatient treatment
fadli ties.
Employment at the beginning of therapy shows some positive effect
on the completion of treatment; 53% of the already employed clients
completed treatment, compared with 24% of the unemployed clients.
The percentage of clients with previous treatment experience who com-
pleted treatment is as high (40%) as the percentage of those who never
attended a previous treatment program and were planned discharges.
This shows that previous therapy experiences do not have an important
influence on the type of termination.
The clients who failed to complete therapy had more frequently
failed to finish educational and vocational training than those who regu-
larly completed the program. In addition, no sex differences were found
with respect to completion of the program.
Pre-Post Comparisons among Planned Discharges

Fourteen of the 46 clients (30%) were planned discharges. For this
subsample, some preliminary data concerning status on enrollment in
therapy and the situation at the time of completion are available. During
the 6 months before entering therapy one client was abstinent, nine
were injecting heroin and opiates, and four took opiates and narcotics
orally. Upon completing therapy, every client had shown drug-free be-
havior for the previous 3 months. This was checked by urine tests.
With the exception of two clients who were still unemployed, all
were regularly employed. The three clients who had found a job during
the first three months of treatment had all been required to enter thera-
py by court order. The housing situation of two clients changed during
the course of therapy. Both had left their family homes and were living
in a close relationship with another person. No client finishing therapy
was living with a drug-using partner or in a group living situation in
which one or more members were using drugs.
DISCUSSION
There are some methodological issues that need to be discussed

before evaluating the results. Because of the lack of a randomized con-
trol group, it is only possible to compare the data with those from
matched groups of clients enrolled in other programs. These com-

parisons have not yet been made because some clients from the total
sample (not included in this report) are still in treatment. Another meth-
odological problem is related to the retrospective nature of measure-
ments on those clients who dropped out. Lack of data regarding the
status of clients at the time of dropout makes a pre-post comparison for
the whole sample very difficult. Therefore, this information must be
assessed in retrospect at the first follow-up.
If we bear in mind that follow-up data are not yet analysed, the
program seems to constitute a lower-cost, additional treatment resource
for opiate dependents. The results from the 30% planned discharges,
based on strict termination criteria, such as a drug-free status, achieve-
ment of social integration, and the solution of individual problems, are
encouraging if compared with data provided by German residential
treatment programs (Klett et al., 1984) and American outcome studies
(Craddock et al., 1982). In addition, 13% of clients were referred on to
other treatment services, mainly residential programs.
Despite all the apprehensions and criticisms surrounding the outpa-
tient treatment approach at the beginning of the research project, there
was no shortage of clients applying for the program. The attraction of
the program probably relates to a less severe intrusion in the lives of
clients who remain part of the social network, to the opportunity to
improve integration during the therapeutic process, and to the chance to
undergo detoxification as an initial part of treatment. The last point is in
contrast to conventional treatment programs, where clients have to be
detoxified before intake.
Detoxification proved to be a critical issue. A high percentage of
clients failed to become detoxified or relapsed into a continuous pattern
of drug abuse. The therapeutic program focuses on the individual stage
in the change process. Clients who are still consuming drugs have first
to develop a motivation for abstinence and this process is quite time-
consuming. On the other hand, there is a danger arising from the com-
bination of therapy and ongoing drug use. The attention and under-
standing of the therapist can become preoccupied with drug abuse and
interfere with the motivation for change. Another critical point is shortly
after withdrawal. Detoxification is connected with the realization of defi-
cits and problems that seem to be overwhelming to the client and are
often cues for relapse. Therapy has to be more clearly structured in
terms of when the client should move to the active stage of detoxification.
It seems that the emphasis on the client's self-responsibility in this phase
has to be more clearly connected with external pressure from the treat-
ment program. Another factor that interferes with moving into the absti-
nence stage is the availability of legally prescribed drugs. Many physi-
cians and pharmacists prescribe addictive substances very easily, which

raises temptation and affects the motivation of dependents.
A further crucial issue in the treatment of drug abuse and addiction
is the lack of motivation shown by drug dependents for entering thera-
py. What is the impact of a court order on treatment? There is a great
deal of controversy about the way in which compulsory treatment can
be used to raise motivation for changing to a drug-free life. The present
data show a positive effect of court order in terms of treatment duration
and planned discharges. This external pressure seems to prevent an
early dropout and improves the probability that intrinsic motivation for
abstinence will be developed. This raises the issue of further improving
interventions with regard to self-motivation.
The individually orientated treatment approach fits the wide range
of differences among the clients enrolled in this program. At this point in
the study, it is difficult to outline indicators for outpatient, drug-free
treatment. Single factors like social integration, duration of dependence,
abstinence at the beginning of therapy, and previous treatment experi-
ences do not have the expected prognostic value. It is likely that more
complex factors have to be considered in determining indicators. These
include hypotheses concerning the influence of the agreement between
the client's expectations and the therapy goals, and the connection be-
tween initial reasons for enrolling in therapy and life events. Further
research and the analysis of the follow-up data will test these hypotheses.
REFERENCES
Biihringer, G. & De long, R. (1980). Manual fUr die stationiire Behandlung (Entwohnung)
von Drogenabhiingigen. In S. Kraemer & R. De long (Eds.), Therapiemanual fur ein
verhaltenstherapeutisches Stufenprogramm zur stationiiren Behandlung von Drogenabhiingi-
gen. (pp. 97-208). Miinchen: Gerhard Rottger Verlag.
Biihringer, G., De long, R., Kaliner, B., Kraemer, S., Ferstl, R & Feldhege, F. I. (1978).
Beschreibung eines stationiiren verhaltenstherapeutischen Programms zur Be-
handlung jugendlicher Drogenabhiingiger. In R De long & G. Biihringer (Eds.), Ein
verhaltenstherapeutisches Stufenprogramm zlIr stationiiren Behandlung von Drogenabhiingi-
gen. (pp. 9-104). Miinchen: Gerhard Rottger Verlag.
Craddock, S. G., Hubbard, R L., Bray, R. M., Cavanaugh, E. R. & Rachal, I. V. (1984).
Client characteristics, behaviors and intreatment outcomes: 1980 TOPS admission cohort. NC:
Research Triangle Institute.
Cummings, c., Gordon, J. R. & Marlatt, G. A. (1980). Relapse: Prevention and prediction. In
W. R MiIler (Ed.), The addictive behaviors. (pp. 291-321). New York: Pergamon Press.
De long, R, & Henrich, G. (1978). Ergebnisse eines stationiiren Programms zur Be-
handlung jugendlicher Drogenabhiingiger: Katamnesen nach einem bzw. zwei Jah-
reno In R De long & G. Biihringer (Eds.), Ein verhaltenstherapeutisches Stufenprogramm
zur stationiiren Behandlung von Drogenabhiingigen. (pp. 281-310). Miinchen: Gerhard
Rottger Verlag.
De long, R., & Henrich, G. (1980). Follow-up results of a behavior modification program
for juvenile drug addicts. Addictive Behaviors, 5, 49-57.
Deutsche Gesellschaft fUr Suchtforschung und Suchttherapie. (1985). Standards fiir die
Durchfiihrung von Katamnesen bei Abhiingigen. Freiburg: Lambertus-Verlag.
DiClemente, C. c., & Prochaska, l. O. (1985). Processes and stages of change: Coping and
competence in smoking behavior change. In S. Shiffman & T. A. Wills (Eds.), Coping
al1d substal1ce abuse. New York: Academic Press.
Feldhege, F. J., Krauthan, G., Schneider, R., Schulze, B., & Vollmer, H. (1977). Ein
ambulantes Breitbandprogramm zur Behandlung jugendlicher Drogenabhiingiger.
Wiener Zeitschrift fiir Suchtforschung, 2, 15-32; 3, 3-14.
Gotestam, K. G., & Melin, L. (1980). A behavioral approach to drug abuse. Drug and
Alcohol Dependence, 5, 5-25.
Kanfer, F. H., & Grimm, L. (1980). Managing clinical changes. A process model of thera-
py. Behavior ModificatiMl, 4, 419-444.
Klett, F., Hanel, E., & Biihringer, G. (1984). Sekundaranalyse deutschsprachiger Katam-
nesen bei Drogenabhangigen. Suchtgefahren, 30, 245-265.
Vollmer, H., & Kraemer, S. (1982). Ambulante Behal1dlung junger Alkoholabhiingiger.
Miinchen: Gerhard Rottger Verlag.
Vollmer, H., Kraemer, S., Schneider, R., Feldhege, F. J., Schulze, B. & Krauthan, G.
(1982). Outpatient behavior therapy for juveniles and young adults with alcohol prob-
lems. In P. Golding (Ed.), Alcoholism: A modem perspective. (pp. 417-434). Lancaster:
MTP Press.
13
Smoking Cessation Strategies
MARTIN RAW
This chapter offers a brief glance at smoking cessation strategies with the
intention of being of practical use to workers in the field now. It is thus
neither academic in tone nor exhaustive and inevitably reflects my own
assessment of the current state of the art. It is my view, for example, that
although we could continue trying to improve techniques for use in
intensive treatment programs, we would be using our resources more
responsibly if we concentrated on developing and disseminating what
we already know. Leventhal and Cleary (1980) suggested more than 5
years ago that refinements of current approaches were unlikely signifi-
cantly to increase success rates and that future work should consider
theories of nicotine dependence. I agree with this and feel that, in the
intervening 5 years, important and useful work has been published that
has put potentially useful tools in the hands of workers.
One of the important tools, which arises directly from considering
the role of nicotine in smoking, is nicotine chewing gum, so that this
chapter will devote considerable attention to it. Another area that has
seen considerable activity is that of the role of health professionals,
especially family practitioners. Since the influential work of Russell and
his colleagues was published (Russell, Wilson, Taylor, & Baker, 1979) on
the potential of CPs as smoking cessation advisors, several studies have
been published developing the theme. As far as intensive treatment
methods go, for use in smokers' clinics or cessation groups, the most
MARTIN RAW • Psychology Department, St. George's Hospital Medical School, London
SW17 ORE, England.
279
280 MARTIN RAW
significant advance has probably been the addition of nicotine gum to

behavioral strategies.
Finally, the least studied smokers of all, those who stop without
formal help, are now beginning to be studied. Although we know rela-
tively little about them yet, they are of obvious interest. Are they able to
stop because they use better strategies than others, or are they different
kinds of smokers (less dependent, for example) whose efforts have few
implications for those unable to stop so easily? The answers are not yet
clear, though Marlatt and Gordon (1985) have already shown something
important and potentially useful: those who do not stop smoking imme-
diately on trying, do not all give in. There is a lot of change between 4
months and one year. This has implications for those helpers, like GPs,
who may be in contact with them.
A selective look at smoking cessation thus follows, preceded by a
simple model of cessation to set cessation in its place. Apologies are
offered to those whose work or approach has been left out. No slight is
implied at all, merely the desire to produce something immediately
useful for the practicing worker needing to counsel smokers about
cessation.
THE CHANGE PROCESS
This conceptual framework represented by the figures that follow

has been drawn loosely from research and from clinical experience.
Although the language is different from that of Prochaska and DiCle-
mente (see Chapter 1), it fits their model quite well, seeing smoking
cessation as a process that goes through various stages. Figure 1
suggests a precontemplation phase (consonant smoking), contempla-
tion (Phases 2 to 5), action (4 to 6) and maintenance (cessation to
eventual success).
A current theme in the smoking cessation field is the need for an
integrated approach. This would recognize the distinct contribution of
different approaches and stress that as many of them should be pursued
as possible in order to reduce smoking prevalence (Kunze & Wood,
1984). It also recognizes (implicitly if not explicitly) a model of smoking
and of the smoker based on at least two underlying principles: that there
is an intimate and dynamic relationship between smoking attitudes and
behavior; and that stopping smoking is a process. The order of events is
not necessarily that shown in Figure 1 and the process can take a long or
a short time.
An example of the dynamic interaction between attitudes and be-
havior is that providing help for someone in stopping smoking may
SMOKING CESSATION STRATEGIES 281
PHASE PROMOTED BY
1. CONSONANT SMOKING J.
Healthy, generally younger smokers Health education of all
~
:::j
kinds
2. DISSONANT SMOKING c
Early attitude change: "It might be a good idea 0
m
to stop." ()
3. DISSONANT SMOKING J:
»z
Continuing attitude change: "I ought to stop."
4. FIRST DECISION
(j)
m ...... -- ..... -- .. -_ ............- .. -_ .. _- ........... -
"I will try to stop." Continuing education
5. SECOND DECISION t Advice of respected
"I will stop." III figure (e.g., GP)
m
6. ACTUAL ATTEMPT TO STOP J: Expectation of success
Translating resolve into action ~ ---- ..... -- ........ - ............... _-_ .............. - ....
7. CESSATION 5JJ Availability of neces-
Eventual success (implying maintenance) () sary resources (e.g.,
probably after several attempts J: willpower, informal
»z
(j)
support, and, occa-
m sionally, formal sup-
port)
1
FIGURE 1. Stopping smoking as a process.
increase their motivation to stop. It is possible for smokers to believe

smoking is dangerous and yet not even try to stop because they believe
they have no chance in succeeding. In that case, it would be no good
merely flooding them with more leaflets, advice, and persuasion. They
may need practical help to reduce their dependence so that they are able
to stop. The availability of help rather than health education would
affect their motivation to stop.
The second underlying principle is also shown in Figure 1. In the
first phase smokers are basically healthy, young (likely to be under 30),
and have no worrying symptoms related to their smoking. Phases 2 to 4
represent the first part of attitude change and are likely to be brought
about by information, education, and, possibly, increasing symptoms. It
should be noted however that the decision to try to stop is not the same
as the decision to stop. We know from survey data that up to 75% of
smokers feel they ought to stop and/or have already tried to do so.
Deciding to stop is a qualitative step further and we know relatively little
about what enables smokers to succeed, except of course for those that
seek help. If health education in the broadest sense has pushed smokers
to the point of wanting to try to stop smoking, then other factors, like
the availability of support, seem likely to convert that intention into real
282 MARTIN RAW
commitment to stopping and to behavior change. Of course, the support

might come from a variety of sources: family, friends, colleagues, health
professionals and even, in some cases, cessation groups.
Figure 1 suggests that the first part of the process of stopping smok-
ing is, broadly speaking, attitude change and the second part behavior
change, and that, on the whole, health education is most relevant to the
first part. There is clearly some confusion about how the second part can
be promoted. Conceptually, translating attitude change into behavior
change depends on the availability of appropriate resources. These
might be internal, like "will-power," or external, like support from oth-
ers. Obviously these are not mutually exclusive. Some commentators
have implied that because many millions of smokers have stopped
smoking "without help" there is no real need for the provision of help.
This is illogical. The existence of smokers who do not need help in
stopping does not deny the existence of those who do. Nor does the fact
that those who need help are in a (large) minority mean they are not
worth helping. In fact, the smokers most in need of help in stopping are
the heaviest smokers and thus the most at risk. Helping them stop may
make a real contribution to the reduction and cost of smoking related
disease (VICC, 1969).
The process of stopping smoking could, in principle, happen almost
overnight or it could take years. It is complex and in each individual it
will be affected by different factors, to different degrees, at different
times. The overall aim of smoking cessation programs is to enhance the
process as strongly as possible at as many points as possible.
MOTIVATION FOR CHANGE AND DEPENDENCE
Figure 2, derived from Russell's (1977) work, shows where the dif-
ferent approaches to smoking cessation might contribute. It charac-
terises smokers according to two major dimensions-their motivation to
stop and their dependence. Although for descriptive purposes they are
presented in the figure as independent dimensions, it seems unlikely in
practice that they are completely independent. Extreme dependence
(real or perceived), for example, might undermine determination to stop
smoking. Those who have run smoking cessation groups will recognise
smokers who claim sincerely to want to stop but, when asked if they feel
they will, answer evasively with "I hope so" or "I'll certainly try," not
because they do not really want to stop but because they doubt their
ability to do so. Again broadly speaking, it is the function of education to
motivate people to stop smoking (to move them from the bottom to the
MOTIVATION
TO STOP
high
(1) (2)
x
X
STOPS?
DEPENDENCE
low high
I
X
~I X
(4) low ( 3)
FIGURE 2. Motivation for change and dependence.
top half of the figure) and the function of support to reduce their depen-
dence (right to left).
In terms of Figure 2, smokers in the top-right quadrant (2) are disso-
nant-they want to stop smoking but are too dependent to achieve this
goal without help. Smokers in the bottom half (3 and 4) do not want to
stop smoking, so the first approach needed with them is educative and
persuasive. Those in the bottom-left quadrant (4), the lighter social
smokers, should respond to health education by stopping, as several
millions have in the United Kingdom over the last 5 to 10 years. Those in
the bottom-right quadrant (3) should find, once they respond to health
education, that they need further help to achieve abstinence. This help
should ideally be offered at various levels according to their needs. In
theory, there should not be any smokers in the top-left quadrant (1), or
they should be in the process of giving up. An integrated approach to
smoking cessation could be developed from this conceptual framework.
Perhaps it should also be noted that this conceptual framework
makes no assumptions about the causes of tobacco dependence. These
are complex and include psychological, social, economic, and phar-
macological factors. The relative importance of these factors in maintain-
284 MARTIN RAW
ing smoking and in predicting how best to help people stop is not yet
understood perfectly. There is certainly strong evidence that nicotine is a
powerful factor affecting the way people smoke and that it is addictive
(Russell, 1976; Russell & Feyerabend, 1980). However, another reason
why it is difficult to remain abstinent after stopping is because smoking is
still so widespread, as are the pressures to smoke, including tobacco
advertising. The steadily increasing social acceptability of not smoking
must be making it easier for many smokers to stop, even those so
"dependent" that they need professional help. Perhaps it should be
emphasized that, just as health education encompasses a wide range of
activities, not all conducted by health education professionals, so also is
support a very broad concept. There are many ways of offering support to
people who want to stop smoking, some of which are now mentioned.
THE MERITS OF SIMPLE INTERVENTION: DOCTORS' ADVICE
Russell et al. (1979) showed that, in response to simple advice to

stop smoking, with a warning of follow-up, 5% of smokers stopped for
at least a year (0.3% in the nonintervention controls). This work was
followed up by a more recent study in which the offer of nicotine chew-
ing gum was added to the stop-smoking advice. The success rates in this
more recent study (using a slightly different but still fairly stringent
criterion of outcome) were 4% in the nonintervention controls, 4% in the
advice group, and 9% in the advice plus nicotine gum group (Russell,
Merriman, Stapleton, & Taylor, 1983). Promising results have also been
shown in general practice by Jamrozik and his colleagues, without Gam-
rozik et al., 1984) and with Gamrozik, Fowler, Vessey, & Wald, 1984)
nicotine gum.
Family practitioners are important not just because of the absti-
nence rates they can promote but also because oftheir access to smok-
ers. Within a year, 75% of the population will visit their GP (90% within
5 years), so that even modest improvements in effectiveness could yield
great benefits in numbers of ex-smokers. Fowler (1983) has suggested
that the GP can offer advice when requested; seek the opportunity to
offer advice in any consultation; advise on how to stop; supplement
advice with appropriate literature; follow-up attempts to stop; and offer
nicotine gum to those who need it. The advice offered should include
reference to presenting medical problems when possible; information
about the health hazards of smoking; emphasis on the benefits of stop-
ping; a reminder that there is no magic cure; a plan to include a target
date for stopping; ways to prepare for stopping, ways to cope with
difficulty after stopping; a warning of the dangers of relapse; and an

explanation of the need for follow-up.
Catford and Nutbeam (1984) have shown that nicotine gum is al-
ready the cessation aid GPs are most likely to offer smokers in the
United Kingdom. The time seems ripe therefore to intensify our efforts
to make a successful treatment aid-nicotine gum (Raw, 1985)-more
widely and effectively used by those who in most countries have access
to it-doctors. And if doctors are important in this regard, other health
professionals may also have important contributions to make to smok-
ing cessation, especially nurses as educators/advisors (Llewelyn & Field-
ing, 1983) and clinical psychologists, probably as helpers (Jerrom &
Simpson, 1983). The latter have been particularly heavily involved in the
development of nicotine gum as a cessation aid in smokers' clinics, and
may prove important in briding the gap between what has been
achieved in this specialist setting and in general practice.
MUTUAL AID IN THE CHANGE PROCESS: CESSATION GROUPS
Smoking cessation groups are, essentially, mutual aid groups (what

used to be called "self-help" groups) (See Robinson, Chapter 13, this
volume). Some of them function without any professional leader,
though most are led by trained professionals, some as part of communi-
ty programs like the North Karelia project (Puska, Koskela, & Bjorkvist,
1979) and some in formal smokers' clinics. Although there is some de-
bate about the role of such groups, they can be useful, not just in
helping people stop smoking but for training others in cessation coun-
seling skills and as a back-up resource for community programs. A
detailed description of United Kingdom smokers' clinics (recruitment,
methods, outcome, history, and role) has been presented elsewhere
(Raw & Heller, 1984) and is beyond the scope of this chapter. A few
comments are in order, however, on the evidence for the effectiveness
of nicotine chewing gum as an aid to smoking cessation.
In specialized smokers' clinics extremely encouraging results have
been achieved with nicotine gum, with one year abstinence rates as high
as 47%, 49%, and 50% in some studies and clear evidence of a specific
effect over and above that of placebo. In these studies the gum was
given with support and encouragement (mostly in groups), expert guid-
ance, and careful monitoring of progress in courses lasting from about 3
to 13 weeks. When nicotine gum is given with minimal or with no
support, as an adjunct to advice, then overall abstinence rates go down.
This is not surprising: They go down as a result of the decreased inten-
sity of the intervention. The effect has nothing to do with the gum itself.
286 MARTIN RAW
Used by doctors as an adjunct to cessation advice, the gum has been

shown to be effective, but the relative contribution of specific and
placebo factors to this have not yet been elucidated. This evidence is
summarised in Raw (1985).
TOWARD AN INTEGRATED CESSATION STRATEGY
Smoking prevalence is falling steadily in many western indus-

trialized countries and, although this trend conceals a variety of pro-
cesses, we know that people are giving up smoking. We also know that
most of these are doing so without the help of formal, professionally run
cessation programs. Clearly then, health education, understood in very
broad terms, is succeeding in persuading people to stop smoking-a
remarkable achievement considering the resources ranged against
health interests. However, we do not know how far the trend will con-
tinue, whether it will accelerate or slow, and what progress we will
make in dissuading children from starting smoking. It would be foolish
to become complacent. In the United States and United Kingdom there
are still some 60 to 70 million smokers, contributing to tomorrow's per-
sonal, social, and health care costs and serving as models for today's
new generation of smokers. We would be wise to continue pursuing as
broad a cessation strategy as possible, advising and helping people to
stop smoking alongside our efforts at primary prevention, and we
would be wise also to continue to provoke social, political, and economic
change.
Smokers' clinics have played a crucial role in developing nicotine
chewing gum as a cessation aid. This research role may continue to
prove valuable in a small number of them but, as an integrated part of a
local community's overall smoking prevention program, they may be
more valuable as a training and back-up resource for health profes-
sionals. And these, especially GPs and community nurses but also hos-
pital doctors and nurses, dental practitioners, psychologists, and many
others, have a crucial role to play. Their daily work provides many
opportunities to counsel and support and this work could be much
improved by what we already know.
REFERENCES
Catford, J. c., & Nutbeam, D. (1984). Prevention in practice: What Wessex general practi-
tioners are doing. British Medical Journal, 288, 832-834.
Fowler, G. (1983). Smoking. In M. Gray & G. Fowler (Eds.) Preventive medicine in general
practice (pp. 133-148). Oxford: Oxford University Press.
Jamrozik, K., Fowler, G., Vessey, M., & Wald, N. (1984). Placebo controlled trial of
nicotine and chewing gum in general practice. British Medical Journal, 289, 794-797.
Jamrozik, K., Vessey, M., Fowler, G., Wald, N., Parker, G., & van Vunakis, H. (1984).
Controlled trial of three different antismoking interventions in general practice. British
Medical Journal, 288, 1499-1503.
Jerrom, D. W. A., & Simpson, R. J. (1983). Overlapping General Practice: Clinical psychol-
ogist. British Medical Journal, 287, 1185-1186.
Kunze, M., & Wood, M. (1984). Guidelines on smoking cessation (UICC Technical Report
Series, vol. 78). Geneva: Union Internationale Contre Le Cancer.
Leventhal, H., & Cleary, P. D. (1980). The smoking problem: A review of the research and
theory in behavioral risk modification. Psychological Bul/etin, 88, 370-405.
Llewelyn, S., & Fielding, G. (1985, April 27). A job for all nurses. Nursing Mirror, pp. 36-
37.
Marlatt, G. A., & Gordon, J. R. (Eds.) (1985). Relapse prevention. New York: Guilford Press.
Puska, P., Bjorkvist, S., & Koskela, K. (1979). Nicotine-containing chewing gum in smok-
ing cessation: A double blind trial with half-year follow-up. Addictive Behaviors, 4, 142-
146.
Raw, M. (1985). Does nicotine chewing gum work? British Medical Journal, 290, 1231-1232.
Raw, M., & Heller, J. (1984). Helping people stop smoking. The development, role and potential of
support services in the UK. London: Health Education Council.
Russell, M. A. H. (1976). Tobacco smoking and nicotine dependence. In R. J. Gibbins, Y.
Israel, H. Kalant, R. E. Popham, W. Schmidt, & R. G. Smart (Eds.) Research advances in
alcohol and drug problems (Vol. 3, pp. 1-47). New York: Wiley.
Russell, M. A. H. (1977). Smoking problems: An overview. In M. E. Jarvik, J. W. Cullen, E.
R. Gritz, T. M. Vogt, & L. J. West (Eds.) Research on smoking behaviour (pp. 13-33).
NIDA Research Monograph 17 Rockville, USDHEW.
Russell, M. A. H., & Feyerabend, C. (1980). Smoking as a dependence disorder. In L. M.
Ramstrom (Ed.) The smoking epidemic, a matter of worldwide concern (pp. 74-80). Stock-
holm: Almqvist & Wiksell.
RusseIJ, M. A. H., Merriman, R., Stapleton, J., & Taylor, W. (1983). Effect of nicotine
chewing gum as an adjunct to general practitioners' advice against smoking. British
Medical Journal, 287, 1782-1785.
Russell, M. A. H., Wilson, c., Taylor, c., & Baker, C. D. (1979). Effect of general practi-
tioners' advice against smoking. British Medical Journal, 2, 231-235.
Union Internationale Contre Le Cancer (1969). Influencing smoking behaviour (UICC Tech-
nical Report Series, vol. 3). Geneva: Author.
14
Mutual Aid in the Change Process
DAVID ROBINSON
INTRODUCTION
Jim Prochaska, in his opening chapter to this book, outlines a route

"towards a comprehensive model of change in the addictive behaviors."
On the way he rightly distinguishes between "change within therapy"
and "individual self-change," points out that improvement in therapy
and improvement outside of therapy appear to involve the same stages
and processes, and concludes that by taking seriously the successful
efforts that individuals make without therapy a "transtheoretical model
of change" can be enhanced.
In this chapter, I discuss neither change within therapy nor indi-
vidual self-change but focus instead on mutual aid-or mutual self-
change. This may constitute a useful footnote to Prochaska and DiCle-
mente's comprehensive model of change.
Although this volume is concerned with addictive behaviors, this
chapter draws on material from a broader range of health concerns in
order to identify the specific processes at work in mutual-aid projects or
in the activities of some particular mutual-aid groups. Much of the liter-
ature in the mutual-aid field is focussed on specific groups set up to
address specific health problems. Much of this literature is written as
I am grateful to the Medical Council on Alcoholism for permission to reproduce in one

section of this chapter material from D. Robinson (1983) The Growth of Alcoholics Anony-
mous, Alcohol and Alcoholism, 18(2), 167-172.
DAVID ROBINSON • Addiction Research Centre, University of Hull, Hull HU6 7RX,
England.
289
290 DAVID ROBINSON
though the processes involved in a specific mutual-aid group are

unique. It is clear, however, from work over the past decade that issues
of identification, coping through activity, de stigmatization, and a vari-
ety of kinds of sharing are common to a wide range of self-help and
mutual-aid enterprises.
Many professionals, who have not had the opportunity of working
closely with mutual-aid groups, imagine that they are "out there" wait-
ing to be used or that they, the professionals, can start, develop, or
"facilitate" mutual-aid groups as part of their professional activity.
Many professionals also believe, or by their actions imply, that all mutu-
al-aid groups are the same in the sense of seeking standardized rela-
tionships with professionals. Part of the purpose of this chapter is to
indicate, in addition to the similarities across mutual-aid groups, the
differences between mutual-aid groups, which professionals must take
into account when seeking to work with them. The two case examples of
Alcoholics Anonymous and the Clubs for Hypertensives illustrate sever-
al of these issues.
THE GROWTH OF MUTUAL-AID GROUPS
Medicine, as practiced in developed countries, is seen by many

people as a threat to health, not merely in the technical sense of malprac-
tive, clinical iatrogenesis, and inappropriate treatment, but in the wider
sense of diverting attention from the social-structural and environmen-
tal causes of ill-health. Not surprisingly, there has been a growing hostil-
ity towards any health care system that undermines the power of indi-
viduals to care for themselves or shape their own environments. Nor is
it surprising that there has been a rapid and substantial growth of mutu-
al-aid groups, often called self-help groups, which now represent a sig-
nificant feature of contemporary life.
A good deal of attention has certainly been given to self-help and
mutual aid by professionals, governments, interested lay people, and
the media. There is now hardly any wide circulation magazine or profes-
sional journal that has not carried an article on some mutual-aid group.
In addition, several World Health Organisation publications have set
out the philosophy, organization, and achievements of many alternative
approaches to meeting basic health needs (WHO/UNICEF, 1978;
Djukanovic & Mach, 1975; Newell, 1975). An excellent collection of arti-
cles was published by WHO EURO in 1983 under the title "Self-help and
Health in Europe" (Hatch & Kickbusch, 1983) whereas from across the
Atlantic, again in 1983, an equally excellent collection of papers re-
MUTUAL AID IN THE CHANGE PROCESS 291
minded us with its title, "Rediscovering Self-Help" (Pancoast, Parker, &

Froland, 1983) that mutual aid is, of course, as old as human history.
WHY THE INTEREST IN MUTUAL-AID GROUPS NOW?
Most answers to the question, Why the increasing attention to mutu-

al aid now? identify it as a reaction to some inadequacy, need, problem, or
changing situation. Some have argued that the growth of mutual aid is a
response to the decline of existing institutions and the need to "fill the
gap." Mowrer (1971), for example, noted the decreasing importance of
the established church and sees mutual-aid groups as "the emerging
church of the 21st century." Others see the decline in the extended family
system and close-knit communities as bringing about a need for new
ways of providing and sustaining emotional and social support. In-
creasingly, the argument is heard that the emergence of mutual aid is a
response to the disillusionment with, and unfulfilled promise of, the
helping professions (Back & Taylor, 1976) and, in the United Kingdom,
the welfare state. Some, for example, assert that mutual-aid groups and
other "consumer-initiated services" arise when "a hiatus exists between
felt need and the existence of available services ... adequate to meet
such a need" (Gillie, Price, & Robinson, 1982). Others suggest, somewhat
more sceptically, that the "gaps" have arisen because of the inevitable
fallability of medical science, with many groups growing up around those
people who feel abandoned by the clinical services either because they
represent its failings or because they have socially unacceptable problems
(Zola, 1975).
In addition to the alleged failure of traditional institutions, there
have been those who locate the growth of mutual aid in relation to
changing philosophical and social ideas: for instance, coincident with a
changing social conscience toward disablements following World War II;
as part of an alternative culture, reflected in a broader decentralization
and debureaucratization of public life; or as part of a "power-to-the
people" movement, itself a product of the cultural shifts of the 1960s.
For Katz and Bender (1976) in their seminal book The Strength in Us a
mixture of these and other social forces has combined to make mutual
aid "the most important social phenomenon in recent years." They say
that industrialization, a money economy, the growth of vast structures
of business, industry, and government have led to the depersonaliza-
tion and dehumanization of institutions and social life; feelings of aliena-
tion and powerlessness; the loss of choices and a loss of identity. Mutual
aid is one of a number of social movements that, according to Katz and
Bender, have arisen to counter this trend.
292 DAVID ROBINSON
Although these global explanations may make sense, another layer

of understanding of the question Why mutual aid now? may be got from
looking at the emergence of particular groups. It is clear from an inspec-
tion of the literature produced by the groups themselves that a whole
range of people and agencies were in some way involved with setting
them up. There were the people who shared the problem, their inti-
mates, various categories of professional and voluntary helpers, govern-
ment departments, local authorities, community and voluntary agen-
cies, and the media. Five major themes in the groups' accounts of their
own origins tend to recur. These are the identification of a shared prob-
lem; the failure of some helping agency; the recognition of the impor-
tance of contact between those who share a common problem; innova-
tions of some sort in the handling of the shared problem; and, finally,
the role of the media in bringing to light the extent of the shared prob-
lem, or some innovative attempt to solve it.
Most mutual-aid groups mention the media as playing a key role in
some way. Depressives Associated, for example, started as a result of
thousands of letters sent spontaneously in response to Nemone Leth-
bridge'S television play Baby Blues, which dealt with postnatal depres-
sion. Open Door started in 1965 after a woman with agoraphobia in
Macclesfield, a small town just south of Manchester, placed a small
advertisement in her local newspaper, received a number of replies from
other agoraphobics and within a year found herself at the heart of a
rapidly expanding national organization.
Some groups, such as the Ileostomy Association, the Society for
Skin Camouflage and the Possum (lung machine) Users Association tie
their origins or development to some professional innovation, whereas a
small number of others say that they were set up to enable some profes-
sional to do what otherwise would have been impossible. Recovery Inc.
is interesting and unusual among well established mutual-aid groups in
its readiness to acknowledge the role of an outsider, and a professional
at that, in its foundation (Antze, 1976). But with the histories of mutual-
aid groups we must remember that we are dealing with just that, histo-
ries; and the purpose of history is, of course, to produce statements
about the past that can be used in the present. And for most groups it
may be more in line with their beliefs about the nature and purpose of
mutual self-help to drop references to the involvement of professionals
and outsiders from their accounts of time past.
THE RANGE OF GROUPS AND THEIR CORE CHARACTERISTICS
Some groups, such as Alcoholics Anonymous, are well known,

long established, and well researched (Bean, 1975; Robinson, 1979). But
there are thousands of newer, some less well known, groups as well-
for schizophrenics, for people with skin diseases, for phobics, for
smokers, the anxious, the depressed, gamblers, people with hyperten-
sion, people with cancer, child batterers, widows, parents of handi-
capped children, people who eat too much and those who refuse to eat
at all, and many more besides, including the delightful-but as yet
unconfirmed-Analysands Anonymous: "open to anyone who has
been in analysis for twelve years or longer and needs the help of a power
greater than their own-or that of their analyst-to terminate the analy-
sis" (Hurvitz, 1970).
Just as the groups themselves have multiplied, so has the number of
directories that attempt to draw together the vast amount of rapidly
changing information about which mutual-aid groups exist, where,
what they do, with whom, and why (Darnborough & Kinrade, 1977;
Knight, 1970; Moorhead, 1975; National Council for Voluntary Organi-
sations, 1982; Patients Association, 1982; Robinson & Robinson, 1979;
Share Community, 1980; Thames Television, 1978; Todd, 1982).
Not surprisingly, given the wide variety of mutual-aid activities,
there have been almost as many attempts to define mutual aid as there
are groups. In 1976 two major international journals, the Journal of Ap-
plied Behavioural Science and the Journal of Social Policy published special
issues devoted entirely to mutual self-help. 1976 was also described by
one commentator as "a bumper year for new books on self-help" (Brig-
gs, 1977). More importantly, however, it was the year in which some
attempt was made to draw together the large number of accounts of
particular mutual-aid groups in order to find their common charac-
teristics (Caplan & Killilea, 1976; Katz & Bender, 1976). This has gone on
ever since (Gartner & Riessman, 1977; Hatch & Kickbusch, 1983; Lieber-
man & Borman, 1979; Richardson & Goodman, 1983; Riessman &
Gartner, 1981; Robinson & Henry, 1977).
One of the best reviews is still that written by Killilea (1976) who, as
well as extracting from the literature 20 different "categories of in-
terpretation" of mutual aid, identified seven characteristics of groups
and their processes to which writers had given particular emphasis.
These are as follows:
1. Common experience of members: the care giver has the same dis-
ability as the care receiver;
2. Mutual help and support: the individual is a member of a group
that meets regularly in order to provide mutual aid;
3. The helper principle: in a situation in which people help others
with a common problem it may be the helper who benefits most
from the exchange;
294 DAVID ROBINSON
4. Differential association: the reinforcement of self-concepts of nor-

mality, which hastens the individual's separation from commit-
ment to their previous deviant identities;
5. Collective will power and belief: the tendency of each person to look
to others in the group for validation of their feelings and attitudes;
6. The importance of information: the promotion of greater factual un-
derstanding of the shared problem as opposed to intrapsychic
understanding; and finally-and most importantly-
7. Constructive action towards shared goals: mutual-aid groups are ac-
tion orientated, their philosophy being that members learn by
doing and are changed by doing.
HOW MUTUAL-AID GROUPS WORK
In order to understand what mutual-aid groups do, we need to look

not just at the descriptions and analyses by outsiders, but also, as with
the accounts of their origins, at the descriptions and explanations of the
groups themselves. Four major themes tend to recur: identification,
sharing, coping with practicalities and stigma, and change through
activity.
IOENTIFICATlON
Great stress is always put on the common problem, position, or

circumstance, often expressed colloquially as "being in the same boat."
Being in the same boat means, first of all, understanding the problems of
others; that is, knowing what it is like. It is said that only those experi-
encing the problem can really understand. As the founder of CARE, the
Cancer Aftercare and Rehabilitation Society, put it:
The organisation consists in the main of cancer patients-people who know
what it is like to have cancer, who know the problems, mental and social,
associated with the disease. These people we feel are best fitted to give moral
assistance and help to patients and families before and after treatment.
(Robinson & Henry, 1979, p. 48).
SHARING
It is this understanding based on common experience, say the

groups, that produces the necessary common bond of mutual interest
and common desire to do something about the problem. And the basic
ingredient of this "doing something" is collectively helping oneself. As
SHARE, a mutual-aid group for the disabled, says: "To help others is to
help yourself."
In addition to helping yourself collectively and helping yourself
through helping someone else, great stress is put on the importance of
example in the sharing and copying of experiences, a point that is suc-
cinctly expressed again by the cancer group CARE: "What better thera-
py then seeing someone who has had exactly what you have got and
who is partcipating in normal activities, work and social life."
Being in the same boat, knowing what it is like, sharing experi-
ences, and helping yourself by helping others all add up to the "fel-
lowship" that Hurvitz (1970) takes to be the key feature of mutual-aid
groups: "Within such relationships and in the presence of members
who acknowledge the help they receive through fellowship," he says,
the members "make it possible and desirable to accept each others
efforts to modify their own and others behaviour." In this fellowship lies
the essence of mutual self-help, which Mowrer (1971) sums up as "You
cannot do it alone, but you alone can do it."
In most groups, sharing means the sharing of information and com-
mon experiences. The mechanics of sharing range from formal group
meetings through the no less important informal meetings between
group members, to telephone contact networks, correspondence, news
letters, tape exchanges, or even radio contacts when the members are
geographically dispersed or prevented by their shared problem from
meeting face to face.
COPING WITH PRACTICALITIES AND STIGMA
Paradoxically, the first stage of getting rid of or coping with the

problem is to concentrate on it. For although it is easy for a group to
proclaim, "We are all special together," it is difficult for newcomers to
share that feeling. New members have to be encouraged to accept that
they "are", or "have", whatever is the focus of the group's concern, and
even encouraged to make public declarations to that effect. In Gamblers
Anonymous, Alcoholics Anonymous, Parents Anonymous, Neurotics
Anonymous, and Smokers Anonymous, members introduce their con-
tribution to the meeting by saying, "My name is Joe and I am a com-
pulsive gambler," or whatever. Parents Anonymous says that the easi-
est method of coming to accept the problem is for new members to
declare, "I've got problems as a parent and I want help. My problem
shows itself in the form of ... ," whatever the form of abuse the new
member feels has been shown-verbal, emotional, physical, sexual or
neglect. It then recommends that the person asks other members for
help to overcome these problems. Once the public declaration has been
296 DAVID ROBINSON
made the feeling of relief can be enormous. Many groups say that the
relief of "publicly" sharing the problem is their members' single most
important experience.
Once the problem is settled on, admitted, and brought out into the
open, group members can begin to cope with it by, first of all, sharing
information about practical solutions to specific difficulties. This may
concern physical aids, procedures, diets, or official agencies and
rights-in short anything that makes it more possible to handle the
practicalities of the shared problem. Clearly, the range of specific prac-
tical aids being used in mutual-aid groups is immense.
The most difficult task for many groups is to cope with the stigma of
their shared condition. One way of destigmatizing the problem is by
changing members' self-perception, a feat partly achieved by meeting
others in the same situation and therefore feeling less odd. The National
Council for One Parent Families, consider that their groups have a dou-
ble value to lone parents and their children in providing the mutual
support that is so helpful, and also helping the children to realise that
there are many lone parents and that they, the children, are not in any
way unusual.
In addition, it is common for all groups to direct their destigmatiz-
ing efforts towards changing those who are seen as the cause of the
stigma-the general public, society, or just "all those who do not under-
stand." The Breakthrough Trust, for example, aims to bring deaf and
hearing people into realistic contact with each other and so alleviate
much of the isolation, apathy, and frustration that deafness imposes. By
working together on equal terms, "deaf and hearing people," says the
Trust, "educate each other in the skills of communication and conse-
quentlya deeper understanding is gained"-because, of course, "deaf-
ness is not just a problem for the deaf, it is a hearing person's problem
too."
Coping with stigma, then, involves first of all the realization that
you are not alone: There are others like you and they understand and
appreciate your problems, ideas, and aspirations. But coping with stig-
ma involves mutual-aid groups in much more than this. People as well
as problems have to be dealt with. Members often have to be encour-
aged to relearn, or even learn for the first time, that they have a value, a
contribution to make, and a full place to occupy in the social world.
Outsiders have to be made to understand the members' problems, both
practical and personal, to give care and support, and to appreciate that
having the problem does not invalidate a person's membership in the
human race.
CHANGE THROUGH ACTIVITY
A mutual-aid group is not just a place where people help each other
to cope with the practicalities and stigma of "their problem." Although
these are very important parts of what groups do, of course, they can
provide much more. As well as helping to diminish the importance of
the problem, mutual-aid groups can enable their members to change-
to begin to build up a new way of everyday life through being involved
in a wide range of group activities.
At one level, group activities are geared to helping to solve the
group members' specific problem-be it having cancer or a mentally
handicapped child, being disabled or depressed. But those who feel that
they have really benefitted from being in a mutual-aid group speak of
"getting involved," "making a contribution," "doing things for the
group," and so on. In most groups there is a whole range of activities in
which most members can become involved.
In addition to the usual offices-chairperson, secretary, treasurer-
there will be members who arrange the meeting place, or send out
notices, or handle publicity, or speak on a particular topic, or make the
tea, or put out the chairs, or reply to queries from interested health
workers or members of the general public, or collect contributions. It is
easy to see how almost everybody in a mutual-aid group, whatever their
physical or mental capabilities, can have their own tasks to perform and
their own things to be responsible for.
It is easy to see also how being involved in group activity can help to
rebuild confidences and help members to realize that in spite of their
problem they still have some value, something to offer, a contribution to
make. Many people with severe and long-standing problems feel this for
the very first time in their lives in a mutual-aid group.
Sharing experiences, giving support to each other, and working
together provides an ideal opportunity for new friendships to develop.
On the basis of friendships made in mutual-aid groups, members begin
to build up a network of relationships and activities outside the groups
that are, nevertheless, still based on the support and understanding that
the group provides. The value of the outside social activities is not just to
pass the time or to have fun, although both of these are important. The
real value comes from the fact that people who help each other to handle
their particular problem can help each other in many other ways as well.
Every group, whatever its "problem," is likely to have a range of skills
and expertise at its disposal. Almost everyone can be a resource for the
group or for a small collection of friends in relation to some aspect of
everyday life. And having been involved together in mutual-aid group
298 DAVID ROBINSON
activities, the framework is there for making these other resources avail-
able to the wider community.
Mutual-aid groups, then, are more than huddle-together sessions
for people who feel discriminated against, or overwhelmed by a com-
mon problem or by some aspect of late 20th-century life. Mutual help
offers most to people when it manages to combine reciprocal support for
those who share a common problem with activities and schemes that
encourage personal change and development, and enable people to in-
fluence the quality of their everyday lives.
FROM MUTUAL AID TO HEALTH: CASES IN POINT
A large proportion of the mutual-aid groups in developed coun-

tries, particularly in the United States, Canada, Britain, Scandinavia and
other parts of Western Europe, operate quite independently of the for-
mal health services. In fact, the impetus for the establishment of many
groups has been the lack of adequate understanding, care, treatment, or
support from the various health professions. This immediately raises the
question for those concerned with services of whether these "indepen-
dent" groups can contribute to a coherent and comprehensive primary
health care system in anything other than a purely ad hoc manner.
Evidence from various parts of the world shows that health services
at the local level do accommodate and respond to the activities of even
the most independent of groups when those groups are clearly provid-
ing an important element of primary health care. Alcoholics Anony-
mous, for example, has developed into an international network of tens
of thousands of groups in over a hundred countries that cooperate with
formal health and social services in an attempt to provide comprehen-
sive care for those with drinking problems, and yet A.A. retains com-
plete control over its own philosophy and group activities.
The A.A. mutual-aid process of "talking out of alcoholism" (Robin-

