Citation
A synthesis of psychologically oriented alcoholism treatment outcome research

Material Information

Title:
A synthesis of psychologically oriented alcoholism treatment outcome research
Creator:
Poage, James David, 1935-
Publication Date:
Language:
English
Physical Description:
viii, 144 leaves : ; 28 cm.

Subjects

Subjects / Keywords:
Alcoholic beverages ( jstor )
Alcoholism ( jstor )
Hospitals ( jstor )
Medical treatment outcomes ( jstor )
Outcome variables ( jstor )
Psychological assessment ( jstor )
Psychological research ( jstor )
Psychology ( jstor )
Research design ( jstor )
Statistical significance ( jstor )
Alcoholism -- Psychological aspects ( lcsh )
Alcoholism -- Treatment ( lcsh )
Counselor Education thesis Ph. D
Dissertations, Academic -- Counselor Education -- UF
Psychotherapy ( lcsh )
Genre:
bibliography ( marcgt )
non-fiction ( marcgt )

Notes

Thesis:
Thesis (Ph. D.)--University of Florida, 1982.
Bibliography:
Bibliography: leaves 132-137.
General Note:
Typescript.
General Note:
Vita.
Statement of Responsibility:
by James David Poage.

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
The University of Florida George A. Smathers Libraries respect the intellectual property rights of others and do not claim any copyright interest in this item. This item may be protected by copyright but is made available here under a claim of fair use (17 U.S.C. §107) for non-profit research and educational purposes. Users of this work have responsibility for determining copyright status prior to reusing, publishing or reproducing this item for purposes other than what is allowed by fair use or other copyright exemptions. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder. The Smathers Libraries would like to learn more about this item and invite individuals or organizations to contact the RDS coordinator (ufdissertations@uflib.ufl.edu) with any additional information they can provide.
Resource Identifier:
028540771 ( ALEPH )
08969917 ( OCLC )

Downloads

This item has the following downloads:


Full Text






A SYNTHESIS OF PSYCHOLOGICALLY ORIENTED ALCOHOLISM TREATMENT OUTCOME RESEARCH

















BY

JAMES DAVID POAGE


A DISSERTATION PRESENTED TO THE GRADUATE COUNCIL
OF THE UNIVERSITY OF FLORIDA IN
PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY


UNIVERSITY OF FLORIDA


1982



























This work is affectionately dedicated to


Peg, a warmhearted humanist, who nevertheless has a real

understanding of carrots and sticks;


Maureen, who knows why;


a few dozen alcoholics, who are my respected teachers;


and my mother and father, who between them showed me

how to be gentle, muleheaded, and logical.












ACKNOWLEDGMENTS


There are several persons who have helped me in the writing of this dissertation. Larry C. Loesch, Ph.D., who served as my committee chairman, provided me with good-humored encouragement and guidance

dispensed with a kindly hand. The other members of my committee, Paul W. Fitzgerald, Ed.D., and Franz R. Epting, Ph.D., patiently stuck with my meandering process for over five years. All have earned my

gratitude.

Mary Kay Hartung, reference librarian at the University of South Florida, was of invaluable assistance in helping me talk to a computer so that it could understand what I wanted in the way of a literature search. Linda Mazzone typed and retyped the manuscript with patience and care.

Margaret Scripps Buzzelli, owner and President of Anabasis, Inc., urged and permitted me to spend company time on the project. She knew when to say, "You can do it!" and she did.

My sons, John, Bob, and David Poage, alternately harrassed and loved me toward finishing the work. Katherine Bailey, my stepdaughter,

an accomplished applied psychologist at age eleven, administered her smile as a selective reinforcement when needed.

Finally, firstly, and in between, my wife, Maureen Ann O'Harra, Ph.D., gave me loving support without stint, served as second reader and

first-rate editorial consultant, and was optimistic enough for both of us.


iii









TABLE OF CONTENTS


PAGE

ACKNOWLEDGMENTS . . . . . . . . . . . . . . . . . . . . . . . . . .iii

LIST OF TABLES. . . . . . . . . . . . . . . . . . . . . . . . . . . . vi

ABSTRACT. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .vii

CHAPTER PAGE


INTRODUCTION . . . . . . . . . . . . . . . . .


Context. . . . . . . . . . .
General Assumptions. . . . . Statement of the Problem . . Need for the Study . . . . . Purpose of the Study . . . . Rationale for the Approach . Research Questions . . . . . Definition of Terms. . . . . Overview of the Remainder of


. 1


.1
.2
. 2 .3
.4
.5
.11
.11 .13


the Study


SURVEY OF RELATED LITERATURE . . . . . . . .

Support for the Delineated Problem . . . .
Support for the Need for the Study . . . .
Support for the Approach of the Study. . .
Summary. . . . . . . . . . . . . . . . . .

METHODOLOGY. . . . . . . . . . . . . . . . .

Overview of the Study. . . . . . . . . . .
Delineation of Variables . . . . . . . . . Description of the Population. . . . . . .
Sampling Procedures. . . . . . . . . . . .
Description of Sample. . . . . . . . . . .
Description of Research Design . . . . . .
Description of Research Procedures . . . .
Description of Data Analysis Procedures. .
Methodological Limitations . . . . . . . .


RESULTS. . . . . . . . . . . . . . . . . .


. . . . . . . .15


.15
.24 .46 .49

.50

.50 .51
.54 .55 .57 .57 .57 .58 .59


. . . . . . . . .61


Description of Studies . . . . . . . . . . .
Subject Variables. . . . . . . . . . . . . .
Outcome Variables. . . . . . . . . . . . . .
Treatment Variables... ...........
Research Models and Statistical Significance Research Models and Practical Significance iv


I


II


III


IV


.61 .81 .83 .85 .87
.94










Research Questions . . . . . . . . . . . . . . . . . . . .99
Interactions . . . . . . . . . . . . . . . . . . . . . . 103

V DISCUSSION . . . . . . . . . . . ............ 109

Generalizability Limitations . . . . . . . . . . . . . . 109
Answers to Research Questions. . .* . .* . . . . . . . . . 110
Conclusions and Interpretations. . . . . . . . . . . . . 116
Implications for Theory, Research,
Training, and Practice . . . . . . . . . . . . . . . . 120
Recommendations. . . . . . . . . . . . . . . . . . . . . 122
Summation. . . . . . . . . . . . . . . . . . . . . . . . 127

LIST OF REFERENCES . . . . . . . . . . . . . . . . . . . . . . . . . 129

APPENDICES

A DATA-GATHERING PROTOCOL . . . . . . . . . . . . . . . . . . . . 132

B INTERJUDGE AGREEMENT. . . . . . . . . . . . . . . . . . . . . . 134

C REFERENCE LIST OF STUDIES ANALYZED. . . . . . . . . . . . . . . 135

BIOGRAPHICAL SKETCH. . . . . . . . . . . . . . . . . . . . . . . . . 138


v










LIST OF TABLES


PAGE


TABLE


Table 1 Table 2 Table 3 Table 4 Table 5 Table 6 Table 7 Table 8 Table 9 Table 10 Table 11 Table 12 Table 13 Table 14


.18


.64 .82


Alcoholism Treatment Articles in Psych Info.

Description of Selected Outcome Studies
Published 1973-1981. . . . . . . . . . . .

Numbers of Studies Reporting on Various
Subject Variables. . . . . . . . . . . . .

Numbers of Studies Reporting on Various
Outcome Variabes . . . . . . . . . . . . .

Numbers of Studies Reporting on Various
Treatment Variables. . . . . . . . . . . .

Major Findings with Statistical Significance
of p-.05 or Better in Studies Utilizing
Correlational Research Models. . . . . . .

Major Findings with Statistical Significance
of p<.05 or Better in Studies Utilizing
Experimental Research Models . . . . . . .