son, 1979) is well understood. Suffice it to say that, at a personal level,
the program aims at transforming the dependent, isolated, drinking
alcoholic into an independent, integrated, sober alcoholic. At an organi-
zationallevel, A.A. aims at being self-reliant, self-sufficient, beholden to
no one and dependent on no one. It remains uninvolved in outside
political or social issues, although it cooperates closely with other bodies
in order to bring as many people as it can to its view of sobriety. A.A.
calls this cooperation "being friendly with our friends" and distinguish
it from "affiliation" -an unacceptable notion.
The question is often raised as to whether Alcoholics Anonymous
can really operate outside the particular sociocultural context in which it
originated. It only requires a glance at the national and international
directories to see that, on a worldwide scale, Alcoholics Anonymous has
groups in catholic and protestant countries, in developed and develop-
ing countries, in beer-producing and wine-producing countries, in
countries with private medical care and in those with state health care
systems (Robinson, 1983). But although A.A. is widespread, its develop-
ment has, naturally, been uneven. It is thin in Africa outside South
Africa and Zimbabwe. It is also thin in Eastern Europe, although there
are the well-known alcoholic clubs in Yugoslavia and elsewhere that
operate on somewhat similar lines to A.A. In the Middle East and India
many of the members are employees of foreign firms, whereas in Asia
many of the groups are started by the United States forces. But in all
areas of the world the number of groups is growing. This has been
particularly the case over the past decade in Central and South America.
In Mexico, for example, there were A.A. members meeting spo-
radically ever since 1941, and a regular English-speaking group was
started in 1946. It was not until 1956, however, that the first Spanish-
speaking group emerged. By 1969 there were 181 groups. Since then the
development of A.A. in Mexico has been very rapid indeed. The 181
groups in 1969 grew to 928 groups in 1974, and to almost 6,000 groups
by the end of 1984.
To give some idea of the spread of A.A., there are now approx-
imately 1,000 groups in Australia, 250 in New Zealand, 30,000 in the
United States, 1,000 in Germany, 115 in Trinidad, 2,000 in Great Britain,
500 in Finland, 1,000 in EI Salvador, 200 in Belgium, 120 in India, 650 in
Ireland, 120 in Iceland, 700 in Guatemala, 2,000 in Brazil, 250 in France,
200 in South Africa, 400 in Nicaragua, 75 in Japan, and 30 in Poland.
Given what we know about how A.A. works and what is required
of members in personal and social terms (Robinson, 1979), it is possible
to identify certain features of the mutual-aid process that may be more
acceptable in some cultures than in others.
Alcoholics Anonymous, like many other mutual-aid groups, is
based, as was pointed out earlier, on a philosophy of independence. The
problem, however it arises, is seen to be the property of individuals and,
as such, is held to be within their own power to overcome, albeit with
the support of fellow sufferers.
The mutual aid process of Alcoholics Anonymous also demands
openness in several crucial ways. First, members have to be open with
each other about their past, their activities, their relationships, and their
300 DAVID ROBINSON
emotions, in order to create the necessary common bond of shared

experience and understanding. Second, AA operates an open mem-
bership policy in which personal and social attributes that normally
distinguish people from each other are played down whereas the one
thing that members share, their alcohol problem, is emphasized. Third,
members of AA have to be open to the possibility of change, because it
is an essential part of mutual aid for members to help each other to
change, to some extent, their self-perception, their network of friends
and relationships, and even the style and content of their everyday life.
Alcoholics Anonymous, then, demands that individuals, with the
support of the group, take responsibility for their condition, their every-
day life and, thus, their destiny. They do this in a mutual-aid process
that requires them to be totally open about themselves with other mem-
bers, who may be of a different sex and very different in terms of
politics, race, socioeconomic status, age, and religion. They also have to
be willing to accept that changes in several aspects of everyday life are
not only desirable but essential. So, clearly, any culture that puts a very
high permium on privacy in emotional or social terms, or in which
people are tightly fixed in complexes of highly differentiated roles and
relationships, will find it less easy than others to accommodate the core
principles and practices of the mutual-aid process of groups like Alco-
holics Anonymous.
In contrast with the most independent groups such as AA., other
mutual-aid groups and organizations are very much the brainchildren of
health workers who remain closely involved with them. This too raises
several important questions. To what extent are these mutual-aid or self-
help groups in the usual meaning of the phrase? What role do the health
workers play in the everyday activities of the group? How able are the
groups to develop structures and procedures that best suit the needs of
their members? What are the implications of being so closely involved
with the formal health services-for the groups, the members, for the
neighborhoods in which they are situated, and for the development of
more satisfactory systems of primary health care?
CLUBS FOR HYPERTENSIVES
The clubs for hypertensives in Zagreb, Yugoslavia, provide a clearl

illustration of both the disadvantages and the advantages of mutual-aid
enterprises in which professional health workers are closely involved
(Hatch & Kickbusch, 1983. pp. 107-117).
The original role of the health workers was to get the clubs started,
to encourage the election of officers, to train the members in self-
monitoring techniques, and to provide gUidance and aid. There is no
doubt that the clubs are now very successful within those strict limits in
which they were established. They have gathered people together who
can now perfectly adequately monitor and control their blood pressure.
But it is also true that, although many new clubs are opening, many of
the older clubs are becoming rather staid and set in their ways; mem-
bership is static, the same people have held office in some clubs ever
since they began, and the same people tend to measure blood pressure
at each meeting. The clubs, although self-governing, are still very much
under the guidance of the associated professional staff.
It was one of the original aims of those who established the clubs
that people should be members for only a limited time-a year or 18
months-after which they would have learned how to control their
blood pressure and established a dietary, exercise, and relaxation regime
which would maintain it at a satisfactory level. But, as the health work-
ers have found out, once people come together and are encouraged to
become a group in order to provide mutual support and encouragement
to each other, they are not just going to fade away as soon as the
problem has been brought under control. Individuals will have become
a group and like it. Fortunately, given the enthusiasm of many club
members and the interest and involvement of various health workers,
there is now the possibility of responding to and building on this
situation.
Both the members and associated health workers have ideas about
how the clubs for hypertensives might develop into more general health
education, care, and maintenance groups. One club has developed pro-
grams for diabetics and those with obesity and heart disorders. The
members of another club have set up teams that do home visits to
provide social care for elderly people and those who are handicapped.
The members of a club that is based in a furniture factory screen their
workmates and are beginning to assume basic health education func-
tions in relation to smoking and other issues. A film about abortion,
which had been shown in the factory, stimulated discussion about
gynaecological problems and was taken by some club members and
shown in their local neighborhood centers.
Once mutual-aid group members have learned the basic skills of
controlling their blood pressure, or whatever their problem is, they can
acquire more and more health skills and techniques in relation to every-
day physical and mental health care. Subgroups of members can devel-
op particular sets of skills and so become resources, not only for other
group members but for their families, neighbors, friends, and work-
mates. In this way some mutual-aid groups, instead of being inward
looking and concerned only with the particular problem that brought
the members together, could, it is claimed, become the ideal settings for
302 DAVID ROBINSON
the education and development of basic health workers. And certainly,

group members, because of their own experience, are well-placed to
understand that there are a great many technical and social skills that
members of the general public can easily learn and that mutual aid and
support is an essential component in the handling of the most modern
health problems.
LIMITS OF MUTUAL AID
Many people believe that mutual-aid groups are paving the way for
a radical change in the way everyday problems are handled, and even
providing a blueprint for the construction of a new political order. But it
does not take long to recognize that, for a variety of reasons, most
mutual-aid groups seem neither inclined, nor likely to be able, to accom-
plish any great social changes.
One of the major limits to mutual aid in the health field is that most
groups tend to operate with the same view of health and illness as
conventional helpers. Problems, however they arise, are seen to be the
responsibility of the individual. The core aim of both conventional help
and most mutual self-help is to do something to, or with, people who
"have" problems, in order that they might be better able to find their
way around the world as it is. Those groups that look beyond the imme-
diate concerns of their members do little more than press for some
adaptation of the current professional or administrative system. They
push for recognition of their problem, or for more humane, accessible,
or competent professional treatment for their problem.
Concentration on individuals and their problems is, of course, an
essential feature of the mutual-aid process. But it means, as well, that
groups rarely focus their attention on any broader political issues. Their
attention is much more likely to be given over to making sure that one is
serviced properly, rather than to raising the question of whether one
needs the service, or of what changes need to be made in order to make
it less likely that the problem that needs servicing will arise at all.
Not only do most groups not look at broad causes of their problems
but they may, by their mutual-aid activities, actually make them worse.
Mutual-aid groups, it could be argued, provide an excuse for govern-
ment authorities to avoid fulfilling their obligations. Suggesting that
people attempt, with inadequate resources, to build up their own com-
munities or provide their own services may divert them from seeking
their full share of the resources of the entire society.
Clearly, everyone involved in mutual-aid groups, however, suc-
cessful they feel they are in alleviating or handling the problems of their
members, should ask themselves the following question: Is what I am

doing likely to increase or decrease the number of people with this
problem? In other words, they must consider the extent to which they
collude with the system that caused, maintained, or accentuated the
problem in the first place. That is the core dilemma for everyone who
gives help, whether they are mutual self-helpers or professionals.
Not surprisingly, some professionals feel very threatened by the
growing number of mutual aid groups. Others, recognizing the value of
particular mutual-aid enterprises, have proposed that professionals
should become directly involved, that professionals should try to set up
groups, and even that universities should train people to do this
(Mowrer, 1971). Others, rather less enthusiastic, recognize that this
could undermine that one value uniquely cherished by the mutual self-
help group-its ability mutually to help itself.
CONCLUSION
Mutual-aid groups are a familiar and recently much discussed fea-

ture of contemporary life, and health workers in many countries have
come to recognize the value of people with particular problems coming
together to help each other to help themselves. Mutual self-help is mis-
perceived, however, if it is seen merely as a temporary expedient or
passing fashion, because mutual aid is, of course, as old as human
history. People have always banded together to solve their common
difficulties and promote their mutual interests in family networks, clans,
tribes, guilds, professions, trade unions, friendly societies, clubs, and
on street corners. Finally, mutual aid is grossly misperceived if it is seen
as a poor second best to "fill the gap" for people who are starved of
"real" services. For it is professional services, of course, that are the
stop-gaps, filling in where basic mutual self-help needs some specific
technical, organizational, or expert assistance.
REFERENCES
Antze, P. (1976). The role of ideologies in peer psychotherapy organizations: Some the-
oretical considerations and three case studies. Journal of Applied Behavioural Science, 12,
323-346.
Back, K. W., & Taylor, R. C. (1976). Self-help groups: Tool or symbol. Journal of Applied
Behavioural Science, 12, 295-309.
Bean, M. (1975). Alcoholics Anonymous. Psychiatric Annals, 5, (2), 7-61.
Briggs, H. (1977). Editorial. Self-Help Reporter, 1, (8), 37-57.
304 DAVID ROBINSON
Caplan, G., & Killilea, M. (1976). (Eds.), Support systems and mutual help: Multidisciplinary
explorations. New York: Grune & Stratton.
Darnborough, A., & Kinrade, D. (1977). Directory for the disabled. Cambridge: Woodhead-
Faulkner.
Djukanovic, c., & Mach, E. P. (1978). (Eds.). Alternative approaches to meeting basic health
needs in developing countries. A Joint UNICEF/WHO Study, Geneva: WHO.
Gartner, A., & Riessman, F. (1977). Self-help in the human services. San Francisco: Jossey-
Bass.
Gillie, 0., Price, A., & Robinson, S. (1982). (Eds.). The Sunday Times self-help directory.
London: Granada.
Hatch,S., & Kickbusch, I. (Eds.). (1983). Self-help and health in Europe, New approaches in
health care. Copenhagen: WHO-EURO.
Hurvitz, N. I. (1970). Peer self-help psychotherapy groups and their implications for
psychotherapy. Psychotherapy: Theory, Research and Practice, 7(1) 44-49.
Katz, A., & Bender, E. I. (1976). The strength in us: Self-help groups in the modern world. New
York: Franklin Watts.
Killilea, M. (1976). Mutual help organisations: interpretations in the literature. In G. Cap-
lan & M. Killilea (Eds.), Support Systems and Mutual Help: Multidisciplinary Explorations.
New York: Grune & Stratton.
Knight, S. (1980). 'Help! I Need Somebody'. London: Henry Kimpton.
Lieberman, M. A., & Borman, L. D. (1979). Self-help groups for coping with crisis: Origins,
members, processes and impact. San Francisco: Jossey-Bass.
Moorhead, C. (1975). Helping, a guide to voluntary work. London: Macdonald & Jones.
Mowrer, O. H. (1971). Peer groups and mediation, the best 'therapy' for professional and
layman alike. Psychotherapy: Theory, Research and Practice, 8, 44-54.
National Council for Voluntary Organisations. (1982). Directory of community health ini-
tiatives. London: NCVO.
Newell, K. (1975). (Ed.). Health by the people. Geneva: World Health Organization.
Pancoast, D. L., Parker, P., & Froland, C. (Eds.). (1983). Rediscovering self-help: It's role in
social care. Beverly Hills, CA: Sage.
Patients Association. (1982). Self-Help and the patient, a directory of national organisa-
tions concerned with particular diseases and handicaps. London: Author.
Richardson, A., & Goodman, M. (1983). Social help and social care: Mutual aid organizations in
practice. London: Policy Studies Institute.
Riesman, F., & Gartner, A. (1981). Help: A working guide to self-help groups. London: Croom
Helm.
Robinson, D. (1979). Talking out of alcoholism: The self-help process of alcoholics anonymous.
London: Croom Helm.
Robinson, D. (1983). The growth of Alcoholics Anonymous. Alcohol and Alcoholism, 18,
167-172.
Robinson, D., & Henry, S. (1977). Self-help and health: Mutual aid for modern problems.
London: Martin Robertson.
Robinson, D., & Robinson, Y. (1979). From self-help to health. London: Concord Books.
Share Community Limited. (1980). Guide to self-help groups. London: Author.
Thames TV Help Programme (1978). Health help. London: Thames Television.
Todd, J. (1982). Someone to talk to: A directory of self-help and support services in the community.
London: Thames Television in association with the Mental Health Foundation.
WHO/UNICEF, Alma Ata, Declaration of (1978). Report of the International Conference on
Primary Health Care, ICPHC/ALA/78.1O. Geneva: Author.
Zola, I. (1975). Helping one another: A brief history of mutual aid groups. Mimeo, Brandeis
University.
15
The Family in the Change Process
BARBARA S. McCRADY
The role of the family in the assessment and treatment of addictive

behaviors has received careful consideration by researchers and clini-
cians. However, careful theoretical models have lagged behind. In this
chapter, a social support framework will be used to organize considera-
tion of the family's role in the treatment of addictive behaviors.
SOCIAL SUPPORT
Social support has been considered from several different perspec-

tives (Colletti & Brownell, 1982). A social support network is basically a
network of persons who serve certain functions for an individual. This
network may include friends, family, neighbors, self-help or mutual-aid
groups, members of the clergy, health care workers, etc. The network
may help a person to feel cared for, loved, and valued; it may provide
feedback to the individual about his or her beliefs or values, and it may
provide a sense of belonging and a feeling of being connected to other
people.
A social support network may serve many positive functions for a
person. For example, having such a support network may attenuate the
negative emotional reactions that an individual may have to life stresses.
Studies have shown that individuals experience less depression associ-
BARBARA 5. McCRADY • Center of Alcohol Studies, Rutgers University, Piscataway, NJ

08854.
305
306 BARBARAS.McCRADY
ated with job loss if they have a positive support network. Positive social
support also has an influence on the outcomes of medical treatments by
shortening the recovery time from surgery, and by facilitating easier
deliveries during childbirth (Colletti & Brownell, 1982).
Conversely, a lack of social support is associated with a number of
problems, including increased morbidity, and more emotional prob-
lems. For example, it is not uncommon for a marital partner to die
within several months of the death of the spouse. In severe emotional
disorders, such as schizophrenia, a disturbed family environment or
lack of a social support network is associated with more rapid relapses
and rehospitalizations after psychiatric treatment (McGill, Falloon,
Boyd, & Wood-Siverio, 1983).
SOCIAL SUPPORT AND EMOTIONAL AND BEHAVIORAL PROBLEMS
Social support appears to have an important general role in the

maintenance of positive mental and emotional functioning. These obser-
vations of the generally positive role for social support have led re-
searchers and clinicians to attempt to modify existing social networks to
increase the supportive capacity of the network. These attempts are
based on the belief that the spouse or family can be helpful to a person in
their recovery from behavioral or emotional problems. It is also believed
that it is possible to teach the spouse or family how to be supportive.
An alternative approach to conceptualizing the family'S role in treat-
ment has been to see the family as an integral part of the individual's
problems. General systems-theory views suggest that an individual's
problems are manifestations of the problems of a larger system, which is
usually the family. The individual's symptoms are seen as symptoms of
the system dysfunction, and treatment of the whole system is seen as
primary. This emphasis is much different from the notion that the family
can learn to help a dysfunctional individual, because the whole family is
seen as being in need of help.
In the mental health field, a number of recent studies have exam-
ined the effectiveness of involving the spouse in the treatment of a
variety of psychological disorders. For example, Emmelkamp and de-
Lange (1983) reported a study of spouse involvement in the treatment of
obsessive-compulsive disorders. They noted that the spouse often gets
involved in the obsessive-compulsive rituals by reassuring the person
that they are all right, or even by taking over some of the ritualistic
activities, such as doing some of the cleaning that the partner requests,
or by checking such things as whether the house is locked or the oven
turned off. Emmelkamp and deLange note that when not involved in
treatment, the spouse may impede the treatment process, and that the
THE FAMILY IN THE CHANGE PROCESS 307
ritualistic behaviors and the stress associated with these may also lead to
significant relationship problems for the couple. Because of this formula-
tion, they involved both spouses in the treatment process. The spouses
were instructed in how to respond to ritualistic behaviors, and were also
taught how to assist their partners in their homework assignments. No
specific treatment interventions were directed at the couple's rela-
tionship. They found greater improvements in the ritualistic behavior at
the end of treatment when the spouse was involved than when only the
ritualistic partner was involved, but one month after the end of treat-
ment these differences had disappeared.
In a similar study, Barlow and his associates (Barlow, O'Brien, &
Last, 1984) considered the role of the spouse in the treatment of agora-
phobia. They noted that the spouse often becomes a "safe" person for
the agoraphobic, becoming the only person with whom the symp-
tomatic partner will leave the house, drive, or perform other feared
activities. Partners often do not know what to do when the demands on
them become greater-they do not know how much they should take
over responsibilities or push their spouses to maintain responsibilities.
They do not know how much to accompany and reassure their spouses,
and they do not know what to do or say during a panic attack. Barlow et
al. involved the spouses in the treatment program primarily to assist
their partners in carrying out homework assignments, and also in-
structed them in how to respond to the panic attacks and demands of
their agoraphobic mates. They found significantly more agoraphobics
improved after spouse-involved than non-spouse-involved treatment
(12 of 14 versus 6 of 14).
In considering the effects of depression on the marital dyad, Cayne
(1984) noted that depressed behavior produces negative moods in oth-
ers. Depressed persons attempt to elicit support from others, and often
induce guilt in those around them. At the same time, people close to a
depressed person may feel that expressing anger is inappropriate, and
they may therefore become impatient, withdraw, and exacerbate the
individual's depression as a result. Couples often feel as though they are
"walking on eggs," and find themselves making unacceptable compro-
mises in order to avoid confrontation. If the partner tries to be helpful,
he or she becomes frustrated and angrier at the depressed mate. Al-
though Cayne does not propose or test a specific treatment model, his
observations lend support to the notion that individual problems result
in severe disruption in the family and marital relationship, and that
spouse-involved treatment is therefore probably appropriate.
Schizophrenia is also a problem that has been considered to be both
an individual and a family problem. McGill and colleagues (McGill et al.,
1983) noted that deinstitutionalization requires schizophrenics and their
308 BARBARA S. McCRADY
families to take more responsibility for managing the illness. However,

families feel handicapped by their lack of knowledge, and they do not
know what actions of theirs may be helpful or harmful. The treatment
includes the family in an educational and therapeutic program. The
family educational program addresses the nature and phenomenology of
schizophrenia, defines the patient as an expert who educates the family
about the phenomenology, and addresses theories of the etiology of
schizophrenia, the role of environmental stresses, the role of the family in
treatment, and the role of chemotherapy. The family therapy component
involves an assessment of the strengths and deficits in family commu-
nication and coping styles and attempts to enhance such communication
skills as expressing feelings, reflective listening, and making positive
requests for change. Although they report few outcome data from their
study, their preliminary results were described as encouraging.
SOCIAL SUPPORT IN ADDICTIVE BEHAVIORS
In considering the role of the spouse or family in the treatment of

addictive behaviors, there is a large literature in the areas of obesity and
alcoholism, and a relatively small empirical literature in the areas of drug
abuse and smoking. A number of early studies on spouse-involved
treatment for obesity found significantly greater weight reductions and
maintenance of weight reduction when the spouse was involved in the
treatment than when only the obese person was involved (Brownell,
Heckerman, Westlake, Hayes, & Monti, 1978). Brownell's early work
was not replicated (Brownell & Stunkard, 1981), and later studies have
found mixed results. One recent study examined the relative effective-
ness of spouse presence during weight reduction treatment, compared
to active behavior change for the spouse (Murphy et al., 1982), and
found that couples-involved treatment was more effective than indi-
vidual treatment over a 2-year period, but that the type of spouse in-
volvement did not matter. In contrast, Dubbert & Wilson (1984) found
no incremental benefit for spouse involvement over individually
focused group behavioral treatment for weight reduction.
In the smoking area, a more naturalistic approach has been taken to
understanding the role of the spouse in successful smoking cessation.
Mermelstein, Lichtenstein, & McIntyre (1983) developed a questionnaire
to assess a variety of partner behaviors that might be related to smoking
cessation. Their 76-item Partner Interaction Questionnaire (PIQ) was
completed by a group of subjects involved in a smoking cessation pro-
gram. A cluster analysis yielded four major clusters-nagging/shunning
the smoker, policing, cooperative participation (including such items as
"talked me out of smoking a cigarette"), and reinforcement. They found
the lowest PIQ scores for those subjects who never quit smoking during
the program, and found lower PIQ scores for relapsers than abstainers
at 3 and 6 months after treatment. Recent work by Coppotelli (1984)
yielded similar results indicating the strong role for the spouse in suc-
cessful smoking cessation.
Thus, in a number of quite diverse disorders, a similar trend has
emerged in considering the role of the spouse and family in the treat-
ment of these problems. Early models emphasized a strictly individual-
oriented etiology and treatment of the disorders. Subsequent models for
many severe problems (such as schizophrenia) implicated family mem-
bers as the main etiological agents in the development of the disorder (as
in the schizophrenogenic mother concept). As systems models became
more influential, the family as a unit began to be seen as the primary
agent responsible for the development, maintenance, and treatment of
severe psychopathology. Family therapy was then recommended as the
treatment of choice for such problems. Recently, more complex models
are emerging as understanding of the interactions among biological,
psychological, and social systems has increased. In many of the problem
areas reviewed here, a strong individual component is now recognized
in the etiology and maintenance of disorders. The individual component
may be biological, as in contemporary theories of schizophrenia and
obesity (e.g., set-point theory, Brownell, 1982), or primarily condition-
ing based, as in obsessive-compulsive disorders. At the same time, fami-
lies are viewed as trying to cope with these problems, but these coping
attempts are seen as exacerbating the already existing problem, as well
as creating further problems. As the family lives with a problem for a
long time, the structure and functioning of the family changes, so that
the family now has major problems that exist independently of the
original problem. To illustrate these concepts, I will now consider the
role of the family in alcoholism.
SPECIFIC ApPLICATIONS OF SOCIAL SUPPORT TO ALCOHOLISM
The family system of the alcoholic may contain a number of different

members, including the spouse, children, parents, siblings, and other
more distant relatives. Determining what family members are involved
with the alcoholic's drinking, and whether their involvement revolves
specifically around alcohol or is more pervasive requires complex clinical
judgments. I try to discriminate between alcohol-focused family involve-
ment and more general connections between family functioning and the
alcoholic's drinking. Although this may seem an artificial distinction to
make, I find it useful in clinical case conceptualization.
Alcohol-Specific Influences
Any family member may engage in behaviors that are directly relat-
ed to the alcoholic's drinking. However, the behavior of the spouse of
the alcoholic has been the focus of the most research and clinical
speculation.
In clinical settings spouses describe a variety of actions that would
appear to cue drinking behavior. Some of these behaviors can be de-
scribed as nagging the alcoholic about drinking, as for exa~ple, telling
him or her to cut down, providing warnings about the bad things that
are happening or will happen because of drinking, or repeatedly bring-
ing up past drinking episodes in an argumentative, angry tone. A sec-
ond category of spouse behaviors that has been described as occurring at
a high frequency among wives of alcoholic men (Orford et al., 1975) are
behaviors that are intended to control the drinking, but in fact appear to
cue further drinking. Examples include going to the bar to bring the
alcoholic home (which the children of the alcoholic may be asked to do),
hiding liquor or throwing it away, taking control of the checkbook or car
keys, or inviting in friends or relatives to try to control the drinking
behavior. Many of these actions are similar to those described as charac-
teristic of spouses of persons with obsessive-compulsive rituals (Em-
melkamp & delange, 1983).
Spouses and other family members may drink with the alcoholic,
and certain family celebrations may have alcohol as an integral part of
the celebration. Although data are scanty about the actual functional
relationship between "nagging" or "control" behaviors and drinking, it
is noteworthy that in a study of alcoholic men seeking treatment, 74% of
their wives reported that they had stopped trying to control the drinking
behavior prior to the man deciding to seek treatment (Djukanovic,
Milosavcevic, & Jovanovic, 1976). It is possible that this decrease in
spouse cues for drinking had some impact on the alcoholic's drinking.
Family-mediated consequences of drinking fall into three catego-
ries: (a) reinforcement for drinking behavior in the form of attention or
care taking, (b) shielding the alcoholic from experiencing the negative
consequences of drinking, and (c) punishing drinking behavior. Reinfor-
cers for drinking may include such behaviors as providing beverage
alcohol to help ease a hangover, providing something to eat or drink
during a drinking bout, drinking together, or engaging in enjoyable
activities together during a drinking episode. Behaviors that result in
protecting the alcoholic from experiencing other naturally occurring
punishers for the drinking may include taking over the alcoholic's re-
sponsibilities, cleaning up after him or her, calling the employer and
covering at work, paying debts or bills accrued by the alcoholic while
drinking, pretending to others that there are no problems, or completing

chores left undone during a drinking episode (Jackson, 1954; Lemert,
1960). Many of these actions are similar to those taken by spouses of
agoraphobic clients.
the third type of family consequence of drinking involves deliver-
ing punishers for drinking. Although it would initially appear that
punishing behaviors should result in decreased drinking, it may be that
such punishers temporarily suppress the response, but ultimately either
cue further drinking or result in avoidance of the family member who
delivers the punishers. Examples of such double-edged punishers may
include refusing to talk to the alcoholic during a drinking episode, leav-
ing the room, or locking the alcoholic out.
General Family Influences

In addition to family interactions that revolve specifically around
alcohol, alcoholic couples have certain patterns of interaction that create
general problems in the family. Alcoholic couples are postulated to have
poor communication and problem-solving skills. They are seen as hav-
ing a low rate of positive exchanges, and evolving over the years a mode
of interacting that involves attempts to control each other coercively,
such as through threats or nagging. As the aversive situation has esca-
lated over time, communication is believed to become more ambiguous,
vague, and inconsistent. As a result of these poor communication skills
and ineffective methods of control, a large backlog of problems should
accumulate.
A number of studies have reported controlled observations of alco-
holic couples' interactions. Several studies have found a high frequency
of hostile and/or coercive verbal interactions (e.g., Billings, Kessler,
Gomberg, & Weiner, 1979; Cvitkovic, 1979). Couples highest on hostile
coercive interactions prior to treatment also have been found to have the
poorest treatment outcomes (Moos, Bromet, Tsu, & Moos, 1979; Orford,
Oppenheimer, Egert, Hensman, & Guthrie, 1976). A paucity of effective
communication skills has been noted, including a low rate of friendly
acts (Billings et al., 1979), a low rate of cognitive acts (e.g., suggestions,
information) (Billings et al., 1979), a low rate of relationship-relevant
messages emitted by the alcoholic (Klein, 1979), and a low rate of verbal
output generally (Billings et al., 1979; Foy, Miller, & Eisler, 1975). A lack
of intimate, positive exchanges and a low frequency of spending free
time together has also been observed (Djukanovic et al., 1976). None of
these studies, however, has directly examined the influence of aversive
marital exchanges on drinking behavior. It is possible that poor commu-
nication and problem-solving skills characterize the relationships that
alcoholics have with children, siblings, and parents. If similar deficits

exist (and there are no data relevant to this question), then problems in
these other relationships might also accumulate and become cues for
drinking.
Family consequences of drinking are varied, and may serve to rein-
force strongly the drinking response. One or both members of an alco-
holic couple may markedly change their behavior after drinking, result-
ing in positive exchanges not present during nondrinking interactions
(e.g., Steinglass, Davis, & Berenson, 1975). Such changes have also been
observed with an alcoholic father and son (Steinglass, Weiner, & Men-
delson, 1971). The alcoholic member of a couple may increase his or her
assertive or aggressive responses, which might reinforce the drinking
(Cvitkovic, 1979). Problem-solving behavior may also increase (Franken-
stein, 1982). Alcoholics also seem to increase the rate and amount of
verbal output while drinking, which also may reinforce drinking in a
marital relationship in which their verbal output is typically low (Billings
et ai., 1979; Foy et ai., 1975). However, some studies have not noted
these improvements in the alcoholic's marital communications during a
drinking episode, but rather have noted a decrease in relevant commu-
nications during drinking (Cvitkovic, 1979), and an increase in spouse,
rather than alcoholic communications during drinking (Cvitkovic, 1979).
In summary, the literature on alcoholic couples provides a micro-
cosm of the larger literature on couples with major emotional and addic-
tive problems. It appears that there is a significant individual component
in the development of alcohol problems. This individual component
may be learning based, and may also have a heritable element. Howev-
er, the family of the alcoholic then becomes involved with the person's
drinking through the mostly ineffective means they employ to cope with
the drinking. Family and marital problems may have contributed to the
original development of the maladaptive drinking, but also develop as a
result of the drinking, and then contribute to the maintenance of the
drinking.
Interventions
Given this kind of an analysis of the family's role in an addictive
behavior, several kinds of treatment interventions might be considered.
Treatment should be directed at three domains: (a) individual behavior
change, (b) the partner's behavior vis-a-vis the addictive behavior, and
(c) the family's interactions. In the individual realm, individual-behav-
ior-change techniques can be used to help the person stop drinking,
stop smoking or lose weight. There are a number of promising behav-
ioral models that clinicians may use for treatment of these problems,
including McCrady's (1985) work on behavioral treatment of alcohol

abuse and alcoholism, recent work on the use of very-low-calorie diets
and behavior therapy in the treatment of obesity (Wadden, Stunkard,
Brownell, & Day, 1984), and recent work on the use of nicotine gum
(Killen, Maccoby, & Taylor, 1984) and nicotine fading procedures
(Beaver, Brown, & Lichtenstein, 1981) in smoking cessation programs.
In considering the partner's behavior, the clinician should consider
ways that the partner can be helpful to the client in the process of
individual change. For example, the partner can learn to provide rein-
forcement for behavior change, and to withdraw reinforcement if the
client relapses (as in Hunt & Azrin's, 1973, community reinforcement
approach to the treatment of alcoholism). The spouse might also learn
how to be supportive to the change process, by listening in a nonjudg-
mental way when the client is having urges to use the substance that he
or she is trying to stop using, or by suggesting alternative behaviors to
use to cope with urges. The spouse might model appropriate behavior,
as has been suggested in spouse-involved weight reduction programs
(Brownell et al., 1978). The spouse might also learn how to change be-
haviors that cue or reinforce the addictive behavior, and might learn to
provide honest feedback about the impact that the addictive behavior
has on the spouse. The spouse might also be helped to recognize the
limits of one's influence on an addictive behavior, and learn ways to
cope with feelings about the addictive behavior and its consequences.
The third domain of clinical intervention is in the interactions of the
couple or family. A behavioral marital-therapy model (e.g., Jacobson &
Margolin, 1979) might be used to help the couple increase positive ex-
changes and improve communication skills. Family or friends might also
provide concrete assistance to the individual, such as advice on how to
cope with practical problems (as in Azrin's, 1976, use of "buddies" in
the treatment of alcoholism). Finally, families or couples can learn to
spend time together in activities that are reinforcing for the entire fami-
ly, and are incompatible with the use of the addictive substance.
These suggestions are based partly on clinical experience and partly
on the empirical literature. In the next section, a research study will be
presented that examines the relative effectiveness of spouse-involved
alcoholism treatments that have differing degrees of emphasis on the
role of the family.
COMPONENTS OF SPOUSE INVOLVEMENT IN ALCOHOLISM

TREATMENT
Subjects for the study were recruited through newspaper advertis-

ing, contact with community agencies, and from the admissions depart-
ment and inpatient units of the psychiatric hospital in which the

research was conducted. To be included in the study, subjects had to be
between 21 and 60 years of age, married, their spouse had to be willing
to participate, they had to have had a drinking problem for at least 2
years, have been drinking in the last 60 days, have a score of 5 or greater
on the Michigan Alcoholism Screening Test (MAST) (Selzer, 1971), and
they had to report that at least four clear problems had occurred because
of drinking in the last 12 months. Subjects were excluded if they abused
drugs, showed evidence of a major depressive disorder, schizophrenia,
or organic brain syndrome, or if the spouse was an alcohol or drug
abuser. Subjects were also excluded if they were involved in any other
form of treatment and were unwilling to discontinue that treatment
while involved in the research treatment program.
Subjects were evaluated prior to treatment, during treatment, in
face-to-face interviews immediately after treatment, and 6, 12, and 18
months later. Subjects and spouses were also interviewed monthly in
separate telephone interviews. Data were collected about daily drinking
behavior, daily marital satisfaction, and daily urges to drink. Em-
ployment status, legal problems, and marital separations or divorces
were assessed. Structured questionnaires were used to assess marital,
social, and psychological functioning, and all couples were videotaped
to provide information about communication skills.
After informed consent and baseline data collection, subjects were
randomly assigned to one of three treatment conditions: minimal spouse
involvement (MSI), alcohol-focused spouse involvement (AFSI), or mar-
ital therapy (MT). In all groups, each couple was seen conjointly for 15
treatment sessions, each 90 minutes long. In all experimental groups,
subjects were taught skills to learn how to maintain abstinence from
alcohol. In the AFSI and the MT group, spouses were taught how to
provide support for abstinence, how to discuss assertively concerns
about drinking, how to respond to any drinking episodes, and how to
respond to drinking situations. In the MT condition, couples also re-
ceived specific treatments to modify their marital relationship, including
treatments to increase the positive exchanges in the relationship, and
communication and problem-solving training.
Fifty-three couples began in the experimental treatments. However,
there was a trend toward a differential dropout rate among the groups
(X 2 (2) = 5.21, p<.08), with more subjects discontinuing the individually
focused treatment than the other two treatments (52.8% completed MSI
treatment, compared to 76.9% in the AFSI group and 84.2% in the MT
group). During treatment, the subjects in the AFSI group did not signifi-
cantly decrease their drinking over the 15 weeks of treatment, whereas
subjects in the other two treatment conditions did significantly decrease
the quantity and frequency of drinking. Most subjects were abstinent by

the end of treatment. There were no marked differences among the
groups on marital satisfaction during treatment. The couples in the MT
group were more compliant with joint homework assignments than
couples in the AFSI group, but there were no differences among the
three groups on homework completed by the subjects alone or by the
spouses alone.
During the 18 months of follow-up, 90.2% of the follow-up inter-
views scheduled with clients were completed. Over the first 6 months of
follow-up, there were significant differences in outcomes for both drink-
ing and marital satisfaction. Subjects in the AFSI group relapsed more
quickly than subjects in either of the other two treatment conditions.
Subjects in both the MSI and the AFSI condition reported marked de-
creases in marital satisfaction, compared to the MT group.
Over the full 18 months of follow-up, there were no differences in
drinking outcomes among the three groups. Subjects reported absti-
nence on approximately 82% of the days during follow-up, with 31.7%
of the subjects being continuously abstinent (28.6% of the MSI group,
27.3% of the AFSI group, and 37.6% of the MT group). However, there
were several separations and divorces in both the MSI (four separations,
mean of 178 days) and the AFSI groups (three separations, mean of 83.7
days), with high marital stability in the MT group (one brief, 6-day
separation). Clients in the MT group reported higher marital satisfaction
on the Locke-Wallace Marital Adjustment Test (MAT) (Locke & Wallace,
1959) when compared to the AFSI group. For those couples in the MSI
group who remained together, their MAT scores were also significantly
higher than those in the AFSI group. Communication skills were also
different in the AFSI couples than the other two groups, with more
negative verbal behavior by subjects, and more negative nonverbal be-
havior by spouses in this group than in the other two groups.
CONCLUSIONS
The research presented here suggests several important issues rele-

vant to family involvement in the understanding and treatment of addic-
tive behaviors. First, there were differences in treatment adherence and
compliance with treatment among the groups. If a couple is willing to be
involved in treatment for alcoholism, it appears that active spouse in-
volvement is beneficial. These results are in contrast to results in the
obesity literature (Murphy et ai., 1982). By the end of 18 months of
follow-up, the differences in drinking among the three groups had dis-
appeared. However, there were major differences in the marriages of
the couples who received the different treatments. There were more
separations among those who did not receive marital treatment, and
there were differences in marital communication and marital satisfaction
among those who stayed together. It may be that there were two differ-
ent pathways to improved marital relationships in these couples-direct
treatment of the marital problems, or intensive treatment to help the
individual cope effectively with his or her drinking problem. In either
case, the marriages that survived seemed to benefit.
At the beginning of the chapter, the positive role of family-involved
treatment was emphasized. However, the research literature is more
mixed than the enthusiastic clinical and theoretical literature. The re-
search literature clearly demonstrates that marital therapy helps mar-
riages, even when an addictive behavior is present. It also appears to
support the notion that spouses can provide support for changing cer-
tain addictive behaviors (e.g., smoking), but there is not a convincing
literature that demonstrates that clinicians can change spouses' behav-
iors to become supportive if they are not already so.
It may be that one of the reasons that family or spouse-involved
treatment has not been as effective as hoped is that the approaches to
family involvement attempted to date have not taken into account the
stage of the addictive behavior (Prochaska, 1979). For example, the fami-
ly's role in helping a person who is in the precontemplation stage may
be quite different than the family's role in treatment for a person who is
at the maintenance stage of behavior change. In the precontemplation
phase, the family may be acutely aware of a problem, whereas the
individual is completely unaware that a problem exists. At this stage,
the family may provide information about the problem, and may estab-
lish contingencies that require that an individual seek help, even though
the individual does not see the behaviors as problematic. At the con-
templation stage, further information and feedback may continue to be
helpful, as well as continuing enforcement of contingencies for behavior
change. During action, the family may provide support, and the whole
family may be involved in treatment to change problems that are present
in the family. During maintenance, the family may continue to reinforce
change, continue to change their own behavior, and an altered family
structure might also support the individual change efforts. The family
appears to be especially important in the contemplation and action
stages when it comes to emotional feedback and contingency setting. By
considering the different roles of the family at different stages of indi-
vidual change, it may be possible to improve the effectiveness of family-
involved treatment for addictive behaviors.
REFERENCES
Azrin, N. H. (1976). Improvement in the community-reinforcement approach to alco-

holism. Behaviour Research and Therapy, 14, 339-348.
Barlow, D. H., O'Brien, G. T., & Last, C. G. (1984). Couples treatment of agoraphobia.
Behavior Therapy, 15, 41-58.
Beaver, c., Brown, R. A., & Lichtenstein, E. (1981). Effects of monitored nicotine fading
and anxiety management training on smoking reduction. Addictive Behavior, 6, 301-
305.
Billings, A. G., Kessler, M., Gomberg, C. A., & Weiner, S. (1979). Marital conflict-resolu-
tion of alcoholic and nonalcoholic couples during drinking and nondrinking sessions.
Brownell, K. D. (1982). Obesity: Understanding and treating a serious, prevalent, and
refractory disorder. Journal of Consulting and Clinical Psychology, 50, 820-840.
Brownell, K. D., & Stunkard, A. J. (1981). Couples treatment of obesity. Archives of General
Psychiatry, 38, 1224-1229.
Brownell, K. D., Heckerman, C. L., Westlake, R. J., Hayes, S. c., & Monti, P. M. (1978).
The effect of couples training and partner cooperativeness in the behavioral treatment
of obesity. Behaviour Research and Therapy, 16, 323-333.
Cayne, J. C. (1984). Strategic therapy with depressed married persons: Initial agenda,
themes and interventions. Journal of Marital and Family Therapy, 10, 53-62.
Colletti, G., & Brownell, K. D. (1982). The physical and emotional benefits of social
support: Application to obesity, smoking and alcoholism. Progress in Behavior Modifica-
tion, 13, 109-178.
Coppotelli, H. C. (1984, August). Spouse support for smoking cessation maintenance by women.
Paper presented at the annual meeting of the American Psychological Association,
Toronto, Canada.
Cvitkovic, J. V. (1978). Alcohol use and communication congruence in alcoholic and non-
alcoholic marriages. Dissertation Abstracts International, 39, A1335 (University Micro-
films No. 7816784).
Djukanovic, B., Milosavcevic, V., & Jovanovic, R. (1976). Drustveni zivot alkoholicaro i
njihovih suprnga. (The social life of alcoholics and their wives.) Alkoholizam, Beograd,
16, 67-75. (From Journal of Studies on Alcohol, 1978, 39, Abstract No. 1141).
Dubbert, P. M., & Wilson, G. T. (1984). Goal-setting and spouse involvement in the
treatment of obesity. Behaviour Research and Therapy, 22, 227-242.
Emmelkamp, P. M. G., & deLange, l. (1983). Spouse involvement in the treatment of
obsessive compulsive patients. Behaviour Research and Therapy, 14, 341-346.
Foy, D. W., Miller, P. M., & Eisler, R. M. (1975, November). The effects of alcohol consump-
tion on the marital interactions of chronic alcoholics. Paper presented at the annual meeting
of the Association for Advancement of Behavior Therapy, San Francisco, CA.
Frankenstein, W. (1982). Alcohol intoxication effects on alcoholics' marital interactions for three
levels of conflict intensity. Unpublished Masters thesis, Rutgers University.
Hunt, G. M., & Azrin, R. H. (1973). A community-reinforcement approach to alcoholism.
Jackson, J. K. (1954). The adjustment of the family to the crisis of alcoholism. Quarterly
Jacobson, N., & Margolin, G. (1979). Marital therapy. New York: Brunner/Maze!'
Killen, J. D., Maccoby, N., & Taylor, C. B. (1984). Nicotine gum and self-regulation
training in smoking relapse prevention. Behavior Therapy, 15, 234-248.
Klein, R. M. (1978). Interaction processes in alcoholic and nonalcoholic marital dyads.
Dissertation Abstracts International, 39, B1484. (University Microfilms No. 7816408)
Lemert, E. M. (1960). The occurrence and sequence of events in the adjustment of families
to alcoholism. Quarterly Journal of Studies on Alcohol, 21, 679-697.
Locke, H. J., & Wallace, K M. (1959). Short marital-adjustment and prediction tests: Their
reliability and validity. Marriage and Family Living, 21, 251-255.
McCrady, B. S. (1985). Alcoholism. In: D. H. Barlow (Ed.), Clinical handbook of psychological
disorders: A step-by-step treatment manual. New York: Guilford Press.
McGill, C. W., Falloon, I. R H., Boyd, J. L., & Wood-Siverio, C. (1983). Family educational
intervention in the treatment of schizophrenia. Hospital and Community Psychiatry, 43,
934-938.
Mermelstein, R, Lichtenstein, E., & McIntyre, K (1984). Partner support and relapse in
smoking cessation programs. Journal of Consulting and Clinical Psychology, 51, 465-466.
Moos, R H., Bromet, E., Tsu, V., & Moos, B. (1979). Family characteristics and the
outcome of treatment for alcoholism. Journal of Studies on Alcohol, 40, 78-88.
Murphy, J. K, Williamson, D. A., Buxton, A. E., Moody, S. c., Absher, N., & Warner, M.
(1982). The long-term effects of spouse involvement upon weight loss and mainte-
nance. Behavior Therapy, 13, 681-693.
Orford, J., Guthrie,S., Nicholls, P., Oppenheimer, E., Egert, S., & Hensman, C. (1975).
Self-reported coping behavior of wives of alcoholics and its association with drinking
outcome. Journal of Studies on Alcohol, 36, 1254-1267.
Orford, J., Oppenheimer, E., Egert,S., Hensman, c., & Guthrie, S. (1976). The co-
hesiveness of alcoholism complicated marriages and its influence on treatment out-
come. British Tournai of Psychiatry, 128, 318-339.
Prochaska, J. (1979). Systems of psychotherapy: A transtheoretical analysis. Homewood, IL:
Dorsey Press.
Selzer, M. L. (1971). The Michigan Alcoholism Screening Test: The quest for a new diag-
nostic instrument. American Journal of Psychiatry, 127, 1653-1658.
Steinglass, P., Weiner, S., & Mendelson, J. H. (1971). A system approach to alcoholism.
Archives of General Psychiatry, 24, 401-408.
Steingiass, P., Davis, D. I., & Berenson, D. (1975, May). In-hospital treatment of alcoholic
couples. Presented at the Annual Meeting of the American Psychiatric Association,
Anaheim, CA.
Wadden, T. A., Stunkard, A. J., Brownell, K D., & Day, S. C. (1984). Treatment of obesity
by behavior therapy and very low calorie diet: A pilot investigation. Tournai of Consult-
16
Cognitive Processes in Addictive
Behavior Change
IAN ROBERTSON
INTRODUCTION
The aim of this chapter is to explore some possible relationships be-

tween self-regulation and behavior change, on the one hand, and higher
level cognitive processes, namely attention, memory, and thinking, on
the other. The emphasis is on drinking behavior, but the relevance to
other addictive behaviors will become obvious. This chapter draws upon
a cybernetic model of self-regulation elaborated by Carver and Scheier
(1981) from the work of Miller, Galanter, and Pribram (1960), Powers
(1973) and others, and is unashamedly speculative in tone. It aims to
draw together theoretical and empirical studies from diverse areas in the
hope that some fruitful testable hypotheses may emerge.
The fundamental unit of a cybernetic system is the feedback loop, and
Miller, Galanter, and Pribram (1960) describe the TOTE loop: namely,
the person or machine tests the sensory input to see whether it conforms
to a predetermined standard, and then operates on the environment in
order to bring the sensory input in line with the standard. It then tests
again, and if input and standard are compatible, it exits and no further
action is necessary during this behavioral phase. If input and standard
are still discrepant, the operate phase must be repeated.
IAN ROBERTSON • Department of Psychology, Astley Ainslie Hospital, Edinburgh EH9

2HL Scotland.
319
320 IAN ROBERTSON
FIGURE 1. Self-regulation loop.
There are parallels in self-control theory, and Figure 1 outlines a

parallel scheme in the language of self-control.
How is this model useful in understanding addictive behavior? In-
dividualloops can be nested into decision trees, algorithms that provide
hypothetical models of the course of self-regulation. One such cyber-
netic model will now be discussed in some detail.
CONTROL THEORY
Carver and Scheier (1981) have elaborated upon Power's (1973)

cybenetic model to produce a model of self-regulation, and this will
serve as a working model in this chapter through which to discuss some
hypothetical mechanisms by which cognitive processes may affect self-
regulation. Figure 2. outlines their basic model, taking for simplicity's
sake only the top six levels of the hierarchy. Next to each level is an
example of how a hypothetical problem drinker might fit into this
framework.
Most behavior is nested within superordinate goal systems, though
there are clearly examples where this is not the case, for instance, where
classical or operant conditioning takes place without awareness. In the
control theory of Carver and Scheier, there are a number of levels of
organization of behavior, consisting of a complex network of feedback
loops. The comparison standard, or goal, for one level of organization is
determined by the output of the next higher level in the hierarchy.
If we take another example in Figure 2, we note, for instance, that
the highest system level may be a higher-order self-concept, such as "I
am a responsible person." The next level down, the principle level,
might be a concept such as "1 try to do what 1 say 1 will do." Now the
standard (or goal) for this principle level is specified by the output from
the system level. In other words, the person judges the adequacy of his
functioning at this level according to the lofty ideals set by the higher
COGNITIVE PROCESSES IN BEHAVIOR CHANGE 321
"I am a sober.
responsible.
family man."
"I will stop

meeting Gerry
and 8ill and go
straight home
after work."
"I will leave work

early this Friday.
I'll tell Gerry
and Bill I won't
seem them. II
"Go to 8ill's
office Friday
morning; arrange
to II!"1IIve work early. "
"I'm not going to

the pub, Bill. II
FIGURE 2. Hypothetical cybernetic hierarchy for a problem drinker-I. Adapted from Car-
ver and Scheier, Psychological Bul/etin, 92, 111-135.
level. Thus, at the program level, one possible output might be, "I must
visit my aunt as I promised." The adequacy of functioning at this level is
determined by the output from the program level, and so on. (Note that
the finer-grained components of the behavioral sequences are necessary
for the output to be perceived by the individual, but for simplicity's sake
let us omit the lower levels of the hierarchy.)
The person illustrated in Figure 2 has reached the stage where he
sees a reason to change his drinking behavior, whether through his own
devices or with the help of a counselor. His self-concept and intentions
can be analyzed in the language of Carver and Scheier. Each level of this
system is regulated and adjusted according to the standards set by the
next highest. Thus the man assesses his principle-level intentions ("I am
going to change my drinking habits") in terms of the system level output
("I am a sober family man"). Alterations in the principle-level output are
compared with the comparison set by the system. So, for instance, if the
man begins to think "I'll go and have a good few drinks with the mates,
now and again," comparison of this with the system output ("I am a
responsible family man") should result in a change at the principle level
to reduce discrepancy between output and standard. Similar events oc-
322 IAN ROBERTSON
cur all down the chain, and the perfectly functioning system ensures
that the final behavioral output is in complete accord with the highest
level system output. So our man goes home early on Friday night and
spends a pleasant time with his family. Or does he?
Carver and Scheier maintain that the network of commands will
operate only below the level upon which attention is deployed. Thus behav-
ioral output will only be compatible with the highest level of control that
is attended to. Take Figure 2, for instance. Suppose attention is de-
ployed only at the relationship level, that is, "1 will leave work early this
Friday, etc."; Suppose he tells them this, and they say, "Look, we're
meeting at 10 tonight-come along then." The man has satisfied the
goals for the relationship level (he told them, didn't he?), yet the low-
level deployment of attention ensures that the actual behavior output is
at odds with higher level systems, principles, and programs.
For self-regulation to be in accord with the highest-order goals,
attention has to be deployed at the highest levels. In other words, self-
awareness, or high-level self-monitoring, is required. Other cognitive
mechanisms are required for self-regulation-memory and thought in
particular-and I will return to these later. First, however, let me turn to
the evidence about self-awareness, particularly in relation to alcohol and
its effects.
SELF-AWARENESS AND SELF-REGULAnON
Jay Hull (1981) published an influential paper that argued that a

major effect of alcohol is to interfere with higher-order processing of
self-relevant information, that is, information both about cues regarding
appropriate forms of behavior, as well as self-evaluative feedback about
past behaviors. Hull also argued that this provides a major motivation
for drinking, and reviewed research from a number of areas to support
the hypothesis, ranging from experimental studies of intoxication where
normals produced fewer spontaneous self-referring words and phrases
than those who had taken no alcohol, to studies of conversations among
intoxicated people, where a failure to acknowledge the content of other
people's speech is observed. I do not intend to debate Hull's thesis here,
as this has been done elsewhere (Hull & Reilly, 1983; Wilson, 1983);
rather, I would like to take it as a working hypothesis and to examine its
implications for self-regulation.
There is some evidence to suggest that the chronic effects of alcohol
on the brain are to some extent similar to its acute effects while still in
the blood. For instance, Butters and Cermack (1980) demonstrated that
alcoholics showed a poorer verbal memory on certain tests, and that it
seemed that their poor performance was in part attributable to a failure
spontaneously to deploy mnemonic or other memorising strategies.

Rosen and Lee (1976) found a similar result with acutely intoxicated
normals; their poor performance on certain memory tasks was attributa-
ble in part to a failure spontaneously to organize the material into se-
mantic categories, so as to make it more easily remembered. Thus there
are some grounds for assuming deficiencies in high-level initiation of
cognitive processes resulting from both the acute and the chronic effects
of alcohol. Could these deficiencies relate to a failure of high-level self-
monitoring and self-awareness? Do intoxicated normals and dried-out
problem drinkers do poorly in part because they fail to ask themselves
questions such as" "How am I going to remember this?", or "What do
these words have in common?". Certainly, the poor performance of
alcoholics on problem-solving tasks requiring this sort of higher-order
process is compatible with such a hypothesis (e.g., Klisz & Parsons,
1979). Furthermore, Tarbox and his colleagues at Houston (Tarbox,
Weigel, & Biggs, 1984) have identified an "internal scanning" factor that
predicts adjustment of alcoholics at a 2-year follow-up. This factor is
described by Tarbox et al. as follows:
It reflected the extent to which the individual made use of information relat-
ed to an act prior to responding. Among the sources of information available
to the scanner and relevant to an act are memory storage, remote associa-
tions, judgmental or logical operations, physiological sensations and self-
concept. (p, 2)
Thus there exists some evidence to suggest deficient high-order

self-monitoring or self-awareness among alcoholics, and perhaps also
among intoxicated normals. Could it be that even among problem drink-
ers who have not suffered long-term brain-damage, a "dulling" of high-
level self-monitoring may persist for several days of abstinence? If so,
then perhaps the implications of dulled self-monitoring go far wider
than to the chronically damaged alcoholic population. One might specu-
late that even drinkers who give up drinking for a week or two are
hindered in these efforts by a persisting, nonpermanent dulling of self-
awareness, and thus never really give themselves the chance fully to
take stock of their lives and problems because of this.
It is being argued therefore that high-level attentional processes
may be disrupted by alcohol, and awareness and attention limited.
Given that Carver and Scheier have argued that the self-regulation con-
trol system (see Figure 2) only operates below the level where attention
is deployed, then a disruption of self-regulation because of the atten-
tion-disrupting effects of alcohol seems possible. If this is the case, then
behavioral misregulation may well be a consequence, because lower-
order control levels may function independently of higher-order com-
mand levels, that is, the hierarchy of control is truncated.
324 IAN ROBERTSON
OTHER REASONS FOR Low-LEVEL DEPLOYMENT OF ATTENTION
I have just argued that self-regulation can become detached from

the output of higher-order principles because attention deployment is
reduced to low and relatively more concrete levels through the effects of
alcohol-both acute and chronic. But other factors can also disrupt at-
tention and focus it on lower levels. First, where a strongly established
behavioral program (or script, or schema, depending on whose lan-
guage you are using) exists, then it is likely that it will be associated with
certain cues or stimuli. If these cues are sufficiently salient and unex-
pected, it is quite possible that they might set in action a given repertoire
of behavior-hence, for instance, the smoker who temporarily forgets
himself after a few drinks, and finds himself launched into the chain of
thoughts and actions comprising the "smoking program."
Second, strong emotions may disorganize higher-attentional pro-
cesses so that regulation according to higher-order principles is aban-
doned and attention is focussed on lower-level processes. The poor
performance of very anxious and depressed people on a variety of cog-
nitive tasks testifies to the disruptive effects of strong emotions on high-
er cognitive processes.
Third, there are considerable individual differences in the extent to
which people show higher-order slips of action and memory. Broad-
bent, Cooper, Fitzgerald, and Parks (1982), among others, have at-
tempted to measure cognitive failures, and have developed a question-
naire that has predictive utility. For instance, they cite one study where
student nurses entering a hospital were followed-up after they had been
working in a variety of wards, some of which were known to be "high-
demand," stressful wards. What was found was that those nurses who
reported high levels of cognitive failures showed markedly higher levels
of anxiety and mental stress in the high-demand wards than did the
low-cognitive-failures group, in spite of showing comparable levels of
anxiety and depression prior to entering the ward. In the low-demand
wards, their levels of mental distress remained similar to that of the low-
cognitive-failures group.
Thus the attentional processes that are necessary for self-regulation
can break down for a number of reasons, or for combinations of these
reasons. For instance, a problem drinker will likely have some very
active schemata for drinking, which are responsive to highly salient
cues. Much of the time the drinker may be able to anticipate the cues
and deploy attentional processes so as to inhibit the prepotent schemas
and substitute some more benign, albeit novel, schemata. Yet if these
attentional processes are disrupted-whether by the acute or chronic
effects of alcohol, powerful emotional states, or constitutional deficits in
the higher-order attentional processes-then such cues may more read-

ily trigger their corresponding behavioral script.
SCHEMATA ACTIVATION
Let us go back to the hypothetical drinker described in Figure 2.

Following attendance at a drinking problems clinic, where he has spent
a considerable amount of time talking over his problems with other
people, he has moved, in Prochaska and DiClemente's terms, through
the precontemplation and contemplation phases of change and now is
considering plans for action on his discharge. This man is working hard
and is a good bet in the eyes of the staff. The sequence of tasks in Figure
2 is a possible scenario for one challenge that faces him.
Let us assume that he has the bones of an alternative behavioral
schema or plan that involves a somewhat altered self-image, adjusted
values and priorities, new behavioral intentions, and a set of salient
memories and arguments for a new plan. In Prochaska and DiCle-
mente's words, his consciousness has been raised, he is to some extent
liberated from his self and from his social environment, and he has
reevaluated his priorities. Figure 2 represents a nested set of intentions
and schemata consistent with the new image. Let us compare this with a
hypothetical set of schemata, currently dormant, yet potentially active,
which were in operation prior to the contemplation phase of his behav-
ior-change process (Figure 3.).
Here we have two sets of schemas, one (Figure 2) as yet depending
mainly on messages from highest order schemas for its activation,
whereas the other (Figure 3) is activated at all levels by a set of well-
established cues and contingencies.
Let us assume that this person's attention is disrupted and focused
at relatively concrete levels because he is still dulled by the persisting
effects of alcohol, even though it is 3 weeks since he had his last drink.
Suppose his attention is focused at the relationship level of the feedback
hierarchy. Figure 4 outlines the possibilities.
Which of the two behavior chains will come into operation will
depend on a number of factors already mentioned, including the
following:
1. Deployment of attention to higher-order levels of feedback
systems
2. Control over potentially disruptive emotions
3. Ready access to relevant memories and schemata
4. Rapid problem-solving and decision making
Thus, in order to sustain a relatively unsupported, low-probability
326 IAN ROBERTSON
"I am I good-time guy.

one of the lad •. "
IIPeopie like me tlke •

good drlnk,lI
"Jill go drinking
tonight."
"Going out for I drink."
"I'll go to the UIUII

pub,"
Goel to the
"Clnny Mann."
FIGURE 3. Hypothetical cybernetic hierarchy for a problem drinker-II. Adapted from

Carver and Scheier, Psychological Bulletin, 92, 111-135.
behavioral schema, the person must not only actively deploy attention
to the highest level schemata (a relatively rare tendency for most of us in
day-to-day life), but he or she must also exert cognitive control over
thought processes associated with strong emotions, resist the impulse to
act on the basis of these thoughts, appraise problem situations, decide
CUES
FIGURE 4. Competing schemata.

on responses, implement them, and check out the results with the per-
sonal and normative standards. He or she must also have ready access
to a wide range of memories, both verbal and visual, relevant to deci-
sions made at feedback loops at all levels of the control hierarchy. The
range of cognitive tasks required of individuals include self-awareness,
self-monitoring, environmental monitoring, stimulus control, goal spec-
ification, self-evaluation, self-reinforcement, rehearsal, cognitive con-
trol, decision making, scheduling and accessing memories, images or
schemata, among others.
THE ROLE OF MEMORY AND THINKING IN SELF-REGULAnON
I have already suggested that the disruption of higher-order atten-

tional processes may, under certain circumstances, disrupt self-regula-
tion, and it is clear that, to the extent that each of the previously
mentioned processes must be uncued and self-initiated, then attentional
deficits may interfere with these self-regulation processes. Where does
memory come into this?
On tests of both verbal and nonverbal memory (though not all
tests), many problem drinkers-as well as acutely intoxicated non-
problem drinkers-show deficits. But why should memory for lists of
words or for pictures have any impact on self-regulation? There is no
direct evidence relating to this question in the problem-drinking liter-
ature, but let us look for a possible analogy in the field of depression.
Teasdale and Fogarty (1981) induced either a happy or a depressed
mood in a group of normal subjects. The induced mood was validated
by well-established measures such as rate of speaking. Subjects were
given a list of words and were asked to retrieve pleasant memories that
were sparked off by each word, and to press a button as soon as such a
memory came into mind. The mean latency for those who had a happy
mood induced was significantly shorter (4.05 sees) than for those who
had a depressed state induced (5.25 sees). In other words, retrieving a
pleasant memory took on average 1.2 sees longer when a depressed
state had been induced in normal subjects. My reason for showing this
finding is to argue that a similar slowing of memory activation for "emo-
tional," behaviorally relevant memories might well be expected in prob-
lem drinkers. What difference does one or two seconds make in the
retrieval of memories? Well, in the complex day-to-day decisions that
are necessary for self-regulation, one second slowness in retrieving a
memory could be the difference between the activation of one schema
over another, given that the amassing of relevant memories may well be
part of the process of activating higher-order schemata. If the person is
328 IAN ROBERTSON
faced with a potent drinking cue, quick retrieval of "good family" mem-
ories may be absolutely crucial in determining whether or not the cue is
resisted. This is, however, speculative and must be subject to proper
experimental study.
With regard to thinking, on the other hand, more direct evidence
pertaining to the addictive behaviors is available. Chaney, O'Leary, and
Marlatt (1978) carried out a study of training in relapse prevention and
problem solving in which problem drinkers had to identify problematic
situations liable to produce relapse and were rehearsed in ways of re-
sponding to these situations. The following is one example:
You are eating at a good restaurant on a special occasion with some friends.
The waitress comes and says, "Drink before dinner?". Everyone orders one.
All eyes seem to be on you.
Now it so happens that the treatment group showed a better out-

come on a number of measures than the control group, but that is not
the main concern here. One of the measures taken was latency of re-
sponse-that is, how long it took a subject to come up with a response
to a role-played problematic situation. This latency measure was un-
affected by training, yet it was the best predictor of a number of one-year
outcome measures, including days abstinent, days hospitalized, and
total amount drunk.
It is quite astonishing that such a measure should outweigh other
demographic and drinking history measures, but that is what Chaney et
al. found (though, of course, correlations do not demonstrate cause). At
least four other studies have shown that cognitive variables predict out-
come in a similar group (Abbot & Gregson, 1981; Bergland, Leijorgust, &
Horlen, 1977; Gregson & Taylor, 1977; Tarbox et al., 1984).
CONCLUSIONS
I have argued that cognitive variables-attention, memory, and

thinking-have a potentially strong influence on self-regulation, and
that self-regulation theory has paid insufficient attention to these vari-
ables. I have focused on the case of alcohol as this most clearly affects
cognitive functions. Yet, given that there are wide individual variations
in cognitive state for reasons ranging from mood to genetic factors, it is
not only with respect to problem drinking that these variables are rele-
vant, as the research quoted by Broadbent et ai. (1982) shows. Perhaps
self-regulation theory should pay more attention to the facilitation of
improved memory, attention, and thinking strategies, that is, on think-
ing processes as opposed to thought content. This might also apply to the
process of counseling, where many assumptions are made about the

ability of clients to translate words into actions. These assumptions may
be far from justified in many cases.
REFERENCES
Abbot, M. W., & Gregson, R A. M. (1981). Cognitive dysfunction in the prediction of
relapse in alcoholics. Journal of Studies on Alcohol, 42, 230-242.
Berglund, M., Leijongust, H., & Horlen, M. (1977). Prognostic significance and rever-
sibility of cerebral dysfunction in alcoholics. Journal of Studies on Alcohol, 38,1761-1769.
Broadbent, D. E., Cooper, P. F., Fitzgerald, P., & Parkes, K. R (1982). The Cognitive
Failures Questionnaire (CFQ) and its correlates. British Journal of Clinical Psychology,
21, 1-16.
Butters, N., & Cermack, L. (1980). Alcoholic Korsakoff's syndrome: An information processing
approach. New York: Academic Press.
Carver, C. S., & Scheier, M. F. (1981). Attention and self-regulation: A control-theory approach
to human behavior. New York: Springer Verlag.
Chaney, E., O'Leary, M., & Marlatt, G. (1978). Skill training with alcoholics. Journal of
Consulting and Clinical Psychology, 46, 1092-1104.
Gregson, R., & Taylor, G. (1977). Prediction of relapse in men alcoholics. Journal of Studies
on Alcohol, 38, 1749-1760.
Hull, J. G. (1981). A self-awareness model of the causes and effects of alcohol consump-
tion. Journal of Abnormal Psychology, 90, 586-600.
Hull, J. G., & Reilly, N. P. (1983). Self-awareness, self-regulation and alcohol consump-
tion: A reply to Wilson. Journal of Abnormal Psychology, 92, 514-519.
Klisz, D. K., til Parsons, O. A. (1979). Hypothesis testing in younger and older alcoholics.
Journal (f Studies on Alcohol, 38, 1718-1729.
Miller, G. A., Galanter, E., & Pribram, K. (1960). Plans and the structure of behavior. New
York: Holt, Rinehart & Winston.
Powers, W. T. (1973). Behavior: The control of perception. Chicago, IL: Aldine.
Rosen, L. J., & Lee, C. L. (1976). Acute and chronic effects of alcohol use on organizational
processes in memory. Journal of Abnormal Psychology, 85, 309-317.
Tarbox, A. R, Weigel, J. D., & Biggs, J. T. (1984, August). A cognitive typology of alcoholism:
Implications for treatment outcome. Paper presented at the Third International Con-
ference on Treatment of Addictive Behaviours, North Berwick, Scotland.
Teasdale, J., & Fogarty, S. (1979). Differential effects of induced mood on retrieval of
pleasant and unpleasant events from episodic memory. Journal of Abnormal Psychology,
88, 248-257.
Wilson, G. T. (1983). Self-awareness, self-regulation and alcohol consumption: An analysis
of J. Hull's model. Journal of Abnormal Psychology, 92, 505-513.
17
Change without Therapists
The Use of Self-Help Manuals by Problem
Drinkers
NICK HEATHER
The main title of this chapter is misleading if it implies that I will be

concerned here with so-called spontaneous remission. Rather, as the
subtitle indicates, I am interested in the extent to which, and the way in
which, problem drinkers may be assisted to achieve and maintain bene-
ficial changes in drinking behavior by using self-help manuals written
by professionals-what might be termed "assisted spontaneous remis-
sion." We all know that problem drinkers can, and frequently do, re-
duce their drinking to non-problem levels without any formal help from
therapists (e.g., Saunders & Kershaw, 1979; Tuchfeld, 1976). But can
self-help manuals, based on the principles of self-management theory,
assist them in this process and, if so, how? This is the principal question
I wish to pose.
NICK HEATHER • Department of Psychiatry, Ninewells Medical School, University of

Dundee, Dundee DDl 9SY Scotland. The main funding for the self-help manual evalua-
tion was provided by the Medical Research Council, but the Scottish Health Education
Group also made a substantial financial contribution. I would like to thank Ian Young, Sam
Docherty, Mike Church, and Ian Thompson of SHEG for their help and advice during the
project, as well as lain Glenn for making available the results of the Highlands DRAMS
evaluation. The DRAMS evaluation is a joint project with the Department of General
Practice, Ninewells Medical School, and is funded by the Scottish Home and Health
Department (Chief Scientist Office). The medical wards research is funded by the Health
Promotion Research Trust.
331
332 NICK HEATHER
This chapter is decidedly oriented toward the Action and Mainte-

nance stages of Prochaska and DiClemente's model of the change pro-
cess; it assumes that the Contemplation and Decision-Making stages
have been successfully negotiated and that the drinker has come to a
decision, however tenuously or inconsistently held, that he or she must
do something about the problem with alcohol. This is implied in the title
of the self-help manual we have used in our own research on the topic,
"So You Want To Cut Down Your Drinking?" (Robertson & Heather,
1983).
The chief advantage in describing the use of self-help manuals as
"change without therapists" is that it thereby implies an attempt to
break free of the mould of the traditional, hospital-based treatment de-
livery system that has been inherited from the medical context in which
clinical psychology originated. To reiterate an obvious point, it is far
from clear from a social-learning perspective whether what we are deliv-
ering in self-help manuals is treatment or education, or whether the
distinction is in fact meaningless from this point of view. Thus, self-help
manuals can constitute one kind of response to the challenge issued by
Jim Orford, at the conclusion of his contribution to this volume, of
attempting to change our intervention practices to match the revolution
in our thinking about alcohol problems and addictive behaviors in gen-
eral (see Orford, Chapter 5, this volume).
In this chapter I will first discuss reasons for the upsurge of interest
in self-help manuals for problem drinkers, before describing the results
of previous research in this area. I will then give a brief account of some
research conducted at the Addictive Behaviours Research Group into
the effectiveness of a controlled drinking self-help manual, including the
results of a recently completed one-year follow-up of a cohort of media-
recruited problem drinkers. I will conclude by mentioning other relevant
research projects just started, about to start or planned, and suggest
some possible directions for future research and practice.
REASONS FOR INTEREST IN SELF-HELP MANUALS
Given that the ideal of self-help has a very long history indeed (see
Robinson & Henry, 1977; Robinson, Chapter 14, this volume), why is so
much attention being paid to self-help manuals at the present time? The
most obvious reason for this interest, certainly on the part of govern-
ments and other holders of purse strings, is their relative cheapness
compared with more labor-intensive forms of intervention.
It now seems clear that, in most industrialized countries of the
world, health services have reached the limits of expansion, although
CHANGE WITHOUT THERAPISTS 333
whether or not this is a political inevitability cannot be debated here. In

Britain, certainly, most commentators would agree that there will never
be enough psychiatrists, clinical psychologists, nurses, or other profes-
sional helpers to respond adequately on a personal basis to all those
problem drinkers who might profit from their expertise. This is es-
pecially true if one abandons a narrow disease perspective on alcohol
problems, thereby considerably expanding the number of drinkers
whose alcohol-related difficulties are considered legitimate targets for
intervention-that "other world" of problem drinkers described by
Room (1980). As Bruce Ritson has pointed out (see Chapter 19), if this is
true of industrialized nations, it is even more relevant to the nonin-
dustrialized parts of the world and this has apparently been recognized
by the World Health Organization (see Grant, Chapter 3, this volume).
At the same time as the need for more cost-effective services has
become an economic imperative, there has emerged a specifically mod-
ern version of self-help in the sense of mutual aid. This phenomenon
has been described in detail by David Robinson (Chapter 14) and need
not be dwelt on here. As Robinson asserts, the expansion of mutual-aid
groups was butressed by the libertarian sentiments of the 1960s, with
their objection to bureaucratic social control, resistance to professional
dominance and imperialism, and the call for a greater degree of self-
determination in the response to life's problems (see, e.g., Illich, Zola,
McKnight, Caplan, & Shaiken, 1977). Related to this is an increasing
awareness of the dangers of iatrogenesis (Illich, 1977) and a rich so-
ciological literature on the deleterious effects of the labeling process and
its resultant stigmatizing properties (see, e.g., Rubington & Weinberg,
1968). It is not clear to what extent the use of behaviorally based self-
help manuals by individuals resonates with this lingering 1960s ide-
ology, but it does largely avoid iatrogenic and labeling possibilities. This
might be claimed as one of its main advantages.
This will be a convenient point to introduce a warning about the
dangers of uncritical enthusiasm for self-help approaches. In one of
those curious twists of which history seems peculiarly fond, it is pre-
cisely the radical arguments of the 1960s against institutionalization and
professional dominance that are now being used by conservative gov-
ernments to justify cutting back on essential medical and social services
(Sedgewick, 1982). With regard to problem drinking, for example, how-
ever successful self-help procedures are found to be, there will still exist
many problem drinkers who will continue to need specialized, face-to-
face help. This may be especially true of those who have incurred some
permanent neurological damage as a result of their drinking, because a
prominent aspect of cognitive impairment caused by alcohol is a reduc-
tion in the capacities for self-regulation and forward planning (see
334 NICK HEATHER
Robertson, Chapter 16), capacities that are obviously central to effective

self-management programs. It may be possible to think of other sound
reasons for retaining an individual, controlled drinking treatment model
for specified classes of problem drinkers. In any event, the conclusion is
that the use of self-help manuals must always be accompanied by critical
evaluative research and must never be employed as an excuse for the
second rate.
Although self-help has a long history, there is another feature of the
current situation that makes self-help manuals attractive, at least to psy-
chologists and allied professionals. This is the emergence over the last 15
to 20 years of a body of self-management theory and practice (see, e.g.,
Karoly & Kanfer, 1982; Thoresen & Mahoney, 1974), which has provided
us with a systematic and empirically based technology of behavior
change with exciting possibilities, not least for the construction of self-
help manuals. It remains to be demonstrated, however, that this rela-
tively sophisticated technology can add significantly to the effectiveness
of the spontaneous and naive boot-strapping efforts of problem drinkers
and other client groups.
Turning now to reasons for interest in self-help manuals more spe-
cific to the problem drinking area, I have already noted that a major
consequence of abandoning the disease perspective on alcohol problems
is that it expands our attention from the relatively small number of
individuals in society with high levels of neuroadaptation to alcohol to
the large number whose lives have in some way been adversely affected
by their drinking. It should be remembered that the preponderant atten-
tion paid to highly physically dependent alcoholics was recognized as a
major limitation of the disease theory by Jellinek (1960). The social learn-
ing alternative to the disease theory (see, e.g., Heather & Robertson,
1986) results in a disaggregative model for the recognition and classifica-
tion of alcohol-related problems (Room, 1981) and a problem-centered
approach to treatment and prevention (d. Thorley, 1980). For many of
the problem drinkers who are now regarded as legitimate targets for
intervention, total abstinence would be unacceptable or would be actu-
ally counterproductive (Polich, Armor, & Braiker, 1980; Sanchez-Craig,
1980). It is also assumed that, because many of these problem drinkers
are relatively less dependent on alcohol and are more likely to be able to
change their drinking behavior, less intensive methods of intervention
are appropriate. A further empirical question, however, concerns the
extent to which standardized self-help programs and other minimal in-
terventions can succeed in encompassing the diversity of specific prob-
lems that exist.
The poor cost-effectiveness of intensive, traditional treatment com-
pared with forms of minimal intervention, not just for low-dependence
problem drinkers but apparently for alcoholics too (Armor, Polich &
Stambul, 1978; Emrick, 1975; Orford & Edwards, 1977), is a well-worn
path that I do not intend to tread again here. (Some of this evidence is
reviewed by Ritson, Chapter 19). Nor do I wish to become embroiled in
the vexed question of whether or not "treatment" for alcoholism can be
said to "work." (For a full discussion of the issues arising from this
question, see Part II in Heather, Robertson, & Davies, 1985). Nev-
ertheless, a study recently completed in Dundee has some relevance to
this question, as least as far as a controlled drinking treatment goal is
concerned (Robertson, Heather, Dzialdowski, Crawford & Winton, in
press).
We compared 16 relatively recent and/or low-dependence problem
drinkers who had received a full program of individually tailored, cog-
nitive-behavioral therapy at a hospital-based, outpatient clinic with 21
who had received two or three sessions of evaluation, advice, and an
individualized set of drinking guidelines to take away. At follow-up an
average of 15.5 months later, there was no statistically significant dif-
ference in the proportions of successful controlled drinkers between the
two groups, but the intensive group showed a significantly greater
number of days of complete abstinence in the month before follow-up,
plus a significantly greater reduction in consumption. These findings
appear to conflict to some extent with those from a number of com-
parisons of intensive and minimal controlled drinking treatment (see,
e.g., Miller & Taylor, 1980; Miller, Taylor & West, 1980; Miller, Grib-
skov, & Mortell, 1981; Vogler, Weissbach, Compton, & Martin, 1977;
Berg & Skutle, Chapter 9, this volume). One might speculate about the
reasons for this apparent discrepancy (see Robertson et al., in press), but
the point for present purposes is that intensive controlled drinking treat-
ment should not be entirely abandoned while it may still prove cost-
effective for those certain types of recent or low-dependence problem
drinkers who are willing to attend hospital-based facilities. This willing-
ness is clearly much more than a question of the degree of seriousness of
the problem. These findings by no means detract from the validity of a
self-help approach; they merely reinforce the point made earlier about
the need for continuing critical evaluation of minimal interventions.
PREVIOUS RESEARCH ON CONTROLLED DRINKING SELF-HELP

MANUALS
It is a surprising fact that, in Glasgow and Rosen's (1978) influential

review of behavioral bibliotherapy, no research with problem drinkers
was listed and the area was mentioned only in passing. This may have
336 NICK HEATHER
been because of the persistence of abstinence as the dominant treatment

goal for alcoholism and the presupposition by psychologists that absti-
nence was not amenable to a bibliotherapeutic approach, although there
appears to be no convincing reason to suppose that this is indeed the
case.
With regard to self-help manuals in general, Glasgow and Rosen
made the important distinction between (a) manuals that were entirely
self-administered; (b) those where minimal contact with a therapist was
present, and; (c) therapist-administered programs where manuals were
employed in the context of regular meetings with a therapist. The au-
thors also complained that not enough evaluative research was being
conducted and argued that manuals should be evaluated under condi-
tions as close as possible to their intended use.
Since Glasgow and Rosen wrote, of course, an extensive program of
research into controlled drinking self-help manuals with low-depen-
dence problem drinkers has been conducted by William R. Miller and
his colleagues, now at the University of New Mexico, Albuquerque. A
first paper (Miller, 1978) reported the "serendipitous" finding that cli-
ents who had been randomly allocated at the termination of treatment to
receive a self-help manual outlining the principles on which their behav-
ioral self-control training had been based fared better at a 3-month fol-
low-up than those who had not. There then followed a series of studies
(Miller & Taylor, 1980; Miller et al., 1980; Miller et al., 1981) that were
designed directly to compare the effects of a self-help manual given on
its own with various versions of intensive behavioural treatment. These
studies all found that clients in the manual groups were at least as
improved as those in the formal treatment conditions. A more recent
report (Miller & Baca, 1983) extended the follow-up of this accumulated
cohort of clients to 24 months and discovered an overall improvement
rate in the manual group of 73%, which was equal to that found among
clients who had received behavioral self-control training on an indi-
vidual basis from a therapist.
By their very nature, the above findings rely on "proving the null
hypothesis" of no difference between minimal and intensive groups,
reflecting the origins of minimal interventions as control groups with
which the hypothesized effects of more intensive interventions could be
compared. The next stage in the research development of minimal inter-
ventions such as self-help manuals is clearly to hypothesize a superiority
of the manual to some control condition. This has in fact been done in
some unpublished research by Buck and Miller (1983). These authors
compared a self-help manual group to one involving only the self-
monitoring of alcohol consumption and a further control group of no-
treatment, waiting-list controls. Follow-up was restricted to the end of
the lO-week treatment phase because of the ethical requirement of offer-
ing control clients some form of treatment, in this case therapist-admin-

istered treatment in a group format. It was found that the bibliotherapy
group was superior to the other two on measures of consumption at the
10-week follow-up point but that this superiority disappeared after con-
trol clients had received treatment. It must be pointed out that there are
well-known problems with the use of waiting-list controls, who may
simply defer a decision to cut down drinking to the time when they
know they will begin receiving treatment.
Despite such problems, this program of research has clearly demon-
strated the general viability of a self-help approach with a controlled
drinking goal and has provided the necessary empirical justification for
offering self-help bibliotherapy to low-dependence problem drinkers.
Nevertheless, it does have certain limitations, mainly being unavoidable
consequences of the conventional service-delivery setting in which self-
help manuals were evaluated. First, most clients in bibliotherapy condi-
tions were self-referrals to an outpatient treatment clinic and were there-
fore presumably highly motivated to change. The same cannot be as-
sumed for all those who buy self-help manuals commercially or who
might be sent them through the post by helping agencies after respond-
ing, for example, to newspaper advertisements.
More importantly from the point of view of Glasgow and Rosen's
(1978) analysis, all clients were seen at least once by a therapist for
assessment purposes and were given self-monitoring cards to be filled in
and posted to the clinic each week. Thus minimal therapist contact was
present and it may well be this factor, rather than the self-help material
itself, that was largely responsible for the impressive results (d. Orford
& Edwards, 1977). In more practical terms, if it is intended to distribute
self-help manuals beyond the clinical setting, for example by making
them commercially available or offering them free through the post, the
research infringed Glasgow and Rosen's recommendation that manuals
be evaluated under conditions as close as possible to their intended use.
In short, despite the methological problems that would inevitably arise,
an evaluation of an entirely self-administered manual, without any ther-
apist contact whatever, is needed.
A further limitation of Miller's research was that the relatively small
numbers of clients taking part did not allow the investigation of such
variables as take-up rate (Le., the proportion of relevant problem drink-
ers who will take up use of the manual after having been exposed to it)
and follow-through rate (the proportion who follow through with advice
and instructions given in the manual after having taken it up). From a
cost-effectiveness perspective, these are clearly issues of some impor-
tance in evaluating the impact of attempts at large-scale self-help
interventions.
There are also unavoidable limitations in Miller's type of design
338 NICK HEATHER
with respect to the control groups used. Although Buck and Miller
(1983) showed that a self-help manual was superior to self-monitoring
alone and to a waiting-list control group, this comparison was restricted
for ethical reasons, as we have seen, to only 10 weeks follow-up. Clear-
ly, a longer-term evaluation is necessary in which the superiority of a
manual to some control intervention is predicted. It is also essential to
compare a self-help manual based on behavioral self-management prin-
ciples to a condition that controls for the nonspecific effects of bibli-
otherapy. In other words, it is necessary to establish whether it is the
self-management ingredients of a self-help manual that make for effec-
tive bibliotherapy or the act of reading any reasonably relevant and well-
intentioned material.
Finally, of course, although it may have been demonstrated that a
self-help manual for problem drinkers works in the United States, it
remains to be demonstrated that this effect is generalizable to other
cultures. It is possible to imagine reasons why such an approach might
not succeed quite as well in Scotland.
EVALUATION OF A SCOTTISH SELF-HELP MANUAL FOR MEDIA-

RECRUITED PROBLEM DRINKERS
For all the reasons that have been discussed, we decided to evaluate
a self-help manual with as large a number of presumptive problem
drinkers as it was possible to recruit through the national and local press
and without any direct contact with therapists. (It will not be possible to
give a full account of this research here and the interested reader is
referred to Heather, Whitton, & Robertson, 1986, and Heather, Robert-
son, Whitton, Allsop, & Fulton, in preparation.) The advertisement shown
in Figure 1 was placed in several Scottish national and local newspapers
during the early part of 1983.
All those who replied to the advertisement were sent, in strictly
alternate order, either a specially prepared self-help manual (Robertson
& Heather, 1983) produced by the Scottish Health Education Group
(SHEG) or a general information and advice booklet on alcohol problems
(Grant, Plant, & Saunders, undated) also available from SHEG. We had
originally intended to use Miller and Munoz's (1982) excellent manual
but eventually decided that something more culturally specific was
needed, reflecting Scottish customs, language, and attitudes. However,
the contents were similar to those of the Miller and Munoz manual and
included information on the effects of alcohol, an analysis of reasons for
drinking, instruction in self-monitoring, a guided functional analysis of
harmful and harmfree drinking occasions, instruction in limit-setting
DRINKING TOO MUCH?

If you have the feeling that you are drinking more alcohol
than you should, you can get free help and advice through
the post now.
This is not for alcoholics, but for people who genuinely feel
they should cut down on their drinking and are having some
difficulty doing so.
The advice which will be sent to you is absolutely free. This
is not a trick to try and sell you something. It's an official
health education project to help people reduce their drinking.
If you think you would benefit from drinking less, just write
your name and address in the space provided below, cut it
out and send it to SCOTTISH HEALTH EDUCATION
GROUP, FREEPOST, DUNDEE 001 9XW. No postage is
required if the word FREEPOST is clearly written on the
envelope
Please note that your name and address will be kept STRICT·
LY PRIVATE AND CONFIDENTIAL. No unauthorised
person will ever know that you have responded to this advert·
isement.
----------------------------,
I would like advice on how to cut down my drinking. SEND I
TO SCOTTISH HEALTH EDUCATION GROUP, FREE· I
I
POST, DON DEE 001 9XW (Block letters pleasel I
I
NAME (Mr./Mrs./Missl . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. I
I
AGE ............................................ :
I
ADDRESS........................................ I
I
.......•..................................•..... I
I
I
I ........................................... S.D.1 I
-----------------------------
FIGURE 1. Advertisement used in self-help manual evaluation.
and self-reinforcement, advice on methods of rate-reduction, an explo-

ration of functional alternatives to drinking, and some instruction on
relapse prevention. As in the advertisement (Figure 1), it was stressed
that the manual was not for "alcoholics," defined crudely as those who
had experienced significant withdrawal symptoms, and other classes of
drinkers who were better advised to attempt or to maintain total absti-
nence (e.g., those with organic damage, "recovering alcoholics," and
pregnant women) were described. A list of addresses of helping agen-
cies was provided for those who might feel the need for more person-
alized help. The control booklet satisfied the requirement of controlling
for the nonspecific effects of bibliotherapy; it contained no specific in-
340 NICK HEATHER
structions on how to cut down drinking but did include a list of ad-
dresses that readers could use to obtain further help if needed.
The possibility that this research was subject to unavoidable meth-
odological problems has already been mentioned. Before proceeding to
the results, therefore, it will be convenient to consider these problems
here. The first stems from the fact that the main type of information on
outcome was derived, of necessity in a large-scale postal project of this
kind, from respondents' self-reports of alcohol consumption and other
measures. The validity of such self-reports is, of course, a contentious
issue and has been reviewed by Midanik (1982). A frequent conclusion is
that, despite a tendency for the heaviest drinkers to underreport con-
sumption, problem drinkers' self-reports are reasonably valid and gen-
erally suitable for research purposes (e.g., Maisto, Sobell, & Sobell,
1979).
This conclusion has been recently challenged by Watson, Tilleskjor,
Hoodecheck-Schow, Pucel, and Jacobs (1984) who, after a review of
some of the relevant literature and a report of some new data, end by
proposing a moratorium on the use of self-reports in follow-up studies
of alcoholism treatment. In my view, this suggestion is greatly overstat-
ed and is not even warranted by the authors' own data. With regard to
the present study, the question becomes, Is there any convincing reason
to suppose that any invalidity in self-reports would differ systematically
between the two groups being compared? particularly in view of the fact
that the study was shown to be double-blind, in the sense that no
subject at 6-month follow-up claimed to be aware of the existence of an
alternative experimental condition. It could conceivably be argued that
control group subjects were less likely to be satisfied with the materials
received and therefore less likely to try to please researchers by report-
ing a favorable outcome. Indeed, a specific question at the 6-month
follow-up point revealed that a significantly higher proportion of man-
ual than control group subjects were "satisfied" with what they had
received. Manual group subjects were also significantly more likely to
say that they had enjoyed their booklet and that it had helped them,
although not more likely to say that they had finished reading it. More-
over, in the sample as a whole, there were low but significant correla-
tions (p < 0.01) between these variables and percentage reductions in
weekly consumption from baseline. Thus, it is conceivable that manual
group subjects were responding to an unknown extent to demand char-
acteristics of the experimental situation.
Despite this problem of interpretation, however, there appears to
be no alternative to the use of self-reports in research of this kind, in
view of the expense of attempting to interview collaterals or obtain
blood samples from subjects so widely geographically dispersed. We did
obtain such confirmatory data on a small sample of subjects living in a

circumscribed area (Glasgow) at the one-year follow-up point, but this
limited evidence does not alter the self-report basis of our main conclu-
sions. Thus, although it is important to remain cautious in interpreting
self-reported data, the only alternative is to give up doing this kind of
research and leave the effectiveness of widely distributed self-help man-
uals unevaluated.
Given that problem drinkers' self-reports in general have created
methodological difficulties, it might be thought that self-reports col-
lected soley through the post, with no personal contact whatever with
research subjects, would be especially contentious. There appears to be
little previous research that is directly relevant to this issue. In a longitu-
dinal study of changes in problem drinking, Clark and Cahalan (1976)
reported little evidence of differences related to alcohol use or drinking
problems between respondents personally interviewed and those who
returned mailed questionnaires. These authors refer to a study by
Hochstim (1967), who found that data collected from personal inter-
views, telephone interviews, and mailed questionnaires were "virtually
interchangeable." Nevertheless, in view of obvious possible objections
to mail-only data, we decided to interview a subsample of respondents
by telephone, on the ground that the rapport presumably present in
such interviews might make for more valid information. The telephone
subsample also provided an opportunity to collect more extensive
data-for example, regarding the degree of alcohol dependence-than
could reasonably be obtained from the mailed questionnaire, which was
deliberately kept as short as possible so as not to tax the respondent's
patience. An inducement of 5 pounds was offered for each of these
telephone interviews. As we shall see, the decision to include a tele-
phone subsample led to some complications in the interpretation of
results, although it also yielded a potentially very interesting finding.
A third problem has to do with the high rates of attrition that must
be expected in this kind of research. Of the 785 individuals who re-
sponded to the advertisement and were sent a self-help manual or con-
trol booklet, 538 either failed to return assessment questionnaires or
refused to be interviewed by telephone. Fortunately, these dropouts
were almost equally distributed among the two groups, with 127 re-
spondents remaining in the manual group and 120 in the control group.
Nevertheless, this rate of attrition would be disastrous if we were con-
cerned with survey research and the statistics of point estimation, where
we wished to estimate the true value of a certain variable, say the pre-
vious week's alcohol consumption in a given population, by obtaining
data from a representative sample of that population. But this, of course,
is not at all the object of the research. Rather, we were concerned to test
342 NICK HEATHER
a specific hypothesis, that receiving a self-help manual after responding

to a newspaper advertisement would be more beneficial than receiving a
nonspecific control booklet. It is therefore the statistical logic of hypoth-
esis testing that is relevant. Given that the two groups did not differ on
any relevant variable in the initial assessment data, the manual and
control groups are statistically comparable. Obviously, we can have no
idea of the impact of self-help materials in the entire sample of 785
respondents and can make no statement regarding the relative effective-
ness of the materials sent among this larger sample. However, confining
attention to the 247 respondents who returned assessment question-
naires, a valid test of the hypothesis under examination is possible. I
have devoted some space to this elementary issue because there appears
to have been some misunderstanding about it.
Of the 247 left in the sample, a further 115 failed to respond or could
not be contacted at the 6-month follow-up. On this occasion, there was a
significant difference (p < 0.05) in the proportions remaining in the two
groups, with 78 (59.1%) in the manual group and 54 (40.1%) in the
control group. At the one-year follow-up point, 137 of the 247 initially
contacted had been lost and, of those remaining, 63 (57.3%) were in the
manual group and 47 (42.3%) were in the control group (p < 0.05). These
follow-up attrition rates create more serious problems of interpretation,
despite the fact that there was no significant difference between groups
on initial measures at either follow-up point and also no significant
interaction between the two variables manual versus control group and
followed-up versus not followed-up. Nevertheless, the fact that signifi-
cantly more respondents were contacted in the manual than in the con-
trol group must be taken into account. If the usual assumption is made
that those lost to follow-up tended to do less well than those remaining
in the sample, it is important to note that this would work against the
hypothesis under examination and would tend to minimize any pOSSible
superiority in effectiveness of the self-help manual. On the other hand,
if it is assumed that the lost respondents did better than the remainder,
this would favour the hypothesis and possibly lead to a spuriously sig-
nificant difference in favor of the manual group. As we shall see, there
are grounds for speculating that the lost respondents may have had a
superior outcome.
It should also be noted that the initial contact rate of 31.3% could be
regarded as the minimum take-up rate of self-help materials and it is
interesting that there was little difference between the two groups in this
respect. On this basis of lessons learned from the past experience, ways
might be devised to increase the numbers of respondents who return
questionnaires. For example, it might be possible to make the provision
of materials conditional on the prior receipt of a completed question-
naire. Another suggestion is to make this equivalent to an entry in a
lottery with an inviting first prize. Similarly, the proportion of 53.4%

who responded to the 6-month and the 44.5% to the one-year follow-up
could be regarded as minimum follow-through rates and, again, ways of
increasing these might be tried in future research. It is a finding of this
research that the differential contact rate was significantly higher in the
manual group at 6 months and thereafter remained roughly constant.
This suggests that, although the initial take-up rate of the self-help man-
ual was no better than that of the control booklet, its follow-through rate
was superior.
A final problem, which is not specific to self-help research but to all
research on controlled drinking treatments, is how satisfactorily to de-
fine the favorable outcome of controlled drinking. There is a specific
implication for self-help manuals, however, because the definition used
for controlled drinking must obviously be reflected in the advice given to
problem drinkers regarding the setting of daily and weekly limits. The
limits recommended in our manual, and the definition for a controlled
drinking outcome used in the research, were 50 standard units of alco-
hol per week for men and 35 for women.
It is apparent from an international perspective that these limits are
often regarded as alarmingly high. One source of confusion here is due
to the fact that different measures are being used for standard units and
this should be deared up without further delay. Our standard unit is
based on one half pint of 4% alcohol-content beer, which contains 1.05
d. pure ethyl alcohol. (Because British and American fluid ounces differ,
it is best to ignore this unit of measurement in calculations.) This is
roughly equivalent to an average British pub measure of spirits (Y6 gill,
0.94 d.) or table wine (1.04 d.) or fortified wine (1.12 d.). It is therefore
convenient to regard these different drinks as equivalent in alcohol con-
tent and as equalling one d. ethyl alcohol or one standard unit. This
contrasts with the American standard unit (e.g., Miller & Munoz, 1982),
which contains 1.5 d. alcohol. Thus our male and female weekly limits
are roughly equal to 33 and 24 American standard units, respectively.
The limits we used for men were roughly equivalent to those pro-
posed in a report which has proved highly influential in Britain, by the
Royal College of Psychiatrists.(1979). It was further employed to define
"heavy drinking" in a large, nationwide survey of British drinking hab-
its by Wilson (1980), who also used 35 units to define female heavy
drinking. The evidence for using these particular limits appears to be
that, for men at least, they represent the point at which risk of drinking
symptoms, particularly alcohol-related tremor, begins to rise signifi-
cantly (Armor et al., 1978; Edwards, Chandler, Hensman, & Peto, 1972)
and the level at which the risk of liver damage begins to increase (Pe-
quignot, Chabert, Eydoux, & Courcoul, 1974).
Although these particular limits are somewhat higher than those
344 NICK HEATHER
given in other advice to the general public (Health Education Council,

undated), they do have the advantage of being achievable, especially in
a sample over which the researchers had so little direct influence. More
conservative goals may have demoralized some problem drinkers and
hence proved counterproductive. It is also important to note that read-
ers of the manual were advised to abstain on at least one day a week but
preferably more, thus decreasing the likelihood that they would drink to
the maximum permitted limits. Moreover, it was stressed that the limits
given were maximum upper bounds and not norms for drinking. Other
guidelines given in the manual for the distribution and frequency of
drinking meant that the maximum limit of 50 or 35 units would be
difficult to achieve if instructions were properly followed and, indeed,
this seems to be borne out by the quantities being consumed by suc-
cessful responders at follow-up.
Despite these arguments, it would appear that our limits are still
higher than other workers in the field would recommend. In a survey of
relevant authorities, Anderson, Cremona, and Wallace (1984) asked
what limits workers felt were appropriate in advice to the public and
plotted a frequency distribution of the replies. There were two modes in
the distribution of limits for men, one at 20 units and one at about 36
units. Bowing to this consensual pressure, therefore, we have lowered
out recommended limits in a revised version of the self-help manual
(Roberton & Heather, 1985) to 35 for men and 20 for women. Nev-
ertheless, we would continue to insist that the placement of recom-
mended limits should be based on empirical evidence and not on moral,
perhaps "neo-prohibitionistic", sentiments or the personal preferences
of the authors of self-help manuals.
SIx-MoNTH FOLLOW-UP RESULTS

Six months after receiving their materials through the post, the 247
respondents remaining in the sample were sent a follow-up question-
naire and 132 replies were received. Measures recorded, which had also
been taken at initial assessment, included standardized, scaled scores on
seven factors relevant to treatment evaluation that had been identified
and refined in an extensive program of statistical analysis at the Fort
Logan Mental Health Center (e.g., Foster, Horn, & Wanberg, 1972) and
the Human Factors Laboratory, University of South Dakota (Ellingstad,
1977; Swenson & Clay, 1980). These factors were Marital Problems; Con-
trol of Drinking Problems; Income/Employment Stability; Physical
Health and Well-being; Residential Stability; Social Interaction; Control
of Drinking. In addition, respondents were asked a series of questions
about their reaction to the booklet sent them-whether it had been what
was expected, how much had been read, whether the respondent had
been helped by it, and so on. Weekly consumption was recorded by a
self-completion method adapted from Chick and Duffy (1981). At initial
assessment, respondents in the telephone subsample had also been
given the shortened Michigan Alcoholism Screening Test (MAST) (Sel-
zer, Vinokur, & van Roisjen, 1975) and the Edinburgh Dependence
Schedule (Chick, 1980). It was established that the evaluation was dou-
ble-blind, in that no respondent claimed to be aware of an alternative
experimental manipulation and no telephone interviewer was aware of
which group the respondent belonged to.
Changes in weekly consumption from initial to 6-month follow-up
for manual and control groups are shown in Figure 2. (This figure is
based only on subjects who provided data at both 6-month and one-year
follow-up points.) On an analysis of covariance, the manual group
showed a significantly greater reduction in consumption than the con-
trol group (p < 0.05, one-tailed test). When changes in drinking were
expressed as percentages of individual initial levels, the difference be-
tween groups was significant (Mann-Whitney, p < 0.05, one-tailed test),
with the manual group showing a mean reduction of 40.2% and the
control group of 25.2%. Moreover, when 30 respondents who had stated
that they had obtained an alternative form of treatment after having
received the self-help materials were excluded from the analysis, on the
ground that their changes in consumption may have been primarily due
to an extraneous influence, the absolute reduction in mean consumption
was significantly different in favor of the manual group (p < 0.05).
70
c 60
0
"i =35)
-------
E= 50 Control Group (N
::I 0
a8
'" .c
40
.",<C....0
(,J-
~
Manual Group (N =51)
81 1I 30
3: .-c
8-
'" :::::I
20
'!
A.
10
Initial Six Month One Year

Assessment Follow-up Follow-up
FIGURE 2. Changes in consumption for manual and control groups.