Major Findings with Statistical Significance
of p<.05 or Better in Studies Utilizing
Quasi-experimental Research Models . . . .


Major Findings with Practical but not Statistical
Significance in Studies Utilizing Correlational Research Models. . . . . . . . . . . . . . . . .

Major Findings with Practical but not Statistical
Significance in Studies Utilizing Descriptive
Research Models. . . . . . . . . . . . . . . . .

Major Findings with Practical but not Statistical
Significance in Studies Utilizing Experimental
Research Models. . . . . . . . . . . . . . . . .


.94 .95 .96


Major Findings with Practical but not Statistical
Significance in Studies Utilizing Quasi-experimental Research Models. . . . . . . . . . . . . . . . . . .97

Major Research Questions of Outcome Studies
of Psychologically Oriented Alcoholism
Treatment. . . . . . . . . . . . . . . . . . . . . .99

Major Interactions Observed. . . . . . . . . . . . . 104



vi


. . . . .84


. . . . .87 . . . . .88 . . . . .90 . . . . .92


W I















Abstract of Dissertation Presented to the Graduate Council of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy


A SYNTHESIS OF PSYCHOLOGICALLY ORIENTED
ALCOHOLISM TREATMENT OUTCOME RESEARCH By

James David Poage

May, 1982

Chairman: Larry C. Loesch, Ph.D. Major Department: Department of Counselor Education


Thirty-four studies published from 1973 through 1981, and evaluating outcomes of psychological treatment for ten or more alcoholics at 12 months or more after treatment, were reviewed. All studies included data on subject, treatment, and outcome variables. Subject variables were reported in 17 categories, with previous drinking history reported in all studies. Outcome variables were reported in 19 categories, with drinking behavior as an outcome variable considered most frequently (in 33 studies). Treatment variables were reported in six categories, with treatment method reported in all studies.

No study treated the theoretical stance of the treatment program as a variable of focal importance.

Subject variables were measured 95 times by nonstandardized methods and 51 times by standardized methods. Outcome variables were measured 148 times by nonstandardized methods and 28 times by standardized methods.


vii










Correlational research models were used in nine studies, descriptive research models in six, experimental research models in ten, and quasi-experimental research models in nine. Every study in which an effort was made to compute statistical significance reported at least one finding significant at the .05 level or better. Findings deemed practically significant but not statistically significant were identified in 21 studies.

Two studies reported subject-treatment interactions. Subjectoutcome interactions were reported in 13 studies. Seventeen studies reported treatment-outcome interactions. One study reported on subject-treatment-outcome interactions.

Implications for theory, research, training, and practice were described. Recommendations were made for further developments in theory, research, training, and practice.


viii













CHAPTER I
INTRODUCTION


The often stodgy and ingrown world of research owes a debt of gratitude to Charles Shultz for a mildly iconoclastic Snoopy cartoon he drew in the 1960's. In the first panel Snoopy is looking at a huge and untidy pile of bones. In succeeding panels he tosses the bones one by one over his shoulder, so that in the last picture he has a large and untidy pile of bones on the other side of the yard from their original location. Snoopy grins: "We call it research."

The present dissertation intends to modify Snoopy's model by the addition of considerable order and refinement. Research in alcoholism does in some ways resemble an untidy bone-pile that needs not just reshuffling, but purposeful sorting.

The specific bone-pile of data to be sorted in this study is composed of selected research reports on the outcomes of psychologically oriented treatment of alcoholism published in English in the years 1973-1981.

Context

There is in the field of alcoholism an information explosion. The literature doubled in the seven years, 1967-1973 (Moll & Narin, 1977). It continues to grow rapidly in various directions, making any unified comprehension more difficult with each new addition. The present study is responsive to the notion that it is more helpful to make some sense of and suggest new direction for the existing corpus of research than to

add one more set of empirical findings to an already less than comprehensible pile of such data.


I






2



General Assumptions

The present author has personal reasons for undertaking this particular study in this particular way. His work is the treatment of

alcoholics, and he deems it important to learn what works, with whom, and, if possible, why. To undertake the "treatment" of "alcoholics" is

to embrace, implicitly at least, two assumptions. The first is that

there are persons to whom the term "alcoholic" is appropriately applied,

and that there is something about the condition of an "alcoholic" which is maladaptive and undesirable. The second assumption is that

"treatment" may ameliorate that which is maladaptive, undesirable, and specific to "alcoholics."

A third assumption is also important to the present study. It is

that something of pragmatic and theoretical value may be learned from the analysis of a relatively thin slice of the research literature on outcomes of psychologically oriented treatment of alcoholism.

Statement of the Problem

Succinctly put, the problem to which this study is addressed may be

stated in three propositions: (a) There are several hundred thousand alcoholics in the United States; (b) Substantial resources, both human and material, are expended on the treatment of these alcoholics; (c)

There exists a huge body of research on the treatment of alcoholics, but

there is no definitive synthesis on that research which has yielded a set of conclusions which serve as reliable guides to the most effective research, theory, practice, and training for the treatment of alcoholics.









JIN






3


Need for the Study

There are more than 9,000 alcoholism treatment programs in the continental United States (NIAAA, 1979). In 2,821 of these programs for

which data are available, approximately 28,000 treatment personnel provide some form of treatment for an estimated 1,712,000 problem drinkers and alcoholics annually at an annual expenditure of approximately $795,183,000 (DeLuca, 1981). Unfortunately, no one seems to know what is being done or how well.

The preceding sentence is, of course, hyperbole. Literally hundreds of published studies provide data on what is done in alcoholism treatment and what the apparent outcomes are. However, no single, clear picture emerges from this embarrassment du richesse. While the attainment of such a single, clear picture seems unlikely at this time, it is also apparent that there are needs which might be met by making a start on the tasks of summarizing and clarifying. Those needs are delineated in the following statements of contingency.

If researchers knew how alcoholism treatment outcomes have been studied, then researchers might identify the more (and less) fruitful avenues of approach and be guided by those identifications in future research.

If researchers knew which models of alcoholism treatment outcome studies have yielded the clearest and most significant results (and the contrary), researchers could then be more usefully selective in choosing models for further research.

If practitioners and treatment planners knew which theoretical stances are positively associated with the most favorable treatment






4


outcomes, then they could be more pragmatic in selecting the best theoretical bases for treatment.

If they knew the treatment variables which are the best predictors for the best prognosis when combined with given client variables, practitioners and treatment planners could match the treatment to the client for maximum effectiveness and efficiency.

If teachers and trainers had a clear and valid understanding of the

issues mentioned above, then they could train alcoholism counselors and other alcoholism workers with greater confidence and precision in pragmatically tested theory, diagnosis, treatment planning, and practice.

Purpose of the Study

This study has several purposes. Of necessity, the first purpose is to determine whether the research considered herein can provide answers to the research questions of this study. If it cannot provide those answers, a second purpose will be to suggest ways of research that might be more effective in supplying the needed information.

The other purposes of this study, all with reference to psychologically oriented treatment of alcoholics, are as follow: (a) To identify the more (and less) fruitful approaches to outcome research;

(b) To describe criteria for selecting useful models for such outcome research; (c) To describe criteria for selecting theoretical models of alcoholism which are related to positive treatment outcomes; (d) To describe criteria for matching treatment to client; (e) To describe theories and practices appropriate for inclusion in the training of alcoholism counselors and other alcoholism workers.





5


Rationale for the Approach

The approach of this study is descriptive and, in a sense, historical. It is descriptive in that the emphasis is on delineating facts

and characterics of a population of interest (Isaac & Michael, 1971), and historical in that the immediate population of interest is composed not of persons but of documentation and interpretation of events already past.