346 NICK HEATHER
Finally, the manual group showed significantly greater improvements

on the Physical Health and Well-being and Control of Drinking Prob-
lems (p < 0.05). Thus, there is evidence that the self-help manual pro-
duced greater reductions in consumption among those remaining in the
sample than the control booklet, accompanied by improvements on ob-
viously relevant variables.
Figure 3 shows changes in consumption separately for the postal
and telephone subsamples. Independently of any differences in changes
between manual and control groups, the telephone subsample showed
significantly greater improvements on Control of Drinking Problems (p
< 0.001), Physical Health and Well-being (p < 0.05), and Residential
Stability (p < 0.05). However, despite the fact that the mean percentage
reduction for the telephone subsample was 48.4% compared with 28.6%
in the postal subsample, there were no significant differences for abso-
lute or percentage reductions in consumption. The telephone subsample
did contain a significantly greater proportion of respondents who had
reduced their drinking by more than 10 units per week (Chi-square, p <
0.05).
There are, of course, at least two ways of explaining the apparently
superior outcome in the telephone subsample. It could be due to the
demand characteristics of the interview situation, in view of the pre-
sumed rapport with the interviewer and the fact that respondents were
being paid to take part. On the other hand, it could be a valid superiority
resulting from a greater degree of contact with researchers. That re-
search interviews in general may have some kind of therapeutic effect is
70
,,
a
.+:;
60
,,
Co
E=
0
50
:::I
88
"
, Postal Subsample (N .. 61)
40 , , Telephone Subsample (N .. 25)

(J-
.. <C -------
i~ 30
..
;t .-
&-
c
:::)
20
.;
I!!
Q.
10

FIGURE 3. Changes in consumption in postal and telephone subsamples.
now well recognized (e.g., Sobell & Sobell, 1981) and, furthermore,
evidence from the weight-reduction and smoking-cessation fields has
suggested that minimal interventions may be made more effective by
some amount of therapist contact (Brownell, Heckerman & Westlake,
1978; Glasgow, Schafer, & O'Neill, 1981), although there is conflicting
evidence here (Jeffrey, Danaher, Killen, Farquhar, & Kinnier, 1982).
Nevertheless, this possibility will be returned to.
A curious, incidental finding at the 6-month point was that, when
those remaining in the. sample were compared with those who had
dropped out, the latter were found to show significantly higher scores
on the factors Income/Employment Stability (p < 0.001) and Residential
Stability (p < 0.01), and were more likely to be married (p < 0.05) and in
current employment (p < 0.05), whereas their scores on Social Interac-
tion were nearly significantly higher. These variables are almost defini-
tive of social stability, a construct with a long history in alcoholism
research (Straus & Bacon, 1951). The finding is surprising because, in
terms of conventional treatment evaluation, it is those showing lower
social stability who are less likely to be successfully contacted at follow-
up (Sobell, Sobell, & Ward, 1980). On a purely speculative basis, it may
be reasoned that, because higher social stability is associated with a
better prognosis in treatment (see, e.g., Gibbs & Flanagan, 1977) and
with a greater probability of spontaneous recovery (Tuchfeld, 1976),
respondents lost to follow-up in this study improved rapidly without
resorting to the materials received and therefore felt no need to remain
associated with the project. On the other hand, if the more usual as-
sumption is made that those lost to follow-up tended to have a poorer
outcome, this suggests that we are in a very different situation from
conventional treatment evaluation, but also raises the possibility that
mailed self-help materials may be especially appropriate for problem
drinkers of relatively low social stability. My own guess is that the low-
response and follow-up rates encountered in this project were due to the
fact that advice advertised in the media is particularly attractive to those
who are most sensitive to the stigmatizing propensities of the formal
treatment process. Thus the more "respectable" respondents, whether
they had benefitted from the materials sent them or not, tended to avoid
any further involvement.
It has already been pointed out that the unobserved outcome of
those lost to follow-up is crucial to the interpretation of the results of this
study, in view of the differential follow-up rates in manual and control
groups. If those lost to follow-up had a poorer outcome, as is usually the
case in conventional treatment evaluation, then the observed superiority
of the manual over the control group is probably a valid result and may
even be based on an underestimate of this superiority. Conversely, if for
348 NICK HEATHER
reasons connected with a higher level of social stability, dropouts had a

better outcome than those who were contacted, then the observed supe-
riority of the manual group may well be spurious.
A last type of finding from the 6-month follow-up must be briefly
mentioned. Although the self-help manual dearly stated that it was not
intended for those experiencing serous problems, some of these did
remain in the sample and it is of interest to know what became of them.
Two criteria were used to define a more serious problem: first, those in
the total sample drinking over 100 units per week at initial assessment (n
= 25) were distinguished from those drinking at or below this level (n =
102); secondly, respondents in the telephone subsample giving evidence
of "late dependence" (n = 10) on the Edinburgh Dependence Schedule
(Chick, 1980), which includes severe restlessness, morning tremor, and
relief drinking, were divided from those showing early or no depen-
dence (n = 22). Preliminary analyses indicated that both these criteria
were associated in the sample with greater alcohol-related problems and
other relevant variables.
At the 6-month point, high consumers showed a significantly great-
er reduction in consumption (139.3 to 63.3) than the low consumers
(44.9 to 32.4) (p < 0.05). When percentage scores were used, the signifi-
cantly greater improvement on the part of high consumers was main-
tained. It must be recognized immediately that these results are dearly
subject to artifacts arising from the way in which groups were formed-
a floor effect that applies to the absolute reduction scores and a regres-
sion towards the mean effect that applies to both absolute and percent-
age reduction measures. However, the same objections do not apply to
the other criterion of seriousness, the stage of dependence on alcohol,
although here the small numbers entering the analysis becomes a prob-
lem. There was no significant difference in reduction measures between
late dependence and early/no dependence respondents. Previous
week's consumption among late dependence respondents was reduced
from 78.3 to 24.8 units, whereas among early/no dependence re-
spondents it was reduced from 56.4 to 43.3 units. (See Figure 4, which
applies only to those giving data at both follow-up points.) These results
cannot be explained by any greater tendency for serious problem drink-
ers, under either definition, to become total abstainers or to seek help
from other sources. Although the numbers on which the analysis was
based are very small, there is also some evidence that high consumers
benefitted relatively more from the manual than the control booklet, in
that a significantly greater proportion reduced drinking from over to
under 100 units per week (Fisher-Yates Exact Test, p < 0.05). Certainly,
the evidence suggests that the beneficial effects of both kinds of material
80
70
g 60
'i
E=
:::I 0
50
1!!i5
o u
= 18)
. Early/No Dependence (N
(J-
.. e(
~
....0 40
.-
----
• fJ
~
c 30 Late Dependence (N = 7)
"'::;)
5-
.~ 20
a..
10
0
FIGURE 4. Changes in consumption in late dependence and early/no dependence groups.
were not limited to those drinking relatively less or showing only early
dependence on alcohol.
ONE-YEAR FOLLOW-UP RESULTS
Altogether 110 individuals responded to the one-year follow-up, 63

from the manual and 47 from the control group. As previously stated,
the difference in proportions between the two groups was statistically
significant (p < 0.05). However, as at the 6-month point, there were no
significant differences between the groups on initial measures and no
significant interaction between group and whether or not respondents
were successfully followed-up. The total of 110 respondents includes 22
who were not contacted at 6 months and 88 who provided data on all
three testing occasions.
There are at least two ways of analyzing the one year outcome data.
The first restricts attention to changes between initial assessment and
one-year follow-up, using all 110 respondents with data at both points.
Mean previous week's consumption had been reduced from 59.6 to 39.9
units in the manual group and from 64.2 to 46.5 units in the control
group, but the difference in absolute reductions was not significant on
an analysis of covariance whereas the difference in percentage reduc-
350 NICK HEATHER
tions between groups was also nonsignificant. As at the 6-month point,

respondents were asked whether they had received any other form of
help since the last contact, and those who had replied positively on
either occasion were excluded from the analysis. When this was done,
the difference in absolute reductions in consumption between groups
was statistically significant (p < 0.01). The manual group also showed a
significantly greater improvement in Social Interaction (p < 0.01) and a
nearly significantly greater improvement in Marital Problems (p < 0.10).
Comparing the two subsamples, there was no significant difference
in consumption measures, but the telephone respondents showed sig-
nificantly greater improvements in Control of Drinking Problems, Phys-
ical Health and Well-being, Residential Stability, and Social Interaction
and a nearly significantly greater improvement in Marital Problems (p <
0.10). Thus the apparently superior outcome for the telephone subsam-
pIe observed at 6 months had spread to a higher number of relevant
variables at one year. .
With regard to the analysis of more serious problems, there were no
significant differences between high and low consumption groups for
both absolute and percentage reduction measures. A similar finding
appeared for the "dependence" criterion of seriousness; there was no
significant difference between early/no dependence and late depen-
dence groups, irrespective of the kind of material they received. (See
Figure 4.) Incidentally, there was no tendency at the one-year point for
respondents who had dropped out since the 6-month follow-up to ap-
pear more socially stable than those who remained in the sample.
The other way to analyze outcome at one year is to consider only
those 88 respondents who provided data at all three points. Figures 2
and 3 give mean previous week's consumption scores over time for
manual and control groups, and postal and telephone subsamples.
These figures show that, for all groups and samples, there is relatively
little change in consumption between 6 months and one year, and that
the gains that had been made at 6 months were retained at one year.
However, on a repeated measures analysis of variance, there were no
significant differences between groups or subsamples. On the other
hand, the manual group showed a significantly greater improvement in
Social Interaction compared with the control group, whereas the tele-
phone subsample showed significantly greater gains in Control of
Drinking Problems, Physical Health and Well-being, and Social Interac-
tion compared with the postal subsample. Changes in consumption for
telephone respondents with all three sets of data are shown according to
late or early/no dependence in Figure 4, which also makes clear the
small numbers on which these data are based.
It has already been mentioned that, as part of the one-year follow-
up, an attempt was made to conduct personal interviews with a subsam-

pie of respondents in which blood samples would be taken and collat-
eral information collected. All those giving addresses in the greater
Glasgow area (n = 52) were sent a letter inviting them to take part in a
personal interview and offering 5 pounds for doing so. Thirty-eight
(73.1 %) replied and, of these, 20 (38.0% of the Glasgow total) agreed,
with 13 coming from the manual group and 7 from the control group. A
fuller account of the results of these interviews is given elsewhere (see
Heather, Robertson, et aI., in preparation).
The 20 respondents eventually interviewed are clearly a very lim-
ited and in many ways unrepresentative subsample of the problem
drinkers who participated in this study. Bearing this firmly in mind,
however, the results give some grounds for confidence in the reliability
and validity of self-reported data collected through the post or by tele-
phone. First, only two respondents, one in each group, showed a
positive blood alcohol concentration (BAC) at interview and, in the case
of the manual group respondent, this was consistent with a high level of
self-reported consumption. Second, although the rank-order correlation
between self-reported consumption and gamma-glutamyl-transpep-
tidase (GTP) was low (0.35) and nonsignificant, in only two cases was a
self-report of under recommended limits ar.companied by a gamma-GTP
of over 50 international units. The limitations of gamma-GTP in the
screening of problem drinkers have been noted by Chick, Kreitman, and
Plant (1981). Third, in no case was a collateral report of recent consump-
tion seriously discrepant from the respondent's self-report and, finally,
self-reported estimates of previous week's consumption showed a high
degree of correspondence with estimates made for the same week in the
personal interview on the basis of intensive questioning. There are ob-
vious grounds for believing that the Glaswegian respondents who
agreed to be interviewed were likely to have been more honest about
their drinking that those who did not agree, but these results at least
suggest that a substantial proportion of the self-reported data gathered
in the main follow-ups was sufficiently valid.
Besides a check on reliability and validity, a further reason for con-
ducting the personal interviews was an attempt to determine whether
the apparent superiority of results in the telephone subsample was a
genuine effect. In the event, the interview subsample was too small to
allow any conclusion in the regard to be reached. This problem must
await a further study, specifically designed with the problem in mind.
The interview data also covered several other areas of interest and a
paper describing different ways in which the self-help manual was put
to beneficial use is currently being prepared (Allsop, Heather, & Fulton,
in preparation).
352 NICK HEATHER
SUMMARY OF FINDINGS
The most important findings of the research concern the comparison

between manual and control groups. There was evidence from the 6-
month follow-up that, among those remaining in the sample, the manual
produced greater absolute and percentage reductions in alcohol con-
sumption and that this was accompanied by greater improvements in
physical health and well-being and in degree of social interaction. When
respondents who said they had received other forms of treatment since
receiving self-help materials were excluded from the analysis, the manual
group showed a significantly greater absolute reduction in mean con-
sumption. At the one-year follow-up, differences in reductions between
groups, in both absolute and percentage terms, were not statistically
significant. However, when those who claimed to have received other
treatment at any time after initial contact were excluded, manual group
respondents again showed significantly greater absolute reductions.
The interpretation of these results is complicated by two factors.
First, because of the reliance on unsupported self-report data, there is a
possibility that manual group respondents were subject to greater de-
mand characteristics of the experimental situation, a possibility that
arises because of their greater degree of satisfaction with the materials
sent them. Second, because of significantly different rates of contact
between the groups, the validity ascribed to the observed superiority of
the manual group depends to a great extent on the unobserved outcome
of those respondents who dropped out of the study. In view of the
greater social stability of these lost respondents at 6-month follow-up, it
is possible that this observed superiority could be spurious.
Another important set of findings concern differences in outcome
between postal and telephone subsamples, which occurred indepen-
dently of differences in changes between groups. Although the greater
mean reductions in consumption in the telephone subsample failed to
reach statistical significance at either follow-up point, telephone re-
spondents showed greater improvement on a variable measuring alco-
hol-related problems at the 6-month point and also greater improve-
ments in physical health and well-being and even residential stability.
At the one-year follow-up, this superiority in outcome had extended to
degree of social interaction. There are two ways of explaining this ob-
served superiority: An artifactual consequence of the greater demand
characteristics associated with a telephone interview and with payment
for taking part; or a valid result reflecting a therapeutic effect of follow-
up interviews. A further experiment is needed to investigate these com-
peting explanations.
Comparisons between outcomes for serious and less serious prob-
lem drinkers also form an important aspect of the results. Under both
criteria of seriousness-initial consumption above 100 units per week
and some evidence of late dependence-serious problem drinkers were
shown to have reduced drinking to an extent at least equal to that of less
serious problem drinkers at the 6-month and the one-year follow-up
points. However, the interpretation of these findings is again compli-
cated, in one case by the method by which groups were formed and, in
the other, by the small numbers entering the analysis. Bearing these
difficulties in mind, it is probably safe to conclude that, on the evidence
of the study, the potentially beneficial effects of self-help materials are
not entirely confined to relatively low-consumption or early dependence
problem drinkers. There was also some tentative evidence to suggest
that the specially prepared self-help manual was superior to the control
booklet in assisting high consumers to reduce consumption.
Finally, in the sample as a whole and irrespective of the kind of
material received, considerable improvements in adjustment were ob-
served. As an illustration of this point, among respondents who re-
ported drinking above recommended limits at initial assessment and
who were successfully contacted at the 6-month follow-up, fully 60%
had reduced their drinking to below these limits and by at least 10 units
per week. The mean percentage reduction in consumption in the total
follow-up sample at 6 months was roughly 34% and significant improve-
ments were observed for all the other measures of adjustment. Ob-
viously, without a no-treatment control group, these gains cannot be
logically attributed to the materials sent. It may be, for example, that
simply taking the trouble to respond to a newspaper advertisement is
symbolic of a shift from contemplation to action and that the process of
spontaneous remission, however it is accomplished, would itself have
resulted in the improvements observed without any additional help.
Nevertheless, the results of the study give some grounds for optimism
in the effectiveness of minimal interventions of this kind and at least
suggest the potential usefulness of further research into controlled
drinking self-help manuals that bridge the gap between alcohol educa-
tion and treatment.
OTHER RESEARCH PLANNED OR IN PROGRESS
We would defend the design of the study, despite the meth-

odological problems which arose, on the ground that it provided a basic
starting point for the evaluation of a controlled drinking self-help man-
ual used in its natural setting. It is clear, however, that some interesting
and unresolved questions remain after the completion of the study.
354 NICK HEATHER
An experiment is now being planned in which greater control

would be exerted over information relevant to the effects of self-help
materials. The sample studied would be more highly selective and con-
fined to a limited geographical area, thus allowing personal interviews
and collateral data to be collected at follow-up. All respondents would
be asked to agree to these conditions before entering the study and a
lottery system could be used to encourage higher rates of successful
follow-up. The opportunity would be also taken to examine what is
perhaps the most interesting hypothesis to emerge from the previous
experiment, that some form of additional, minimal contact significantly
enhances the effectiveness of a self-help manual. Thus a group that
received only assessment telephone interviews at follow-up would be
compared with one that received monthly, supportive telephone inter-
views in which progress was systematically reviewed. Given sufficient
numbers, a further group could simply be asked to provide monthly
reports of progress to a telephone answering service. This project is still
in the planning stage.
Meanwhile, other projects investigating the effects of various
modes of self-help in different contexts are underway or about to begin
at the Addictive Behaviours Research Group. One involves the evalua-
tion of the DRAMS Scheme in a general practitioner setting. DRAMS,
which stands for Drinking Reasonably and Moderately with Self-Con-
trol, was developed by the Scottish Health Education Group on the basis
of our self-help manual (Robertson & Heather, 1983) to provide GPs
with an interactive and cost-effective method of responding to low-
dependence problem drinkers encountered in their practice. At the ini-
tial consultation, the possibility of a problem with alcohol is raised with
the patient, a blood test is taken for the measurement of gamma-GTP
and mean cellular volume, and the patient is given a self-monitoring
card to take away. At the second consultation 2 weeks later, both sets of
data are reviewed with the patient and a shortened version of the self-
help manual is provided with instructions on how best to use it. Further
appointments are then made as necessary. .
The viability and attractiveness of the scheme to GPs and patients
has been demonstrated in a pilot project in the Scottish Highland region
but this was not a controlled evaluation. In our study, after excluding
individuals with too high levels of dependence and others who are not
suitable for a controlled drinking goal, patients are randomly allocated
to one of three conditions: (a) DRAMS; (b) an advice group in which the
GP gives strongly worded advice on the need to cut down but no further
contact is initiated; (c) a nonintervention control group. So far, 104 pa-
tients have entered a trial and a blind, 6-month follow-up including
biochemical and collateral data has begun.
Another project about to start concerns the relative effectiveness of

various levels of minimal intervention among low-dependence problem
drinkers whose alcohol consumption has either caused or aggravated
their medical condition. Estimates of the proportion of such individuals
found on medical wards range from 10% to 30% (see, e.g., Jarman &
Kellet, 1979; Quinn & Johnston, 1976). Research by Chick, Lloyd, and
Crombie (1984) found that patients given a 30- to 60-minute counseling
interview by an experienced nurse before discharge from hospital fared
better at one-year follow-up than those who had not (see Ritson, Chap.
18, this volume). Our study will compare the following conditions: (a) a
group that receives an assessment interview only; (b) a group that re-
ceives a structured interview lasting about 45 minutes and based on self-
management principles; (c) a group that, in addition, receives a self-help
booklet to take away that encapsulates the conclusions arrived at in the
structured interview; (d) a group that receives, as well as the preceding
procedures, monthly supportive home visits. Blind follow-up, which
again will include data confirmatory of self-report, will be at 6 months
after discharge.
A novel feature of this study is that it will involve parallel investiga-
tions of minimal interventions directed at a controlled drinking and an
abstinence outcome, taking both medical opinion and the patient's
wishes into account. If medical opinion allows some continued drinking
but the patient prefers abstinence, the latter goal will be paramount. If,
on the other hand, medical opinion demands abstinence but the patient
insists despite strongly worded warnings on continuing to drink, the
patient is excluded from the study and offered controlled drinking treat-
ment on an individual basis, allowing the goal to be revised in the
future. In the majority of cases, of course, no disjuction between medical
advice and the patient's preference is anticipated. The decision as to goal
will be arrived at before random allocation to the minimal intervention
condition. This is the first time, to our knowledge, that a minimal inter-
vention has been directed at total abstinence.
These projects only skim the surface of possible research in the area
of self-help manuals and minimal interventions generally. (For a fuller
discussion, see Heather, 1986.) We know very little about the kinds of
problem drinker who are most suited to this kind of approach and the
best methods by which they may be recruited and encouraged to take
advantage of self-help procedures. At the same time, we need more
information on the types of problem drinker for whom self-help ap-
proaches are likely to be ineffective and who are best channeled into
traditional, face-to-face therapeutic situations. The variables of degree of
seriousness of alcohol-related damage and of degree of alcohol depen-
dence are most obviously relevant to this issue but other variables may
356 NICK HEATHER
be involved. The limits of application of minimal interventions cannot

merely be assumed on a priori grounds but must be empirically
determined.
A whole collection of variables bear on the issue of the cost-effec-
tiveness of the means by which the problem drinker's own efforts at self-
help can be encouraged and assisted. These variables include the
number of clients who can be reached; the proportions who take up,
follow through with, and respond positively to different vehicles for the
promotion of self-help; the cost of preparing, distributing, or commu-
nicating self-help materials; and the level of training and amount of time
spent by professional or paraprofessional workers who may be involved
in the dissemination of self-help procedures. As well as written mate-
rials, a range of other media, including audiotapes, videotapes and
home computer programs, can be envisaged (see Christenson, Miller, &
Munoz, 1978).
Finally, it is attractive to imagine a series of self-help programs
ordered rationally according to Prochaska and DiClemente's model of
the change process, with qualitatively different types of material, with
different aims, being directed at the Precontemplation, Contemplation,
Action and Maintenance stages.
ACKNOWLEDGEMENTS
I am grateful to Ian Robertson, Barbara Whitton, Steven Allsop,

Archibald Fulton, and all my other colleagues who made valuable con-
tributions to the research described in the chapter.
REFERENCES
Allsop,S., Heather, N., & Fulton, A. (in preparation). Case studies in the use of a controlled
drinking self-help manual. Alcohol Studies Centre, Paisley College of Technology.
Anderson, P., Cremona, A., & Wallace, P. (1984). What are the safe levels of alcohol
consumption? British Medical Journal, 289, 1657-1658.
Wiley.
Brownell, K D., Heckerman, C. L., & Westlake, R. J. (1978). Therapist and group contact
as variables in the behavioral treatment of obesity. Journal of Consulting and Clinical
Buck, K, & Miller, W. R. (1983). Why does bibliotherapy work? A controlled study.
Unpublished manuscript, Department of Psychology, University of New Mexico.
Chick, J. (1980). Alcohol dependence: Methodological issues in its measurement; reliability
of the criteria. British Journal of Addiction, 75, 175-186.
Chick, J., & Duffy, J. (1981). Drinking survey. Edinburgh: Alcohol Research Group, Univer-
sity of Edinburgh.
Chick, J., Kreitman, N., & Plant, M. (1981, June 6). Mean Cellular Volume and Gamma-
Glutamyl-Transpeptidase as markers of drinking in working men. Lancet, pp. 1249-

125l.
Chick, J., Lloyd, G., & Crombie, E. (1984). Counselling problem drinkers in medical
wards: A controlled study. British Medical Journal, 290, 965-967.
Christenson, A., Miller, W. R., & Munoz, R. F. (1978, May). Paraprofessionals, peers,
partners, paraphernalia and print: Expanding mental health service delivery. Profes-
sional Psychology, pp. 249-269.
Clark, W. B., & Cahalan, D. (1976). Changes in problem drinking over a four-year span.
Edwards, G., Chandler, J., Hensman, c., & Peto, J. (1972). Drinking in a London suburb:
II. Correlates of troubles with drinking among men. Quarterly Journal of Studies on
Alcohol 33 (Suppl. No.6), 94-119.
Ellingstad, V. S. (1977). LAIICSQ scoring manual. Vermillion, SO: Human Factors Laborato-
ry, Department of Psychology, University of South Dakota.
Emrick, C. D. (1975). A review of psychologically oriented treatment of alcoholism: II. The
relative effectiveness of different treatment approaches and the effectiveness of treat-
ment versus no treatment. Quarterly Journal of Studies on Alcohol, 36, 88-108.
Foster, F. M., Horn, J. L., & Wanberg, K. W. (1972). Dimensions of treatment outcome: A
factor-analytic study of alcoholics' responses to a follow-up questionnaire. Quarterly
International Journal of the Addictions, 12, 1097-114l.
Glasgow, R. E., & Rosen, G. M. (1978). Behavioral bibliotherapy: A review of self-help
behavior therapy manuals. Psychological Bul/etin, 85, 1-25.
Glasgow, R. E., Schafer, L., & O'Neill, K. (1981). Self-help books and amount of therapist
contact in smoking cessation programs. Journal of Consulting and Clinical Psychology, 49,
659-667.
Grant, M., Plant, M., & Saunders, W. (undated). Drinking and alcohol problems in Scotland.
Edinburgh: Scottish Health Education Group.
Health Education Council. (undated). That's the limit: A guide to sensible drinking. London:
Health Education Council.
Heather, N. (1986). Minimal treatment interventions for problem drinkers. In G. Edwards
(Ed.), Current issues in clinical psychology. London: Plenum Press.
Heather, N., & Robertson, I. (1986). Problem drinking. Harmondsworth: Penguin Books.
Heather, N., Robertson, I., & Davies, P. (1985). (Eds.). The misuse of alcohol: Crucial issues in
dependence, treatment and prevention. London: Croom Helm.
Heather, N., Robertson, I., Whitton, B., Allsop, S., & Fulton, A. (in preparation). Evalua-
tion of a controlled drinking self-help manual: One year follow-up results. Addictive
Behaviours Research Group, Department of Psychiatry, University of Dundee.
Heather, N., Whitton, B., & Robertson, I. (1986). Evaluation of a self-help manual for
media-recruited problem drinkers: Six month follow-up results. British Journal of
Hochstim, J. R. (1967). A critical comparison of three strategies of collecting data. Journal of
the American Statistical Association, 62, 976-989.
IIIich, I. (1977). Limits to medicine. Harmondsworth: Penguin Books.
IIlich, I., Zola, I. K., McKnight, J., Caplan, J., & Shaiken, H. (1977). Disabling professions.
London: Marion Boyars.
Jarman, c. M. B., & Kellett, J. M. (1979). Alcoholism in the general hospital. British Medical
Journal, 285, 469-472.
Jeffrey, R. W., Danaher, B. G., Killen, J., Farquhar, J. W., & Kinnier, R. (1982). Self-
358 NICK HEATHER
administered programs for health behavior change: Smoking cessation and weight
reduction by mail. Addictive Behaviors, 7, 75-63.
Jellinek, E. M. (1960). The disease concept of alcoholism. New Haven, CT: Hillhouse Press.
Karoly, P., &'Kanfer, F. (1982). Self-management and behavior change: From theory to practice.
New York: Pergammon Press.
Maisto, S. A., Sobell, L. c., & Sobell, M. B. (1979). A comparison of alcoholics' self-reports
of drinking behavior with reports of collateral informants. Journal of Consulting and
Midanik, L. (1982). The validity of self-reported alcohol consumption and alcohol prob-
lems: A literature review. British Journal of Addiction, 77, 357-382.
Miller, W. R. (1978). Behavioral treatment of problem drinking: A comparative study of
three controlled drinking therapies. Journal of Consulting and Clinical Psychology, 46, 74-
86.
directed controlled drinking training for problem drinkers. Behavior Therapy, 14, 441-
448.
Miller, W. R., & Munoz, R. F. (1982). How to control your drinking: A practical guide to
responsible drinking (rev. ed.). Albuquerque, NM: University of New Mexico Press.
Miller, W. R., & Taylor, C. A. (1980). Relative effectiveness of bibliotherapy, individual
and group self-control training in the treatment of problem drinkers. Addictive Behav-
iors, 5, 13-24.
Miller, W. R., Taylor, C. A., & West, J. C. (1980). Focused versus broad spectrum behavior
Miller, W. R., Gribskov, c., & Mortell, R. (1981). The effectiveness of a self-control manual
for problem drinkers with and without therapist contact. International Journal of the
Addictions, 16, 829-839.
Orford, J., & Edwards, G. (1977). Alcoholism. London: Oxford University Press.
Pequignot, G., Chabert, c., Eydoux, H., & Courcoul, M. A. (1974). Increased risk of liver
cirrhosis with intake of alcohol. Review of Alcoholism, 20, 191-196.
treatment. Santa Monica, CA: Rand Corporation.
Quinn, M. A., & Johnston, R. V. (1976). Alcohol problems in acute male medical admis-
sions. Health Bulletin, 34, 253-256.
Robertson, I., & Heather, N. (1983). So you want to cut down your drinking? A self-help manual
for controlled drinking. Edinburgh: Scottish Health Education Group.
Robertson, I., & Heather, N. (1985). So you want to cut down your drinking? A self-help guide to
sensible drinking (rev. ed.). Edinburgh: Scottish Health Education Group.
Robertson, I., Heather, N., Dzialdowski, A., Crawford, ]., & Winton, M. (in press). A
comparison of minimal versus intensive controlled drinking treatment for problem
drinkers. British Journal of Clinical Psychology.
Robinson, D., & Henry, S. (1977). Self-help and health: Mutual aid for modern problems.
London: Martin Robertson.
Room, R. (1980). Treatment-seeking populations and larger realities. In G. Edwards & M.
Grant (Eds.), Alcoholism treatment in transition (pp. 205-224) London: Croom Helm.
Room, R. (1981). The case for a problem-prevention approach to alcohol, drug and mental
problems. Public Health Reports, 96, 26-33.
Royal College of Psychiatrists. (1979). Alcohol and alcoholism. London: Tavistock.
Rubington, E., & Weinberg, M. S. (Eds.). (1968). Deviance: The interactionist perspective.
New York: Macmillan.
Sanchez-Craig, M. (1980). Random assignment to abstinence or controlled drinking in a
cognitive behavioral program: Short term effects on drinking behavior. Addictive Be-
haviors, 5, 35-39.
Saunders, W. M., & Kershaw, P. W. (1979). Spontaneous remission from alcoholism: A
community study. British Journal of Addiction, 74, 251-265.
Sedgewick, P. (1982). Psycho politics. London: Pluto Press.
Selzer, M. L., Vinokur, A., & van Roisjen, L. (1975). A self-administered short MAST.
Sobell, L. c., Sobell, M. B., & Ward, E. (Eds.). (1980). Evaluating alcohol and drug abuse
treatment effectiveness: Recent advances. New York: Pergammon Press.
Sobell, M. B., & Sobell L. C. (1981). Frequent follow-up data gathering and continued care
with alcoholics. International Journal of the Addictions, 16, 1077-1086.
Straus, R., & Bacon, S. D. (1951). Alcoholism and social stability: A study of occupational
integration in 2023 male clinic alcoholics. Quarterly Journal of Studies on Alcohol, 12,231-
245.
Swenson, P. R, & Clay, T. R (1980). Effects of short-term rehabilitation on alcohol con-
sumption and drinking-related behaviors: An eight-month follow-up study of drunk-
en drivers. International Journal of the Addictions, 15, 821-858.
Thorley, A. (1980). Medical responses to problem drinking. Medicine, 135, 1816-1822.
Thoresen, C. E., & Mahoney, M. J. (1974). Behavioral self-control. New York: Holt, Rinehart
& Winston.
Tuchfeld, B. (1976). Changes in patterns of alcohol use without formal treatment. Research
Triangle Park, NC: Research Triangle Institute.
Vogler, R E., Weissbach, T. A., Compton, J. V., & Martin, G. T. (1977). Integrated
Watson, C. G., Tilleskjor, c., Hoodecheck-Schow, E. A., Pucel, J., & Jacobs, L. (1984). Do
alcoholics give valid self-reports? Journal of Studies on Alcohol, 45, 344-348.
Wilson, P. (1980). Drinking habits in the United Kingdom. Population Trends, 22, 14-18.
18
Dependence and Compulsion
Experimental Models of Change
HOWARD RANKIN
INTRODUCTION
Prochaska and DiClemente (1982) suggest that behavioral techniques are

most influential and valuable to the action and maintenance stages of
their proposed model of change. The work reported on here suggests
that they do indeed have great value at these stages, but also suggests
that benefits can be derived at all levels and phases of realizing, coming
to terms with, and really doing something about, the perceived prob-
lem. It is likely that, like the research reported here, behavioral treat-
ments have effects on attitudes, expectations, and cognitions, and have
utility not only in the action and maintenance phase, but in the early
contemplative stages of the process.
ALCOHOL DEPENDENCE
In recent years the notion of alcohol dependence has begun to re-

place more specifically disease-oriented concepts of alcoholism. Numer-
ous models of dependence and ways of measuring it have been pro-
duced, but the one that has attracted most widespread interest and
generated most research has been Edwards' Alcohol Dependence Syn-
HOWARD RANKIN • 5t. Andrew's Hospital, Billing Road, Northhampton NN1 5DG
England.
361
362 HOWARD RANKIN
drome. This syndrome, as posited by Edwards (1977), represents an

"observable coincidence of phenomena" and, overall, a hypothetical,
psychophysiological condition. Most of the features of the proposed
syndrome are now well known and are included in Table 1.
At the time of its presentation, the Alcohol Dependence Syndrome
as described had several advantages over contemporary thinking. It
consisted of empirical, testable assumptions that seemed more coherent
and consistent than previous definitions and took into account those
facts suggesting that the disorder is not a single, monolithic entity but a
continuous and individually colored process.
However, the Alcohol Dependence Syndrome is not without its
critics. First, there is resistance in some quarters to the medico-political
overtones of the model. Critics argue that the term syndrome gives the
model a degree of medical respectability that is not warranted. More-
over, such a term is seen by some as an attempt to keep what is ostensi-
bly a psychosocial problem clearly in the realm of the medical profes-
sion. These critics (e.g., Shaw, 1980) continue to argue that there is no
need to make generalized quasi-medical statements when the problem is
an individual, psychological one. It is also argued that what is really
being described is merely the escalation of drinking and that, to all
intents and purposes, the degree of dependence merely corresponds to
the amount drunk.
Most of these criticisms can be answered by appeal to research
findings. Hodgson (1980), for one, has pointed out that the model is
worth preserving if it has utility and value and a number of studies do,
indeed, show that the model has predictive utility. For example, using
the Severity of Alcohol Dependence Questionnaire (Stockwell, Hodg-
son, Edwards, Taylor, & Rankin, 1979) and Hodgson, Stockwell, and
Rankin (1979) were able to show behavioral differences between those
rated severely and those rated moderately dependent in their response
to priming doses of alcohol. Similar work has been conducted by Fun-
TABLE 1.
Features of the Alcohol
Dependence Syndrome
Narrowing of repertoire
Salience of drink seeking behavior
Increased tolerance to alcohol
Repeated withdrawal
Relieve/avoidance of withdrawal
Subjective awareness of compulsion to drink
Rapidity reinstatement after abstinence
DEPENDENCE AND COMPULSION 363
derburk and Allen (1977), showing different behavioural responses in a

drinking situation between those rated severely and moderately depen-
dent. Further studies have shown differences in drinking speeds
(Rankin, Hodgson, & Stockwell, 1980) differences in response to drink-
ing cues (Rankin, Stockwell, & Hodgson, 1982) and to internal states
(Laberg, 1984) between those differing in their degree of dependence. In
addition to these empirical studies conducted in the laboratory, outcome
studies suggest that degree of dependence is a predictive factor. Both
Orford, Oppenheimer, and Edwards (1976) and Polch, Armor, and
Braiker (1980) demonstrated degree of dependence as a predictive vari-
able in treatment outcome.
One difficulty with the measure of dependence used by the pre-
viously cited studies is that they rely heavily on only some aspects of the
proposed dependence syndrome, notably those related to tolerance and
withdrawal. One criticism of these studies therefore is that they focus
almost exclusively on the neurobiological elements of the proposed syn-
drome, whereas the cognitive and behavioral elements are largely ig-
nored and unmeasured. The relationship between the psychological and
physiological components of the dependence syndrome has always
been a thorny problem. On the one hand, it is unreasonable to stress the
separateness of one from the other, whereas joining them together in
one concept de-emphasizes the individual contribution of each. Studies
have demonstrated, for example, that tolerance and withdrawal are
themselves environmentally determined (e.g., Siegel, 1975) and it is
clear that psychological and behavioral determinants have their physio-
logical correlates.
A more refined version of the dependence syndrome has been
posited in a World Health Organization (1980) memorandum. Here the
term dependence is substituted by the term neuroadaptation. In this, the
physiological and behavioral dimensions of dependence are seen to be
independent but related correlates that co-vary with each other. How
they co-vary is a matter of research.
Even given these somewhat academic limitations of measurement
and conception, the question remains as to how dependence, as cur-
rently assessed, relates to other variables at a clinical level. Some re-
search now suggests that dependence is a dimension which is indepen-
dent of the social problems and consequences of alcohol abuse. For
example, Meyer, Babor, Esselbrock, Esselbrock, and Kaplan (1985) have
produced factor analytic studies that demonstrate that the degree of
dependence is distinct from the consequences of heavy alcohol abuse.
Because it is the consequences and problems that arise as a result of
heavy use that represent the main clinical difficulties, it is thus impor-
tant to realize that dependence is just one part of the presenting com-
364 HOWARD RANKIN
plaint. Dependence per se is not necessarily a clinical problem. Equally,

people who are in clinical difficulties with alcohol use may not be depen-
dent. Similar research (e.g. Skinner, personal communication) suggests
that the degree of dependence may be partly independent of the quan-
tity and frequency of drinking. In short, these studies suggest that the
dependence syndrome as currently assessed represents only one dimen-
sion of the clinical presentation of alcohol-related problems.
Although it may be that the overall dimension of dependence is
independent of one scale of alcohol-related problems, it may also be that
extreme degrees of dependence are related to difficulties and that severe
dependence is related to loss of flexibility or plasticity of response and
behavior. From a behavioral viewpoint, Rankin et al. (1982) demon-
strated that people rated severely dependent endorsed the fact that
more cues elicited their drinking behavior and that these cues more
frequently and intensively exerted an effect on behavior. This is sup-
ported by research and other clinical surveys which suggest, for exam-
ple, that many severely dependent individuals are easily "primed up"
by a drink or two, and once blood alcohol rises above a particular level,
difficulties in stopping are encountered. This indeed suggests that the
blood alcohol level itself is the potent internal cue for those who are
rated severely dependent, in a way it is not for other individuals differ-
ing in their dependence.
DEPENDENCE AND COMPULSIONS
The fact that more cues frequently and more intensively exert their
effect on behavior is an important definition, not only of severe depen-
dence, .but also of compulSiveness. This raises the whole issue of the
relationship between dependence and compulsion and focuses attention
on the individual's inability to exert control, the development of stereo-
typed behavior, and the loss of flexibility in response. This raises further
questions about dependence and compulsion. Clearly, they are not the
same phenomenon, but do actually share similar consequences, notably
a loss of flexibility in behavior. The search for a generic concept of
dependence would seem to rest on this varying dimension of plasticity
and consequent compulsion. However, a word of caution is necessary
here about relabeling behaviors that have compulsive elements as de-
pendence behaviors. It may be reasonable to call those who suffer the
consequences of their alcohol consumption "alcohol dependent" and
those who suffer consequences as a result of heroin use "opiate depen-
dent," but is it reasonable to call someone who is suffering an obsessive-
compulsive hand-washing ritual "soap dependent"? Among other con-
ditions that have been subsumed under the catch phrase of dependence
are eating disorders, and in particular anorexia nervosa, which clearly
has compulsive elements; anorexia is, in my view, not a dependence at
all.
TREATMENT STRATEGIES
If severity of dependence does feature compulsive elements, one

way of reversing this degree of dependence and treating it where it is
problematic would be to use those methods known to be of value in
deconditioning compulsive behavior. One way of viewing continued
drinking in the face of priming doses of ethanol and rising blood alcohol
levels is to see this as an avoidance response aimed to put off or delay
the onset of minimum withdrawal symptoms consequent on the blood
level falling towards zero. This does place continued drinking and in-
ability to stop once started, which is an essential difficulty of those who
are severely dependent, in the same category of response as hand wash-
ing is to the obsessive-compulsive. In short, it is a discriminated oper-
ant. The development and subsequently successful use of cue ex-
posure/response prevention methodology for the treatment of
compulsive behaviors has been well researched and documented (e.g.,
Rachman & Hodgson, 1980). The possibility remains that such ex-
posure/response prevention treatment could be used with severely de-
pendent individuals in order to discuss the intense relationship between
cues and inflexible behaviors, thus restoring more control to the
individual.
Some initial work on the cue exposure/response prevention meth-
odology with clients suffering alcohol related problems has showed
some promise (e.g., Hodgson & Rankin, 1976) and there exist in the
literature a number of case studies (e.g., Hodgson & Rankin, 1981;
Rankin, 1982). More recently Rankin, Hodgson, and Stockwell (1983)
reported a controlled experiment using such methodology and it is that
experiment on which I now wish to focus. I will do so because some
further light can be shed on the process by which experimental changes
were effected.
CUE EXPOSURE: A CONTROLLED STUDY
The subjects of this experiment were 10 inpatients, all assessed as

suffering from a severe degree of dependence and all of whom reported
being "primed up" by initial doses of ethanol. All 10 subjects received
366 HOWARD RANKIN
six sessions of cue exposure, although five received six initial control
sessions consisting of imagined resistance to alcohol. Cue-exposure (CE)
sessions consisted of subjects drinking an initial amount of ethanol,
which typically raised their blood alcohol concentration to between 65-
100 mg. %, and then resisting an available third drink for 45 minutes.
The amount of temptation in the cue exposure session was maximized
by asking subjects to continually interact with the third drink to be
resisted by, for example, holding the glass in their hand, putting it to
their lips, and smelling the alcohol. Details of the design, the experimen-
tal procedure, and the control sessions are given in Figures I, 2, and 3.
For a more detailed assessment of the procedures, the reader is referred
to Rankin et al. (1983).
MEASURES
The measures taken included not only actual objective ratings of the
subject's current feelings but measures of how the subject expected to
feel. These expected scores were elicted at the beginning of each session
when the subject was informed of the content of the up-coming experi-
mental period and what was expected of him. In addition, physiological
measures of pulse, tremor, and blood alcohol concentration were
recorded.
•
2 Control 2 Experimental
• • Group 2
1 Experimental
• - - - - - - - . Group 1
Behavior Behavior Behavior

Test 1 Test 2 Test 3
Time
FIGURE 1. Summary of experimental design. Reprinted with permission from Behaviour

Research and Therapy, 21 (3), "Cue exposure and response prevention with alcoholics: A
controlled tria!." Pergamon Press, 1983.
Subjective
Ratings:
Expected ALCOHOL Subj. CUE Subj. CUE Subj. CUE Subj.
Subjective-- CONSUMPTION ) Ratings ) EXPOSURE ) Ratings ) EXPOSURE ) Rating~ EXPOSURE ) Ratings
Ratings Pulse 3 minutes Pulse 3 minutes Pulse 3 minutes Pulse
Pulse Tremor each: Tremor each: Tremor each: Tremor
Tremor eye contact SAL eye contact SAL eye contact SAL
holding holding holding
sniffing sniffing sniffing
plus misc. plus misc. plus misc.

chat chat chat
= 15 minutes = 15 = 15
minutes minutes
FIGURE 2. Flow chart of operations in the experimental condition. Reprinted with permission from Behaviour Research and Therapy, 21 (3), "Cue
exposure and response prevention with alcoholics: A controlled trial." Pergamon Press, 1983.
Subjective
Ratings:
Expected Subjective Subjective Subjective

Subjective Discussion Imaginal Ratings Imaginal Ratings Imaginal Ratings
Ratings of Resistance Pulse Resistance Pulse Resistance Pulse
Pulse - - - - - - - l• Resistance ~ 3 scenes, ) Tremor ) 3 scenes ) Tremor ) 3 scenes ) Tremor
Tremor Scenes 3 minutes 3 minutes 3 minutes
BAL each each each
plus misc. plus misc. plus misc.
chat = chat = chat =
15 minutes 15 minutes 15 minutes
FIGURE 3. Flow chart of operations in the control condition. Reprinted with permission from Behaviour Research and Therapy, 21 (3), "Cue exposure
and response prevention with alcoholics: A controlled trial." Pergamon Press, 1983.
RESULTS
The key measures used to determine the experimental effect were

behavioral, subjective, and physiological in nature. The behavioral mea-
sure was time taken to consume standard amounts of alcohol in a test
that was administered before and after each condition. Such a measure
has previously been validated as ·an indicator of degree of desire for
taking an alcoholic drink (e.g., Rankin et al., 1980). In addition, a variety
of subjective measures, including the desire for a drink, ability to resist,
anxiety, and temptation were also recorded at various frequent times
throughout each session. Physiological readings, in particular pulse,
hand tremor, and blood alcohol concentrations, were also taken
throughout the sessions. A significant treatment effect was found on the
main behavioral measure, as demonstrated in Table 2. This shows that
the cue exposure condition resulted in much greater and significant
decrements on this measure than the control condition. In addition,
there were significant reductions on the measures of Desire for a Drink
and Difficulty to Resist across the cue exposure condition and, in gener-
al, the findings were of significant decrements in the experimental con-
dition on all subjective measures.
Despite modifications of the subjective and behavioral measures, no
significant effects were found at all across the control or experimental
conditions on the physiological measures thus recorded. For a more
detailed exposition of the results, the reader is once again referred to
Rankin et al. (1983).
DISCUSSION
The results presented here suggest that cue exposure methodology

does indeed produce significant decrements in behavioral and subjec-
TABLE 2.
Time Taken to Consume Standard Dose of Alcohol on Behaviour Test (in Sec)
Total time 1st-2nda ,b

probabilities 2nd-3rd
1st Behav. 2nd Behav. t test 3rd Behav. Behav. c
test test (transformed) test test
Group 1 X 220.0 507.8 <0,02

SD 24D.4 355.8 «0.01)
Group 2 X 243.4 303.0 NS 682.2 0.01
SD 189.6 226.6 (NS) 301.4 (0.01)
a1st-2nd Behav. test = experimental effect Group 1.

"1st-2nd Behav. test = control effect Group 2.
'2nd-3rd Behav. test = experimental effect Group 2.
370 HOWARD RANKIN
tive measures. One of the interesting aspects of this experiment is to ask

how the observed changes were effected. In the relatively short space of
time, was this primarily physiological, cognitive, or behavioral change?
It has been posited, for example, that behavioral treatments work by
effecting changes in the following order. First, physiologically condi-
tioned responses are changed, which then results in a reduction of
avoidance behavior and, some time thereafter, subjective and attitudinal
changes take place. This suggestion was made by Watson, Gaind, and
Marks (1971), with the delay in attitudinal-subjective change being con-
sidered to be something of a "cognitive lag." Along this tack, it might be
interesting to speculate that a more conventional psychotherapeutic ap-
proach would affect its change in a different order. With such a treat-
ment, one might expect subjective changes first, leading to behavioral
changes and ultimately physiological change.
In this experiment, no observed physiological change took place
whatsoever. It would have been exciting to have been able to demon-
strate modification to either pulse rate or tremor, even in the presence of
a standard dose of alcohol across sessions, but this was not observed. As
far as can be ascertained, there were no significant physiological
changes at all. An alternative view, therefore, is that this is a basically
cognitive procedure and that the repeated experience of being able to
resist available alcohol when "primed-up" is actually changing the sub-
jects' negative expectations about their ability to cope with the situation.
In short, one might be modifying expectations to the point where cue
exposure subjects, who did not formerly see themselves as being able to
cope satisfactorily, have cognitions changed as a result of the treatment
itself. Some confirmation of the notion of modified expectancies can be
derived by looking at the data on subjects' expectations on the subjective
measures of Desire for a Drink and Difficulty to Resist. These demon-
strate that, for both measures, expectancies are significantly reduced
within the cue exposure sessions for Group 2 (CE only) but not for
Group one (control sessions). Comparisons between cue exposure and
control conditions barely miss statistical significances (probability typ-
ically being around 0.07). More support for the suggestion that the cue
exposure effect is mediated by subjective change can be found in the fact
that a rankorder correlation of 0.86 (p < .01) was found between the
change on the measures of time taken to consume alcohol in the behav-
ioral test and changes on the subjective measure (in this case Difficulty
to Resist). What this demonstrates is that those subjects who showed
the most change on their subjective scores also tended to demonstrate
the most change in their behavioral tests.
In addition, from an anecdotal point of view, what some subjects
were actually reporting was a change of attitude about their inability to
cope in such a situation. After the experimental debriefing, many sub-
jects reported initially being very apprehensive about the procedure and
their ability to cope, but by the end were feeling more confident and
optimistic.
It is not unreasonable to suggest, therefore, that the cue exposure
methodology described here is working through a largely cognitive ef-
fect. Despite the artificiality of the hospital environment in which the
treatment was conducted, the fact that the subjects actually had to drink
alcohol and exercise their coping strategies in the presence of alcohol
and, indeed, under the influence of it, makes this treatment more real-
istic than conventional verbally oriented treatments. Because one of the
major problems of any therapy is getting a generalization from the
therapeutic setting to real life, any such simulation would seem to be
valuable. In the case of addiction, of course, the problems of state-
dependent learning come into play and it is not unreasonable to suggest
that individuals would need to learn coping responses in the same state
as they will need to actually implement them. In this case, that involves
being under the influence of a small amount of alcohol and learning
relapse prevention procedures in that state, rather than in a sober one.
Moreover, given the evidence about possible cognitive deficits in sub-
jects who use alcohol regularly (see Robertson, Chapter 16, this vol-
ume), there may be real advantages in actually getting subjects practic-
ing coping strategies and implementing them in this sort of simulated
setting, rather than merely talking about proposed strategies.
Finally, further information needs to be elicited about the mecha-
nism of action of any therapeutic cognitive change. Data from the cur-
rent experiment show that expectations were modified, particularly in
the Group 2 experimental subjects. Moreover, not only were changes in
expectation on one measure related to changes in expectation on others,
but they were also positively correlated (in some cases significantly) to
changes in actual subjective scores.
The exact relationship between the actual and predicted scores re-
mains to be elucidated. Figure 4 shows the relationship for the experi-
mental condition of Group 2. Given the fact that the predictions were
made at the beginning of each session, it could be hypothesized that
discrepancies between actual and expected scores in anyone session
would be related to changes in the following sessions. One hypothesis
flowing from a cognitive model of therapeutic change would be that
where actual scores are lower than expected scores, decrements on actu-
al subjective ratings will accrue in the following session. Where actual
scores are higher than expected, an increment in the following session
would be hypothesized. Given this model, group data that showed
significant decrements in actual response, as was the case here, should
be characterized by graphs that, by and large, show actual responses
lower than expected ratings. The graphs in Figure 4 broadly support this
372 HOWARD RANKIN
Max 10 Immediately After Alcohol 10 15 Mins After
8 8
6 6
Group
Mean
Ratings
4 4
2 2
Min
1 234 5 6 2 3 4 5 6
Max 10 30 Mins After 10 45 Mins After
8 8
6 6
Group
Mean
Ratings
4 4
2 2
Min
2 3 4 5 6 1 2 345 6
Experimental Sessions
- - - Expected - Actual
FIGURE 4. The relationship between expected and actual scores for Group 2 experimental
condition. Reprinted with permission from Behaviour Research and Therapy, 21 (3), "Cue
exposure and response prevention with alcoholics: A controlled trial." Pergamon Press,
1983.
view. However, the graphs also show, against the general hypothesis,
that response decrements in actual ratings occur in sessions following
occasions where actual ratings were higher than predicted. To further
explore this line of thinking, individuals' data were examined across
each reading of the actual and expected ratings of desire in the experi-
mental condition. This examination revealed that on only 32% of occa-
sions did the actual expected discrepancies in one session lead to the
predicted changes on actual ratings in the following sessions. In short,
the discrepancy between actual and expected scores on one occasion did
not satisfactorily predict changes in actual ratings from that occasion to
the next.
Equally plausible is that actual/expected differences result in modi-
fications of expectancies. Using the same technique as on the data men-
tioned previously, it was found that on 61 % of occasions expectations
changed in the predicted direction from one session to the next, depend-
ing on their relationship with the actual rating in the previous session.
In other words, the expected/actual discrepancies were better related to
subsequent changes in expectation than to subsequent changes in actual
scores. Of course, much of what determined expectations in the current
experiment is not available for analysis. Quite apart from individual
daily variations and fluctuations, hard data are lacking on what hap-
pened after the experimental session. How subjects coped in the hours
following the experimental session might reasonably be expected to in-
fluence their predictions about any future sessions.
Although the specific, no doubt complex, relationship between ex-
pectations and actual realities remains to be untangled, the evidence
here suggests that, as in other studies, both are modified in the response
prevention setting.
In conclusion then, cue exposure treatment of the nature described
here seems to be a useful adjunct to treatment and has demonstrable
effects on subjects' behavior and cognitions. It is likely that these effects
are mediated by cognitive changes that help to change the client's view
towards a more positive, confident conception of both his coping ability
and ultimate treatment outcome.
ACKNOWLEDGMENTS
With many thanks also to Mrs. Jean Cox who typed the manuscript.
REFERENCES
Edwards, G. (1977). Alcohol Dependence Syndrome: Usefulness of an idea. In G. Edwards

& M. Grant (Eds.). Alcoholism: New knowledge and new responses (pp. 16-28). London:
Croom Helm.
374 HOWARD RANKIN
Funderburk, F., & Allen R. (1977). Alcoholics' disposition to drink: Effects of abstinence
and heavy drinking. Quarterly Journal of Studies on Alcohol, 38, 410-425.
Hodgson, R. (1980). The Alcohol Dependence Syndrome: A step in the wrong direction?
Hodgson, R., & Rankin, H. (1976). Cue exposure in the treatment of alcoholism. Behaviour
Research and Therapy, 14, 305-307.
Hodgson, R., & Rankin, H. (1981). Cue exposure and relapse prevention. In P. Nathan &
W. Hay (Eds.). Clinical case studies in the behavioral treatment of alcoholism. New York:
Plenum Press.
Hodgson, R., Stockwell, T., & Rankin, H. (1979). Alcohol dependence and the priming
effect. Behaviour Research and Therapy, 27, 379-387.
Laberg, J. (1984, August). Alcohol and expectancies: Somatic and subjective responses to expectan-
cies and intake of alcohol in severely, moderately and non-independent drinkers. Paper pre-
sented at the Third International Conference on the Treatment of Addictive Be-
haviours, North Berwick, Scotland.
Meyer, R. E., Babor, T. F., Esselbrock, M., Esselbrock, V., & Kaplan, F. F. (1985). New
directions in the assessment of the alcoholic patient. In L. Tolle (Ed.), Proceedings of
NIAAAIWHO collaboration centre designation meeting: Alcohol research seminar. Wash-
ington DC: U.S. Department of Health and Human Sciences.
409-418.
Polich, M., Armor, D., & Braiker, H. (1980). The course of alcoholism: Four years after treat-
ment. Santa Monica, CA: Rand Corporation.
Prochaska, J., & DiClemente, C. (1982). Transtheoretical therapy: Toward a more inte-
grative model of change. Psychotherapy: Theory, Research and Practice, 19, 276-288.
Rachman, S., & Hodgson, R. (1980). Obsessions and compulsions. Englewood Cliffs, NJ:
Pren tice-Hall.
Rankin, H. (1982). Cue exposure and response prevention in South London. In P. Nathan
& W. Hay (Eds.), Clinical case studies in the behavioral treatment of alcoholism (pp. 227-
249). New York: Plenum Press.
Rankin, H., Hodgson, R., & Stockwell, T. (1980). The behavioural measurement of alcohol
dependence. British Journal of Addiction, 75, 43-47.
Rankin, H., Stockwell, T., & Hodgson, R. (1982). Cues for drinking and degrees of alcohol
dependence. British Journal of Addiction, 77, 287-296.
Rankin, H., Hodgson, R., & Stockwell, T. (1983). Cue exposure and response prevention
with alcoholics: A controlled trial. Behaviour Research and Therapy, 21, 435-446.
Shaw, S. (1980). A critique of the concept of the alcohol dependence syndrome. British
Siegal, S. (1975). Evidence from rats that morphine tolerance is a learned response. Journal
of Comparative and Physiological Psychology, 89, 498-506.
Stockwell, T., Hodgson, R., Edwards, G., Taylor, c., & Rankin, H. (1979). The develop-
ment of a questionnaire to measure severity of alcohol dependence. British Journal of
Watson, J., Gaind, R., & Marks, I. (1971). Rapid exposure in the treatment of phobic
disorders. British Medical Journal, 68, 247-251.
World Health Organisation. (1981). Nomenclature and classification of drug-and alcohol-
related problems: A WHO memorandum. Bulletin of the World Health Organisation,
59(2), 225-242.
19
Merits of Simple Intervention
BRUCE RITSON
INTRODUCTION
This chapter concerns the growth of an idea about the merits of simple
intervention. At the outset, I should own up to the bias of my own
vantage point, that of a psychiatrist working within the United Kingdom
National Health Service. It is important to acknowledge this particular
perspective because it has determined the clinical influences that I feel
have been important. Shedding firmly held beliefs about treatment is a
disconcerting and unnerving process, particularly when new beliefs
come to be held with equal tenacity. Are the new beliefs about simple
intervention based on convincing evidence or are they simply the latest
fashion waiting to be discarded for the next season's model? In 1977, The
Lancet commented on current approaches to services for alcoholics thus:
This treatment approach owes its existence more to historical process than to
science. It is possible to discern the deposits, akin to geological layers, of a
sequence of therapeutic fashions-the residue of almost forgotten enthusi-
asms for in-patient psychotherapy units, for group processes and the
therapeutic community, for family therapy and later for community psychia-
try. To say that treatment for alcoholism is only an accretion of fads and
fashions would be too harsh, for it is also built on much clinical experience;
but it must be admitted that we have not done enough to assess scientifically
the effectiveness of treatment methods. (Lancet, 1977, p. 489)
The history of the evolution of alcohol treatments during the past 20
BRUCE RITSON' Department of Psychiatry. University of Edinburgh, Royal Edinburgh

Hospital, Edinburgh EHlO, Scotland.
375
376 BRUCE RITSON
years in Britain is now very familiar. In 1962 the Ministry of Health

encouraged Health Authorities to establish specialized units. By 1975, 21
such specialized units with 434 beds were in existence. They were con-
cerned with providing a service for a population of alcoholics who were
at that stage estimated to number 86,000 in England and Wales alone.
Therefore to any casual observer it was very evident that the resources
were totally inadequate to meet the needs of the population. This statis-
tic alone encouraged some to seek for simpler, briefer, and less intensive
modes of intervention.
Ten years ago, the model for an alcoholism treatment unit com-
monly involved an emphasis on an inpatient stay of 4 to 6 weeks or even
longer. Patients were commonly selected for their motivation and capac-
ity for intrapsychic and interpersonal exploration in a group setting.
Group treatments, often borrowing from the influences of Alcoholics
Anonymous, were the rule. Many clinics provided follow-up services,
again in groups or on an individual outpatient basis. The unspoken
assumption often appeared to be that the best treatment was to be found
in intensive inpatient communities. Although many patients appeared
to benefit enormously from these experiences, a number of evaluation
studies began to question first the real need for prolonged stay and then
the need for inpatient treatment at all. The emphasis in treatment began
to move toward earlier intervention within frontline agencies, such as
the primary health care team and social work (Department of Health and
Social Security, 1978).
This change in emphasis was facilitated by the evidence of a series
of evaluation studies. These compared different durations of inpatient
stay; intensive as opposed to simpler inpatient programs; and inpatient
contrasted with outpatient treatments (Clare, 1980; Edwards & Guthrie,
1967; Levinson & Sereny, 1969).
Wherever such evaluation studies have been properly conducted,
the more intensive, prolonged, and expensive treatment has not been
shown to have significant advantages over its simpler counterpart. Prob-
ably the most influential study for British clinicians, and a subsequent
source of much therapeutic nihilism amongst psychiatrists, was that by
Orford and Edwards (1977).
They randomly allocated 100 married male alcoholics to either ad-
vice or treatment. In the advice group, the patient was carefully assessed
and, along with his wife, was given approximately half an hour's specif-
ic advice during which it was made clear that "responsibility for the
attainment of abstinence lay in his own hands" rather than anything
more which could be done by others. The treatment group, on the other
hand, were provided with whatever specialist help seemed best suited
to their needs and included all the specialist resources of a major psychi-
MERITS OF SIMPLE INTERVENTION 377
atric teaching hospital-counseling, marital therapy, Alcoholics Anony-

mous, citrated calcium carbimade, and admission if needed. A research
social worker visited both groups weekly for a year and maintained a
94% contact rate. No statistical difference in outcome was observed be-
tween the treatment and control groups after one year.
This was an extremely carefully designed study but it is important
to note that it concerned only married men from whom psychiatrically
disturbed alcoholics had already been excluded. Some have questioned
whether the follow-up itself did not become a powerful therapy for both
groups, although the social worker took care to avoid moving outside
the role of research interviewer. No attempt was made to match patient
with therapy in any systematic way as suggested by Glaser (1980), al-
though there was presumably an intuitive matching of patients' needs
with available resources in the treated group. The particular point to
note at this stage, however, is that the advice group was not offered "no
treatment." They had a morning of devoted attention focused on their
drinking and their problems, probably more intensive interest than
many individuals receive in a lifetime, and then were given carefully
chosen advice in a prestigious specialist center in the presence of their
wives. What many wrongly interpret as no treatment was a very signifi-
cant experience in the life of the patient. The authors themselves caution
against nihilism, stating that the research had not proposed:
an overthrow or negation of all established effort, the compassion which it
witnesses, and the community support which it has won. What is proposed
is only that, in terms of what should be seen as a process of evolution rather
than a static treatment model, there is now need for further evolution. (Or-
ford & Edwards, 1977, pp. 113-114)
It may well prove that the single advice session largely confirmed
attitudes and decisions for patients already in the contemplation stage
and gave them a simple action plan after the fashion already described
by Prochaska & DiClemente (1982).
A further study of this kind is underway in Edinburgh. On this
occasion both men and women have been included, as are single per-
sons provided they have a key informant. Subjects are being randomly
allocated to three treatment categories: intensive therapy; advice given
in one session but tailored to the needs of that particular patient; and
finally an extremely simple advice session in which, during the course of
3 minutes, the patient is advised to abstain and that the responsibility
for action lies in his or her own hands. The follow-up process has tried
to avoid establishing any therapeutic relationship between the follow-up
social worker and the contact person, and follow-up has been main-
tained over a 2-year period. The results of this study are not yet
available.
378 BRUCE RITSON
These evaluation studies have cast doubt on the justification for

continuing treatment for alcohol-related problems by intensive and ex-
pensive therapies and on the belief that more treatment is necessarily
better treatment. The case for continuing with elaborate residential treat-
ment at a time of mounting health costs would have to be very strong
indeed and yet programs with lengthy and intensive inpatient compo-
nents continue. However, there are equally no grounds for closing the
door of treatment agencies and investing our resources solely in educa-
tion and political action, important as both are in reducing the long-term
burden of alcohol-related harm. Treatment does not need an apologist
but it is now free to rethink appropriate methods. The clinician can
move toward simple treatment without feeling that second-best or sub-
standard goods are on offer. Now that evaluation studies have legit-
imized simple intervention on clinical grounds, some of its additional
merits can be examined.
ECONOMIES
Health services, particularly those based on item-of-service costs for

treatment, are concerned about escalating health expenditure. The pres-
sure to evaluate the effectiveness and value of these interventions is
evident both in private insurance and state-funded services. Concern is
particularly appropriate when there is little evidence that intensive treat-
ments are significantly better than simpler measures.
As already indicated, the number of specialist, trained personnel is
very small when compared with the task in hand. Grounds for deploy-
ing such resources in time-consuming treatment for the few would have
to reside in strong evidence that such interventions can be amply justi-
fied. Such evidence is not forthcoming.
The equation between resources and needs becomes even more
unbalanced in Third World countries where alcohol problems are giving
increasing cause for concern. Early identification and simple interven-
tion techniques have become a priority in the World Health Organiza-
tion's (WHO) current research program towards health for all by the
year 2000 (WHO, 1983b; and see Marcus Grant's contribution to this
volume, Chapter 3).
Quite apart from the costs in health resources, earlier intervention
should also minimize the social and economic cost to the individual
drinker and his family by reducing the time involved in treatment.
Bringing treatment nearer to home and the work place also minimizes
the disruption in the family and work.
STIGMA
A sense of shame is common among individuals who have alcohol

problems. There is corresponding reluctance to admit the problem, both
to oneself and particularly to those around. This feeling of shame is
equally evident in very different settings. A study conducted in commu-
nities in countries as disparate as Zambia, Mexico, and Scotland all
revealed that shame was a prominent reason for being reluctant to admit
to having a drinking problem (WHO, 1983a). This blow to self-esteem
was particularly evident among women and prevented many of them
from seeking help from agencies, particularly when that agency itself is
labeled as offering treatment for alcoholics or problem drinkers.
There is also the stigma involved in feeling that one has somehow
lost control of oneself and needs someone else to take charge. This
giving up of self-control can seem very demeaning, particularly in
cultures where independence is a highly prized virtue. With simple
intervention, the client/patient is still very much in control and is simply
given the tools with which to effect personal change.
One of the features of precontemplation (Prochaska & DiClemente,
1982) is that it is very difficult for the patient to contemplate going to
treatment for help with problems because that involves admitting some
significant aspect of life is out of control. Individuals have the need to
believe that they are in control of their own destiny, therefore simple
advice goes a long way to alleviating the problem of handing over con-
trol to others and also avoids the equally common pitfall of rendering
the patient/client dependent on the therapist.
PROXIMITY
In most systems of care for alcoholics, there has been a gradation in

the decision-making and treatment structure that involves a passage
TABLE 1.
Levels of Recognition and Intervention for the Problem Drinker
Individual
Levell: Family, friends
Level 2: Workmates, employer, barman, social welfare worker
Level 3: Primary health care team, area social work team, probation officers, police,
clergy, casualty department
Level 4: AA, (AI Anon), council on alcoholism, alcohol treatment unit
380 BRUCE RITSON
towards ever increasingly specialised services. Table 1 illustrates a com-

mon pattern of levels of intervention. There is ample evidence that the
life of the excessive drinker is replete with incidents that draw attention
to his or her problem at an early stage. The first hint that something is
amiss is usually detected in the family and later incidents occur at work
or bring the drinker into contact with a range of institutions, such as
those outlined in the second and third layers of Table 1. If we take each
level in turn, it is evident that appropriate intervention at any stage may
turn a crisis into an opportunity for a positive change in life-style. At the
first level in Table 1, the drinker or his family may be instrumental in
effecting change. This is probably the basis of many spontaneous remis-
sions and is also the kind of level at which the bibliotherapy described
elsewhere by Miller and Taylor (1980) is taking effect.
At the second level, we find individuals who do not have a desig-
nated responsibility to care or a therapeutic role and yet may be ex-
tremely influential. The most promising interventions at this level are in
Alcohol in Employment programs; a recognition of impaired work per-
formance due to alcohol can lead to involvement in an alcohol counsel-
ing program. At the third level are professional care givers. They are in
key positions and in the future will, it is hoped, prove the main pur-
veyors of simple intervention to problem drinkers. Unfortunately there
is already evidence that, as a group, such frontline workers are often
profoundly pessimistic about their capacity to help. This is an issue that
will be discussed later. The fourth level contains specialist alcohol coun-
seling and treatment agencies. Where simple intervention is being pro-
moted, we can hope that some of these staff will change their role to
providing support, information, and consultation to the primary-level
workers.
It is obviously preferable if intervention can be offered near to
home. The patient or client is seen in a familiar setting, for instance at
the Health Center, which is nonstigmatizing and where the family doc-
tor and nurse or other primary worker may already know a lot about his
or her background.
Many treatment centers for problem drinkers comment on a very
high failed first-attendance rate, which is sometimes as great as 40%,
and also on the high dropout rate that follows first contact. If simple
intervention can be offered nearer to home and without undue delay, it
is much more likely that patients will attend. Where counseling services
have been recently made available in Health Centers in the Lothian area,
the attendance rates for first appointments improved to over 80%. Apart
from the issue of stigma described above, it is also physically easier to
come to a nearby facility, particularly for those who have problems of
looking after young children or difficulties in getting away from work.
There is also a ripple effect concerning the skills that are being
learned. The primary level worker (and perhaps the client too) will
acquire and gain confidence in the necessary coping skills and this fos-
ters the development of a repertoire of skills, rather than perpetuating a
dependency on the specialist to whom they must make referrals. There
are many merits in avoiding any process that appears to deprive the
primary-level worker of skills, and this has been one of the most unfor-
tunate and undesirable consequences of the growth of specialization.
ATTITUDES
Ideally, simple interventions should be at least as effective in the

hands of a primary level worker as they are when offered by specialists.
There is evidence that many primary level workers, for instance, general
practitioners and social workers, feel profoundly pessimistic about their
capacity to help problem drinkers (Ritson & De Roumanie, 1984; Shaw,
Cartwright, Spratley, & Harwin, 1978). Part of this pessimism resides in
a feeling that little can be done to help the problem drinker and that, in
any case, they themselves lack the skills to offer any effective help. One
hoped-for outcome from research into simple intervention is the devel-
opment of simple tools that can be used with confidence in the frontline.
One interesting and encouraging development in this respect is the
DRAMS project described elsewhere in this book (Heather, Chapter 17).
This project involved giving general practitioners a pack containing sim-
ple instructions about recognizing alcohol problems, giving patients
written material about modifying drinking habits, and then structuring
subsequent monitoring of progress by the use of diaries and general
practitioner follow-up. This combination of straightforward advice and
self-help would appear to be a promising way of combatting therapeutic
nihilism and at the same time reaching a large number of problem
drinkers.
EVALUATION OF SIMPLE INTERVENTION
Although evaluation studies of the past 20 years described earlier

have certainly prompted and encouraged the search for simpler inter-
ventions, there are very few studies that give any substantial answer to
the question, Do simple interventions work?
A few examples that do exist derive exclusively from industrialized
countries. WHO is currently engaged in a study that will explore the
382 BRUCE RITSON
feasibility and value of early identification and simple intervention in

some Third World countries, where resources are even more precious.
Many of the diseases that cause most concern in the Western world,
such as lung cancer, hypertension, and liver cirrhosis, are thought to be
a consequence of chosen life-style behaviors, such as cigarette smoking,
overeating, lack of exercise, and excessive drinking. This evidence has
fostered an interest in education and advice aimed at health promotion
as an alternative to costly medical care. In response to this acknowledg-
ment, there are now a number of studies that have examined the impact
on unhealthy behaviour of simple advice, usually given by a physician.
Research in Norway, Finland, and the United States provides evidence
of the benefits of advice about healthy life-style given to individuals at
risk to coronary heart disease (McAlister, Puska, Salonen, Tuomilehlo,
& Koohela, 1982). Until recently a pall of therapeutic pessimism has
hung over attempts to give simple advice to those who abuse addictive
substances-as if their very addictive properties demanded subtle ap-
proaches and put the patient beyond reach of reasonable discussion!
This pessimism was countered by Russell, Wilson, Taylor & Baker
(1979), who showed that it was possible to have a significant impact on
cigarette smoking by simple advice. They randomly assigned smokers
who attended their general practitioners to four categories: a noninter-
vention control group; a questionnaire only group; a simple advice to
stop smoking group; an advice plus leaflet plus follow-up group. They
found that the last group was significantly more successful in stopping
smoking and sustaining this over one year. Although only 5% achieved
this goal, the authors point out that if all the family doctors in the United
Kingdom adopted this simple approach, the yield would exceed half a
million ex-xmokers a year. This target could not be matched by increas-
ing the present 50 or so special withdrawal clinics to 10,000.
As far as alcohol-related problems are concerned, Kristenson (1982),
in Malmo, Sweden, reported the beneficial effects of advice given to
men who had been identified as heavy drinkers as part of a general
health screening project. Among those identified as having a raised
Gamma-GT on two occasions 3 weeks apart, 76% were found to be
either heavy or moderate drinkers. Heavy drinking was defined as con-
suming more than 40g of alcohol a day and moderate as 20g. to 40g of
alcohol per day.
Those who were observed to have a raised Gamma-GT were ran-
domly allocated to an intervention and a control group. Subjects in the
control group were informed by letter that test results revealed they had
an impaired liver, were advised to restrict their drinking, and invited to
attend for further blood tests after 2 years. In contrast, subjects in the
intervention group were given a detailed physical examination and in-
terviewed about their drinking histories, symptoms of alcohol depen-
dence, and evidence of alcohol-related problems. This group was then

offered appointments with the same physician every 3rd month and
monthly contacts with the same nurse, who repeated the Gamma-GT
assessments. Subjects were advised about moderating their drinking.
Progress was monitored by regular feedback of Gamma-GT levels and
general encouragement to attain normal drinking levels. Once these
results had achieved an acceptable level, the frequency of clinical contact
was reduced.
The subjects' progress was evaluated 2 and 4 years after the initial
screening. The Gamma-GT values of both groups decreased signifi-
cantly. There was, however, an important difference between the two
groups in sick absenteeism, hospitalization, and mortality. In the inter-
vention group, the mean annual sick days per individual increased from
24 to 29, whereas in the control group the corresponding rise was from
25 to 52. The control group had 1,644 hospitalized days whereas the
intervention group had a total of 808 in the hospital over the 4-year
period. If alcohol-related conditions were isolated from other causes of
hospitalization, the difference was even more striking, with 482 days in
the control group and 133 in the intervention group.
This study showed that simple intervention with regular feedback
based on a biochemical marker could have significant effects on the
drinking habits and physical health of the population. It should howev-
er be noted that this intervention did involve quite a lot of contact and
the use of a skilled physician. It was again in a medical setting and a
health orientation predominated.
In France, the National Health Ministry recommended the estab-
lishment of Centres d'Hygiene Alimentaire as part of a national program to
prevent alcoholism (Chick, 1984). These clinics have directed their
efforts primarily towards the nondependent, excessive drinker.
Staff members are instructed to do the following:
1. Give the drinker proof of his chronic alcohol misuse
2. Gain the drinker's confidence
3. Persuade him or her that a radical change in drinking habits is
necessary
4. Show by feedback processes that the reduction or elimination of
alcohol leads to an improvement in health
This process is accomplished by a series of outpatient visits to these
centers, which provide a combination of clinical diagnosis, medical
treatment, dietary counseling, health education, and family counseling.
No randomized control study of these centers has been undertaken, but
they seem to offer a promising and simple approach that does not call
upon elaborate resources of counseling and psychotherapeutic skills.
A recent study in the Royal Infirmary in Edinburgh has attempted
384 BRUCE RITSON
to assess the effectiveness of brief intervention with problem drinkers

identified in a general hospital. The presence of such problems was
established by a trained nurse using a structured interview of 10 minutes
duration covering drinking habits, recent and previous medical history,
and social background. The mean corpuscular volume and gamma glu-
tamyl transpeptidase were recorded in each case. The criteria for inclu-
sion in the study were that the patient should not have received prior
treatment for an alcohol problem and should have some degree of social
stability to facilitate follow-up. The patient also had to give evidence of
heavy drinking or alcohol problems from the categories shown in Table
2.
TABLE 2.
Criteria for Inclusion as a Problem Drinker
Points
Consumption
More than 12 units a in a day on 10 or more occasions in the last year 1
More than 50 units in typical week 1
More than 12 units in 24 hours in typical week 1
Alcohol related problems
Current medical problem
Present illness potentially alcohol related 1
Present illness definitely alcohol related 2
Weight problem due to alcohol 1
Medical problems in past 2 years
Peptic ulcer aggravated by drinking 1
Liver disease due to alcohol 1
Accident due to drinking 1
Alcohol-related social problems in past 2 years
anti-social behavior 1
problems at work (inc. absence) 1
domestic arguments 1
violence 1
family rupture-threatened or actual 1
financial 1
police 1
Dependence on alcohol in past 2 years
Difficulty in reducing consumption 1
Restlessness without alcohol 1
Tremor (more than 1 day per week) 1
Morning relief drinking (more than 1 day per week) 1
Hallucinations 1
Withdrawal seizure 1
al unit = 1 oz of 40% (by volume) spirits, VI pint of 3.6% (by volume) beer, 1 glass of wine etc. (i.e.,
approximate 8g ethanol).
The sample of identified patients was then randomly divided. No

comment was made to control group subjects although all agreed to
follow-up one year later. The intervention group received a further 30 to
60 minutes counseling from the nurse in the presence of the patient's
spouse, where possible (although this was rarely the case). Finally, the
patient was given a booklet containing advice about techniques for re-
ducing drinking. After one year both groups were interviewed by a
nurse who did not know the design of the original study. Preliminary
findings showed that fewer of the intervention group had alcohol prob-
lems at the end of one year, 52% were categorized as definitely im-
proved as opposed to 34% of the control group. These results suggest
that simple intervention by an experienced nurse giving clear advice
about drinking does have a positive effect on subsequent drinking prob-
lems during the course of the ensuing year (Chick, Lloyd, & Crombie,
1984).
These studies, although very far from providing conclusive evi-
dence, do offer encouragement to those who are seeking to provide
simple tools that can be given to frontline workers for everyday use in
helping problem drinkers. The tools at present are rather blunt and
require to be honed to greater precision by further research. Clinical
psychology has an important task to perform in designing simple self-
help techniques for achieving behavioral change. They need to be of
proven reliability and capable of being given away for use by the client
or by a primary-level worker who has no specialist training.
QUESTIONS
There are, however, some questions and doubts that must temper
enthusiasm before the wholesale adoption of these simple techniques. If
adopted in an unthinking way, they may become a prescription for low-
cost or shoddy interventions and a means of hampering the develop-
ment of resources for this client group. Obviously, they could make a
very attractive recipe for any government keen to reduce expenditure on
health services.
There is also the anxiety that simple intervention may not confront
the issues of the precontemplation and contemplation stages. How, for
instance, does simple advice accord with Prochaska's view (see Chapter
I, this volume) that the minimal requirements for effective psycho-
therapy are the ability to help clients become aware of their defenses
against change? We are still in need of techniques for overcoming the
understandable resistances to change within the client group.
The timing and time scale of simple advice is also poorly under-
stood. The need to recognize the stages involved in decision making and
386 BRUCE RITSON
the process of change suggests that there are times when the client is
ripe for advice and others when it would appear unwelcome and be
rejected. Timing requires skill and sensitivity, and a blanket prescription
of self-help manuals and words of advice may be just as wasteful as the
overuse of psychodynamic approaches have been on other occasions.
Simple interventions also need to have a sense of time as well as timing.
Many alcohol-related problems wax and wane over years and resolve
may require to be strengthened and relapses discussed. Very little is
known about the longitudinal perspective of this approach to therapy,
although it is acknowledged by Prochaska and DiClemente (1982) in
their maintenance stage. Brief but repeated advice may prove preferable
to a few prolonged assessment interviews.
Edwards (1982), in his book on helping alcoholics, says:
The relationship between patient and therapist is fundamental both to what
can be achieved in anyone therapeutic session and to what changes can be
achieved over time. (p. 198)
In these simple strategies, what has happened to the importance of the

relationship?
Much work also requires to be done on the importance of the status
of the advice giver. Most of the studies that have been described have
been in health settings; there is a need to extend these to other settings
and for a closer examination of the importance of the status and cred-
ibility of the therapist. There are, of course, other important issues con-
cerning assessment and matching, both in terms of the client charac-
teristics and the stage of change they have reached. Miller has ably
reviewed these elsewhere in this book (see Chapter 8).
There is also resistance among professionals themselves. Pessimism
and resistance of colleagues in primary-level agencies has already been
acknowledged. This resistance can be considerably softened by provid-
ing them with clearly formulated techniques and instruments that en-
hance self-confidence in managing alcohol-related problems. Among
specialists, there is also an understandable resistance to changing
therapeutic techniques-skills that have been costly and hard to
acquire.
In conclusion, it is evident that simple advice has many merits and
is amply justified, but the details of timing, presentation, content, and
follow-up need much more definition.
REFERENCES
Chick, J. (1984). Secondary prevention of alcoholism and Centres d'Hygiene Alimentaire.

Chick, J., Lloyd, G., & Crombie, E. (1984). Counselling problem drinkers in medical
wards: A controlled study. British Medical Journal, 290, 965-967.
Clare, A. W. (1980). How good is treatment? In G. Edwards & M. Grant (Eds.), Alcoholism:
New knowledge and new responses (pp. 279-289). London: Croom Helm.
Department of Health and Social Security. (1978). The pattern and range of services for problem
drinkers. Report by Advisory Committee on Alcoholism. London: HMSO.
Editors. (1977). The Alcoholism treatment package. Lancet, 2, 488-490.
Edwards, G. (1982). The treatment of drinking problems. London: Grant McIntyre.
Edwards, G., & Guthrie, S. (1967). A controlled trial of in-patient and out-patient treat-
ment of alcohol dependence. Lancet, 1, 555-559.
Glaser, F. B. (1980). Anybody got a match? Treatment research and the matching hypoth-
esis. In G. Edwards & M. Grant (Eds.), Alcoholism treatment in transition. London:
Croom Helm.
Kristenson, H. (1982). Studies on alcohol-related disabilities in a medical intervention program: in
middle-aged males. Unpublished doctoral thesis, University of Lund.
Levinson, T., & Sereny, G. (1969). An experimental evaluation of 'insight therapy' for the
chronic alcoholic. Journal of Canadian Psychiatric Association,S, 14-18.
Miller, W., & Taylor, C. A. (1980). Relative effectiveness of bibliotherapy, individual and
group self-control training in the treatment of problem drinkers. Addictive Behaviours,
5,13-24.
McAllister, A., Puska, P., Salonen, J. T., Tuomilehlo, J., & Koohela, K. (1982). Theory and
action for health promotion. American Journal of Public Health, 72, 43-50.
Orford, J., & Edwards, G. (1977). Alcoholism. Oxford: University Press.
integrative model of change. Psychotherapy: Theory Research and Practice, 19, 276-288.
Ritson, B., & De Roumanie, M. (1984). Community attitudes and responses. In N.
Krasner, J. Madden, & R. Walker (Eds.), Alcohol related problems (pp. 267-279). Lon-
don: Wiley.
Russell, M. A. H., Wilson, c., Taylor, c., & Baker, C. D. (1979). Effect of general practi-
tioners' advice against smoking. British Medical Journal, 285, 231-235.
Shaw, S., Cartwright, A., Spratley, T., & Harwin, J. (1978). Responding to drinking problems.
London: Croom Helm.
World Health Organisation. (1983a). Community response to alcohol related problems. Geneva:
WHO MNH/83.17. Author.
World Health Organisation. (1983b). Seventh general programme of work. global medium-term
programme. Geneva: Programme 10. MNH.MTP/83.19. Author.
IV
Maintenance
Preventing Relapse
20
Alcoholism Survival
The Prevention of Relapse
GLORIA K. LITMAN
INTRODUCTION
Although there is still a dearth of knowledge about the mechanisms of

relapse in alcoholism, there is an even greater dearth of research into
how individuals who have been treated for alcoholism survive-that is,
do not resume heavy drinking. In our studies of relapse, we have been
investigating not only the causes of relapse, but, equally important,
looking at the ways patients have found to avoid relapse and maintain
survival.
Our early model of survival hypothesized that there is an interac-
tion among (a) situations perceived to be dangerous for the individual in
that they may precipitate relapse; (b) the coping strategies available
within the individual's repertoire to deal with these situations; (c) the
perceived effectiveness of these coping behaviors; (d) the individual's
self-perception and self-esteem and the degree of learned helplessness
with which they view their situation. If individuals regard themselves as
helpless victims of their feelings, their situation, or their personality,
This chapter is a version of a paper first presented at a conference on Relapse in
Alcoholism organized by the Northamptonshire Alcohol Counselling and Information
Service at Northampton, England in November, 1983.
GLORIA K. LITMAN· Addiction Research Unit, Institute of Psychiatry, London SE5

8AF, England. The work reported in this chapter was funded in part by generous grants
from the Medical Research Council, the Nuffield Foundation, the Scotch Whiskey Associa-
tion and the Brewers' Society.
391
392 GLORIA K. LITMAN
they will be less likely to take appropriate action to avoid relapse. On the
other hand, if a person learns coping behaviors and strategies and per-
ceives them to be effective, this may lead to positive changes in self-
perception and self-esteem. Concurrent with these changes is feedback
from the social environment. Therefore, the process of enhancing self-
esteem is generated from external as well as internal sources.
THE ORIGINAL STUDY
Our original study developed formal questionnaires designed to

elicit information on responses, attitudes, and strategies that charac-
terize clients' attempts to avoid relapse. The questionnaires cover the
following areas: (a) situations that were dangerous to the individual-
the Relapse Precipitants Inventory (RPI); (b) the coping strategies the
individual had used to avoid relapse-the Coping Behaviors Inventory
(CBI); (c) how effective these coping strategies were perceived to be-
the Effectiveness of Coping Behaviors Inventory (ECBI); and (d) how
dependent on alcohol individuals saw themselves-the Dependence In-
ventory (01). These questionnaires were given individually to outpa-
tients at the Maudsley Hospital and Withington Hospital and to inpa-
tients at the Bethlem Royal and Warlingham Park Hospitals. We also
sent questionnaires to former patients who were definitely known to
have been abstaining for 6 months or more. One hundred and twenty-
six patients were included in the sample.
When we analyzed the results of these questionnaires (Litman,
Eiser, Rawson, & Oppenheim, 1977), we found that the Relapse Precipi-
tants Inventory could be summarized by four components, which ac-
counted for 60% of the variance:
Factor 1. Unpleasant Affect (e.g., "When I feel depressed"; "When
I feel tense")
Factor 2. External Events (e.g., "When I pass a pub or off-licence")
Factor 3. Social Anxiety (e.g., "When I have to meet people and
feel afraid")
Factor 4. Lessened Cognitive Vigilance ("When I start thinking that
one drink would do no harm")
The Coping Behaviors Inventory (CBI) could also be summarized by
four factors, which accounted for approximately 48% of the variance:
Factor 1. Positive Thinking (e.g., "Stopping to examine my motives
and eliminating the false ones").
Factor 2. Negative Thinking (e.g., "Thinking of the mess I've got
myself in through drinking")
ALCOHOLISM SURVIVAL 393
Factor 3. Distraction/Substitution (e.g., "Start doing something in

the house")
Factor 4. Avoidance (e.g., "Keeping away from people who
drink").
The Effectiveness of Coping Behaviors Inventory could be summa-
rized by three components, accounting for approximately 47% of the
variance.
Factor 1. Cognitive Control (which included both Positive and
Negative Thinking
Factor 2. Avoidance (as in the CBI)
Factor 3. Distraction/Substitution (as in the CBI)
We looked at the association between relapse and coping behaviors
and found that, for this sample, Unpleasant Affect seemed to be associ-
ated with Distraction/Substitution and Negative Thinking. Not surpris-
ingly, these were perceived to be ineffective. There did not seem to be a
coping style associated with External Events, and Distrac-
tion/Substitution was perceived as being particularly ineffective. Social
Anxiety as a dangerous situation was found to be associated with Distrac-
tion/Substitution and Negative Thinking, both of which were perceived
to be ineffective. However, contrary to the clinical view that situations
should be confronted, Avoidance was perceived to be an effective coping
behavior for both Social Anxiety and Lessened Cognitive Vigilance (Lit-
man et al., 1977).
When we compared relapsers and survivors in this sample (Litman,
Eiser, Rawson, & Oppenheim, 1979) some interesting differences
emerged. Relapsers' scores on the RPI were significantly higher than
those of survivors, indicating that the more situations individuals per-
ceived as being dangerous, the more likely they were to relapse. High
scores on the Unpleasant Affect component of the RPI discriminated
between relapsers and survivors, suggesting that those more prone to
depression, anxiety, etc., are more likely to relapse. External Events also
discriminated between the groups. The results indicated too that indi-
viduals who adopt a multiplicity of coping styles-a flexibility that en-
ables them to cope with a variety of dangerous situations-are more
likely to survive. The strongest discriminator between relapsers and
survivors was the perception of Cognitive Control as an effective coping
behavior.
A CONCEPTUAL FRAMEWORK
At this point, we had some information about the process of re-

lapse, but more information on survival was required. Therefore, we
interviewed in depth 25 patients who had been successful either in

abstaining or in controlling their drinking after treatment. These inter-
views were recorded and complete transcripts were examined for key
concepts, recurring themes, and individual variations. From this exam-
ination of the transcripts, we developed what we called the "Conceptual
Framework for Alcoholism Survival" (Litman, 1980, 1982). This concep-
tual framework served as the basis for hypothesizing stages of survival
rather than survival as an all-or-none phenomenon. One hypothesis
was that there may be a gradient of coping strategies that survivors
develop over time after the treatment phase. Although Avoidance may
be an effective coping strategy in the initial reentry phase after treat-
ment, it is the development of more complex cognitive coping strategies
that may determine whether individuals treated for alcoholism will re-
lapse or survive.
We also developed hypotheses regarding what we have called the
"Critical Perceptual Shift"-the point at which the individual is con-
fronted with the choice of either a drastic change in drinking habits and
life-style or self-destruction-and a concommitant shift in their percep-
tion of the therapist, who may be invested with magical qualities in the
early stages of successful treatment. We hypothesized also that survival
necessitated a shift in the locus of control (Rotter, 1966) from external
attributions of success and failure to internal responsibility. Because our
original model included external influences as well, we also hypoth-
esized that a stable social support network would be more likely related
to survival than to relapse.
THE PROSPECTIVE STUDY
The next stage in our work was to launch a prospective study to test
some of the hypotheses formulated on the basis of our previous work
with known groups of relapsers and survivors. We modified the in-
ventories used in the earlier study on Dependence, Relapse Precipitants,
Coping Behaviors, and the perceived Effectiveness of Coping Behaviors
to include only those items that had the highest loadings on the original
factors and that, on the basis of a discriminant function analysis, signifi-
cantly discriminated between relapsers and survivors in our original
study. We also developed and piloted further inventories designed to
obtain a comprehensive drinking and relapse history, and to measure
Critical Perceptual Shift, Self-Efficacy, Self-Esteem, Locus of Control,
Perceived Social Supports, Commitment and Motivation, and the indi-
vidual's Perception of the Therapist. The full questionnaire consisted of
326 items comprising 12 inventories.
Two hundred and fifty-six patients who presented for treatment for
alcoholism were selected for the prospective study. The sample included
patients from Bexley Hospital, Warlingham Park Hospital, Queen Eliz-
abeth Military Hospital, ACCEPT, St. Andrew's Hospital and the
Maudsley and Bethlem Royal Hospitals. Although we tested all con-
secutive admissions to each of these centers, in our final sample we
excluded any individuals who were diagnosed as brain damaged or
psychotic by the centers, or who had histories of multiple drug abuse.
We also eliminated from the sample any individuals who had no fixed
abode and, because of confidentiality requirements, any individual who
was discharged from the Army. Approximately one third of patients
admitted for treatment to those centers over the one-year intake period
of the study were excluded by one or more of these criteria.
The full 326-item questionnaire was administered to patients by
members of the hospital staff at intake (or as soon as withdrawal from
alcohol was complete). One hundred and ninety-eight patients or ap-
proximately 77% of the sample were located for follow-up and question-
naires were administered to these patients approximately 6 weeks after
discharge and 6 to 15 months subsequently by members of the research
team. All subjects were informed that they were participating in a re-
search program and that the results would be kept confidential and
would not appear on their clinical records.
RELAPSE PRECIPITANTS INVENTORY (RPI)
One of the factors involved in relapse is the fact that treated alco-
holics return to a world that holds many dangers, both internal and
external, that may precipitate the resumption of excessive drinking. The
literature on relapse precipitation is not extensive. Hore (1971) asked 22
patients to keep a record of their daily anxiety, craving, and depression
levels over a 6-month period and concluded that there was no rela-
tionship between mood state and subsequent relapse. However, Marlatt
(1979), analyzing the responses of a group of 70 patients, found that 38%
of his subjects reported negative affect prior to relapse, a finding that
concurs with the results of our initial study. Hodgson and Rankin (1982)
theorized that excessive drinking behavior could be conceptualized as a
discriminant operant that may be modified by exposure to drinking
cues. Their single case study (Hodgson & Rankin, 1976) reports modest
success in using cue exposure to modify drinking behavior.
Analysis of the RPI