The descriptive-historical approach has some disadvantages. Compared to experimental approaches, the data used are secondary in the sense that the author is not working directly with first-hand sensory data or such sensory data enhanced by instrumentation. Hence,

descriptive-historical studies may be seen as less empirical than experimental studies. The conclusions of such descriptive-historical studies may be viewed as more subjective and less verifiable than those reached via "brass instrument" approaches. The corresponding advantage of the descriptive-historical approach in the case at hand lies

precisely in the fact that so many empirical, if not purely experimental, data exist without organization into meaningful Gestalts. Information is not the same as knowledge, and knowledge is sorely needed in the field of alcoholism. Human beings gather data; computers store but do not integrate data. For data to become useful, it is necessary that they become integrated, organized under categories relevant to human interests, which interests in the final analysis determine what is "useful." Some "subjectivity," in the sense of value-informed





1The remainder of this section, "Rationale for the Approach," represents the current writer's own thoughts.






6


choices which constitute organizing principals for the integration of information, is a necessity if information is to become useful. Thus, while the conclusions of a descriptive-historical study are more

"subjective" than those of other approaches, that very "subjectivity," as described above, renders the work useful, albeit a price is extracted in terms of the lowered verifiability by independent observers.

Whatever is lost in empiricism may well be regained in pragmatism.

A second disadvantage of the descriptive-historical approach versus

a more empirical approach is that the former has less formidable safeguards against author bias. In the present instance, one of the

author's major biases is a preference for the practical; he is in search of that which works. This interest is not an unworthy one. Why, after all, is research undertaken if not finally to lead to workable solutions to significant problems?

Perhaps the most weighty disadvantage of the descriptive-historical approach to the problem at hand lies in the nature of the problem. It is possible that the materials to be studied are so diverse in method, focus, assumptions, and treatment that they simply are not amenable to summarization and integration in a manner that is both meaningful and logically valid. The counter to this disadvantage is that the need is so great that some attempt to address the problem is justified, even if ultimately it should fall far short of desired success.

A descriptive-historical method of studying research on alcoholism treatment outcomes has the advantage of being responsive to the need for a sort of Baedeker's Guide to Researchland. There is a finite but huge number of paths that might be taken (or hacked out) in the wilderness of potential alcoholism research. If the present writing can identify only





7


a few demonstrably dead-end trails and one or two routes that may lead to the useful, the project will have been worthwhile.

While the "Baedeker Factor" points to a clear advantage of the descriptive-historical approach, the approach has also an inherent disadvantage for such purposes. The author might through bias, ineptitude, or ignorance miss the mark and encourage ultimately fruitless lines of exploration or inappropriately dismiss as not worth the effort approaches which could in fact yield valuable results. The

descriptive-historical approach has fewer inherent safeguards against error due to bias than more rigorously empirical methods--although no known methodology is proof against determined ineptitude or ignorance.

For all of that, these risks are counterbalanced by the facts that the author's conclusions will be based not only on bias but also on data

in the form of published research, that alcoholism researchers are not famous for being easily discouraged from following their own bents, and that a Baedeker is needed and will be subject to critical review which could reveal its errors.

The current study makes minimal use of statistical analysis. The decision to work in this way has the disadvantages of rendering the findings less precise than they might otherwise be and of overlooking subtle differences and relationships visible only through more sophisticated statistical analysis. However, the primarily non-statistical summary and integration undertaken here is more compatible with the scope of the author's intent and capabilities. The reader should keep in mind that this study is intended to be of practical value for front-line alcoholism workers, many of whom do not hold college degrees and have had no exposure to more than the simplest statistics (e.g., averages).





8


The commitment to practicality brings with it a limitation; the current study is not apt greatly to advance alcoholism theory per se. The advantage of the pragmatically oriented approach is that it lends itself well to improvement in research, practice, and training. Moreover, it is hoped that the observation of correlations of programs' theoretical stances with treatment outcomes may lead toward reality-based theory selection in program design and training.

The current project is not exhaustive in its scope. The vast

volume of published research on alcoholism and the span of human life preclude the practical possibility of any one person even reading all the material, much less synthesizing it. The focus must be narrowed in both longitude and latitude.

The longitudinal limitation of this study is roughly demarcated by the years 1973-1981. One reason for this is that to examine a nine-year

slice of the literature is obviously more manageable than to survey references to alcohol problems from the early Babylonion epoch to the present. Also, other writers (Baekeland, 1977, Costello, 1975 a, b, Emrick, 1973) have carried out somewhat similar studies of the relevant literature into the early 1970's.

The selection of the more recent literature also reflects the assumption that each generation of researchers builds on the work of its

predecessors, screening out gross error and retaining proven fact and useful knowledge, so that more recent productions are a kind of

distillate and collectively constitute at least the beginnings of integration and synthesis of what the research community has learned.

The disadvantage of a nine-year longitudinal slice of the literature is the possibility of missing longer-term trends in research,






9


but, since it is not a central emphasis of this paper to analyze such trends, and since literature reviewing older literature is considered herein, this is not a telling disadvantage.

Three latitudinal limitations of note have been chosen for the current study. Firstly, it is limited to published work in English. This carries with it the possibility of missing some important reports in other languages. However, a computer-assisted literature search, per

the descriptors used, turned up only five non-English language reports which appeared to be of even remote interest and which on the basis of their abstracts could be excluded as not of interest.

Secondly, the current study is restricted to reports dealing with psychologically oriented treatments of alcoholism. This limitation of focus is disadvantageous in that alcoholism has biological complications, may well have biological causal components, and has often been treated as a physical illness. The literature on medical and

especially pharmaceutical treatment is voluminous, highly technical, intriguing, and beyond the scope of both the author's capabilities and his primary interest. The choice to exclude nonpsychological treatment of alcoholics is predicated on four facts: (a) Beyond detoxification and the use of chemical antagonists of alcohol (e.g., disulfiram), alcoholism per se is usually regarded as requiring primarily psychological treatment, although concomitant conditions such as cirrhosis of the liver or alcoholic gastritis of course call for medical treatment; (b) Enormous amounts of money and energy are expended on psychologically oriented treatment, and present knowledge about the efficacy of such treatment is insufficient; (c) The literature on

alcoholism treatment regardless of treatment orientation is too great in






10


size to be encompassed in the present study; (d) The context of this dissertation is psychological, not medical, penal or religious.

The mention of the words "penal" and "religious" above leads into the third latitudinal limit of this dissertation. While punitive

measures have been evoked as a means of social control or "treatment" of

alcoholics, to consider this aspect of response to alcoholism would go beyond the intended scope of this study, even though such measures surely have psychological dimensions. In a similar vein, there is a time-honored tradition of viewing alcoholism as a spiritual-religious problem to which the spiritually oriented approaches have been myriad and powerful.

Foremost among those spiritually oriented approaches, certainly in the twentieth century, has been the fellowship of Alcoholics Anonymous. There is no doubt that AA may properly be described as a psychological treatment (broadly defined) of alcoholism. However, Alcoholics Anonymous as a treatment will not receive major attention in the present

work for three reasons. (a) AA is infrequently treated as a sole

treatment modality in the research literature, partially because a group

whose cornerstone is anonymity is very difficult to study scientifically. (b) While many treatment programs encourage client exploration of and involvement with the principles and program of AA, well defined AA tradition prohibits that there should be a formal, institutional AA treatment program extant anywhere save in the voluntary and anonymous fellowship of AA. (However, in instances in which formal, institutional treatment program researchers reviewed herein have

reported AA attendance as a client or treatment variable, such data are






11


included in this study); (c) A study of the treatment outcomes of

Alcoholics Anonymous would constitute in and of itself a monumental undertaking far beyond the resources and primary interest of the author.

Hence, while AA is seen as a potent force in the psychologically oriented treatment of alcoholism, AA per se must be by and large excluded from consideration herein.