The items in the Relapse Precipitants Inventory are given in Table I,
along with the appropriate instructions (see Litman, Stapleton, Op-
penheim, Peleg, & Jackson, 1983).
TABLE l.
Relapse Precipitants Inventory
Instructions:
Here are some situations which some people have experienced as being dangerous to
their staying off drink. Which of these may be dangerous for you? There are four
boxes, "Very dangerous," "quite dangerous," "a little dangerous," "not at all." Please
tick that box which comes closest to your feelings about those situations which may be
dangerous to your staying off drink. There are no right or wrong answers or trick
questions. We want to know how you feel.
1. When I pass a pub or off-licence (liquor store)
2. When I'm with other people who are drinking
3. When I feel no one really cares what happens to me
4. When I feel tense
5. When I have to meet people
6. When I start thinking that just one drink would cause no harm
7. When I feel depressed
8. When there are problems at work
9. When I feel I'm being punished unjustly
to. When I feel afraid
11. When I'm on holiday
12. When I feel happy with everything
13. When I have money to spend
14. When I remember the good times when I was drinking
15. When there are rows and arguments at home
16. When I'm full of resentments
17. When I feel irritable
18. When I'm at a party
19. When I start thinking I am not really hooked on alcohol
20. When I feel myself getting very angry
21. When there are special occasions like Christmas, birthdays, etc.
22. When I start feeling frustrated and fed up with life
23. When I feel tired
24. When I feel disappointed that other people are letting me down
25. When I have already taken some drink
When we analyzed the results of the RPI for our present sample at
intake by means of principal components analysis using Varimax rota-
tion, we found that the first three factors, which account for 55% of the
variance, summarized the data adequately. These three components
were the following: Factor I, Unpleasant Mood States; Factor 2, External
Events; Factor 3, Lessened Cognitive Vigliance. The items on the fourth
factor in our previous study that referred to Social Anxiety now loaded
in the first factor, thus relating to a more generalized anxiety or
depression.
In order to compare the factor structure with our previous work, we
reanalyzed the previous data using only the 25 items that were included
in the present Inventory. (The original RPI contained 41 items). We
found that with the renalysis of these data, three factors emerged, ac-
counting for 57% of the variance. The items on these factors were almost
identical to those found in the present data. When we performed a
Kaiser analysis to obtain a more objective measure of the similarity of the
two factor structures, we found that the resultant interfactor coefficients
for Factors 1, 2, and 3 were .98, .99 and .98, respectively, indicating a
high degree of stability. Having established three stable factors, we then
proceeded to look at the relationship between scores on these factors at
intake and subsequent outcome 6 to 15 months later.
In our previous work, the definition of relapse was not an issue,
because we categorized as relapsers those patients who had returned to
hospital for further treatment for alcohol abuse. Because the present
work is a prospective study, we were faced with the dilemma of what
constitutes relapse and how survival is defined. Although we are con-
cerned with the quality of posttreatment adjustment in other areas, for
the purposes of this analysis we defined relapse and survival in terms of
the amount of alcohol consumption, which seems reasonable in view of
the fact that all the questions on the RPI were directed to "situations
dangerous to staying off drink." Although drinking outcome was based
on self-report, we obtained corroborative data from various sources.
Abstinence was relatively simple to define. If the subjects maintained
they had not been drinking at all since their discharge from the hospital
and if there were no further evidence to suggest that the subjects were
drinking during the 30 days preceding the final questionnaire, nor that
the subject was drinking during the earlier follow-up period, they were
put in the abstinence category.
Distinguishing between light or moderate drinking and heavy
drinking was more difficult. As the Rand Report researchers (Armor,
Polich, & Stambul, 1978) have noted, any cutoff point is essentially
arbitrary. Schmidt (1976) proposes a 150 ml cutoff (approximately the
daily equivalent of 50 oz of absolute ethanol) whereas Pequinot, Tuyno,
and Berta (1978) and Lieber (1979) suggest that less than 5 oz but more
than 2 oz may still imply substantial risk.
In view of the way the alcohol consumption questionnaire was set
out, we categorized outcome into (a) Abstinence (as discussed) (77 sub-
jects); (b) Light/Moderate Drinking, which we defined as less than the
daily equivalent of 5 pints of beer, 1 half bottle of fortified wine, or, a
half bottle of spirits. These amounts do not represent simple averaging
over the follow-up period. If there were any evidence that subjects had
exceeded these amounts during the follow-up period, they were ex-
cluded from this category (31 subjects); (c) Heavy Drinking, the daily
equivalent of 5 or more pints of beer, one and a half or more bottles of

wine, one or more bottles of fortified wine, one half bottle or more of
spirits, or a combination thereof (90 subjects).
We accept that these categories are arbitrary and, with the state of
the art in alcoholism as it now stands, would be equally open to criticism
by some who feel that our definition of light-moderate drinking is too
generous and by others who feel our definition of heavy drinking to be
too stringent.
A score was derived for each subject on each of the three factors and
also on the number of situations seen as dangerous. When we compared
the two extreme groups, we found that there were significant dif-
ferences between relapsers and survivors on the total number of relapse
precipitants, the scores on Factor 1 (Unpleasant Mood States), and the
scores on Factor 2 (External Events and Euphoric States), with relapsers
scoring significantly higher. Again we had replicated the findings of the
previous study (Litman et al., 1979), in which we found that the total
number of relapse precipitants and the mean factor scores on Unpleas-
ant Mood States and External Events and Euphoria significantly discrim-
inated between known groups of relapsers and survivors. There were
no significant differences between relapsers and survivors with respect
to the factor scores on Lessened Cognitive Vigilance.
The means for the light drinking group on the four scores were
11.20, 0.88, 0.95, 1.10. Only on Lessened Cognitive Vigilance did the
mean score for the light drinkers not fall between the mean scores for
abstinent and heavy drinking groups. The Linear F Statistic for the three
groups was significant for the first three scores.
Gender Differences in Alcoholic Relapse

In our previous work there was some suggestion that women were
more likely to perceive mood states as being more dangerous to staying
off drink, whereas men were more likely to view external events and
euphoric states as the more dangerous relapse precipitants. We com-
pared the men and women in the present sample in terms of the total
number of relapse precipitants and their mean factor scores on the
components.
The results indicated that there were significant differences be-
tween men and women in their factor scores on Factor 2, External
Events and Euphoria, with men having slightly higher means. Although
the difference between women and men on Factor I, Unpleasant Mood
States, did not reach the accepted level of statistical significance, there
seems to be a tendency for women to score higher on this factor. There
were no differences between men and women in their total scores on the
Inventory, nor in the total number of situations perceived as dangerous.
COPING BEHAVIORS INVENTORY (CBI)
According to Lazarus (1966), coping can be defined as some form of

action to reduce a danger, correct a harm, or achieve a gratification (see
also Lazarus, Averill, & Opton, 1974). Coping can thus be seen as a form
of response to a given situation, whether this situation be internally
generated as in mood states, or externally imposed. Because coping has
been regarded as a highly individualized, intrapsychic defence against
threat, there has been little scientific scrutiny of this area until recently.
Social scientists have concentrated heavily on external stimuli that pre-
cipitate distress, without regard for the individual's capacity to amelio-
rate or avoid this distress.
In the field of alcoholism too, it has often been assumed that pa-
tients succeeded in avoiding relapse either because the external con-
straints of their environment were conducive to sobriety or because they
possessed sufficient "willpower" to avoid relapsing. More recently, So-
bell and Sobell (1973) included as part of their broad-spectrum treatment
program sessions involving the identification of discriminative stimuli
for drinking and the generation of more appropriate alternative behav-
iors. Marlatt's (1979) categorization of relapse situations provided the
basis for skill training in relapse prevention (Chaney, O'Leary, & Mar-
latt, 1978). Sanchez-Craig and Walker (1982) attempted to teach coping
skills to chronic alcoholics in a halfway house setting, but concluded that
their subjects were unable to recall strategies one month after the end of
the program, and attributed this failure of retention to the lack of per-
ceived relevance and applicability of the strategy. Much of the work
cited here has either been concentrated on situation-specific behaviors
or laboratory-derived alternatives. To our knowledge, little attention has
been paid to the alcoholic's own ability to devise strategies to cope with
situations that triggered heavy drinking in the past.
Analysis of the CBI

We analyzed the results of the CBI for our present sample by means
of a principal components analysis with Varimax rotation. The first four
factors, which account for 54% of the variance, were thought to ade-
quately summarize this Inventory. These four components were, Factor
I, Positive Thinking; Factor 2, Negative Thinking; Factor 3, Avoid-
ance/Distraction; Factor 4, Seeking Social Supports. Table 2 shows the
TABLE 2.
Coping Behaviors Inventory
Instructions:
1£ there are times when you want to start drinking again, how do you try to stop
yourself? Here are a list of ways some people have tried to stop themselves. Which of
these ways have you tried? There are four boxes "Usually," "often," "sometimes," and
"never." Please tick that box which comes closest to how often you have used these
ways to try to stop yourself from starting to drink again. There are no right or wrong
answers or trick questions. We want to know what you have tried.
1. Thinking about how much better off I am without drink
2. Telephoning a friend
3. Keeping in the company of non drinkers
4. Thinking positively
5. Thinking of the mess I've got myself into through drinking
6. Stopping to examine my motives and eliminating the false ones
7. Thinking of the promises I've made to others
8. Staying indoors-hiding
9. Pausing and really thinking the whole alcoholic cycle through
10. Leaving my money at home
11. Recognising that life is no bed of roses but drink is not the answer
12. Going to an AA meeting
13. Knowing that by not drinking I can show my face again without fear of what
others will think
14. Cheering myself up by buying myself something special instead
15. Facing up to my bad feelings instead of trying to drown them
16. Working harder
17. Realizing that it's just not worth it
18. Waiting it out until everything is shut
19. Remembering how I've let my friends and family down in the past
20. Keeping away from people who drink
21. Going for a walk
22. Looking on the bright side and trying to stop making excuses for myself
23. Realizing it's affecting my health
24. Start doing something in the house
25. Considering the effect it will have on my family
26. Reminding myself of the good life I can have without drink
27. Getting in touch with old drinking friends who are better now
28. Making up my mind that I'm going to stop playing games with myself
29. Eating a good meal
30. Avoiding places where I drank
31. Thinking about all the people who have helped me
32. Saying I am well and wish to stay so
33. Going to sleep
34. Remembering how it has affected my family
35. Forcing myself to go to work
36. Trying to face life instead of avoiding it
items in the CBI, along with the appropriate instructions (see Litman,
Stapleton, Oppenheim, & Peleg, 1983).
In order to compare the factor structure of this Inventory with our
previous work, we reanalyzed the previous data including only those 36
items that were included in the present Inventory. In the reanalysis of
these data, the first four factors accounted for 49% of the variance, rather
than the previous 40%. When we compared the results, we found that
the factors for the present study and the previous study were very
similar. The Kaiser analysis interfactor coefficients for Factors I, 2, 3,
and 4 were .91, .81, .65, and .75. We found no significant relationship
between the scores on the CBI at intake and subsequent relapse and
survival 6 to 15 months later.
EFFECTIVENESS OF COPING BEHAVIORS INVENTORY (ECBI)
The items in the ECBI are the same as in the CBI. Whereas the
instructions in the CBI ask respondents to tick how often certain types of
coping behaviors are used, the ECBI asks them to tick how well these
coping behaviors work for them.
Analysis of the ECBI

When we analyzed the ECBI for our present sample using principal
components analysis with Varimax rotation, we found that four factors
emerged, accounting for 59% of the variance: Factor I, Positive Think-
ing; Factor 2, Negative Thinking; Factor 3, Avoidance/Distraction; Factor
4, Seeking Social Supports. The factor structure was almost identical to
that of the CBI, indicating that the factor structure remains stable even
when the instructions change (see Litman, Stapleton, Oppenheim, Pel-
eg, & Jackson, 1984).
When we compared relapsers and survivors on the ECBI, we found
that there were significant differences on their total score, on their scores
on Factor 1 (Positive Thinking), and Factor 3 (Avoidance). In other
words, at intake, individuals who 6 to 15 months later were abstaining
from alcohol were more likely to perceive themselves as having more
effective coping behaviors and to perceive Positive Thinking and Avoid-
ance as effective coping behaviors than individuals who were later to
relapse.
When we examined the differences between relapsers and ab-
stainers with regard to the relationship among Relapse Precipitants,
Coping Behaviors, and the perceived Effectiveness of Coping Behaviors,
we found small, but statistically significant, differences between the two
groups. For the group that was found to be abstaining, there was a
negative relationship between their scores on the RPI components and

their perception of the Effectiveness of Coping Behaviors. Because we
have already ascertained that abstainers have lower scores on the total
number of relapse precipitants, this finding seems to mean that the
more effective abstainers perceive their coping behaviors to be, the
fewer dangerous situations they perceive. The only exception to this is
Avoidance as an effective coping behavior, which is positively related to
the total number of dangerous situations and to internal mood states.
When we look at the heavy drinking outcome group, there seems to
be a positive relationship between Relapse Precipitants and Coping Be-
haviors. However, they do not perceive any significant relationship,
either positive or negative, between the situations they perceive as dan-
gerous and the effectiveness of their coping behaviors.
Although we found no differences between relapsers and survivors
on their scores on the CBI at intake, we also found that, when we look at
crude change scores between intake and 6 weeks after discharge, sur-
vivors tend to increase their use of Positive Thinking as a coping behav-
ior (p < 0.05). We found also that survivors tend to decrease their use of
Avoidance as a coping behavior, whereas relapsers tend to increase their
use of Avoidance (p < O. 0001).
SUMMARY AND CONCLUSIONS
Our model of relapse assumes complex stage-by-stage interactions

between relapse precipitants, coping behaviors, their perceived effec-
tiveness, and intraindividual characteristics. However, although we
have not yet looked at the full complexities of our data, there are trends
emerging that may begin to shed some light on some of the mechanisms
underlying alcoholism relapse and survival.
The findings reported in this chapter suggest that even at the time
of admission to a hospital, there are differences between individuals
who will relapse 6 to 15 months after hospital treatment and those who
will survive in terms of the situations they perceive to be dangerous and
the perceived effectiveness of their coping behaviors. Those who relapse
subsequent to treatment see more situations as dangerous to their stay-
ing off drink and are particularly vulnerable to their own unpleasant
affective states and to external events. They do not perceive their coping
behaviors to be effective and, in fact, do not see any relationship be-
tween the situations they perceive to be dangerous and the effectiveness
of their own ability to deal with these situations.
On the other hand, those individuals who survive initially perceive
both Positive Thinking and Avoidance as effective coping behaviors and
perceive fewer situations as being dangerous to their staying off drink.

As they continue to abstain, they tend to decrease their use of Avoid-
ance and increase their Positive Thinking as effective coping behaviors.
The more they experience their coping behaviors as effective, the less
they perceive situations as being dangerous.
These results argue against blanket treatment programs for alco-
holic patients and suggest that more attention needs to be paid clinically
to the particular resources and vulnerabilities of individual patients (d.
Miller, Chapter 8, this volume). Although the RPI, the CBI, and the
ECBI have not yet been developed as precision clinical instruments, they
may be of some value in directing attention to those patients who may
be particularly vulnerable and who should be monitored more closely.
Patients who perceive their affective states as making them particu-
larly vulnerable to relapse could be taught more effective coping behav-
iors, with an emphasis on positive cognitive styles, and the work of Beck
(1976) details procedures as to how this can be carried out. On the other
hand, patients who perceive external events as particularly dangerous
can be taught Avoidance as an effective short-term coping strategy, and
then more positive confrontation techniques, such as cue exposure
Hodgson & Rankin, 1982).
Although we have demonstrated that there is a direct relationship
between the perception of efficacy of coping behaviors and subsequent
outcome, we have not demonstrated the direction of causality. It may be
that the survivors in our study accurately perceived the efficacy of the
coping behaviors that work for them, based on past experience. It may
equally be that it is their belief in the efficacy of these behaviors that
leads to a positive outcome. Until this is tested in controlled clinical
trials, it seems reasonable to suggest that in setting up clinical programs
to teach effective coping behaviors, not only should reinforcement of
these behaviors be built into the program, but patients should be made
consciously and consistently aware of the efficacy of what they are
doing. Our study begins to suggest how perceptions and beliefs may be
utilized along with behavioral methods to effect survival rather than
relapse.
ACKNOWLEDGMENTS
John Stapleton, A. N. Oppenheim, Michelle Peleg, and Paul Jack-

son made valuable contributions to the preparation of this paper. Grate-
ful acknowledgments also to Professor Griffith Edwards, Robert
Stewart, Robert West and Dr Tim Stockwell for their invaluable com-
ments on earlier drafts of this chapter and to Jan Gallagher for her
efficient and accurate secretarial assistance. Grateful acknowledgment,
404 GLORIA K LITMAN
too, to the consultants, staff, and patients of Warlingham Park, Bexley,

Queen Elizabeth Military, the Maudsley, Bethlem Royal, and St. An-
drew's hospitals for their generous participation in this study.
REFERENCES
Armor, J. D., Polich, J. M., & Stambul, H. B. (1978). Alcho/ism and treatment. New York:
Wiley Interscience.
Beck, A. (1976). Cognitive therapy and the emotional disorders. New York: International
Universities.
Chaney, E. F., O'Leary, M. R, & Marlatt, G. A. (1978). Skill training with alcoholics.
Hodgson, R J., & Rankin, H. J. (1976). Case histories and shorter communications: Modi-
fication of excessive drinking by cue exposure. Behaviour Research and Therapy, 14, 305-
307.
Hodgson, R J., & Rankin, H. J. (1982). Cue exposure and relapse prevention. In P. Nathan
& W. Hay (Eds.), Case studies in the behavioral modification of alcoholism (pp. 207-225).
New York: Plenum Press.
Hore, B. D. (1971). Factors in alcoholism relapse. British Journal of Addiction, 66, 89-96.
Lazarus, R S. (1966). Psychological stress and the coping process. New York: McGraw-Hill.
Lazarus, R S., Averill, J. S., & Opton, E. M. (1974). The psychology of coping: Issues of
research and assessment. In G. V. Goehlo, D. A. Hamburg & J. C. Adams (Eds.),
Coping and adaptation (pp. 249-291). New York: Basic Books.
Lieber, C. S. (1979). Ethanol and the liver: a decreasing "threshold" of toxicity. American
Journal of Clinical Nutrition, 32a, 1177-1180.
Litman, G. K (1980). Relapse in alcoholism: Traditional and current approaches. In G.
Edwards & M. Grant (Eds.), Alcoholism treatment in transition. London: Croom Helm.
Litman, G. K (1982). Personal meanings and alcoholism survival: Translating subjective
experience into empirical data. In E. Shepherd & J. Watson (Eds.), Personal meanings
(pp. 129-139). New York: Wiley.
Litman, G. K, Eiser, J. R., Rawson, N. S. B., & Oppenheim, A. N. (1977). Towards a
typology of relapse: A preliminary report. Drug and Alcohol Dependence, 2, 157-162.
Litman, G. K, Eiser, J. R., Rawson, N. S. B., & Oppenheim, A. N. (1979). Towards a
typology of relapse: Differences in relapse and coping behaviours between alcoholic
relapsers and survivors. Behaviour Research and Therapy, 17, 89-94.
Litman, G. K., Stapleton, J., Oppenheim, A. N., & Peleg, M. (1983). An instrument for
measuring coping behaviours in hospitalized alcoholics: Implications for relapse pre-
vention and treatment. British Journal of Addiction, 78, 269-276.
Litman, G. K, Stapleton, J., Oppenheim, A. N., Peleg, M., & Jackson, P. (1983). Situa-
tions related to alcoholism relapse. British Journal of Addiction, 78, 381-389.
Litman, G. K, Stapleton, J., Oppenheim, A. N., Peleg, M., & Jackson, P. (1984). The
relationship between coping behaviours, their effectiveness and alcoholism relapse
and survivial. British Journal of Addiction, 79, 283-291.
Marlatt, G. A. (1979). A cognitive-behavioural model of the relapse process. In N.
Krasnegor (Ed.), Behavioral analysis and treatment of substance abuse (NIDA Research
Monograph 25). Washington, DC: US Government Printing Office.
Pequinot, G., Tuyno, A. J., & Berta, J. L. (1978). Ascitic Cirrhosis in relation to alcohol
consumption. International Journal of Epidemiology, 7, 113-120.
Rotter, J. B. (1966). Generalized expectations for internal versus external control of rein-
forcement. Psychological Monographs, 80, 1-28.
Sanchez-Craig, M., & Walker, K. (1982). Teaching coping skills to chronic alcoholics in a
coeducational halfway house. 1. Assessment of programme effectiveness. British Jour-
nal of Addiction, 77, 35-50.
Schmidt, W. (1976). Cirrhosis and alcohol consumption: An epidemiological perspective.
In G. Edwards & M. Grant (eds.), Alcoholism: New Knowledge and new responses. Bal-
timore, MD: University Park.
21
A Relapse Prevention Model for
Treatment of Alcoholics
HELEN M. ANNIS
One of the few areas of consensus in the alcoholism treatment field

involves the recognition that alcoholism is a chronic condition with a
high risk of relapse. Treatment outcome studies have reported rates of
80% or more by 6 months posttreatment discharge (Armor, Polich, &
Stambul, 1978; Gottheil, Thornton, Skolada, & Alterman, 1979), and
drinking outcomes of individual clients have been found to be highly
unstable over time (Annis & Ogborne, 1983; Finney, Moos, & Newborn,
1980, Litman, Eiser, & Taylor, 1979). It is not surprising, therefore, that,
increasingly, relapse is being recognized as an important phenomenon
for study.
Although follow-up results have typically been poor, alcoholism
treatment programs have been, on the whole, highly successful at ini-
tiating behavior change in their clients. Indeed, reports of alcoholics in
the community suggest that many alcoholics have little difficulty initiat-
ing periods of abstinence on their own. The problem in alcoholism, as in
other addiction behaviors, is one of maintaining change over time. In
the wake of disappointing long-term remission rates following treat-
ment, there has been a tendency to assume that what is needed is more
intensive programming or the development of more comprehensive
HELEN M. ANNIS • Addiction Research Foundation and University of Toronto, Toronto,

Ontario, Canada. The views expressed in this Chapter are those of the author and do not
necessarily reflect those of the Addiction Research Foundation.
407
408 HELEN M. ANNIS
multimodal treatment packages. This solution to reducing relapse rates

is likely to be seriously flawed unless it is recognized that the additional
treatment components must be designed specifically to enhance the
maintenance of behavior change.
THE NEED FOR A THEORETICAL FRAMEWORK
The magnitude of the observed relapse phenomenon among alco-

holics calls for the design of intervention strategies that address directly
the problem of the durability of treatment effects. What theoretical
framework can be drawn on to assist in the development of intervention
strategies designed to produce greater maintenance of behavior change?
A major tenet of social-learning approaches, and self-efficacy theory
(Bandura, 1977, 1978, 1981) specifically, is that the most powerful pro-
cedures for inducing behavior change may not be the most effective
techniques for producing generalization and maintenance of treatment
effects. No treatment-outcome study testing this proposition has ap-
peared to date in the alcoholism field, although a controlled trial of
relapse prevention strategies derived from self-efficacy theory is cur-
rently underway (Annis, Davis, & Levinson, 1981). In the remainder of
this chapter, it will be argued that Bandura's theory of self-efficacy pro-
vides both (a) a testable framework for conceptualizing the phenomenon
of alcoholic relapse, and (b) a basis for the design of relapse prevention
strategies that may have greater potential for the maintenance of
change.
SELF-EFFICACY THEORY
According to self-efficacy theory, treatment interventions are effec-

tive to the extent that they increase the client's expectations of personal
efficacy. An efficacy expectation is defined as a judgment that one has
the ability to execute a certain behavior pattern. (This is distinguished
from an outcome expectation, which involves a judgment of the likely
consequences such a behavior will produce.) Efficacy expectations are
hypothesized to playa major role not only in the initiation, but also in
the generalization and maintenance of coping behavior. The critical pre-
diction of the theory is that the strength of a client's efficacy expectations
will determine the nature of coping behavior and how long it will be
maintained in the face of obstacles and adverse experiences.
RELAPSE PREVENTION MODEL IN ALCOHOLISM TREATMENT 409
SELF-EFFICACY THEORY AND ANALYSIS OF ALCOHOLIC

RELAPSE
Wilson (1978a,b, 1979, 1980) has provided an excellent analysis of

how self-efficacy theory may be useful in explaining the maintenance of
alcoholism treatment effects and the alcoholic relapse process. Most
traditional alcoholism treatment programs, including that of Alcoholics
Anonymous, inculcate the belief that the alcoholic has an irreversible
disease that renders him or her uniquely vulnerable to the addiction
effects of alcohol, such that lifelong abstinence is essential. Because the
alcoholic is qualitatively different from nonalcoholics, the alcoholic will
never be able to exercise voluntary control over consumption once
drinking has been initiated. In terms of self-efficacy theory, such a
therapeutic philosophy can vitally affect relapse by deliberately mini-
mizing the alcoholic's efficacy expectations about his or her ability to
cope with alochol if any drinks were to be taken. Self-efficacy theory
would predict that avoidance of drinking (abstinence), even for a period
of years, would not develop a sense of self-efficacy about coping with
drinking. Given low-efficacy expectations, coping behavior would be
easily extinguished in the face of difficult experiences encountered in
remaining sober in the natural environment. In addition, outcome ex-
pectations emphasizing the certainty of a return to uncontrolled drink-
ing after a single drink would tend to function as a self-fulfilling proph-
ecy, drastically increasing the severity of any relapse episode.
A related factor that would be expected to increase the severity of a
relapse episode is that of "catastrophizing" (d. Bandura, 1978; Ellis,
1970). Individuals with low-efficacy expectations, or acute self-doubts
about their ability to handle drinking, would be expected to "cata-
strophize" the consequences and feel that all will be lost following a
single failure experience. Rarely do traditional alcoholism treatment pro-
grams systematically teach the alcoholic self-regulatory and social skills
to cope with the consequences typically associated with relapse. Cog-
nitive preparation for "slips" (from either an abstinence or a controlled-
drinking goal), and instructions in appropriate coping strategies would
be expected to minimize the effects of potential relapse episodes.
In summary, self-efficacy theory would appear to have important
implications for the analysis of the phenomenon of alcoholic relapse.
Extrapolating from self-efficacy theory to the area of alcoholism, we
would predict that durable treatment effects (i.e., effects that will gener-
alize across time and settings) would be a function of the development
of strong efficacy expectations with respect to coping with alcohol-relat-
ed situations in the natural environment. Because the majority of alco-
holic clients engage in sporadic posttreatment drinking, it would appear
410 HELEN M. ANNIS
to be critical that clients develop a sense of personal capability in dealing

with drinking incidents. Self-efficacy theory would suggest that it is not
the behavior per se of ingesting alcohol that is responsible for a full-
blown relapse in the posttreatment period; rather it is the meaning the
act of drinking has for the client, the coping strategies the client has
available, and the persistence with which the client engages in coping
behavior, which in turn is dependent on the presence of strong efficacy
expectations.
TREATMENT IMPLICATIONS OF SELF-EFFICACY THEORY
Empirical testing of self-efficacy theory across different types of

behavioral dysfunctions, including alcoholism, is very much in its infan-
cy. Some promising results have begun to appear in the literature on the
predictive power of self-efficacy ratings in relation to posttreatment
smoking behavior (Condiotte & Lichtenstein, 1981; Prochaska, Crimi,
Lapsanski, Martel, & Reid, 1982) and drinking behavior (Condra, 1982;
Rist & Watzl, 1983; Stiemerling, 1983). However, the use of self-efficacy
theory to design systematically a program of treatment elements specifi-
cally directed at the maintenance of change has been largely restricted to
the work of Bandura and his colleagues in relation to snake phobia and
more recently agoraphobia (Bandura, Adams, Hardy, & Howells, 1980;
Hardy, 1976).
What principles may be derived from self-efficacy theory to design a
relapse prevention program for alcoholics? What treatment strategies
that have been employed with other behavioral dysfunctions have
shown promise in the maintenance of change over time? It is proposed
that empirical work to date from social-learning laboratory studies and
clinical trials of self-efficacy theory support the extrapolation of the fol-
lowing principles for the development of a relapse treatment model:
1. A client's judgment of perceived self-efficacy will be the best
predictor of future drinking behavior in high-risk situations for relapse.
For good predictive accuracy, self-efficacy judgments on the part of the
client should be made in relation to highly specific drinking situations
(see the Situational Confidence Questionnaire, following).
2. Although cognitive mechanisms mediate behavior, the most
powerful methods of changing drinking behavior will be performance
based. Therefore, treatment should focus on having the client engage in
performance assignments with regard to specific high-risk situations for
alcoholic relapse (see the Inventory of Drinking Situations, following).
(For examples of the demonstrated superiority of performance tasks
with phobic behaviors, see Bandura, Blanchard, & Ritter, 1969, Bandura
et al., 1980; Blanchard, 1970; with obsessive-compulsive disorders see

Rachman & Hodgson, 1979; and with sexual dysfunctions see Kockott,
Dittmar, & Nusselt, 1975; and Mathews et al., 1976.)
3. Performance tasks should be ordered in therapy from easier to
more difficult (d. Bandura, 1978). This can be accomplished through the
use of client's ratings of perceived self-efficacy in relation to different
risk situations for drinking.
4. In structuring performance tasks in relationship to dangerous
drinking situations during treatment, it is critical that the therapist ar-
range conditions so that clients can perform successfully despite their
incapacities (d. Bandura, 1977). In addition to the use of graduated tasks
this may be facilitated by the use of a variety of response-induction aids.
These include the use of modeling and rehearsal of activities (Chaney,
O'Leary, & Marlatt, 1978; Marlatt & Gordon, 1985), joint performance
with the therapist or a responsible collateral (Bandura et al., 1980), pro-
grammed relapse (Marlatt & Gordon, 1985), alternative coping strategies
(Beck, 1976; Sanchez-Craig, 1975; Sobell & Sobell, 1973), and the use of
protective aids such as anti-alcohol drugs (Peachey & Annis, 1983).
5. A two-phase treatment plan should be followed with Phase 1
concentrating on initiation strategies and Phase 2 on maintenance strat-
egies. Powerful strategies for the initiation of a change in drinking be-
havior include the response-induction aids previously outlined. Howev-
er, it is important that such external aids be gradually withdrawn in
Phase 2 when the emphasis shifts to ensuring that the client's cognitions
or self-inferences from mastery experiences are consistent with those
known to facilitate strong, generalized behavior change (see following).
This goal may also be facilitated in Phase 2 by having the client take a
more active role in the designing of performance tasks leading to self-
directed mastery experiences.
6. Unfortunately, successful experiences in controlling drinking be-
havior will not always produce the improvement in the client's percep-
tion of self-efficacy in relation to future drinking situations, which is
necessary for treatment effects to be maintained over time. Therefore, it
is necessary in therapy to monitor the client's cognitions in relation to
successful performances. Although research on determinants of efficacy
judgments is in its infancy, Bandura (1978) has commented on four
factors that engender strong efficacy expectations in a client following a
successful experience in a high-risk situation. These involve a percep-
tion on the part of the client that (a) the situation was challenging (Le., at
one time the situation would have been highly risky), (b) to succeed in
mastering the situation, only a moderate degree of effort was needed, (c)
little external aid was involved (Le., the client himself or herself was
responsible for the success), and (d) the success was part of an overall
412 HELEN M. ANNIS
pattern of improved performance. Additional factors derived from the

literature on self-perception and attribution theory that have been found
to enhance positive self-inferences include a perception on the part of
the client that (e) an increase in personal control was demonstrated, and
(f) the successful performance was highly relevant to problematic situa-
tions frequently encountered. These six cognitive factors influencing the
formation of judgments of self-efficacy should be carefully monitored
throughout treatment (see the Cognitive Appraisal Questionnaire,
following).
7. In monitoring a client's cognitions, it may be found that self-
defeating ideation is interfering with the enhancement of the client's
self-efficacy following a successful experience in controlling drinking
behavior. In such cases, direct cognitive manipulations (e.g., Meichen-
baum's, 1977, self-instructional training; Beck, Rush, Shaw, & Emery's,
1979, recording of dysfunctional thoughts) may be necessary to foster
gains in self-efficacy that will lead to greater maintenance of change in
drinking behavior.
In summary, a number of principles derived from the literature on
self-efficacy theory in relation to the successful maintenance of change
in other behavior dysfunctions are outlined above. It is proposed that
these principles provide guidelines for the development of a treatment
model aimed at reducing the frequency and severity of alcoholic relapse
(d. Annis et al., 1981).
A RELAPSE PREVENTION MODEL
The essence of the proposed relapse prevention model for alco-

holics, derived from Bandura's theory of self-efficacy, involves a highly
individualized microanalysis of drinking behavior within what are, for
that client, high-risk situations for alcoholic relapse. The model is pre-
sented in Figure 1. Because it has been repeatedly demonstrated that
global measures in personality or cognitive functioning do not have the
same predictive power for particular behaviors as do situation-specific
person measures (e.g., Endler, 1975), self-efficacy judgments by the
client must be taken with respect to highly specific drinking situations.
Similarly, cognitive appraisals of past successes and failures must be
made in relation to these same specific drinking situations. What is
needed in treatment, therefore, is a microanalysis of each high-risk
drinking situation for a client in terms of the client's cognitive appraisal
of that situation, his or her resulting expectation of coping behavior in
that situation, which, in turn, will be a strong predictor of the client's
actual future drinking behavior in that situation.
High Risk SHuation Cognitive Appraisal
•
Efficacy Drlnk1ng
for Alcoholic Relapse of Past Perfomance . . . . Expectation Be IV or
ASSESSMENT INSTRUMENTS
Inventory of Drinking Cognlti ve Appra; s a1 Situational Confidence

Situations Ques t1 onnai re Ques tionnai re
FIGURE l. Microanalysis of drinking behavior in a high-risk situation.
It should be noted that assessment and treatment planning are

highly interrelated in this model. Assessment of the three central ele-
ments in the model (risk situations for relapse, cognitions associated
with past performance, and efficacy expectations) form the basis for
designing and ordering homework assignments or performance tasks to
be undertaken by the client in treatment. Because appropriate instru-
mentation for this purpose has been largely lacking in the alcoholism
literature, the present author designed three behavioral assessment in-
struments to assess the central elements of the model. The development
of each of these instruments and their use in implementing relapse
prevention strategies with alcoholic clients is described in the following.
INVENTORY OF DRINKING SITUATIONS (IDS)
DESCRIPTION AND DEVELOPMENT
The Inventory of Drinking Situations, IDS, (Annis, 1982a) is a 100-

item questionnaire designed to assess situations in which a client drank
heavily over the past year (see Appendix A). Suggestions for item con-
tent were taken from a number of sources, including Litman's Dan-
gerousness Questionnaire (Litman, Eiser, Rawson, & Oppenheim,
1979), Chaney'S Situational Competency and Situational Difficulty Tests
(Chaney et ai., 1978), Marlatt's Drinking Profile (Marlatt, 1976), Wilkin-
son's Self-Efficacy Inventory (Wilkinson & Martin, 1979), Deardorff's
Situations for Drinking Questionnaire (Deardorff, Melges, Hout, & Sav-
age, 1975) and discussions with clinicians, former alcoholics, and alco-
holic clients. A draft of the resulting questionnaire was sent to five
clinicians who had extensive experience working with alcoholics to solic-
it comments on item clarity and item coverage of the universe of com-
mon alcoholic relapse situations. Similar feedback was also solicited in a
414 HELEN M. ANNIS
pilot testing of the questionnaire on alcoholics admitted to an inpatient

employee assistance program.
The final 100 items of the questionnaire are designed to assess eight
categories of alcoholic relapse divided into two major classes: (a) Person·
al States, in which drinking involves a response to an event that is
primarily psychological or physical in nature; and (b) Situations Involv·
ing Other People, in which a significant influence of another individual
is involved. Personal States are further subdivided into five categories:
Negative Emotional States (20 items); Negative Physical States (10
items); Positive Emotional States (10 items); Testing Personal Control (to
items); and Urges and Temptations (to items). Situations Involving
Other People are subdivided into three categories: Interpersonal Con·
flict (20 items); Social Pressure to Drink (to items); and Positive Emo·
tional States (10 items). This classification system is based on the work of
Marlatt and his associates, in which chronic male alcoholics were inter·
viewed concerning the circumstances surrounding their first relapse epi·
sode following discharge from two abstinence·oriented inpatient pro·
grams; content analysis of the responses lead to the derivation of the
eight·category classification system (Marlatt, 1978, 1979a,b; Marlatt &
Gordon, 1980).
In order to test for the reliability with which the 100 items of the
Inventory of Drinking Situations could be placed in categories, three
raters were instructed in the classification system and asked to sort the
items into the eight categories. High reliability of item placement was
obtained (interrater reliability of 92% to 99%). Further psychometric
evaluation of the questionnaire is ongoing, and normative data are being
collected.
SCORING AND PLOITING OF CLIENT PROFILE
The client is asked to indicate on the questionnaire the frequency

with which he or she drank heavily over the past year in each of 100
situations. Responses are scored: 1 = never, 2 = rarely, 3 = frequently, 4
= almost always (see Appendix A). Eight subscores are calculated, one
for each category, by a simple addition of scores obtained for each ques·
tion within the category.
A client profile is plotted showing the client's areas of greatest risk
for drinking (see Figure 2). The length of a bar graph for each category,
plotted as a Problem Index varying from 0 to 100, indicates the magni·
tude of the problem experienced over the past year across situations in
that category. The Problem Index for each category is calculated as
follows:
P bl I d = Category Score - No. of Items in Category x 100
ro em n ex Maximum Possible Score - No. of Items in Category
100;--
90-
80-
70-
60-
Problem
Index 50-
40-
301-
I
20f-
10~
o'-----;-:un::::.~.:::.n:-'- - ; ;:PhI"''::;.'--;;;Ple~
.. ~Y' •• ~.n:-'----;T~,~,ng-;-----;iUrg~.st-:;------;Con~fllct-:;--~P;;;:~U="'--;;;PI~.nl
•• .... ...
Emotions Oiscomfort Emotions
.....
Control
Alcohol
Temptanons
to Drink
with Other. from Others
to Orlnk
Times with
Others
I I
Personal States Situations Involving
Other People
FIGURE 2. Client profile.
ApPLICA nON OF TREATMENT

At the inception of treatment, clients are asked by the therapist to
complete the Inventory of Drinking Situations. Based on the responses,
a profile is plotted showing the client's areas of high-risk situations for
drinking. This profile forms the basis for a discussion with the client of
the importance of antecedent events in understanding and modifying
his or her drinking behavior. Further specification and clarification then
takes place between therapist and client of those particular antecedents
(personal states and situations involving other people) that will need to
be worked on in treatment. The strategy of beginning with easier situa-
tions in terms of homework assignments, and progressing to more diffi-
cult tasks is explained, and the patient's involvement in the treatment
plan is solicited.
SITUATIONAL CONFIDENCE QUESTIONNAIRE (SCQ)

The Situational Confidence Questionnaire (SCQ) (Annis, 1982b) is a
lOO-item questionnaire designed to assess Bandura's concept of self-
416 HELEN M. ANNIS
efficacy in relation to a client's perceived ability to cope effectively with

alcohol. The 100 items parallel the 100 drinking situations employed in
the Inventory of Drinking Situations. (For a description of the develop-
ment of these situations, see IDS preceding.) As with the IDS, the eight-
category classification system derived from the work of Marlatt and his
associates is employed to categorize high-risk drinking areas.
SCORING
Clients are asked to imagine themselves in each of the 100 situations

and indicate on the scale provided how confident they are that they will
be able to resist the urge to drink heavily in that situation. Scale re-
sponses are, 0 = not at all confident, 20 = 20% confident, 40 = 40%
confident, 60 = 60% confident, 80 = 80% confident, and 100 = 100% or
very confident. Following Bandura's framework, three self-efficacy
scores may be calculated for each of the eight categories. These are (a)
level/magnitude of expectancies, calculated as the proportion of items
given a confidence rating of 60 or above by the client; (b) strength of
expectancies, calculated as the sum of confidence ratings across items,
and (c) generality of expectancies, calculated as the correlation of
strength scores across categories.
ApPLICATION IN TREATMENT
Clients complete the total questionnaire at the beginning of treat-

ment. The strength-of-expectancy stores provide the basis whereby
therapist and client develop the broad outline of a hierarchy of drinking
situations in which the client feels increasingly less confidence to cope
effectively (nonproblematically) with alcohol. To help ensure a series of
graded mastery experiences over the course of treatment, initial tasks or
homework assignments are designed around situations with moderately
high strength or confidence ratings; as the client begins to cope effective-
ly with alcohol in these situations, progression is made to drinking
situations with lower strength or confidence ratings until these, too, are
mastered. Throughout treatment, clients are periodically reassessed on
items of the SCQ relevant to their homework assignments. This is done
to check whether experiences in coping effectively with alcohol are re-
sulting in strong efficacy expectations. If it is found that self-efficacy is
not being strengthened despite successful performances on homework
tasks, the client's cognitions in relation to those experiences must be
explored (see following).
COGNITIVE APPRAISAL QUESTIONNAIRE
The Cognitive Appraisal Questionnaire (Annis, 1982c) is a struc-

tured interview designed to explore, in depth, cognitions associated
with a past successful experience in controlling drinking behavior in a
particular drinking situation. The operation of six cognitive factors hy-
pothesized to influence the formation of judgments of self-efficacy is
assessed. Questions explore whether or not the client perceived that (a)
the situation was challenging, (b) only a moderate degree of effort was
needed, (c) little external aid was involved, (d) the success was part of an
overall pattern of improved performance, (e) an increase in personal
control was demonstrated, and (f) the success was highly relevant to
problematic drinking situations frequently encountered.
ApPLICATION IN TREATMENT
Because success experiences do not automatically raise judgments

of self-efficacy but are subject to cognitive interpretation on the part of
the client (Bandura 1978, 1981), the client's appraisals of success experi-
ences are monitored over the course of therapy to ensure that appropri-
ate cognitions are being developed. When a client's efficacy expectations
are not strengthened despite the successful completion of homework
assignments, the Cognitive Appraisal Questionnaire provides a useful
interview procedure to explore cognitive factors that may be interfering
with the enhancement of self-efficacy. Any report of self-defeating idea-
tion on the part of the client on this questionnaire may be used by the
therapist to adjust the tasks assigned in an attempt to generate more
constructive cognitions. Such adjustments may involve the introduction
of a more graduated series of tasks to promote a perception on the part
of the client that only a moderate degree of effort was necessary to bring
about control over drinking; the adoption of more difficult or realistic
assignments to promote a perception that the task was challenging or
highly relevant to drinking situations frequently encountered; the re-
moval of external aids, such as the involvement of collaterals in drinking
assignments, or the use of an antialcohol drug, to promote self-attribu-
tion of improved performance; or the use of direct cognitive manipula-
tions (e.g., self-instructional training) in an attempt to directly alter dys-
functional self-statements. Unless a client's experiences in drinking
situations can be arranged so as to foster gains in self-efficacy, it is
418 HELEN M. ANNIS
unlikely that the changes brought about in drinking behavior during

treatment will be maintained after discharge.
SUMMARY
A relapse prevention model for the treatment of alcoholics is pro-

posed based on Bandura's theory of self-efficacy. The model involves a
microanalysis of the drinking behavior of a client in high-risk situations
for alcoholic relapse. The analysis is highly individualized and is situa-
tion specific. It is predicted that a client's cognitive appraisal of his or her
past performance in a particular drinking situation will produce efficacy
expectations about his or her coping ability, which, in turn, will deter-
mine future drinking behavior in that situation and its maintenance over
time. Instruments are described for the assessment of three central ele-
ments of the model: risk situations for relapse (Inventory of Drinking
Situations), cognitions associated with past performance (Cognitive Ap-
praisal Questionnaire), and efficacy expectations (Situational Confi-
dence Questionnaire). Client responses on the assessment instruments
form the basis for the design and ordering of homework assignments to
be undertaken by the client in treatment. An individualized gradient of
performance tasks is the central component in the treatment process.
Homework assignments, graded in terms of increasing difficulty over
the course of treatment, focus on promoting successful performance in
high-risk drinking situations as the primary vehicle of behavior change.
This relapse prevention model is currently being evaluated in two
randomized control outcome trials being conducted at the Addiction
Research Foundation of Ontario. In one study (Annis et al., 1981), the
relapse prevention model is being compared with more traditional
methods of aftercare; it is hypothesized that relapse prevention strat-
egies will result in higher levels of client self-efficacy with respect to
drinking situations, and greater generalization and maintenance over
time of treatment effects. In the second study (Peachey & Annis, 1983),
the efficacy of the relapse prevention procedures in teaching alcoholics
to use a short-acting alcohol-sensitizing drug (calcium carbimide) in an-
ticipation of high-risk drinking situations is being compared with the
more traditional use of this drug in medical primary care; clients receiv-
ing the relapse prevention procedures are encouraged to use the drug
with decreasing frequency in the final phase of therapy. Foilow-up re-
sults for the first study are expected to be available in 1985 and for the
second study in 1986.
REFERENCES
Annis, H. M. (1982a). Cognitive Appraisal Questionnaire. Toronto: Addiction Research Foun-

dation of Ontario.
Annis, H. M. (1982b). Inventory of Drinking Situations. Toronto: Addiction Research Foun-
dation of Ontario.
Annis, H. M. (1982c). Situational Confidence Questionnaire. Toronto: Addiction Research
Foundation of Ontario.
Annis, H. M., & Ogborne, A. C. (1983). The temporal stability of alcoholism treatment outcome
results. Unpublished manuscript, Addiction Research Foundation of Ontario.
Annis, H. M., Davis, c., & Levinson, T. (1981). The prevention of alcoholic relapse: A research
proposal. Unpublished manuscript, Addiction Research Foundation of Ontario.
Wiley.
Bandura, A. (1977). Self-efficacy: Toward a unifying theory of behavioral change. Psycho-
logical Review, 84, 191-215.
Bandura, A. (1978). Reflections on self-efficacy. Advances in Behavioral Research and Therapy,
1, 237-269.
Bandura, A. (1981). Self-referent thought: A developmental analysis of self-efficacy. In J.
H. Flavell & L. Ross (Eds.), Social cognitive development: Frontiers and possible futures (pp.
200-239). New York: Cambridge University Press.
Bandura, A., Blanchard, E. B., & Ritter, B. (1969). Relative efficacy of desensitization and
modeling approaches for inducing behavioral, affective, and attitudinal changes. Jour-
nal of Personality and Social Psychology, 13, 173-199.
Bandura, A., Adams, N. E., Hardy, A. B., & Howells, G. N. (1980). Tests of the generality
of self-efficacy theory. Cognitive Therapy and Research, 4(1), 39-66.
Beck, A. T. (1976). Cognitive therapy and emotional disorders. New York: International Univer-
sities Press.
Beck, A. T., Rush, A. J., Shaw, B. F., & Emery, G. (1979). Cognitive therapy of depression: A
treatment manual. New York: Guilford Press.
Blanchard, E. B. (1970). Relative contributions of modeling, informational influences, and
physical contact in extinction of phobic behavior. Journal of Abnormal Psychology, 76,
55-61.
Journal of Consulting and Clinical Psychology, 46(5), 1092-1104.
Condiotte, M. M., & Lichtenstein, E. (1981). Self-efficacy and relapse in smoking cessation
programs. Journal of Consulting and Clinical Psychology, 49(5), 648-658.
Condra, M. St. John. (1982). The effectiveness of relapse-prevention training in the treatment of
alcohol problems. Unpublished doctoral dissertation, Queen's University, Canada.
Deardoff, C. M., Melges, F. T., Hout, C. N., & Savage, D. J. (1975). Situations related to
drinking alcohol: A factor analysis of questionnaire responses. Journal of Studies on
Alcohol, 36(9), 1184-1195.
Endler, N. S. (1975). A person-situation interaction model of anxiety. In C. D. Spielberger
& I. G. Sarason (Eds.), Stress and anxiety (Vol. 1, pp. 145-164). Washington, DC:
Hemisphere Publishing Corporation.
Ellis, A. (1970). The essence of rational psychotherapy: A comprehensive approach to treatment.
New York: Institute for Rational Living.
Finney, J. W., Moss, R. H., & Newborn, C. R. (1980). Posttreatment experiences and
tion. Journal of Consulting and Clinical Psychology, 48(1), 17-29.
420 HELEN M. ANNIS
Gottheil, E., Thornton, c., Skolada, T., & Alterman, A. (1979). Follow-up study of alco-
holics at 6, 12 and 24 months. In M. Galanter (Ed.), Currents in alcoholism Vol. IV:
Treatment, rehabilitation and epidemiology (pp. 91-109). Toronto: Grune & Stratton.
Hardy, A. B. (1976). Agoraphobia: Symptoms, causes, treatment. Menlo Park, CA: Terrap.
Kockott, G., Dittmar, F., & Nusselt, L. (1975). Systematic desensitization of erectile impo-
tence: A controlled study. Archives of Sexual Behavior, 4, 493-500.
Litman, G. K., Eiser, J. R, Rawson, N. S. B., & Oppenheim, A. N. (1979). Differences in
relapse precipitants and coping behavior between alcohol relapsers and survivors.
Behavior Research and Therapy, 17, 89-94.
Litman, G. K., Eiser, J. R, & Taylor, C. (1979). Dependence, relapse and extinction: A
theoretical critique and a behavioral examination. Journal of Clinical Psychology, 35(1),
192-199.
Marlatt, G. A (1976). The drinking profile: A questionnaire for the behavioral assessment
of alcoholism. In E. J. Mash & L. G. Terdal (Eds.), Behavior therapy assessment: Diag-
nosis, design and evaluation. New York: Springer.
Marlatt, G. A (1978). Craving for alcohol, loss of control, and relapse: A cognitive-behav-
ioral analysis. In P. E. Nathan, G. A. Marlatt and T. Laberg (Eds.), Alcoholism: New
directions in behavioral research and treatment (pp. 271-314). New York: Plenum Press.
Marlatt, G. A. (1979a). Alcohol use and problem drinking: A cognitive-behavioral analysis.
In P. C. Kendall & S. D. Hollon (Eds.), Cognitive-behavioral interventions: Theory, research
and procedures (pp. 319-355). New York: Academic Press.
Marlatt, G. A. (1979b). A cognitive-behavioral model of the relapse process. In N. A
Krasnegor (Ed.), Behavioral analysis and treatment of substance abuse. (National Institute
on Drug Abuse Research Monograph, 25, pp. 191-200). Rockville, Maryland.
Marlatt, G. A., & Gordon, J. R (1980). Determinants of relapse: Implications for the
maintenance of behavior change. In P. Davidson & S. Davidson (Eds.), Behavioral
medicine: Changing health lifestyles. New York: Brunner/Maze!'
Marlatt, G. A, & Gordon, J. R (Eds.). (1985). Relapse prevention: Maintenance strategies in the
treatment of addictive behaviors. New York: Guilford Press.
Mathews, A., Bancroft, J., Whitehead, A., Hackmann, A., Julier, D., Bancroft, J., Garth,
D., & Shaw, P. (1976). The behavioral treatment of sexual inadequacy: A comparative
study. Behavior Research and Therapy, 14, 427-436.
Meichenbaum, D. (1977). Cognitive-behavior modification: An integrative approach. New York:
Plenum Press.
Peachey, J. E., & Annis, H. M. (1983). Studies on the efficacy of calcium carbimide in Alcoholism
Treatment: II A controlled study of self-efficacy therapy with calcium carbimide in chronic
alcoholics. Research proposal in progress, Addiction Research Foundation of Ontario.
Prochaska, J. 0., Crimi, P., Lapsanski, D., Martel, L., & Reid, P. (1982). Self-change
processes, self-efficacy and self-concept in relapse and maintenance of cessation of
smoking. Psychological Reports, 51, 983-990.
Rachman, S., & Hodgson, R (1979). Obsessions and compulsions. Englewood Cliffs, NJ:
Prentice-Hall.
Rist, F., & Watzl, H. (1983). Self assessment of relapse risk and assertiveness in relation to
treatment outcome of female alcoholics. Addictive Behaviors, 8, 121-127.
Sanchez-Craig, M. (1975). A self-control strategy for drinking tendencies. The Ontario
Psychologist, 7, 25-29.
Stiemerling, N. (1983). Relapse in alcohol abusers: A short term longitudinal study. Unpublished
doctoral dissertation, Queen's University, Canada.
Wilkinson, D. A., & Martin, G. W. (1979). Comparison of a brief intervention with a broad
spectrum behavioural treatment applied under three conditions of reinforcement of therapeutic

progress: Application to young drug abusers. Unpublished research proposal, Addiction
Research Foundation of Ontario.
Wilson, G. T. (1978a). Booze, beliefs and behavior: Cognitive processes in alcohol use and
abuse. In P. E. Nathan, G. A. Marlatt, & T. Leberg (Eds.), Alcoholism: New directions in
behavioral research and treatment (pp. 315-339). New York: Plenum Press.
Wilson, G. T. (1978b). The importance of being theoretical: A commentary on Bandura's
"Self-efficacy: Towards a unifying theory of behavioral change". Advances in Behavior,
Research and Therapy, 1, 217-230.
Wilson, G. T. (1979). Perceived control and the theory and practice of behavior therapy. In
L. C. Perlmuter & R. A. Monty (Eds.), Choice and perceived control. (pp. 175-183). NJ:
Hillsdale, N.J.: Lawrence Erlbaum.
Wilson, G. T. (1980). Cognitive factors in lifestyle changes: A social learning perspective.
In P. O. Davidson & S. M. Davidson (Eds.), Behavioral medicine: Changing health life-
styles (pp. 3-37). New York: Brunner/Maze!'
Appendix A
Inventory of Drinking Situations
Listed below are a number of situations or events in which some people drink heavily.
Read each item carefully, and answer in terms of your own drinking over the past year.
If you "NEVER" drank heavily in that situation, circle "1"

If you "RARELY" drank heavily in that situation, circle "2"
If you "FREQUENTLY" drank heavily in that situation, circle "3"
If you" ALMOST ALWAYS" drank heavily in that situation, circle "4"
I DRANK HEA VILY

Almost
Never Rarely Frequently always
1. When I had an argument
with a friend. 1 2 3 4 6
2. When I was depressed about

things in general. 2 3 4
3. When I felt that things were

going to work out well for
me at last. 1 2 3 4 3
4. When I felt shaky and sick. 1 2 3 4 2
5. When I would decide to test

my willpower by showing
that I really could stop after
one or two drinks. 2 3 4 4
6. When good friends would

drop by and I would be full
of good feelings. 1 2 3 4 8
(continued)
423
424 HELEN M. ANNIS
ApPENDIX A (Continued)
I DRANK HEAVILY
Almost
7. When I would see an
advertisement for my
favourite booze. 1 2 3 4 5
8. When I felt uneasy in the

presence of someone. 1 2 3 4 6
9. When someone criticized me. 2 3 4 6
10. When I would be invited to

someone's home and they
would offer me a drink. 1 2 3 4 7
11. When I would have trouble

sleeping. 1 2 3 4 2
12. When I wanted to heighten

my sexual enjoyment. 1 2 3 4 8
13. When I would get a bottle of

my favourite booze as a prize
or present. 1 2 3 4 5
14. When I was enjoying myself. 2 3 4 3
15. When I would be in a social

situation in which I had
always drunk in the past. 1 2 3 4 7
16. When I would become sad at

the memory of something
that had happened. I 2 3 4 I
17. When I would start to believe

that alcohol was no longer a
problem for me. I 2 3 4 4
18. When other people around

me made me tense. 2 3 4 6
19. When I would be out with

friends and they would stop
by a bar for a drink. I 2 3 4 7
(continued)
APPENDIX A 425
I DRANK HEA VIL Y

Almost
20. When I would begin to think
how cool and satisfying a
drink would be. I 2 3 4 5
21. When I wanted to feel closer

to someone I liked. I 2 3 4 8
22. When someone in the same

room would be drinking. I 2 3 4 7
23. When I felt that there was

nowhere left to turn. 2 3 4 I
24. When I felt that I had let

myself down. 2 3 4
25. When I felt sexually rejected. 2 3 4 6
26. When I was bored. I 2 3 4
27. When I was unable to

express my feelings to
someone. I 2 3 4 6
28. When other people treated

me unfairly. I 2 3 4 6
29. When I would remember

how good it tasted. I 2 3 4 5
30. When I felt rejected by

friends. I 2 3 4 6
31. When I felt confident and

relaxed. 2 3 4 3
32. When I would see something

that reminded me of
drinking. I 2 3 4 5
33. When I would begin to feel

fed up with life. I 2 3 4 I
(continued)
426 HELEN M. ANNIS
I DRANK HEAVILY
Almost
34. When I was troubled and I

wanted to think more clearly. 2 3 4 1
35. When I was lonely. 1 2 3 4 1
36. When I would convince

myself that I was a new
person now and could take a
few drinks. 1 2 3 4 4
37. When I was feeling on top of

the world. 2 3 4 3
38. When I would pass by a

liquor store. 1 2 3 4 5
39. When I would be in a

situation in which I was in
the habit of having a drink. 1 2 3 4 5
40. When I felt drowsy and

wanted to stay alert. 1 2 3 4 2
41. When I was tired. 1 2 3 4 2
42. When I was in physical pain. 1 2 3 4 2
43. When I would feel confident

that I could handle a few
drinks. 2 3 4 4
44. When someone close to me

was suffering. 2 3 4 6
45. When I would start thinking

that I would never know my
limits with alcohol unless I
tested them. 1 2 3 4 4
46. When I would be out with

friends lion the town" and
wanted to increase my
enjoyment. 1 2 3 4 8
(continued)
APPENDIX A 427
I DRANK HEAVIL Y
Almost
47. When I would unexpectedly
find a bottle of my favourite
booze. 2 3 4 5
48. When I was having a good

conversation with someone
and wanted to recount some
really good stories. 1 2 3 4 8
49. When I would be offered a

drink and would feel
awkward about refusing. 2 3 4 7
50. When other people didn't

seem to like me. 1 2 3 4 6
51. When I felt nauseous. 1 2 3 4 2
52. When I felt unsure that I

could measure up to other
people's expectations. 1 2 3 4 6
53. When I felt under a lot of

pressure. 1 2 3 4 1
54. When I would wonder about

my self-control over alcohol
and would feel like having a
drink to try it out. 1 2 3 4 4
55. When nothing I did seemed

right to me. 2 3 4 1
56. When other people interfered

with my plans. 1 2 3 4 6

that I was finally cured and
could handle alcohol. 1 2 3 4 4
58. When everything was going

well. 1 2 3 4 3
(continued)
428 HELEN M. ANNIS
I DRANK HEAVILY
Almost
59. When I felt no one really

cared what happened to me. I 2 3 4 I
60. When I would be at a party

and other people would be
drinking. I 2 3 4 7
61. When I felt unsure of myself

and wanted to function
better. I 2 3 4 I
62. When pressure would build

up at work because of the
demands of my supervisor. 2 3 4 6
63. When I couldn't seem to do

things I tried to do. I 2 3 4 I
64. When I was afraid that things

weren't going to work out. I 2 3 4 I
65. When I felt satisfied with

something I had done. 1 2 3 4 3
66. When I felt jealous over

something someone had
done. 2 3 4 6
67. When I would pass by a bar. I 2 3 4 5
68. When I felt empty inside. I 2 3 4 I
69. When I would be in a

restaurant and the people
with me would order drinks. 2 3 4 7
70. When I felt exhausted. I 2 3 4 2
71. When everything was going

badly for me. I 2 3 4 I
72. When I wanted to celebrate

with a friend. 1 2 3 4 8
(continued)
APPENDIX A 429
I DRANK HEAVIL Y
Almost
73. When someone would
pressure me to "be a good
sport" and have a drink. 2 3 4 7
74. When I would start to feel

guilty about something. I 2 3 4 I
75. When I felt jumpy and

physically tense. 2 3 4 2
76. When I was angry at the way

things had turned out. I 2 3 4 I
77. When I would feel under a

lot of pressure from family
members at home. I 2 3 4 6
78. When something good would

happen and I would feel like
celebrating. I 2 3 4 3
79. When I was feeling content

with my life. 2 3 4 3

that I wasn't really hooked
on alcohol. I 2 3 4 4
81. When I would start to think

that just one drink could
cause no harm. I 2 3 4 4
82. When I would be having fun

with friends and wanted to
increase our enjoyment. I 2 3 4 8
83. When I felt confused about

what I should do. I 2 3 4
84. When I would meet a friend

and he/she would suggest
that we have a drink
together. I 2 3 4 7
(continued)
430 HELEN M. ANNIS
I DRANK HEAVILY
Almost
85. When I would want to
celebrate special occasions
like Christmas or birthdays. 1 2 3 4 3
86. When I had a headache. 1 2 3 4 2
87. When I was not getting along

well with others at work. 1 2 3 4 6
88. When I would be enjoying

myself at a party and wanted
to feel even better. 2 3 4 8
89. When I would suddenly have

an urge to drink. 1 2 3 4 5
90. When I would think of the

chances I had missed in life. 1 2 3 4 1
91. When I wanted to prove to

myself that I could take a few
drinks without becoming
drunk. 1 2 3 4 4
92. When there were fights at

home. 1 2 3 4 6
93. When I was enjoying a meal

with friends and felt that a
drink would make it even
more enjoyable. 1 2 3 4 8
94. When there were problems

with people at work. 1 2 3 4 6
95. When I would be relaxed

with a good friend and
wanted to have a good time. t 2 3 4 8
96. When my boss would offer

me a drink. 1 2 3 4 7
97. When my stomach felt like it

was tied in knots. 1 2 3 4 2
(continued)
APPENDIX A 431
I DRANK HEAVILY
Almost
98. When I felt happy at the
memory of something that
had happened. 1 2 3 4 3
99. When I felt that I needed

courage to face up to
someone. 2 3 4 6
100. When I felt that someone

was trying to control me and
I wanted to feel more
independent. 1 2 3 4 6
H. M. Annis, Ph.D., Addiction Research Foundation © 1982. Permission to use this test may be
obtained by writing Helen M. Annis, Clinical Institute, 33 Russell Street, Toronto, Ontario, Canada
M5S 2S1.
Appendix B
Situational Confidence Questionnaire
Listed below are a number of situations or events in which some people experience a
drinking problem.
Imagine yourself as you are right now in each of these situations. Indicate on the scale
provided how confident you are that you will be able to resist the urge to drink heavily in
that situation.
Circle 100 if you are 100% confident right now that you could resist the urge to drink
heavily; 80 if you are 80% confident; 60 if you are 60% confident. If you are more unconfi-
dent than confident, circle 40 to indicate that you are only 40% confident that you could
resist the urge to drink heavily; 20 for 20% confident; 0 if you have no confidence at all
about that situation.
I would be able to resist the urge to

drink heavily
not at all very
confident confident
1. If I had an argument with a friend, 0 20 40 60 80 100 6
2. If I were depressed about things in
general, 0 20 40 60 80 100 1
3. If I felt that things were going to work
out well for me at last, 0 20 40 60 80 100 3
4. If I felt shaky and sick, 0 20 40 60 80 100 2
5. If I would decide to test my willpower
by showing that I really could stop
after one or two drinks, 0 20 40 60 80 100 4
100.
H. M. Annis, Ph.D., Addiction Research Foundation © 1982. A copy of the complete test and permis-
sion to use the test may be obtained by writing Helen M. Annis, Addiction Research Foundation, 33
Russell Street, Toronto, Ontario, Canada M5S 2S1.
433
22
Aftercare in Alcoholism Treatment
A Review
JOANNE R. ITO AND DENNIS M. DONOVAN
Successful treatment of a chronic condition, such as alcohol depen-

dence, often involves extended rehabilitation. As such, the goal of re-
habilitation is to move the patient back into normalized interpersonal
and community life. Alcohol dependence, like other addictive behav-
iors, poses a special challenge to the treatment community. Treatment
programs are successful in helping patients achieve sobriety, but are
often less successful in helping patients maintain sobriety. Alcohol de-
pendence has a rate of recidivism second only to schizophrenia (Na-
tional Institute of Mental Health, 1973) and has a high rate of relapse
following treatment (Hunt, Barnett, & Branch, 1971; Litman, Eiser, &
Taylor, 1979). This suggests that inpatient treatment may be necessary
but not sufficient for full recovery. In fact, Hunt, Barnett, and Branch's
(1971) data indicate that patients are at great risk for relapse in the first
month following inpatient discharge.
An earlier version of this chapter was presented at the Third International Conference
on the Treatment of Addictive Behaviors, North Berwick, Scotland, 1984.
JOANNE R. ITO • Department of Psychology, University of Washington, and Health

Services Research and Development, Seattle Veterans Administration Medical Center,
Seattle, Washington 98108. DENNIS M. DONOVAN • Department of Psychiatry and
Behavioral Sciences, University of Washington, and Alcohol Dependence Treatment Pro-
gram, Seattle Veterans Administration Medical Center, Seattle, Washington
98108. Preparation of this manuscript was partially supported by the Health Services
Research and Development Training Program at the Seattle Veterans Administration Medi-
cal Center and the University of Washington.
435
436 JOANNE R. ITO AND DENNIS M. DONOVAN
Tuchfeld (1981) and Litman and Oppenheim (unpublished manu-

script) describe the experiences of spontaneous and treated long-term
recoveries, respectively. These authors outline the nature of life changes
and coping strategies important in successfully maintaining sobriety.
They suggest that a crisis breaks the alcoholic's defenses and thus pre-
cipitates the alcoholic's journey into sobriety. Recovering alcoholics
must then seek or create an environment and life-style supportive of
their sobriety. This includes avoiding external situations associated with
drinking, developing ways of refusing drinks offered, and avoiding for-
mer drinking companions. It is also important for them to develop non-
drinking leisure activities to fill the void left when they stop drinking.
These reports are rich in clinical information, but reflect only the begin-
ning of a growing interest in the processes that underlie successful
maintenance of sobriety.
Aftercare treatment has been held up as an important component of
treatment that is oriented towards maintenance of therapeutic gains
Ooint Commission on Accreditation of Hospitals, 1974; World Health
Organization, 1955). There is considerable evidence that briefer hospi-
talization followed by outpatient aftercare appears cost-effective when
compared to longer hospitalization (Costello, 1975; Finney, Moos, &
Chan, 1981; Page & Schaub, 1979; Pittman & Tate, 1969; Pokorny, Miller,
& Cleveland, 1968; Walker, Donovan, Kivlahan, & O'Leary, 1983). Some
authors contend that aftercare may be the active ingredient in successful
alcohol treatment. For example, Baekelulld (1977) suggests that once
detoxification is complete, patients may be discharged and successfully
followed in an aftercare program rather than continuing in further inpa-
tient treatment. Page and Schaub (1979) suggest that inpatient treatment
be called an "orientation program," because the goals for short-term
hospitalization are modest. The term orientation program, as they de-
scribe it, refers to a shortened inpatient stay. Orientation would then be
followed by a carefully planned and extended aftercare treatment.
As will become clear in the present chapter, the term aftercare has
assumed a number of different meanings and has encompassed a vari-
ety of intervention strategies. Harmon, Latinga, and Costello (1982)
have proposed a definition of aftercare based on the functions that it
appears to entail. There are two important components. First, aftercare
is conceptualized as those therapeutic activities that function to maintain
those gains achieved in an earlier phase of treatment as opposed to
procedures that attempt to promote or develop new skills. Second, after-
care is conceptualized as being appropriate to every form of primary care
modality that has preceded it. Thus, it would be possible to develop
aftercare plans to accompany initial involvement in outreach activities,
emergency treatment, inpatient treatment, intermediate care, or outpa-
AFTERCARE IN ALCOHOLISM TREATMENT 437
tient treatment. The aftercare goals will vary depending on the compo-
nent in the treatment continuum that they accompany. Whereas this
definition promotes a broad perspective across the spectrum of treat-
ment, most available literature has focused on aftercare services pro-
vided following an inpatient hospitalization. Despite the conviction that
aftercare is crucial to successful recovery, there has been relatively little
research reported in the alcoholism literature that evaluates the effec-
tiveness of aftercare.
Aftercare services make sense. Aftercare can provide a variety of
therapeutic and social supports to the recovering alcoholic. First, after-
care may allow for early detection of relapse and thus, early intervention
in the relapse process. Second, following a return to the community,
aftercare may provide a means for patients to evaluate new behaviors.
This function of aftercare may be especially important because recovery
often requires the patient to make major life-style changes. Third, Pro-
chaska and DiClemente (1983; see also Chapter 1 in this volume) assert
that people at different stages of change (precontemplation, contempla-
tion, decision making, action, maintenance, and relapse) selectively ap-
ply various change processes (e.g., consciousness-raising, countercon-
ditioning, helping relationships). They note that relapsers may use
coping strategies similar to those of recent quitters, and that mainte-
nance is an active stage of change, rather than an absence of change.
This suggests that aftercare services focusing specifically on mainte-
nance and providing appropriate support for relapsed patients are indi-
cated for the recovering alcoholic.
The present chapter critically reviews the alcoholism aftercare liter-
ature. Included in the review are articles reporting on outpatient and
social support aftercare services following inpatient alcohol treatment or
detoxification, but does not include reports examining residential treat-
ment. First, the chapter reviews research addressing the relationship
between aftercare attendance and treatment outcome. Anecdotal and
descriptive, correlational, and cross-lagged correlational reports are pre-
sented in three separate sections. Second, research that identifies factors
affecting aftercare attendance is reviewed. Experimental and correla-
tional studies are critiqued. The last part of the chapter includes an
overview and methodological critique of the research, and offers sug-
gestions for future studies.
AFTERCARE EFFECTIVENESS
ANECDOTAL AND DESCRIPTIVE STUDIES
Five studies offer descriptive or poorly controlled data on the contri-

bution of aftercare in treatment outcome.
Pittman and Tate (1969) report one-year follow-up data from a study
in which a treatment package (3 to 6 weeks inpatient care followed by
outpatient aftercare) was compared to detoxification only (7 to 10 days
inpatient care), for a sample of 237 low socioeconomic status (SES) male
and female subjects. Aftercare was provided by inpatient treatment staff
and Alcoholics Anonymous (A.A.). The effectiveness of the entire treat-
ment package can be assessed, but the effects of aftercare cannot be
separated out. The authors observed that 18 of the 19 abstinent patients
in their treatment package group had extensive outpatient contact with
treatment staff. Of the three abstinent patients from the detoxification-
only group, one had frequent contact with A.A., and one had become
highly involved with a local church. The base rate of aftercare use and
information about nonabstinent patients who were heavy aftercare
users is not reported. The authors suggest that following hospitaliza-
tion, strong supports for sobriety may be necessary to maintain absti-
nence in low SES populations.
Chvapil, Hymes, and Delmastro (1978) examined continuity of care
in aftercare. They report on a post hoc comparison of 20 matched sub-
jects receiving group psychotherapy aftercare. Aftercare was provided
at either the same facility where they had received inpatient care or
elsewhere. The authors report that the former group of patients had
"better" aftercare attendance and abstinence rates, and contend that
inpatient and outpatient follow-up care should be offered at the same
treatment site. This report has several major limitations: (a) the follow-
up period is unspecified, (b) the patient population and matching pro-
cedures are not described, (c) the comparison is post hoc, and (d) no
statistical analyses are presented. This report can offer only very weak
evidence supporting the effectiveness of aftercare or the influence of
continuity of care on aftercare attendance.
Finney, Moos, and Mewborn (1980) include a brief discussion of
aftercare in their report of 2-year follow-up data. They report that sub-
jects who attended at least one A.A. meeting or outpatient session (at-
tenders) reported less depression in follow-up than those who attended
no A.A. meetings or outpatient sessions (nonattenders). Attendance
status was not related to drinking outcome. The base rate of aftercare
use was not reported.
Kirk and Masi (1978) examined community mental health center
(CMHC) data on a sample of 395 patients who received inpatient alcohol
treatment. Readmission for inpatient alcohol treatment during a 3-year
follow-up period is reported. Almost half (47.8%) of their sample at-
tended at least one aftercare session following the index admission. The
median number of CMCH outpatient sessions attended was 6, and the
median number of days of care was 12 (e.g., a few weeks). More atten-
ders than nonattenders were readmitted during the follow-up period.

Attenders also reported significantly more chronic drinking problems.
The authors qualified their conclusions because of (a) the low rate of
aftercare use in CMHC, (b) aftercare may improve functioning in areas
not measured, (c) the apparent confound between aftercare attendance
and chronicity of alcohol problems in the study (Le., 40% of subjects
with one hospitalization used aftercare services, whereas 60% of sub-
jects with at least 3 hospitalizations used aftercare services), and (d)
failure to assess involvement in other aftercare programs.
Katz, Morgan, and Sherlock (1981) report an uncontrolled follow-up
of 36 alcohol patients who received inpatient medical treatment for liver
disease. This report is a single-group case study. Two out of three pa-
tients kept at least 75% of their individually tailored aftercare appoint-
ments. A variety of outpatient treatment was available to these patients
(e.g., behavioral therapies, family/marital therapy, sex therapy, voca-
tional counseling). Clinical observations are provided in their discus-
sion. The authors note that outpatient therapy can be an important
source of support for the relapsed patient. This report, although rich
with clinical observations, offers no evidence on the effectiveness of
aftercare. It does suggest, however, that involvement in aftercare al-
lowed closer contact between patient and provider, and allowed for
earlier intervention into the relapse process by provision of rehospitali-
zation. Like other studies presented in this section, this report is limited
by its descriptive and uncontrolled nature.
CORRELATIONAL STUDIES
Seven studies providing correlational data exammmg the rela-

tionship between aftercare participation and treatment outcome are re-
viewed in this section.
Dubourg (1969) presents 12- to 30-month follow-up data on 76 male
patients. Aftercare was provided in a variety of settings, such as mental
health, probation, A.A., social work, and employment and hOUSing
assistance offices. Dubourg reports that aftercare use was not associated
with an improved status at outcome. Two thirds of the patients not
receiving aftercare did well. Patients receiving aftercare were twice as
likely to relapse. Dubourg suggests that support provided by family and
friends may be more important to recovering alcoholics than therapeutic
or social service support. Dubourg notes that roughly half of patients
with good outcome report that they relied only on support provided by
their relatives, but does not report similar data for patients with poor
outcome. This study is limited by the low base rate of aftercare use
reported. Only 15% of this sample had at least one aftercare contact, and
8% made more than 3 aftercare visits. In addition, Dubourg included

contact with probation officers and contacts for employment and hous-
ing assistance (9 of 76 subjects). These services may be important in
helping patients enter normalized community life, but their connection
to abstinence and reduced drinking would be, at least, indirect.
Ritson (1969) reported 6- to 12-month follow-up data on 50 male and
female inpatients. Outpatient group aftercare was provided at the same
treatment facility. In the year following hospital discharge, both fre-
quent aftercare attendance (attended 12 or more weekly outpatient
groups in the follow-up year) and disulfiram use were associated with
favorable drinking outcome. Because the correlation between aftercare
attendance and disulfiram use was not reported, the aftercare-drinking
outcome relationship may be mediated by disulfirarm use. Ritson also
points out another possible explanation for the aftercare-outcome cor-
relation: relapsed patients were reported to avoid treatment. This sug-
gests a need to recontact relapsed patients.
Pokorny, Miller, Kanas, and Valles (1973) present one-year follow-
up data on 91 V.A. inpatients. Group therapy aftercare was provided by
the same program. Patients attending at least eight aftercare sessions
were identified as aftercare participants (25% of the sample). Aftercare
participation was associated with higher probability of both abstinence
and an improved drinking status. An attempt was made to identify
subject variables predicting aftercare participation. A better pretreat-
ment work history, better psychiatric adjustment, and posttreatment
marital and residential stability were associated with more frequent use
of aftercare.
Van Dijk and Van Dijk-Koffeman (1973) reported outcome in fol-
low-up of at least 30 months on 200 male patients who had been treated.
Most of the aftercare the patients received was provided by the Dutch
Bureau for Alcoholism. Aftercare was also provided by family physi-
cians, social work services, and private psychiatrists. Forty-five percent
of this sample actively sought and attended aftercare and 99% had spe-
cific aftercare referrals. The authors compared patient "attitude toward
aftercare," classifying attitudes as active, compared to passive and with-
drawing attitudes toward aftercare. Both groups were equally likely to
be abstinent. There was, however, a significant association between
active participation in aftercare and improved drinking status. A
strength of this study lies in the relatively large sample size and in the
high base rate of aftercare use.
Davidson (1976) reports 6-month follow-up data on 100 male and
female subjects. Aftercare consisted of A.A., outpatient group therapy
provided by the treatment unit, and monthly patient-reunion meetings.
The base rate of aftercare contact was not reported. Formal or informal
contact with former alcoholics was associated with greater abstinence.

Whether that contact came in therapy groups, A.A., or reunion meet-
ings did not seem important. It was suggested that the postdischarge
receipt of social support from recovering peers is important. The need
for support from professional staff members, however, was called into
question. Davidson notes that the lack of adequate and systematically
gathered pretreatment patient data limits the interpretation of the
results.
Walker et al. (1983) investigated the effects of length of inpatient
hospital stay in an alcoholism program and patients' neuropsychological
status on treatment outcome. Aftercare, consisting of weekly outpatient
group therapy meetings provided by the treatment program, was ana-
lyzed as a nonrandomized covariate. Follow-up of 245 male V.A. pa-
tients over a 9-month interval indicated that aftercare involvement was
strongly related to aspects of treatment outcome. The results of regres-
sion analyses indicated, in particular, that the duration of involvement
in aftercare added significantly to the prediction of the number of heavy
drinking days, the level of stability in residential status, and the average
number of drinks consumed per day. The contribution made by after-
care involvement to the prediction of outcome was significant even after
pretreatment demographic, drinking-related, and functional charac-
teristics of patients had been taken into account. The salience of after-
care was noted further in a post hoc analysis. Of the original sample of
245 patients, approximately 35% of the patients completed their commit-
ment to attend aftercare for 9 months. Aftercare completers were three
times more likely to remain abstinent than those who dropped out of
aftercare before the contracted time (70.2% vs. 23.4% abstinence, respec-
tively). Similarly, the former individuals were also characterized by sig-
nificantly better drinking dispositions and residential stability when
compared to aftercare dropouts.
Walker et al. (1983) also attempted to explore those patient charac-
teristics that predicted duration of aftercare involvement. Neither the
randomly assigned length of hospital stay (2 weeks or 7 weeks) nor the
level of neuropsychological impairment predicted aftercare involve-
ment. Duration of aftercare involvement and completion of the aftercare
contract were associated with average monthly income prior to admis-
sion, age, and verbal intelligence. In each case, higher levels of these
variables were associated with either longer aftercare involvement or
contract completion. It is of interest to note that there was no overlap
between those characteristics that predicted drinking outcome and those
pretreatments associated with aftercare involvement. This finding, in
conjunction with the significant additional contribution made by after-
care to outcome, provides evidence consistent with that of Costello
(1980), and Vannicelli (1978) discussed later. In particular, this constella-

tion of results suggests that aftercare involvement has a positive effect
on treatment outcome and that this effect is independent of the influ-
ence exerted by patients' pretreatment level of general adjustment.
Using a survival analysis, Siegel, Alexander, and Lin (1984) examine
the relationship between aftercare utilization and the probability of read-
mission for 2 years subsequent to the index admission. Results are re-
ported for a cohort of 325 male and female subjects from a middle-class
suburban community. Aftercare services were provided in the commu-
nity mental health system and included services such as individual,
group, and medication treatment; day treatment; and vocational re-
habilitation services. Subjects were classified as compliers (those who
used aftercare services at least once) and noncompliers (those who used
no aftercare services). Forty-three percent of the sample were classified
as compliers, 53% as noncompliers. It is reported that the risk for read-
mission was greatest in the first 4 months after the index admission, that
aftercare compliers were less likely to be readmitted than noncompliers,
and that this was only true when the index admission was the subject's
first admission for alcohol treatment. The authors speculate that subjects
with multiple admissions have an established pattern of chron-
icity/relapse and are more resistent to the effects of aftercare services.
The authors suggest that vigorous efforts to facilitate entry into aftercare
should be focused on first admissions because (a) 40% of first admis-
sions in the sample were noncompliers, and (b) aftercare appears to be
most effective in preventing relapse for first admission compliers.
These correlational reports present suggestive but uncontrolled evi-
dence regarding the usefulness of aftercare. Posthospital support of any
nature (e.g., outpatient groups, family support, A.A.) appears associ-
ated with improved outcome and drinking status as well as decreased
risk of readmission. The relationship of aftercare attendance to absti-
nence is less dear. The strength and direction of the aftercare-outcome
relationship can be estimated from these reports. This correlation can be
explained in a variety of ways. Ritson, for example, suggests that relapse
may precipitate aftercare dropout. Davidson suggests that seeking social
support may be related to both aftercare use and treatment outcome.
Pokorny et al.'s (1973) data suggest that the patient's level of general
adjustment may mediate the aftercare/outcome relationship, whereas
Walker et ai. (1983) suggest that the effects of aftercare and independent
of the patient's pretreatment level of adjustment. Based on the studies
reviewed above, the importance of aftercare in treatment outcome can-
not be determined. Aftercare attendance may instead reflect other crit-
ical factors, such as motivation, general adjustment, or drinking status
itself.
CRoss-LAGGED CORRELATIONAL STUDIES
The cross-lagged correlational design is a nonrandomized design

that allows causal inference. In a simple correlational design, the vari-
ables of interest are measured at a single point in time, and the rela-
tionship between these variables is examined. The primary limitation of
the simple correlation is that no single explanation of a significant cor-
relation (e.g., that A causes B, B causes A, or another variable causes
both A and B) can be ruled in or out. In a cross-lagged correlational
design, variables of interest are measured for at least two points in time.
It is then possible to compute simple correlations for bivariate pairs and
examine them for temporal assymetry. For example, a model that sug-
gests that A causes B would predict that the A (time l)/B (time 2) correla-
tion would be stronger than the B (time l)/A (time 2) correlation. By
examining the pattern of simple correlations, causal interpretations may
be made. Two studies examining the relationship between aftercare at-
tendance and treatment outcome in a cross-lagged correlational design
have been published to date.
Vannicelli (1978) examined aftercare attendance and treatment out-
come at 3 and 6 months after discharge for 100 male and female subjects.
The measure of aftercare was the total number of meetings attended of
all kinds, including religious resources, women's, and open aftercare
groups. Outcome was assessed using an 8-point self-report scale that
measured amount of drinking and its impact on social, vocational, mar-
ital, and physical functioning in the past month. The 3-month after-
care/6-month outcome correlation was higher than either the 3-month
aftercare/3-month outcome or the 3-month outcome/6-month aftercare
correlations. This pattern of correlations is consistent with aftercare par-
ticipation leading to better drinking outcome, and is inconsistent with
better drinking outcome leading to more participation in aftercare. Van-
nicelli also notes that this study cannot rule out motivation as an expla-
nation of results. The generalizability of this study is limited because
patients self-selected aftercare, so enforced aftercare attendance may
produce a less favorable outcome.
Costello (1980) also used a cross-lagged correlational design for 37
male subjects, improving on Vannicelli's study by extending follow-up
to 12 and 24 months and indexing social stability and inpatient behav-
ioral adjustment as prognostic indicators. A composite score that takes
residential stability, interpersonal relationships, social activity, health,
employment, and drinking status into account was used as the outcome
measure. Aftercare included all postdischarge visits to the treatment
unit for social gatherings, disulfiram, or verbal therapy. Costello found
two systems predicting outcome. First, general adjustment at intake
predicted 1- and 2-year outcome. Second, aftercare attendance during

the first year post discharge was directly related to outcome at year 1 and
2. Intake variables did not predict aftercare use, which indicated that the
effects of aftercare were independent of measured patient prognostic
variables. Costello suggests that program planners should give priority
to achieving frequent aftercare attendance soon after discharge to mini-
mize aftercare drop out and maximize positive outcome.
The results of these cross-lagged correlational studies suggest that
aftercare attendance leads to improved treatment outcome. Costello's
findings, like Walker et al.'s (1983) data, also suggest that the influence
of aftercare is independent of general adjustment (i.e., functioning at
intake). The correlational studies equivocally suggest that aftercare con-
tributes to treatment outcome and this conclusion is strengthened by the
positive findings of the cross-lagged correlational studies. Aftercare at-
tendance provides protection against relapse, and apparently operates
independently of intake prognostic indicators.
FACTORS EFFECTING AFTERCARE PARTICIPATION
EXPERIMENTAL STUDIES
In treatment-outcome research, the effectiveness of a treatment

package must be established first. Once effectiveness has been demon-
strated, the critical and nonspecific elements of a treatment can be re-
searched. From a clinical perspective on aftercare, predictors of effec-
tiveness and attendance would provide valuable guidelines in program
planning. Four experimental studies that attempted to identify factors
affecting aftercare attendance are presented in this section.
Intagliata (1976) randomly assigned 40 male V.A. patients to one of
the two "outreach" conditions. The control group received no telephone
contact from staff after their discharge from inpatient treatment. The
other 20 experimental subjects received 6 telephone calls from staff in the
first 10 weeks after discharge. These phone calls were used to express
staff concern about the patient's welfare and to attempt to increase the
patient awareness of and encourage use of the supportive outpatient
services provided by the treatment program. The nature of outpatient
services provided is not described in the report. Intagliata found that the
experimental group made significantly greater use of outpatient services
than the control groups in the 3-month follow-up period. A median break
of 20 units of service was used to assign patients into high- and low-use
groups. High use of outpatient services was positively related to absti-
nence. Although limited by its small sample size and short follow-up
period, this study underscored the effectiveness of low-cost outreach

practice, such as telephoning patients, in increasing aftercare attendance.
Panepinto, Galanter, Bender, and Strochlic (1980) report on the
quasi-experimental evaluation of a programmatic change designed to
reduce attrition in the transition from inpatient to aftercare treatment.
Briefly, the intervention cohort received, as inpatients and as outpa-
tients, orientation lectures that emphasized the importance of outpatient
treatment in recovery. Lectures were followed by a meeting where out-
patients were encouraged to share with the inpatients their experiences
in making the transition to sobriety and life in the community. The
comparison cohort consisted of patients admitted during the same 6-
month period as the intervention cohort, during the prior year. During
the comparison period, inpatients and outpatients received separate ori-
entation sessions. Then, following inpatient discharge, comparison co-
hort patients received the outpatient orientation series and continued to
receive individual counseling. Results are reported for 313 subjects in
the intervention cohort and 257 subjects in the comparison cohort. Out-
patient services are not well described in the paper. Attendance as out-
patients, in the first 30 days after inpatient discharge, was the outcome
measure. The analysis showed that the preparatory outpatient orienta-
tion improved aftercare attendance in the first postdischarge month. In
the discussion of the results, the authors attribute increased aftercare
attendance to the orientation's reducing anxiety associated with enter-
ing a new program and increasing the patients' sense of affiliation with
the outpatient aftercare program.
Ahles, Schlundt, Prue, and Rychtarik (1983) evaluated the effects of
two different scheduling practices on aftercare attendance and drinking
outcome. Follow-up data at 12 months were presented for 50 male V.A.
patients. Aftercare consisted of weekly individual problem-oriented
counseling sessions provided by the treatment unit. The control group
consisted of the unit's usual practice of scheduling aftercare session by
session. This meant that during a session an appointment for the next
session would be scheduled. The experimental group was involved in a
behavioral contract scheduling procedure. These patients received a cal-
endar with eight aftercare appointments marked in red for the first 6
postdischarge months. They also were instructed to display the calender
prominently, attend aftercare regardless of drinking status, and re-
schedule missed appointments. As an attendance incentive, experimen-
tal patients contracted with a significant other to do something special
(e.g., go out to eat) within a week of a kept appointment. It was found
that behavioral contract scheduling increased aftercare attendance and
that behavioral contract subjects were more likley to be abstinent and
improved in their drinking at one-year follow-up (39.9% vs. 11.1%). It
was noted that experimental subjects tended to do better in the first 6

months of follow-up, when aftercare was provided, than in the next 6
months, when it was not provided.
Caddy, Addington, and Trenschel (1985) evaluated several kinds of
aftercare services that were offered to subjects in addition to the group
based aftercare offered by the hospital. This study compares no contact,
monthly telephone contact only, and two levels (four and eight sessions)
of either disease-model supportive therapy or cognitive-behavioral sup-
portive therapy following inpatient treatment. Data on 60 subjects (10 in
each condition) are reported for a 12-month follow-up period. The dis-
ease-model supportive therapy consisted of an "abstinence reinforce-
ment sequence" which stressed the following: an alcoholic identity, al-
cohol allergy, alcholism as a progressive and incurable disease, and
abstinence as a necessary condition for recovery. The behavioral-cog-
nitive supportive therapy consisted of the following: a commitment to
self-regulation, a cognitive behavioral view of relapse, cognitive coping
skills, relaxation, assertive and/or anxiety management training. Each of
these two aftercare approaches was delivered in the form of home visits
at the patient's residence. Unlike other studies described here, the
number of aftercare sessions was an independent variable. Results are
reported for three sets of outcome measures: drinking outcome, general
adjustment, and attitude. General adjustment (e.g., occupational and
marital status) showed no group differences. For drinking outcome mea-
sures on which differences were found, group differences tended to
favor either the eight-session cognitive-behavioral therapy (days func-
tioning well) or both the four- and eight-session cognitive-behavioral
conditions (days functioning well, days intoxicated, and self-described
drinking). The attitudinal measures show subjects in the eight-session
behavioral condition as more likely to endorse a habit (versus disease)
view of alcoholism, less likely to characterize their drinking as loss-of-
control drinking, and as rating themselves more likely to use aftercare
services at 6 months postdischarge. The authors conclude that the con-
tent of aftercare makes a difference in outcome. It was noted that the
learning perspective was acceptable to most subjects, where the disease
perspective often was not, and that the learning perspective increased
patient receptiveness to ongoing aftercare. Although the cognitive-be-
havioral approach to alcohol treatment is often associated with a con-
trolled drinking treatment goal, these authors conceptualize their cog-
nitive-behavioral therapy as a "behavioral inoculation" in which
approximations to sobriety are shaped. The authors note, paradoxically,
that the shaping process seemed to reinforce a decision to be abstinent
for the cognitive-behavioral patients rather than to undermine a commit-
ment to sobriety.
One strong conclusion that can be drawn from the research re-
viewed in this section is that aftercare attendance can be manipulated
directly and with relatively low-cost and simple interventions (follow-up
phone calls, orientation lectures, and behavioral contract scheduling). In
addition, the resulting increase in aftercare attendance is associated with
a more favorable treatment outcome. The Caddy et al. report suggests
that content may contribute significantly to the effectiveness of aftercare:
the learning perspective was especially useful in reinforcing and main-
taining improvements that initially occurred over the course of inpatient
treatment.
CORRELATIONAL STUDIES
The three studies reported in this section use correlational designs