Research Questions

The general research question addressed in this study is: "How

have the outcomes of psychologically oriented alcoholism treatment programs been studied and what has been learned?"

The more specific research questions are six in number. (a) What client, treatment, and outcome variables have been studied; (b) To what

extent have theoretical stances of treatment programs been taken into account as variables contributing to outcomes; (c) What are the ways in which client, treatment, and outcome variables have been measured; (d) What research models have been used; (e) What research models have

yielded significant results; (f) What interactions of client, treatment, and outcome variables have been observed?

Definition of Terms

The history of attempts to define alcoholism has been fraught with ambiguity, frustration, conflict, and some might say, failure. Definitions abound, their number sometimes seeming closely to approach the number of persons using the term, but no one definition has ever captured the loyalty of all concerned parties, and perhaps not even of a majority. The lack of a firm, universally accepted definition is no minor inconvenience but has at times constituted a serious impediment to understanding and treating the condition (Keller, 1960). In a





12


popularized discussion of alcoholism, Johnson (1973, p. 1) characterizes

alcoholism as a disease which "involves the whole man: physically, mentally, psychologically, and spiritually" and which is "primary, progressive, chronic, and fatal." It should be noted that Johnson founded and guided the development of The Johnson Institute in Minneapolis, one of the more prestigious alcoholism treatment centers in

the United States, and that his definition approximates a widely held conception of alcoholism.

A more formal definition is offered by a source of equal or greater prestige, the National Council on Alcoholism/American Medical Society on Alcoholism:

Alcoholism is a chronic, progressive and potentially fatal disease. It is characterized by tolerance and physical
dependency or pathological organ changes, or both--all the direct or indirect consequences of the alcohol ingested
(DeLuca, 1981, p. 36).

The most recent (1977) World Health Organization definition of the syndrome of alcohol dependency stresses the "compulsion to take alcohol on a continuous or periodic basis in order to experience its psychic effects, and sometimes to avoid the discomfort of its absence" (DeLuca, 1981, p. 36).

There exists a wide range of definitions-by-example of alcoholism, from any use of ethanol as a beverage (Gordon, 1958) to the chronic intoxication of the stereotypical skid-row wino. Since personnel who make the diagnostic judgments which determine the inclusion or exclusion of subjects for research reports range from non-degreed

para-professional counselors to rigorously trained research scientists, it is reasonable to believe that research reports span a correspondingly wide range of definitions of alcoholism. Moreover, Keller (1960) argued






13


convincingly that there is no escape from the use of the diagnostician's subjective judgment in diagnosing alcoholism.

For purposes of the present writing, it is necessary to recognize that the literature surveyed neither uses a single definition nor restricts itself to a single set of diagnostic criteria. Therefore, however circular and semantically unsatisfactory it may be, in this study, the word "alcoholism" refers to those conditions believed by the various

diagnosticians represented in the several studies to be present in persons treated for alcoholism. "Alcoholic," herein, refers to a person who "has alcoholism." More simply, in the present writing, "alcoholism"

is whatever is so named by an investigator and an "alcoholic" is whomever an investigator designates as such. Whenever investigators

have indicated which definition of alcoholism was used for diagnosing subjects, that fact will be noted.

As used herein "psychologically oriented treatment" refers to actions undertaken with the intentions that the client: (a) cease and desist from, or exhibit a lower frequency and/or intensity of, maladaptive behaviors associated with alcoholism and (b) exhibit a higher frequency and/or intensity of adaptive behaviors. "Behaviors" in this context is broadly construed to include such constructs as attitudes, thoughts, feelings, beliefs and values, whether or not manifest in "observable behavior." Excluded from this category are treatments which have no specific psychological orientation, such as incarceration, general medical treatment, or the administration of non-psychoactive pharmacological preparations.






14


Overview of the Remainder of the Study

The remainder of this study is organized in four chapters. Chapter

II consists of a review of the related literature on reviewing and analyzing alcoholism treatment outcome literature. Chapter III sets

forth the methodology used herein. Chapter IV contains the summary and analyses of the data examined. The final chapter is a discussion of the

results described in Chapter IV, including answers to the research questions, conclusions and implications and recommendations for theory, research, practice, and training.












CHAPTER II
SURVEY OF RELATED LITERATURE


Support for the Delineated Problem

The problem to which this study is addressed has been outlined in the preceding chaptec in three propositions, which are restated and supported below from relevant literature. Size of Alcoholic Population

First, there are many alcoholics in the United States--and, of course, in other nations, but the present study is focused primarily on this country. A highly regarded resource for information on the

prevalence of drinking problems in the United States is the series of special reports to the U.S. Congress on alcohol and health from the Secretary of Health and Human Services. The latest such report (DeLuca,

1981) describes a pattern of increasing apparent alcohol consumption in the U.S. from two gallons per capita per year in the late 1950's to 2.5 gallons by 1970. Thus, in a single decade there was a 25% increase in apparent per capita consumption. It should be noted that these figures

are in terms of ethanol, which in the report is used synonymously with the term absolute alcohol or pure ethyl alcohol regardless of whether it

was consumed in its more diluted forms (beer, wine) or in its less diluted form (distilled spirits). Apparent consumption rates were derived by dividing the number of gallons of ethanol sold in the form of

alcoholic beverages by the number of persons in the population 14 years of age and older, with the assumption that what was sold as a beverage was consumed as a beverage.


15






16


By 1978, apparent consumption stood at slightly more than 2.7 gallons per year per person aged 14 years or more. By 1981, apparent daily consumption averaged one ounce of ethanol (roughly two drinks) per

person 14 years of age or older; the annual apparent consumption figure would be 2.85 gallons. However, one-third of the adult population

reports not drinking alcohol at all, forcing the conclusion that the mean daily ingestion is higher (about 1.5 ounces or three drinks) for those who do drink. Even this average can be misleading, since a small segment of the adult population drinks much more than the average while the majority drink far less. The report offers an estimate that

approximately 11% of the adult population consume about half of all alcoholic beverages sold.

The estimate that 11% of the drinking age population consume one-half of the beverage alcohol sold is especially interesting in light

of the conclusion in the same Fourth Special Report to Congress that approximately one out of ten adult American drinkers are likely to be either alcoholics or problem drinkers at some time in their lives (DeLuca, 1981). Taking as a conservative estimate that 100 million Americans drink alcohol, and using a low cutting point for problem drinking or alcoholism, it would be concluded that about twenty million fall into the alcoholism or problem drinking category, while a much higher cutting point would still yield an estimate of ten million alcoholics or problem drinkers in the United States.

Even if one presumes that heavy consumption, problem drinking, and alcoholism are overreported, one is still led to the unavoidable conclusion that there are in the nation large numbers of persons described as alcoholic.






17


Extensive Expenditure of Resources on Alcoholism Treatment

The second proposition to be supported in this chapter is that there are vast resources spent in alcoholism treatment. The evidence that great amounts of human and financial resources are expended in the treatment of alcoholics is even clearer than the evidence supporting the

existence of large numbers of alcoholics, and further supports the assertion that there are many alcoholics. Reference has already been made above (Chapter I) to the 1979 NIAAA Directory's listing of over 9,000 alcoholism treatment programs. In surveying that directory, the author found that the majority of private alcoholism treatment

facilities personally known by him to be extant in 1979 or earlier were not listed. One must conclude that there are many more than 9,000 such facilities. It should be noted that the $795 million 1979 expenditure figure for alcoholism treatment mentioned in Chapter I was derived from only 4,073 alcoholism treatment units which reported funding information

in the 1979 National Drug and Alcohol Treatment Survey (DeLuca, 1981). Since fewer than one-half of the known alcoholism treatment units reported, it is clearly the case that more than $800 million is being spent annually on alcoholism treatment in the United States. Paucity of Knowledge About Treatment Outcomes

The third proposition stated in the delineation of the problem at hand is that too little is known about the effect of the efforts to treat alcoholism. The problem is not primarily that too little alcoholism research has been done and reported. Rather, the difficulty has to do with the fact that so much has been reported and so little synthesized.





18


To determine the accuracy of his impression that the research

literature on alcoholism is indeed increasing rapidly in volume, the

current writer queried a computerized data retrieval system known as

Psych Info, which accesses materials referenced in Psychological

Abstracts. The question put to the system was: "By year, how many articles on alcoholism treatment are in the Psych Info data pool?" The

printout received included all articles so categorized from the 1967 inception of the system to the query date, September 21, 1981. The information yielded, plus derived percentages, is incorporated in Table 1.



Table 1
Alcoholism Treatment Articles in Psych Info




Projected no.
No. alcoholism % Citations on alcoholism
treatment Total alcoholism treatment
Year citations citations treatment articles



1981 6 1,518 .4 (195)
1980 48 16,092 .3 (195)
1979 98 23,447 .4 (195)
1978 195 30,366 .6
1977 161 30,029 .5
1976 147 28,258 .5
1975 144 27,546 .5
1974 159 28,350 .6
1973 148 25,753 .6
1972 107 22,658 .5
1971 99 22,141 .4
1970 76 20,034 .4
1969 89 21,142 .4
1968 109 21,032 .5
1967 81 18,115 .4

(TOTALS) 1,714 336,481 (mean %=.5)

Average number alcoholism treatment articles per year = 114 Average total citations per year = 22,432





19


The increase in the number of articles on alcoholism treatment over the 12 year period 1967-1978 is impressive. From 81 articles in 1967 the number grew to 195 articles in 1978, an increase of 240%. Not only has there been an increase in the number of such articles, but also the proportion of alcoholism treatment articles to the total number of citations in Psychological Abstracts has increased from .45% in 1967 to .64% in 1978. It is inferred from the total citation figures in Table 1 that for the years 1979-1981 not all of the citations in

Psychological Abstracts had been entered in Psych Info, so that the listed percentages of alcoholism treatment articles for those three years ought not to be regarded as trustworthy. Therefore, both the

total number of citations (30,366) and the percent of alcoholism

treatment articles (.64%) for 1978 were chosen as constants to yield a projected 195 alcoholism treatment articles per year for the years 1979-1981.

By conducting some manual literature search and accessing the computer data pool of NIAAA via a computer search, the author discovered

that in regard to alcoholism research Psych Info is exhaustive neither in its descriptor system nor in its scope of publications. It is

reasonably concluded, then, that the literature on alcoholism treatment since 1967 far exceeds the 1,714 citations turned up by the single query of Psych Info.

Further corroboration of the fact that the literature on alcoholism research is voluminous and expanding is found in Moll and Narin's (1977) effort to characterize that body of literature. They found that alcohol

research publications doubled in the seven year period (1967-1973)






20


studied. In 1967, 1,134 articles were noted. By 1970 the number was

1,220, and by 1973, 2,348, with the publications scattered through more than 1,000 different journals. Moll and Narin categorized the

literature in three sectors: biomedical, biosocial, and psychosocial. Their psychosocial sector would seem to include the majority of works which in the present study would be classed as dealing with

psychologically oriented treatment of alcoholism. While that sector

occupied some 14% of the literature found in their search, the category was not exclusively restricted to treatment articles, so that it would seem inappropriate to conclude that the scope of the literature of interest here was as large as 14% of their totals, i.e., 160 for 1967, 170 for 1970, and 328 for 1973. Nevertheless, it is reasonable to hold that the volume of alcoholism treatment literature is even greater than that implied by Table 1 above.

Thus far it has been demonstrated that there are a significant number of alcoholics, that they are treated at great expense, and that there is a large and rapidly growing body of research on that treatment.

The remaining assertion in the delineation of the problem is not so easily demonstrable, because it is the negative proposition that there is no definitive synthesis of the research on alcoholism treatment. "Empirical proof" of a negative proposition is, in the strictest sense, impossible, since it would require an exhaustive inspection of all of the relevant data along with proof that the inspection was exhaustive. Hence, it must serve to state here that among the literature surveyed by the present author, there have been found only four recent major reviews

of treatment outcome reports, and no one of them appears to have been accepted by the alcoholism research community as a definitive synthesis.






21


The earliest of the four reviews is that of Hill and Blane (1967), in which they examined 49 studies published in the United States and Canada from 1952 through 1963. The focus of interest in their survey was on the methodological aspects of the studies reviewed, with attention primarily to the designs of the studies and the methods of reporting findings. They found that the various authors represented in their sample generally failed to meet one or more basic requirements of scientific evaluative research and that many reported their findings in ways that were vague, incomplete, or confusing. Hill and Blane did not in their article make statements attempting to set forth a synthesis of the findings of the research evaluating psychotherapy with alcoholics. Indeed, given the breadth and cogency of their methodological

criticisms, one would feel safe in deducing that they would maintain that no nontrivial synthesis could be made of those findings.

Emrick (1974, 1975) reviewed studies published in English from 1952 through 1973 which reported patient outcomes of psychologically

oriented alcoholism treatment. He located 384 such studies published over the 22 year period. With regard to outcomes following treatment he concluded (Emrick, 1974) that approximately 36% of patients were

abstinent during follow-up and about 5% were controlled drinkers, so that the abstinent or controlled drinking segment constituted roughly 40% of the sample. Using other categories, he found that while less than 20% were totally abstinent, about one-third were improved although not necessarily totally abstinent or controlled; two-thirds were at least somewhat improved; one-third were unimproved, and between one-tenth and one-twentieth were in worse condition. While these

findings may appear to be so general as to be trivial, something





22


else that Emrick noted in process of searching them out is of significance.

Emrick (1974, p. 533) established statistical guidelines for

evaluating the commonness of results, specifically, that for rates of abstinence, abstinence-or-controlled-drinking, much-improved,

somewhat-improved, much-or-somewhat-improved, total-improved,

total-unimproved, and deteriorated, specific percentage ranges can be used to judge with objectivity whether the results of a given study are common or are especially noteworthy as being unusually high or low.

Later, Emrick (1979) analyzed 90 studies of randomized controlled trials of alcoholism treatment published between 1952-1978, seeking evidence for the relative effectiveness of various treatment approaches.

He stated that his most salient findings were (a) when nonbehavioral treatment is applied to heterogenous groups of alcohol abusers, its effectiveness is not increased by rendering more than brief care and (b) certain behavioral approaches have been shown to be relatively effective in diminishing problem drinking.

While Emrick's work is useful in that it explicates a number of issues vital to research on alcoholism treatment and provides some clear

guidelines for future research, it does not constitute a definitive or universally acceptable synthesis.

Costello (1975 a) collated 58 research reports published between 1951-1972 on outcomes of alcoholism treatment over a one-year followup period. He categorized outcomes of each study in terms of case fatality rate, problem-drinking rate, success rate, and followup unavailability rate. By a hierarchical grouping method he clustered the studies in groups of maximum multivariate similarity. He noted some similarities





23


within clusters of studies characterized as having good outcome profiles

and some consistencies within those groups of studies described as having poor outcome profiles.

Following the same procedure, Costello (1975 b) collated the findings of 23 studies with two-year followup published between 1952-1972. His results were similar to those of his earlier effort (1975 a) and suggested to him some general "baselines" for measuring treatment effectiveness. He proposed that overall treatment program goals should be relative to the expected outcomes for various prognostic

subgroups of patients as those prognoses may be expected to interact with the varieties of treatments which each treatment program is able to offer.