to identify predictors of aftercare attendance.
Pratt, Linn, Carmichael, and Webb (1977) administered the Ward
Atmosphere Scale (Moos, 1974) to 35 male V.A. patients within a week
of inpatient admission. Aftercare consisted of weekly outpatient groups
provided by the treatment unit. Aftercare attendance was measured in
the first 3 months after discharge. Drinking status and other outcome
measures were not assessed. The results showed that those patients
who attended at least one aftercare session (N = 13) perceived signifi-
cantly more autonomy on the ward than did nonattenders (N = 22). The
authors interpreted these findings to suggest that inpatient treatment
that maximizes patient autonomy and active decision making might
increase participation in aftercare. Although this report is interesting, it
is limited by a short follow-up period and a lack of information concern-
ing outcome.
Prue, Keane, Cornell, and Foy (1979) examined the effect of trans-
portation variables on 3-month aftercare attendance. Data for 40 V.A.
patients living within 40 miles of the hospital are presented. The after-
care program consisted of individual problem-oriented counseling pro-
vided by the treatment program. In a multiple regression analysis, the
number of miles to a freeway and miles on a freeway (to the hospital)
were found to be significant predictors of 3-month aftercare attendance.
Miles to the freeway was significantly more important than miles on a
freeway. This suggests that travel time or travel effort may be more
critical than distance per se. The authors point out that their findings
have implications for the arrangement of aftercare services: for patients
living far away from the hospital, a more conveniently located aftercare
referral might be considered.
Erwin and Hunter (1984) examined cognitive functioning and de-
mographic variables as predictors of aftercare attendance. Data are re-
ported for 80 male and female subjects over a 10-week period following
discharge from inpatient treatment. Aftercare consisted of weekly group
sessions in which patients discussed their ongoing efforts to rehabilitate
themselves. Cognitive functioning was measured on abstract reasoning
and field dependence tasks. Demographic variables examined were age,
occupation, and education. The authors found that subjects charac-
terized by field dependence and preoperational and concrete-opera-
tional cognitive styles dropped out of aftercare at a rate of 70%, com-
pared to a 20% dropout rate for field independent and formal
operational subjects. Occupation and education correlated significantly
with aftercare attendance and cognitive functioning, but accounted for
only one third as much variance as the measures of cognitive function-
ing. This study is limited by a brief follow-up and a lack of information
on drinking and other outcome measures. Erwin and Hunter, noting the
prevalence of field-dependent and concrete-operational thinking (54%
in this solidly middle-class sample), suggest that rehabilitation programs
should maximize concrete training and minimize verbal symbolic mate-
rial in order to increase the likelihood that patients will participate in and
benefit from treatment.
The variety of predictors examined in the studies reviewed in this
section is great. This makes it difficult to distill any themes or conclu-
sions about the prediction of aftercare attendance. Instead, there are
three quite separate conclusions. First, the autonomy perceived by pa-
tients during the course of inpatient treatment may affect aftercare atten-
dance. Second, distance traveled to aftercare should be a consideration
when aftercare referrals are made. Third, it may be wise to take the
cognitive functioning of patients into account when planning aftercare
services.
SUMMARY AND CONCLUSIONS
Our review suggests that aftercare for alcoholics contributes signifi-

cantly to positive treatment outcome. The effects of aftercare are sub-
stantive and appear to be independent of the general adjustment system
that also influences treatment outcome (Costello, 1980; Walker et al.,
1983). Aftercare attendance seems to be more strongly related to im-
proved drinking status than abstinence rates (see Table 1). This suggests
that after achieving abstinence during inpatient treatment, patients
must learn how to stay sober in a learning process that might involve
some testing of limits and making mistakes (i.e., violating abstinence).
Aftercare probably has as one of its important functions the early
detection of and intervention in relapse. Aftercare does not appear relat-
TABLE 1.
Aftercare Efficacy: Summary Table
Aftercare related to Aftercare related
abstinence decreased drinking
N of
Type of data Citations Yes No NA Yes No NA
Descriptivel Anecdotal 5 2 3 1 1 3
Correlational 7 2 3 2 5 1 1
Cross-lagged Correlational 2 2 2
Totals 14 4 3 7 8 2 4
ed to preventing a first slip, because aftercare attendance is not related

to increased rates of abstinence. It is, however, related to improved
drinking status. It may be that at the beginning of a full relapse (i.e., a
slip) patients may drink but in a limited manner (e.g., 2 beers or 1 glass
of wine). After a slip they may be at risk in two areas that are important
to their recovery:
1. They may be at risk for a full-blown relapse and a quick return to

a destructive pattern of drinking (Marlatt & Gordon, 1980).
2. They may be at risk for cutting themselves off from social sup-
ports that are essential to their recovery. They may be more
likely to drop out of treatment, stop going to A.A., and stop
calling or seeing their sober friends.
In aftercare, patients may get help in coping with the stresses that
lead up to the slip and the stress created by the slip itself. If they can
discuss their slip soon after it occurs and get support, they may be able
to regain their sobriety and maintain their social support network for
sobriety. In this way their recovery may continue, in spite of the slip,
rather than end. In order for aftercare to prevent a slip from becoming a
full-blown relapse, patients must let their support network know of the
slip. Providing clear programmatic expectations that slips will be ac-
knowledged immediately, the rationale for this expectation, and reas-
surance to patients that the consequences of the disclosure will not
necessarily be punitive would help to increase the liklihood that a slip
will be identified early and appropriate interventions made (Donovan &
Chaney, 1985).
It appears that aftercare attendance can be increased by relatively
inexpensive interventions, such as telephone calls (lntagliata 1976), ori-
entation lectures (Panepinto et al., 1980), and behavioral contract sched-
uling (Ahles et al., 1983). Some guidelines for increasing aftercare atten-
dance can be made based on the studies reviewed here.
1. Some patient characteristics are associated with aftercare utiliza-
tion. Patients who use aftercare services are those with more stable work
and psychiatric-adjustment histories, who perform better on indicators
of cognitive functioning, (Erwin & Hunter, 1984; Walker et al., 1983) and
have postdischarge marital and residential stability. It may be helpful for
clinicians to be aware of these predictors so that, for example, an un-
monitored change of residence does not prevent outreach efforts. In
addition, Erwin and Hunter recommend that aftercare programs make
the content and presentation of their programs concrete rather than
verbal and symbolic to make material more understandable and useful
to patients. McLachlan (1972, 1974), however, has found that for alco-
holic patients, a patient-therapist match on conceptual level (CL) led to
better outcome and mismatch to poorer outcome. CL is a measure of
interpersonal development that entails both cognitive complexity, and
development along a dependency-interdependency dimension. Seen in
this context, it seems that matching patients and treatment on cognitive
complexity might help to reduce attrition from aftercare.
2. Pratt et aI.'s (1977) findings suggest that when patients participate
in decisions effecting their treatment as inpatients, they are more likely
to participate in aftercare.
3. Chvapil et al.'s (1978) data suggest that there may be a benefit,
both in terms of aftercare attendance and abstinence, from aftercare
services being provided at the same facility where inpatient treatment
had been received.
4. For patients living far from the treatment facility, an aftercare
referral closer to where they live should be considered (Prue et al., 1979).
It might also be helpful in this situation to ask the patient to make
arrangements for aftercare prior to inpatient admission.
Other research is beginning to delineate 'parameters that will aid
clinicians and program planners in selecting optimal referral patterns
(d. Prue et al., 1979; Siegel et al., 1984) and perhaps in the identification
of patients at risk for aftercare dropout (Erwin & Hunter, 1984).
This review examined studies reporting on the efficacy of aftercare
attendance. It appears from correlational studies of efficacy that post-
hospitalization support of any kind (e.g., outpatient groups, family or
church support, A.A.) is associated with improved outcome and drink-
ing status, especially for first admission patients (Siegel et al., 1984)
Vannicelli (1978) and Costello (1980), in their cross-lagged correlational
reports, both found that aftercare attendance leads to improved drink-
ing outcomes. Social support may be important because it may provide
the recovering alcoholic with (a) a nondrinking social network, and (b) a
place for the patient to get support, help, and assistance in coping with
stress in new and nondrinking ways. These two functions of posthospi-
talization social support can reduce the chances that the recovering alco-
holic will relapse when coping with social pressure to drink or with
negative mood states. Marlatt and Gordon (1980) found that these two
categories of high-risk situations accounted for 56% of the alcohol re-
lapse in their sample. Marlatt and Gordon's relapse data suggest that the
development of a non drinking social world and of non drinking methods
of coping with emotional stress are critical elements in recovery.
Cronkite and Moos (1980) found that posttreatment stress accounted for
30% of the variance in alcohol consumption outcomes and for more than
50% of the variance when abstinence and depression are outcome crite-
ria. Of the posttreatment factors investigated, stressors and coping be-
havior account for most of the variance. What these findings suggest is
that for drinking- and mood-related outcomes, patients may be es-
pecially vulnerable to relapse after inpatient treatment. Taken a step
further, Cronkite and Moos' results underscore the need for outpatient
aftercare services, because posttreatment stress and coping are the most
important determinants of treatment outcome as they measured it. Lit-
man and Oppenheim (unpublished manuscript) suggest that avoiding
temptation and seeking social support, as methods for coping with
urges to drink, are very important to survival, especially early in recov-
ery. In aftercare, patients gain contact with sober people, and can get
support from others who are also coping with the loss of a major aspect
of their life, alcohol.
Although aftercare participation appears associated with improve-
ment, methodological flaws and inconsistencies in reporting research
make integrating findings difficult. Some excellent general gUidelines
for the evaluation of alcohol treatment research are available (Emrick &
Hansen, 1983; Maisto & McCollam, 1980; Nathan & Lansky, 1978) and
are applicable to aftercare research as well. Briefly, these authors sug-
gest reporting of diagnostic criteria, control, and comparison groups;
and handling of treatment and follow-up dropouts in describing sample
characteristics. Adequate description of the intervention, the design,
therapist qualifications and training, treatment variables, and follow-up
procedures should be included. Outcome measures should control for
investigator bias, include multiple sources of data and multidimensional
outcome criteria. In the reporting of results, caution should be exercised,
the difference between clinical and statistical significance be remem-
bered, and the generalizability of the results be discussed.
Following are some additional recommendations applying specifi-
cally to aftercare research. The following paragraphs highlight aftercare
research design and reporting issues.
1. Definition of Aftercare: A wide variety of treatment services fall

unoer the rubric of "Aftercare." Services provided vary by (a) modality
(l·IYvdual, family/couples, group); (b) organization (referral and staffing
patterns, continuity of care); (c) time parameters (time limited or open-
ended); (d) therapeutic orientation (behavioral, insight, process, etc.); (e)
purpose (psychotherapy or social support, social services); (f) attendance
expectations (required versus optional attendance, outreach efforts, etc.).
The description of aftercare services provided should include the
nature of these parameters. Reporting of such information would be
especially useful in understanding conflicting or divergent results.
2. Degree of Aftercare Involvement: The research reviewed here indi-
cates a wide range of patient involvement in aftercare. For example,
Ritson (1969) reported that 64% of his sample attended 12 or more after-
care sessions compared to Dubourg's (1969) sample where 85% used no
aftercare services. This large discrepancy suggests a number of issues
important to future reporting. First, the base rate of aftercare use in the
entire sample should be reported. Second, if the sample is divided into
groups (e.g., attenders and nonattenders), the criteria used for classifica-
tion (median split, upper and lower quartiles, completing/not complet-
ing treatment contracts) should be reported. Third, if aftercare involve-
ment can be analyzed as both a noncontinuous and continuous variable,
this would be advantageous (d. Walker et al., 1983).
3. Pretreatment Prognostic Variables: Costello (1980) and Walker et al.,
(1983) report that the influence of aftercare is independent of general
patient adjustment. Both sets of variables, however, affect treatment
outcome. The measurement of pretreatment prognostic variables is crit-
ical in the identification of the unique contribution of aftercare in treat-
ment outcome. In addition, a report of the percentage of patients using
disulfiram would be useful in the description of a sample.
4. Directions for Future Research: It seems clear that aftercare is effec-
tive and that its contribution to outcome is independent of general pa-
tient prognostic variables. Research is needed that identifies critical ele-
ments of successful aftercare treatment. Researchers have begun to turn
attention to factors that affect aftercare attendance. The following list
illustrates some areas where parametric research is needed.
1. What constitutes the minimum length of inpatient treatment

necessary for favorable outcome, when inpatient care is followed
by a strong aftercare program?
2. Patient, staffing, and organizational risk factors in aftercare attri-
tion need identification, and effective outreach strategies for pa-
tients at risk for attrition can be developed and evaluated.
3. What manipulable primary treatment factors predict aftercare

attendance?
4. Are various kinds of aftercare services differentially effective?
For example, are there differences between social support and
skills training, individual and group therapy, patient only and
patient plus spouse or family? Are certain of these approaches
more effective with certain types of patients?
5. What are the attendance effects of different scheduling and refer-
ral practices?
6. How cost-effective are aftercare enabling factors, such as trans-
portation and prompts or incentives?
7. What are the critical time parameters in aftercare? How quickly
must patients enter aftercare and where is the point of diminish-
ing returns?
A final focus for aftercare research is model testing. Models of re-
lapse, such as Litman's (Litman et al., 1979) and Marlatt's (Marlatt &
Gordon, 1980), have gained prominence in the last 5 years. Treatment-
outcome research is needed to further elaborate these models. The-
oretical models can be used to generate hypotheses, develop interven-
tions, and focus data collection (Donovan & Chaney, 1985).
The present review focuses on outpatient aftercare that follows in-
patient treatment for alcohol dependence/abuse. The research reviewed
suggests that outpatient aftercare plays a useful role in maintaining the
gains made during inpatient treatment. This review also highlights the
need for further aftercare research. We know little about how to improve
aftercare attendance and little about the critical elements of successful
aftercare. As recovering alcoholics attempt to find a new social role,
aftercare may provide needed social and psychotherapeutic supports for
such extensive life changes. In addition, aftercare research could make
substantial contributions to our understanding about what former alco-
hol abusers do in the maintenance phase of their recovery, because little
is known about successful coping strategies during maintenance (Pro-
chaska & DiClemente, 1983). Finally, increasing our knowledge about
relapse and factors that protect against and shorten the negative impact
of relapse would increase our understanding about patient needs during
aftercare.
ACKNOWLEDGMENTS
The authors would like to thank William Carter, Daniel Kivlahan,

and G. Alan Marlatt for their helpful comments on an early draft of this
chapter.
REFERENCES
Ahles, T. A, Schlundt, D. G., Prue, P. M., & Rychtarik, R. G. (1983). Impact of aftercare
arrangements on the maintenance of treatment success in abusive drinkers. Addictive
Behaviors, 8, 53-58.
Baekelund, F. (1977). Evaluation of treatment methods in chronic alcoholism. In B. Kissin
& H. Begleiter (Eds.), The biology of alcoholism: Treatment and rehabilitation of the chronic
alcoholic (pp. 385-440). New York: Plenum Press.
Caddy, G. R., Addington, H. J., & Trenschel, W. R. (1985). A comparative evaluation of
aftercare technologies in the management of alcohol dependence. Unpublished manuscript,
Nova University, Department of Psychology, Fort Lauderdale, FL.
Chvapil, M., Hymes, H., & Delamstro, D. (1978). Outpatient aftercare as a factor in
treatment outcome; A pilot study. Journal of Studies on Alcohol, 39, 540-554.
Costello, R. M. (1975). Alcoholism treatment evaluation: In search of methods II. Collation
of two-year follow-up studies. International Journal of the Addictions, 10, 857-867.
Costello, R. M. (1980). Alcoholism aftercare and outcome: Cross-lagged and path analyses.
Cronkite, R. c., & Moos, R. H. (1980). Determinants of post treatment functioning of
alcoholic patients: A conceptual framework. Journal of Consulting and Clinical Psychol-
ogy, 48, 305-316.
Davidson, A F. (1976). An evaluation of the treatment and aftercare of a hundred alco-
holics. British Journal of Addiction, 71, 217-224.
Donovan, D. M., & Chaney, E. F. (1985). Alcoholic relapse prevention and intervention:
Models and methods. In G. A Marlatt & J. R. Gordon (Eds.), Relapse prevention:
Maintenance strategies in the treatment of addictive behaviors (pp. 351-416). New York:
Guilford Press.
Dubourg, G. O. (1969). Aftercare for alcoholics: A follow-up study. British Journal of Addic-
tion, 64, 155-163.
Emrick, C. D., & Hansen, J. (1983). Assertions regarding the effectiveness of treatment for
alcoholism: Fact or fantasy? American Psychologist, 38, 1078-1088.
Erwin, J. E., & Hunter, J. J. (1984). Prediction of attrition in alcoholic aftercare by scores on
the EFT & two Piagetian tasks. Journal of Consulting and Clinical Psychology, 52, 354-
358.
Finney, J. W., Moos, R. H., & Mewborn, C. R. (1980). Posttreatment experiences and
tion. Journal of Consulting and Clinical Psychology, 48, 17-29.
Finney, J. W., Moos, R. H., & Chan, D., (1981). Length of stay and program component
effects in the treatment of alcoholism: A comparison of two techniques for process
analysis. Journal of Consulting and Clinical Psychology, 49, 120-131.
Harmon, S. K, Latinga, L. J., & Costello, R. M. (1982). Aftercare in chemical dependence
treatment. Bulletin of the Society of Psychologists in Substance Abuse, 1, 107-110.
Hunt, W. A, Barnett, L. W., & Branch, L. G. (1971). Relapse rates in addiction programs.
Journal of Clinical Psychology, 27, 455-456.
Intagliata, J. (1976). A telephone follow-up procedure for increasing the effectiveness of a
treatment program for alcoholics. Journal of Studies on Alcohol, 37, 1330-1335.
Joint Commission on Accreditation of Hospitals (1974). Accreditation Manual for Alco-
holism Programs. Chicago, IL: Author.
Katz, A., Morgan, M. Y., & Sherlock, S. (1981). Alcoholism treatment in a medical setting.
Kirk, S. A, & Masi, J. (1978). Aftercare for alcoholics: Services of community mental health
centers. Journal of Studies on Alcohol, 39, 345-547.
Litman, G. K, & Oppenheim, A. N. (n.d.). A model for alcoholism survival. Unpublished
manuscript, Addiction Research Unit, Institute of Psychiatry, Maudsley Hospital,

London, England.
Litman, G. K., Eiser, J. R., Rawson, N. S., & Oppenheim, A. N. (1979). Differences in
relapse precipitants and coping behavior between alcoholic relapsers and survivors.
Litman, G. K., Eiser, J. R., & Taylor, C. (1979). Dependence, relapse, and extinction: A
theoretical critique and behavioral examination. Journal of Clinical Psychology, 35, 192-
199.
Maisto, S. A., & McCollam, J. B. (1980). The use of multiple measures of life health to
assess alcohol treatment outcome: A review and critique. In L. C. Sobell, M. B. Sobell,
& E. Ward (Eds.), Evaluating alcohol and drug abuse treatment effective/less (pp. 15-76).
New York: Pergamon Press.
McLachlan, J. F. C. (1972). Benefit from group psychotherapy as a function of patient
therapist match on a conceptualleve!. Psychotherapy Theory Research and Practice, 9,
317-323.
Marlatt, G. A., & Gordon, J. R. (1980). Determinants of relapse: Implications for the
maintenance of behavior change. In P. O. Davidson & S. M. Davidson (Eds.), Behav-
ioral medicine: Changing health lifestyles (pp. 410-452). New York: Brunner/Maze!'
Moos, R. (1974). Ward Atmosphere Scale Manual. Palo Alto: Consulting Psychologists Press.
Nathan, P. E., & Lansky, D. (1978). Common methodological problems in research on the
addictions. Journal of Consulting and Clinical Psychology, 46, 713-726.
National Institute of Mental Health. (1973). Utilization of Mental Health facilities by per-
sons diagnosed with alcohol disorders (DHEW Publication no. HSM 73-9114). Wash-
ington, DC: U.S. Government Printing Office.
Page, R. D., & Schaub, L. H. (1979). Efficacy of 3 vs 5 week alcohol treatment program.
Illtemational Journal of the Addictions, 14, 697-714.
Panepinto, W., Galanter, M., Bender, S. H., & Strochlic, M. (1980). Alcoholics transition
from ward to clinic: Group orientation improves retention. Journal of Studies 011 Alcohol,
41, 940-945.
Pittman, D. J., & Tate, R. L. (1969). A comparison of two treatment programs for alco-
holics. Quarterly Journal of Studies on Alcohol, 30, 888-899.
Pokorny, A., Miller, B., & Cleveland, S. (1968). Response to treatment of alcoholism: A
follow-up study. Quarterly Journal of Studies 011 Alcohol, 29, 364-381.
Pokorny, A., Miller, B., Kanas, T., & Valles, J. (1973). Effectiveness of extended aftercare
in the treatment of alcoholism. Quarterly Journal of Studies 011 Alcohol, 34, 435-
443.
Pratt, T. c., Linn, M. W., Carmichael, J. S., & Webb, N. L. (1977). The alcoholic's percep-
tion of the ward as a predictor of aftercare attendance. Journal of Clillical Psychology, 33,
915-918.
Prochaska, J. 0., & DiClemente, C. C. (1983). States and processes of self-change of
smoking: Toward an integrative model of change. Journal of Consultillg alld Clillical
Prue, D. M., Keane, T. M., Cornell, J. E., & Foy, D. W. (1979). An analysis of distance
variables that affect aftercare attendance. Commullity Melltal Health Journal, 15, 149-
154.
Ritson, B. (1969). Involvement in treatment and its relation to outcome amongst alcoholics.
British Journal of Addictioll, 64, 23-29.
Seigel, c., Alexander, M. J., & Lin, S. (1984). Severe alcoholism in the Mental Health
sector: II. Effects of service utilization on readmission. Journal of Studies on Alcohol, 45,
510-516.
Tuchfeld, B. S. (1981). Spontaneous remission in alcoholics: Empirical observations and

theoretical implications. Journal of Studies on Alcohol, 42, 626-641.
Van Dijk, W. K, & Van Dijk-Koffeman, A. (1973). A follow-up study of 211 treated male
alcoholics. British Journal of Addiction, 68, 3-24.
Vannicelli, M. (1978). Impact of aftercare in the treatment of alcoholics: A cross-lagged
panel analysis. Journal of Studies on Alcohol, 39, 1875-1886.
Walker, R. D., Donovan, D. M., Kivlahan, D. R., & O'Leary, M. R. (1983). Length of stay,
neuropsychological performance and aftercare: Influences on alcohol treatment out-
World Health Organization. (1955). Report on the alcoholism subcommittee (Technical Report,
series No. 42). Geneva: Author.
Index
Action stage Alcoholics Anonymous (Cant.)

bulimia and, 232 change model and, 103
change model and, 10 description of, 298-300
movement to, 23, 51-57 effectiveness of, 135-136
self-help manuals and, 332 individualized alcoholism treatment,
Addictive behaviors 182
change model and, 3-27 relapse and, 409
cognition and, 59 spiritual aspects of, 98, 99
See also entries under specific addictions Alcoholism aftercare, 435-456
Affect, 77 conclusions in, 448-453
Aftercare. See Alcoholism aftercare definitions in, 436-437
Age differences, 211 effectiveness of, 437-444
Agoraphobia, 307 factors affecting participation in, 444-
Alcohol abuse and alcoholism 448
avoidance of decision making, 72-77 treatment and, 436
change and, 92, 93-94 Alcoholism survival model, 391-405
cognition and, 59 conceptual framework in; 393-394
conflict and stress factors in, 77-80 early hypothesis in, 391-392
contemplation stage and, 61-65 original study in, 392-393
criteria in determination of, 384 prospective study in, 394-402
dependence and, 361-373 Alcoholism treatment, 121-174
determination stage, 65-72 aftercare in, 435-456
judgment shortcomings, 69-72 Alcoholics Anonymous, 135-136
relapse prevention model for, 407- aversion therapies, 139-144
433. See a(so Relapse prevention broad-spectrum approaches to, 149-
model 154
self-help manuals and, 335-353 client matching in, 158-161, 175-203.
self-regulation and, 30, 38-39, 43, See also Individualized alcoholism
322-327 treatment
social costs of, 52 conclusions in, 161-163
World Health Organization (WHO) confrontation in, 134-135
and,51-57 controlled drinking goal and,
Alcohol dependence syndrome, 361-364 145-148
Alcoholics Anonymous, 93 early interventions in, 204-220. See
aftercare and, 438 also Early alcoholism treatment
457
458 INDEX
Alcoholism treatment (Coni.) Aversion therapy (Cont.)

educational strategies, 136-138 individualized alcoholism treatment,
evolution in, 375-376 183
family involvement and, 309-315 See also Electrical aversion therapy
hallucinogens in, 130
length and setting for, 154-158
Behavior therapy
marital/family therapy in, 138-139
outpatient treatment, 264
operant methods in, 148-149
residential treatment, 263
pharmacotherapy, 123-130
Behavioral self-control training (BSCT)
psychotherapy/counseling in, 131-133
early alcoholism intervention, 206
psychotropics in, 128-130
individualized alcoholism treatment,
research in, 121, 126
185
self-efficacy theory and, 410-412
Behaviorists, 17
simple intervention, 375-387
Bibliotherapy
varieties of, 121-122
early alcoholism intervention, 206,
Anorexia nervosa, 227-228
209, 210, 218-219
bulimia and, 230
See also Self-help manuals
cultural differences and, 229
Biofeedback techniques, 152
Antabuse (disulfiram)
Brain. See Neuropsychological
aftercare and, 440
impairment
community reinforcement approach
Broad-spectrum treatment
with, 153
alcoholism treatment, 149-154
effectiveness of, 124-126
192-193
181
Bulimia, 228-236
operant conditioning techniques with,
action stage and, 232
148
contemplation stage and, 231-232
Antianxiety drugs, 128
definitions of, 228-229
Anticipatory regret, 77-78, 79
maintenance stage and, 232-233
Antidepressants
management of, 233-236
effectiveness of, 128-129
precontemplation stage and, 231
individualized alcoholism treatment
syndrome development, 229-231
and, 181
Antidipsotropic drugs, 124-127
Antipsychotics, 128 Cancer, 110-111
Anxiety Catastrophizing, 409
alcoholism treatment and, 152 Change
anticipatory regret and, 77 critical conditions for, 91-108
individualized alcoholism treatment, morality and, 98-99
184 nonspecialist professionals in, 94-96
Apnea aversion therapy, 140-141 prediction of, 92
Assertiveness training situation and, 93-94
alcoholism treatment with, 151 specialist treatment for, 96
self-regulation model, 41 Change model, 3-27
Attention, 324-325 action stage, 10
Attitudes, 381 commonalities among, 4
Attributional failure, 59-61 contemplation stage, 9
Autonomy, 10 decisional balance and, 11-12
Aversion therapy demands on, 3-4
alcoholism treatment, 139-144 development of, 99-105
effectiveness of, 97 experimental models of, 361-374
INDEX 459
Change model (COl1t.) Community reinforcement approach

maintenance stage and, 10-11 (CRA) (COl1t.)
patterns in, 13-16 individualized alcoholism treatment,
prediction in, 12-13 185
process in, 7-8 Compulsion
relapse and, 5-6 change model, 361-365
smoking programs, 4-5, 280-282 dependence and, 364-365
stage process interaction in, 8-9 Conceptual level, 189-190
stages in, 5 Confessional process, 232
strategy in, 19 Conflict. See Stress
therapist/client relationship and, 6-7, Confrontation, 134-135
20-25 Conjunction fallacy, 69-70
transtheoretical approach, 17-18, 20, Consciousness-raising, 15, 23
24-25 Contemplation stage
Change process anticipatory regret in, 77
cognition in, 319-329 bulimia and, 231-232
family in, 305-318 change model and, 9, 19-20,23
Chemical aversion therapy judgment and, 69
alcoholism treatment, 139-144 movement to action and, 51-57
individualized alcoholism treatment, movement to determination and, 59-
183 89
Choice, 80, 193 thinking/reasoning and, 61-65
Citra ted calcium carbimide, 126-127 Control theory, 320-322
Client choice. See Choice Controlled drinking
Client/therapist relationship. See Thera- alcoholism survival model, 394
pist/client relationship alcoholism treatment goals and, 145-
Cognition 148
addictive behavior and, 59 behavioral self-control training, 185
affect and, 77 relapse prevention model and, 411
avoidance of decision making, 72-77 severity of problem and, 187
change and, 92, 319-329 Coping Behaviors Inventory, 399-402
contemplation stage and, 61-65 Coping skills training, 206, 210
decision-making assistance, 81-85 Cost containment, 378
early alcoholism intervention, 207 Counseling
expectancy-value model and, 111-112 alcoholism treatment, 131-133
judgment shortcomings, 69-70 confrontation strategies, 134-135
multiple drug use treatment, 244 Courts. See Legal system
self-regulation and, 327-328 Covert sensitization, 142-144
Cognitive Appraisal Questionnaire, 417- Cues
418 controlled study of, 365-373
Cognitive change decision-making assistance, 82-83
control theory, 320-322 dependence/compulsions and, 364
cybernetic system and, 319-320 feedback and, 71-72
memory and thinking in, 327-328 inattention and, 76
self-regulation and, 322-327 Culture, 229
Cognitive psychology, 59-89 Cybernetic system, 319-320
Cognitive style, 188-190
Community psychology movement, 106 Decisional balance, 11-12
Community reinforcement approach Decision making
(CRA) alcohol abuse and, 65-72
alcoholism treatment, 152-153 assistance in, 81-85
460 INDEX
Decision making (Cont.) Differential approach (Cont.)

avoidance of, 72-77 problem severity and, 186-188
change model and, 100-104 self-esteem and, 191
conflict and stress in, 77-80 social stability and, 192
elimination-by-aspects model, Disease model, 91, 105
68-69 alcoholism treatment, 123
judgment and, 69 controlled drinking goals and, 185
optimizing model, 65-66 motivation and, 207
practice implications, 105-106 Disulfiram (Antabuse)
satisficing model, 66-68 aftercare and, 440
smoking and, 109-118 community reinforcement approach
Deductive reasoning. See Cognition; with, 153
Reasoning effectiveness of, 124-126
Delay, 73-77 individualized alcoholism treatment,
Deliberation, 74-75 181
Demography, 211 operant conditioning techniques with,
Dependence 148
change models and, 361-374 Drug abuse. See Multiple drug abuse;
compulsions and, 364-365 Opiate-dependency treatment
treatment strategies and, 365 Drug Avoidance Inventory, 243-244
Depression
self-regulation and, 327 Early alcoholism treatment, 205-220
treatment for, 307 bibliotherapy, 209
Desensitization coping skills training in, 210
change model and, 23-24 discussion of, 218-219
See also Systematic method in, 206-208
desensitization programs in, 206
Determination stage results in, 210-218
alcohol abuse and, 65-72 Eating disorders, 221-238
judgment shortcomings, 69-72 anorexia nervosa, 227-228
movement from contemplation stage, bulimia, 228-236
59-89 obesity, 222-227
Detoxification self-regulation model, 38-39
importance of, 275-276 Educational achievement, 270-271
opiate-dependency treatment and, 266 Educational strategies, 136-138
Diagnostic and Statistical Manual (DSM- Electrical aversion therapy
Ill) alcoholism treatment with, 141-142
anorexia nervosa, 227-228 relaxation training with, 151
bulimia, 229 See also Aversion therapy
Dieting Elimination-by-aspects model, 68-69
bulimia and, 320-321 Employment history, 270-271
obesity and, 222-224, 226 EV model. See Expectancy-value models
Differential approach (optimizing model)
alcohol dependence and, 194 Expectancies, 77-78
client choice and, 193 Expectancy-value models (optimizing
cognitive style and, 188-190 model)
individualized alcoholism treatment, decision making, 65-66
179 smoking and, 109-112
life problems and, 192-193 study examples in, 112-117
neuropsychological status and, 191 Expectations, 65-66
INDEX 461
Family, 305-318 Individualized alcoholism treatment

addictive behaviors and, 308-309 (Coni. )
aftercare and, 439 neuropsychological impairment and,
alcoholism treatment and, 160, 309- 158-159, 191
315 perceived choice and, 160-161
emotional/behavioral problems and, predictor approach variai;1;:s, 193-195
306-308 predictor studies, 180-186
See also Family therapy; Social support problem severity and, 186-188
networks; Spouse involvement research strategies in, 178-180
Family history, 194 self-esteem and, 191
Family therapy severity level and, 159
alcoholism treatment with, 138-139 social stability and, 192
individualized alcoholism treatment, undifferentiated treatment contrasted,
184-185 175-176
Feedback Inferential learning, 71-72
confrontation, 135 Inventory of Drinking Situations, 413-
control theory and, 320 415, 423-431
cybernetic system and, 319-320
early alcoholism treatment, 207 Judgment, 69-72
failures in, 71-72
Key-level strategy, 19
overconfidence and, 71
Freedom, 80 Legal system
drug abuse treatment programs, 266
Gender differences. See Sex differences opiate-dependency program, 271,
General practitioners. See Physicians 273-274
Groups. See Mutual aid groups Life experiences
alcoholism treatment and, 160
Hallucinogens, 130 individualized alcoholism treatment
Heredity and, 192-193
alcoholism treatment and, 160 Lithium, 129-130
individualized alcoholism treatment, Locus of control
194 alcoholism treatment and, 159
Hypersexuality, 92 cognitive style and, 189
Hypertension, 300-302 Lung cancer, 110-111
Hypnosis, 144 Lysergic acid diethylamide, 130
Inattention, 75-77 Maintenance stage

Individual differences bulimia and, 232-233
cognitive change and, 324 change model and, 10-11
self-regulation model and, 36-38 self-help manuals and, 332
Individualized alcoholism treatment, See also Relapse
158-161, 175-203 Marital therapy, 138-139
client choice and, 193 Maximum-impact strategy, 19
cognitive style and, 188-190 Memory
conceptual level for, 158 alcohol abuse and, 322-323
differential studies, 186-195 self-regulation and, 327-328
family history and, 160 Metronidazole, 127
life problems and, 160, 192-193 Michigan Alcoholism Screening Test
locus of control and, 159 (MAST), 215, 314, 345
matching hypothesis in, 176-178 Modeling approach, 179-180
462 INDEX
Moderate drinking. See Controlled Obsessive-compulsive rituals, 306-307

drinking Operant conditioning techniques, 148-
Morality 149
change and, 98-99 Opiate-dependency treatment, 263-277
overconfidence and, 71 assessment in, 268
satisficing model and, 67-68 discussion of, 274-276
Motivation method in, 267-268
alcoholism treatment and, 160 philosophy and goals in, 264-267
early alcoholism treatment and, 206- results of, 270-274
207 subjects in, 268-269
multiple drug use treatment and, 241- Optimizing model. See Expectancy-value
242, 253-255, 358 models (optimizing model)
self-regulation model, 41-42 Overconfidence, 71
smoking cessation strategies, 282-284
Multiple Cue Probability Learning, 71- Personality. See Cognitive style
72 Pharmacotherapy
Multiple drug use treatment, 239-261 alcoholism treatment, 123-130, 131-
brief interventions in, 239-240 133, 140, 148
discussion of, 256-260 antidipsotropics, 124-127
method in, 242-245 bulimia, 233-235
motivation and, 241-242 drug dependency and, 276
pretreatment drug use, 240 hallucinogens, 130
problem multiplicity in, 240-241 individualized alcoholism treatment,
results in, 245-255 181, 183
self-reliance and, 241 obesity and, 226
treatment progress, 242 psychotropics, 128-130
Mutual aid groups, 289-304 See also entries under names of specific
Alcoholics Anonymous, 298-300 drugs
dynamiCS of, 294-298 Physical exercise and, 226-227
growth of groups in, 290-291 Physicians
hypertensive clubs, 300-302 change process and, 94-95
interest in, 291-292 smoking cessation strategies, 284-285
limits of, 302-303 Pledge taking, 98-99
range and characteristics of, 292-294 Postdecisional regret, 78-80
smoking cessation strategies, 285-286 Precontemplation stage
bulimia and, 231
Nausea change model and, 19-20
alcoholism treatment, 140 Prediction
individualized alcoholism treatment, alcoholism treatment and, 159
183 change and, 92
Neuropsychological impairment individualized alcoholism treatment,
alcoholism treatment and, 158-159 176-178
individualized alcoholism treatment Predictor strategy, 180-186
and, 191 Alcoholics Anonymous, 182
Nicotine gum, 285-286 behavioral self-control training, 185
chemical aversion therapy, 183
Obesity community reinforcement approach,
change strategies in, 222-226 185
classifications of, 226-227, 236 covert sensitization, 183-184
treatment for, 308 disulfiram, 181
See also Eating disorders family therapy, 184-185
INDEX 463
Predictor strategy (Cont.) Relaxation training

individualized alcoholism treatment, alcoholism treatment with, 151-152
178-179 individualized alcoholism treatment,
psychotherapy, 182 184
psychotropic medications, 181 Religion, 94, 98-99
relaxation training, 184 Residential treatment
social skills training, 184 evaluation of, 263-264
variables in, 193-195 legal system and, 271
Problem solving, 61-62 self-regulation model, 42-44
Proximity, 379-381 Resistance, 31-32
Psychoactive Drug Use History, 242, 246-
247 Satisficing model, 66-68
Psychoanalysis, 96 Schizophrenia, 307-308
Psychotherapy Self-concept, 9-10
alcoholism, 131-133 Self-control
bulimia and, 235-236 self-regulation model and, 29-32
change process in, 8 situation and, 30-31
focus of, 17 Self-efficacy theory
individualized alcoholism treatment, change model and, 10, 11
182 relapse prevention model and, 408-
integration of methods in, 3 412
self-regulation model and, 31
Rationality temptation and, 16
change and, 92 Self-esteem, 191
expectancy-value model and, 111-112 Self-help manuals, 331-359
Reasoning, 63-65 evaluation of, 338-353
Refusal, 73 future research directions in, 353-356
Regret. See Anticipatory regret; Postdeci- interest in, 332-335
sional regret research on controlled drinking and,
Rehabilitation 335-338
goals of, 435 See also Bibliotherapy
See also Relapse Self-regulation model, 29-47
Relapse alcoholism and, 333-334
alcoholism treatment aftercare, 435-456 cognitive change and, 322-327
change model and, 5-6 control theory and, 321-322
freedom and, 80 cybernetic system and, 319-320
opiate-dependency treatment, 267, implications summarized, 45-46
271, 273-274 individual differences and, 36-38
rates of, 407, 435 memory and thinking in, 327-328
survival model, 391-405 obstacles in, 31-32
Relapse Precipitants Inventory, 395-399 process in, 32-36
Relapse prevention model, 407-433 research support for, 38-44
Cognitive Appraisal Questionnaire, residential treatment and, 42-44
417-418 self-control in, 29-32
described, 412-413 unresolved issues in, 44-45
Inventory of Drinking Situations and, Self-reliance, 241, 257-258
413-415, 423-431 Severity levels, 255
self-efficacy theory and, 408-412 Sex differences
Situational Confidence Questionnaire, aftercare and, 438
415-416, 433 early alcoholism treatment, 211
theoretical framework for, 408 obesity and, 225
464 INDEX
Sex differences (COllt.) Social systems perspective, 29-32

relapse and, 398-399 Social workers, 95-96
simple intervention studies and, 377 Socioeconomic class
Sexuality, 92 aftercare and, 438
Shifting-levels stategy, 19 early alcoholism treatment, 211
Significant others Spiritual elements, 98-99
anticipatory regret and, 77-78 Spouse involvement
See also Family; Social support net- addictive behaviors, 308
works; Spouse involvement alcoholism treatment, 308-315
Simple intervention, 375-387 treatment effectiveness, 306
attitudes and, 381 See also Family; Family therapy
economics of, 378 Stigma
evaluation of, 381-386 mutual-aid groups, 295-296
proximity, 379-381 simple intervention and, 379
stigma and, 379 Stress
studies in, 375-378 change models and, 100
Situation decision making and, 77-80
early alcoholism intervention, 217 management techniques for, 151-152
self-control situations, 30-31 Subjective expected utility (SEU) theory
Situational Confidence Questionnaire, smoking and, 109, 110-112
415-416, 433 study examples in, 112-117
Smoking Surgery, 224-226
change model and, 4-5, 6, 13 Systematic desensitization, 152
decision-making perspective on, 109-
118
Smoking cessation strategies, 279-287 Temperance movement, 98-99
change process in, 280-282 Temptation, 16
integrated, 286 Therapist/client relationship
motivation in, 282-284 action stage and, 10
mutual aid in, 285-286 change model and, 6-7, 20-25
overview of, 279-280 early alcoholism treatment, 206-207
simple interventions in, 284-285 self-help manuals and, 336
spouse involvement in, 308-309 self-regulation model, 37-38
Social learning theory, 408 TOTE loop, 319-320
Social skills training
alcoholism treatment, 150-151
individualized alcoholism treatment, Values
184 optimizing model, 65-66
Social stability, 192 See also Expectancy-value models
Social support networks Videotape self-confrontation, 135
addictive behaviors and, 308-309
aftercare and, 439 Weight reduction programs, 308. See also
alcoholism treatment and, 309-315 Eating disorders
defined, 305 WHO. See World Health Organization
emotional/behavioral problems and, (WHO)
306-308 World Health Organization (WHO),
functions of, 305-306 51-57, 290, 333, 363, 378,
See also Family 379,381

Treating Addictive Behaviors: Processes of Change

Uploaded by

Document Informationclick to expand document information

Document Informationclick to expand document information

Copyright:

Available Formats

Treating Addictive Behaviors: Processes of Change

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Treating Addictive Behaviors: Processes of Change

Uploaded by

Copyright:

Available Formats

Treating Addictive

Current Volumes in this Series

BEHAVIORAL ASSESSMENT AND REHABILITATION OF THE

HANDBOOK OF ASSESSMENT IN CHILDHOOD PSYCHOPATHOLOGY

HANDBOOK OF BEHAVIORAL GROUP THERAPY

ISSUES IN PSYCHOTHERAPY RESEARCH

A PRIMER OF HUMAN BEHAVIORAL PHARMACOLOGY

THE PRIVATE PRACTICE OF BEHAVIOR THERAPY

RESEARCH METHODS IN APPLIED BEHAVIOR ANAL YSIS

SEVERE BEHAVIOR DISORDERS IN THE MENTALLY RETARDED

SUBSTANCE ABUSE AND PSYCHOPATHOLOGY

TREATING ADDICTIVE BEHAVIORS

Plenum Press • New York and London

Treating addictive behaviors.

(Applied clinical psychology)

© 1986 Plenum Press, New York

All rights reserved

Helen M. Annis, Addiction Research Foundation, and University of

Claus-Peter Appel, Department of Applied Psychology, University of

Geir Berg, The Hjellestad Clinic, Hjellestad, Norway

Gerhard Biihringer, Addiction Research Group, Department of Psychol-

Sabine Dehmel, Addiction Research Group, Department of Psychology,

Carlo C. DiClemente, Texas Research Institute of Mental Science, Hous-

Dennis M. Donovan, Department of Psychiatry and Behavioral Sciences,

Christopher P. Freeman, Department of Psychiatry, University of Edin-

Marcus Grant, Division of Mental Health, World Health Organization,

Nick Heather, Department of Psychiatry, Ninewells Medical School, Uni-

Reid K. Hester, Behavior Therapy Associates, Albuquerque, New Mexico

Joanne R. Ito, Department of Psychology, University of Washington, and

Frederick H. Kanter, Department of Psychology, University of Illinois,

Franz Klett, Addiction Research Group, Department of Psychology,

Simonne LeBreton, Addiction Research Foundation, Toronto, Ontario,

Gloria K. Litman, Addiction Research Unit, Institute of Psychiatry, Lon-

Barbara S. McCrady, Center of Alcohol Studies, Rutgers University, Pis-

William R. Miller, Department of Psychology, University of New Mexico,

Jim Orford, Department of Psychology, University of Exeter, Exeter,

James O. Prochaska, Department of Psychology, University of Rhode Is-

Howard Rankin, St. Andrew's Hospital, Northampton, England

Bruce Ritson, Department of Psychiatry, University of Edinburgh, An-

lan Robertson, Department of Psychology, Astley Ainslie Hospital, Edin-

David Robinson, Addiction Research Centre, University of Hull, Hull,

Arvid Skutle, The Hjellestad Clinic, Hjellestad, Norway

Stephen Sutton, Addiction Research Unit, Institute of Psychiatry, Lon-

D. Adrian Wilkinson, Addiction Research Foundation, Toronto, Ontario,

About a decade ago, psychologists began exploring the commonalities

ogy in general, and decision-making models in particular, to problems

Donovan evaluate research on approaches to aftercare following the

We wish to express our gratitude to those who made special contribu-

PART I. CHANGE IN THE ADDICTIVE BEHAVIORS

CHAPTER 1. Toward a Comprehensive Model of Change 3

CHAPTER 2. Implications of a Self-Regulation Model of Therapy for

PART II. CONTEMPLATION: MOTIVATION FOR CHANGE AND

CHAPTER 3. From Contemplation to Action: The Role of the World

CHAPTER 4. From Contemplation to Determination: Contributions

CHAPTER 5. Critical Conditions for Change in the Addictive

CHAPTER 6. Trying to Stop Smoking: A Decision-Making

PART III. ACTION: ASPECTS AND PROCESSES OF CHANGE