Baekeland (1977) undertook a more general approach in his critical review of English-language literature for the years 1953-1972 on

evaluation of treatment methods in chronic alcoholism. (The present

writer has found no way to account for the rather late publication of this survey.) Having reviewed a 20 year segment of the literature, Baekeland concluded that (a) multifactorial outcome measures are superior to abstinence alone as treatment success criteria; (b) patient,

rather than treatment, factors play a dominant role in outcomes; (c) no great differences appear in the effectiveness of different kinds of treatment regimens. Baekeland's survey and conclusions are helpful, and

all three conclusions mentioned above are frequently reflected in research carried out since his publication. However, there is a need to discover and synthesize what has been happening in the research literature more recently and what it means for the future.





24


The present writer judges that while the reviews by Hill and Blane (1967), Emrick (1973, 1974, 1975), Costello (1975 a, b), and Baekeland (1977) are both admirable and useful; they do not constitute, individually or collectively, a definitive or universally accepted synthesis of the findings on outcomes of alcoholism treatment. Furthermore, since he has found in surveying the literature no reference

to any other work that seems to be so regarded, he concludes that there is extant no definitive synthesis.

In summary, the literature cited in this section supports the propositions which bound the problem to be addressed herein.

Support for the Need for the Study

The need for the study was delineated in the preceding chapter in five statements of contingency. Support from the literature is provided for those statements below.

Need to Identify Fruitful Avenues of Approach

It was stated in the first chapter that there is a need to identify

the more (and less) fruitful avenues of approach to the study of alcoholism treatment. More explicitly, in order for research on

alcoholism treatment outcomes to go forward in useful fashion, it is important to know what questions to ask of the phenomena being studied. "What are the fruitful avenues of approach?" may be reduced to "What sorts of questions are most likely to yield significant and useful information?" When one begins to probe for an understanding of whence research questions arise, it soon seems obvious that scientific

objectivity is somehow inextricably related to subjectivity of what could be termed an artistic nature. In alcoholism research, as in other

human science disciplines, there is an artistic struggle to discern the






25


concrete and human in the intangible and abstract statistic (Keeley, 1979). There is the quest for that knowledge which will help, and thus the valid, primitive reason for research both gives research its direction and sows the seeds of potential error. The human interest of the researcher gives limits to the research questions and thus at the same time renders the endeavor possible (because one cannot ask everything and get anywhere) and biased (because if one asks question A,

one may miss the valuable information that would be gained by asking question B).

The paradoxical necessity of researcher bias in choosing avenues of approach is especially relevant to alcoholism research. Pattison,

Sobell, and Sobell (1977, p. 36) have claimed that the traditional concepts of alcoholism distort treatment outcome evaluation. They

assert that the traditional concept is that alcoholism is a unitary phenomenon, discontinuous with normal drinking patterns, frequently

marked by an apparently irresistible craving for ethanol and loss of control of drinking, representing a permanent and irreversible

condition, and constituting a progressive disease which proceeds on an inexorable course through a distinct series of phases. Such a

conception, of course, dictates to a significant degree what questions the researcher will ask about alcoholism treatment. It may lead to an oversimple classification of treatment outcomes into the dichotomous grouping of "drunk" or "sober," with no attention given to intermediate drinking outcomes. Fortunately, many researchers have gone beyond the simple dichotomy; unfortunately, they have often used vague and

arbitrary categories such as "occasional slips" or "improved drinking" rather than explicitly defined and quantified classes of outcomes. One










result is that direct comparisons between outcome studies are difficult and of doubtful worth (Pattison et al., p. 36).

Moreover, Jacobson (1976, pp. 15-16) fias pointed out that there is evidence that alcoholism is not a unitary clinical phenomenon but a complex multidimensional problem, and that to attempt to measure the problem in terms of the presence or absence of a single common symptom may lead to faulty understanding. He advocates the use of the concept of "alcoholisms" which might be identified and amenable to treatments specific to each. Whether or not his notion will prove useful, it is clear that he understands and takes serious account of the fact that the

researcher's conception of alcoholism is a crucial determinant of the questions to be put to the phenomenon and its data.

While much research on the effectiveness of alcoholism treatment has been published, reviews of the literature (Baekeland, 1977; Emrick, 1973, 1974, 1975; Hill & Blane, 1967) afford no clear consensus with regard to expected rates of relapse or successful treatment (Gottheil, Thornton, Skoloda, & Alterman, 1979, p. 91). Gottheil et al. (p. 92) note that the reported results of treatment vary greatly between studies

and may generally be adjudged good or bad depending whether one's criteria are stringent, such as absolute and lasting abstinence, or more lenient, such as decreased drinking or improved psychosocial functioning. They stress the fact that while the varying stringencies of outcome judgment criteria affect statistics, they do not really alter

the outcome for the patient, and that, nevertheless, the controversy over abstinence versus moderate drinking as treatment goals is no mere semantic debate. The opposed stances are representative of theoretical postures based on real world experiences and having potent implications






27


not only for treatment goals and methods but for the kind of research that should be undertaken. The present author would add that the theoretical position of the researcher also implies or even dictates the research approach that will be taken.

While it is true that the traditional conception of alcoholism holds sway over most alcoholism treatment personnel and many researchers

(Pattison et al., 1977, pp. 37-39), new conceptions are being developed (Jacobson, 1976, Pattison et al., 1977), and each tends to generate approaches to research. Reviewers cited above (Hill & Blane, 1967; Emrick, 1973, 1974, 1975; Baekeland, 1977) have stated or implied in one

way or another that research reports on alcoholism treatment outcomes are so varied and often inexplicit in their approaches that cross-study comparisons are difficult to achieve at best, and that the results of such comparative efforts are problematic. Hence, the present author

concludes that the literature reviewed herein supports the contention that there is a real need for clarification of research approaches taken and for identification of the more and less fruitful avenues of approach to research on alcoholism treatment outcomes. Need for Identification of Productive Research Models

That productive research models are needed is a proposition with which few would contend. However, it is important that the question, "Productive of what?" be addressed. By "a productive model for

research" in the present context is meant a way of studying and reporting on the phenomena of interest which yields clear and significant results. "Clear" in this context is not far from the ordinary dictionary meaning of the term. Clear results are those which are relatively unambiguous and understandable. "Significant" here is






28


used with a pragmatic connotation to refer to those results which are useful for guiding and improving the processes of treatment.

"Statistical significance" is included in the notion of significance here insofar as statistical significance constitutes a trustworthy indicator of real world validity and reliability.

Of the four reviews of the literature on psychologically oriented treatment of alcoholics referred to thus far, the work by Hill and Blane

(1967) is the most attentive to the issue of choice of research model. Their contention was that any effort in evaluative research should attempt to meet some specific basic requirements of scientific behavior. The basic requirements which they set forth include the use of controls,

either matched control and treatment groups or random assignment of subjects to treatment or control conditions, careful selection and definition of the behavior to be evaluated, the use of reliable measurement instruments and methods, the establishment of pretreatment baselines, and the application of the same measures before and after treatment (Hill & Blane, 1967, p. 77). These standards apply to

conducting an evaluative study. Hill and Blane lay out other requirements for the reporting of such studies. They maintain that the setting

in which the treatment took place should be described clearly; the type of therapy should be specified; descriptions of the population, sampling procedures, and characteristics of the samples should be included, along with the facts concerning "lost" subjects; the instruments of measurement used, their reliability, and the timing and manner of their

administration should be specified; findings and attendant statistical applications should be presented in enough detail to permit readers to make their own interpretations of the data (p. 78).






29


The fact that Hill and Blane were so careful to describe requirements for conducting and reporting evaluative research on psychotherapeutic treatment of alcoholics indicates their belief that the issue of research model is a vital one. What they discovered about the selection and use of research models was, by their standards, disappointing.

In reviewing 49 studies published 1952 through 1963 in the United States and Canada, they found only two which were prospective (Hill & Blane, p. 78) and none which perfectly met their criteria for good scientific research. In their sample 43 articles used the patient as the patient's own control, most of them inadequately. None of the

studies made use of a nontreatment control group. None controlled for motivation, although Hill and Blane (1967) regarded it as the one single factor most necessary to control. No study reviewed by them claimed to have used a representative sample of alcoholics-in-general. Sampling

procedures were seldom clearly described and often not reported at all. About one-half of the studies failed to report even the gender, or number of subjects of each sex, in their sample. Only three studies

used sampling procedures developed prior to treatment; in the other 46 random selection and representative sampling were precluded by the use of after-the-fact sampling procedures.

Hill and Blane found serious faults in the selection and definition of criterion variables. All of the studies used drinking behavior as a major criterion; almost half used it as the sole criterion for improvement. This happened despite the fact that the notion that "a one-to-one correspondence exists between the amount of alcohol consumed and severity of problems" was deemed by Hill and Blane (p. 87) to






30


represent a position then known to be untenable. The few studies

reviewed which did undertake to use multiple outcome criteria were inadequate in their methodological management of those criteria.

Almost all of the studies reviewed were weak on the point of reliability. The most frequently-occurring source of data was the interview, and in many instances the basic data sources were clinical interviews which had not been designed with evaluation in mind. Specific

instruments such as attitude scales or other psychological tests were seldom used. On the whole, the studies were found lacking in attempts to deal with the reliability issue. For example, although all of the studies involved judgments on "improvement," only two reported interjudge reliability coefficients and all others were vague in

reporting how ratings were conducted or the extent of agreement between rating sources (p. 92).

On the whole, the articles reviewed failed the requirements for measurement before and after treatment. Hill and Blane found an

acceptable use of pre-post measurement comparisons in only three instances. They held that because of the various authors' failure to provide sound or sufficient pretreatment information, the results of most of the studies could not easily be interpreted.

While posttreatment data collection was more systematic, it frequently resulted in biases related to inadequate follow-up procedures

and failures either to control for or report temporal relationships between date of admission, duration of therapy, and the interval between discharge and evaluation (p. 97).

Hill and Blane found their 49 studies to be inadequate in regard to

the requirements for reporting evaluative research almost as often as





31


they were remiss in meeting criteria for conducting research. Most of the articles did include clear descriptions of the setting of the work described, but several made only vague reference, e.g. "psychiatric treatment," to the type of treatment. Generally their descriptions of the population and sample characteristics were less than optimum, and shrinkages of as much as 30% of the sample were frequently left unexplained and without discussion. Rarely did the studies report

clearly on the definition of the evaluation criterion behavior or its measurement.

Because of the above-noted inadequacies, along with the wide variations in settings and types of treatment conducted, Hill and Blane did not even attempt to make a summary regarding "improvement" (p. 100). Their decision underlines and emphatically supports the need for the identification (and use) of research models that will produce clear and significant results.

Baekeland (1977) reviewed the English-language literature for the years 1953-1973 on methods for the treatment of chronic alcoholism. His conclusions were in some ways similar to those of Hill and Blane, although his emphasis of inquiry was somewhat different. Laying less stress on criticism of the studies themselves and more on the evaluation

of the actual treatment reflected in the reports, Baekeland went in search of solid conclusions about the indications for various methods of

treatment for alcoholism and the relative efficacy of such treatment methods. He chose to look at alcoholism treatment programmatically, in terms of inpatient versus outpatient and in terms of three currently popular single mode approaches, i.e., Alcoholics Anonymous, pharmacological treatment, and behavioral psychotherapy. Baekeland reached eight major conclusions as follow:









(a) Multifactorial outcome measures are superior to abstinence

alone as success criteria;

(b) The absolute minimum acceptable- follow-up interval is six

months;

(c) Patient variables rather than treatment factors play a

dominant role in treatment outcomes;

(d) Outpatient programs had higher improvement rates than

inpatient programs;

(e) The population served by Alcoholics Anonymous is markedly

different from that served by hospitals and clinics, and when population differences are taken into account, it may be that

clinic treatment has a higher success rate than AA;

(f) Behavioral approaches appear to produce about the same results

as other treatment methods, if account is taken of the fact that they are usually applied to carefully selected

volunteers;

(g) Because of high dropout rates and uncontrolled factors in the

studies conducted, the usefulness of antidepressants or

tranquilizing drugs in the treatment of alcoholism has not

been fairly tested;

(h) "Patients who do well on drugs, psychotherapy, or rehabilitation programs seem to have different characteristics and success rates go up with the number of treatment options given

the patient" (p. 428).

Support for the present writer's contention that there is a need for the identification of productive models for alcoholism research is to be found in the first three of Baekeland's conclusions as listed.





33


Moreover, Baekeland's emphasis throughout on the dominance of patient variables over treatment factors in predicting outcomes leads clearly to the inference that there is a need for research models which incorporate patient variable data and demonstrate interactions of patient, treatment, and outcome variables.

The intent of Emrick's 1974 article is similar to that of Baekeland's. Emrick reviewed 271 evaluative studies on outcome of psychologically oriented treatment of alcoholics published in English from 1952 through 1971 with the stated goal of "collecting data relevant to the nature and value of alcoholism treatment" (p. 523). He focused on three areas: (a) outcome criteria used to evaluate treatment; (b) how those criteria were related to one another; (c) the effect that treatment had on drinking behavior.

Emrick found that a great number of outcome measures had been employed. He combined those measures into 19 criteria clusters, one of which was frequency or amount of alcohol ingestion. Dropping from

consideration six studies which were single-case reports, he analyzed the remaining 265 in terms of the relationships between outcome criteria. (It is a fact worthy of note that only one study of the 265 reported data on such interrelationships.) He found that for more than two-thirds of the reports, drinking outcome related positively with other outcome criteria. He tested those relationships by applying the binomial test and found that the null hypothesis could be rejected at the .05 level or better.

In short, although Emrick's primary concern was not with research methodology per se, he appears to have found the state of the art in better shape than did Hill and Blane some seven years previous. They





34


expressed concern about researchers' using total abstinence as the sole criterion on the drinking dimension (Hill & Blane, 1967, p. 85). Finding

as he did that many evaluators used more than the abstinent-not abstinent dichotomy in measuring drinking outcome, Emrick suggested that

the concern of Hill and Blane, and, by extension, that same concern expressed later by Baekeland (1977) seemed more theoretical than actual (Emrick, 1974, p. 534).

From a methodological standpoint, Emrick's most important contribution coming out of his question about the impact of treatment on drinking behavior was his devising a means (p. 533) for judging objectively whether the drinking outcome results of a new study are normal or especially worthy of note as being uncommonly high or low.

That Emrick was able to make sense of a highly varied mass of data contained in 265 reports is a tribute to his ingenious and patient re-analysis and meta-analysis of the reports, not to the methodological excellence or homogeneity of the reports themselves. The fact that his analysis had to be undertaken almost from the ground up is further testimony to the need for alcoholism researchers to be able to identify and use research models which yield easily understood results.

The following quotation from the introduction to a careful,

exhaustive, and controversial research report, The Course of Alcoholism (Polich, Armor, & Braiker, 1981), popularly known as "the second Rand Report," can serve as a summary of the support reviewed literature provides for the need for productive models of research.

A great deal of the existing evidence on the extent of change in alcoholism comes from follow-up studies of clinical populations. This literature, which is now so extensive that the careful review of it has become almost a profession in itself, does not readily lend itself to generalization. . . . Comparison across studies is perilous






35


because of the lack of explicit measurement procedures and the frequent use of broad clinical judgments of what is "improved," "unimproved," etc. Added to this are the numerous methodological deficiencies to which applied research is usually subject, such as the use of convenient samples rather than random ones, failure to measure all sample members at follow-up, and lack of randomization or
statistical control in analysis. (pp. 7-8)


Need to Identify Theoretical Stances Associated with Specific Treatment Outcomes

It is the objective of this section to demonstrate that there is support in the literature for the notion that there is a need to identify those theoretical positions which appeared to contribute to desirable or undesirable treatment outcomes. In order to accomplish

this, an attempt is made below to show that the literature supports the premises that theoretical stance is an important variable in determining

treatment outcome and that too little is known about the relationship between the theoretical positions of treatment agents and the outcomes of treatment.

From the bare fact of their having published a lengthy book entitled Emerging Concepts of Alcohol Dependence, Pattison, Sobell, and Sobell (1977) imply their strongly held belief that, in the field of alcoholism treatment, theory makes a difference. They specifically

assert that the traditional model of alcoholism has dictated much of the

design of alcoholism treatment services, determining that treatment should make alcoholics "aware of their permanent physiological

abnormality and the necessity for them to be permanently abstinent" (p. 3). Furthermore, they claim, the traditional theory contains the implication that treatment should have as its main concern those problems which result from learning to live without alcohol.





36



Linsky (1972) states that in the case of deviant behavior such as alcoholism the theories of etiology will govern the strategies of intervention. This viewpoint is somewhat- parallel to the concern of Pattison (1979) that while progress in therapeutic methods depends upon the development of new models of the illness, alcoholism models have been loosely structured and their associated epidemiological definitions ill-formed.

While it certainly appears to be the case that theory may well be expected to have a potent influence on the processes and results of treatment, the present author has not found the theoretical stance of treatment personnel considered as a weighty variable in the reviews of Baekeland (1977), Costello (1975 a, 1975 b), Emrick (1973, 1974, 1975, 1979), or Hill and Blane (1967) cited earlier as major surveys of the literature on treatment outcomes. This may be because the studies examined did not report clearly on this aspect. In any case, the

absence of systematic consideration of theoretical posture as a variable suggests the need for such consideration. Need to Identify Combinations of Client and Treatment Variables Predictive of Positive Outcomes

In this section, the author will set forth support from the literature for five propositions. First, alcoholism clients vary on several dimensions. Second, treatments for alcoholics vary in a number of ways. Third, there is variation in treatment outcomes. Fourth, some kinds of clients do better in some kinds of treatment than in others. Fifth, to work toward means of matching client and treatment is reasonable and needful.

Emrick (1973, p. 23) shows that researchers have reported data on at least the following list of patient variables: age, term of problem





37


drinking, gender, marital status, employment status, level of vocational

skill, socioeconomic status, arrest record, prior treatment, type of alcohol problem, and psychiatric diagnosis. All of these items were noted by Emrick to have been thought by at least one researcher to relate in some way to alcoholism treatment outcome. Polich et al. (1981, p. 132) examined relationships between outcomes and the following classes of subject characteristics at admission: level of alcohol dependence symptoms, age, social stability, socioeconomic status, previous alcoholism treatment, and ethnicity. Pattison (1979, pp. 137-139) asserted that alcoholic populations may vary in personality structure, social class, gender, and ethnicity, all of which variables should be examined for interactions with treatment variables and outcome variables. He also noted (p. 205) that alcoholics entering treatment vary in terms of areas of disability, degrees of impairment, potentialities for change, and individual preferences regarding goals and methods of treatment.

There is variability in the alcoholism syndrome itself, and hence in the manifestations of that syndrome in the individuals who present for treatment. Pattison (1979, pp. 134-137) points out that there appear to be multiple subtypes of alcoholism composed of complex sets of

drinking, social, and personality variables. Moreover, the course of the syndrome is variable. Individuals may move in and out of

symptomatic drinking or may exhibit a linear, progressive worsening of problem drinking. It is possible for the severity of problematic drinking to remain constant and nonprogressive. The processes of remission or progressive deterioration may significantly vary with time,

place, and circumstance. Pattison held that for optimal matching of





38


treatment to client, some or all of these complex sets of variables must be taken into account.

Given that alcoholism is a highly variable phenomenon and that alcoholics vary widely on several dimensions, it should come as no surprise that treatment for alcoholism is also varied. Polich et al. (1981, p. 135) point out that whereas "treatment" in an experimental study may be expected to be well-defined, unitary, and explicitly bounded in time, actual clinical treatment in its naturalistic environment has none of these characteristics. Treatment varies as to duration; Emrick (1973) recorded outpatient treatment ranging from four sessions to ten years and inpatient treatment varying between one day and three months. There are, of course, inpatient versus outpatient treatment settings to be taken into account (Baekeland, 1977; Pattison, 1979). Treatment varies in terms of coerced versus voluntary, the stage

of alcoholism at which clinical intervention is made, and the behaviors targeted as treatment goals (Pattison, 1979, pp. 155-156).

There are variations in the theoretical postures and ideologies held by treatment programs and differences in treatment personnel as to degree and kind of preparation, ideological position, preferred techniques, and prior personal experience of alcoholism (Pattison, 1979, pp. 164-169). Individual differences among personnel contribute to differences in definition of treatment goals including predilections for abstinence, "social drinking," "attenuated drinking," "controlled drinking," and "normal drinking" (Pattison, pp. 194-202).

Techniques for the treatment of alcoholism are numerous. Pattison (1979, pp. 156-164) discusses specific and nonspecific individual psychotherapies, group psychotherapies, family therapies, behavioral





39


therapies, and four categories of drug therapies. Emrick (1973, p. 17) found in use "a staggering array" of therapies including but not exhausted by "psychodrama; self-confrontation through videotape; conditioned aversion by electric shock, emetics and muscular paralysis; systematic desensitization; antidepressants, tranquilizers and antipsychotic drugs used in conjunction with psychotherapy; deterrent drugs (disulfiram, metronidazole, calcium cyanamide); hallucinogenic drugs; insulin shock; and individual and group insight-oriented analysis." He noted (p. 19) that these approaches were used, singly and in

combination, and for varying terms, in inpatient, outpatient, halfway house, industrial, and prison settings.

Since alcoholism patients and treatments vary, it is to be expected

that the outcomes of treatment vary as well, and this is the case. As Pattison (1979, p. 132) has observed, "different types of alcoholics present themselves at different facilities, receive distinctly different

treatments, and achieve different treatment outcomes." Polich et al. (1981, p. 101) studied treatment outcomes in terms of eight separate psychosocial variables and observed (p. 8) that the outcomes of treatment vary significantly along several continua for judging improvement.

So varied are the outcomes of treatment, or at least the variables examined to measure outcome, that Emrick (1974, p. 526) found that his best effort to collapse variables into categories that might be at once parsimonious and inclusive resulted in no less than 19 clusters of outcome variables. Drinking behavior was the rubric for only one of these criterion clusters, and Emrick, in order to analyze data produced under varying classifications of drinking outcomes, formed nine subcategories under that major heading. Emrick (1973, pp. 80-81) found a






40


number of statistically significant differences ( p. 05 or better) on the drinking outcome variables in studies reviewed by him.

The profusion of therapeutic techniques and approaches implies a widely held belief that some treatments are better than others, or, put more cautiously, that some alcoholics do better in some kinds of treatment than in others. Emrick (1975) reviewed 384 studies of

psychologically oriented treatment of alcoholics and judged that it was possible that differences in treatment outcome may have been due to a specific treatment having unusually beneficial and perdurable effects (p. 94), although the possibility could not be documented because the studies reviewed did not control for the negative experiences of patients in treatment.

Smart (1978) studied 1091 alcoholics from seven treatment

facilities, for each of whom full intake, treatment, and followup information was available. His purpose was to discover whether some alcoholics do better in some types of treatment than others. He concuded that:

The data indicate that patient characteristics are most important in predicting outcome, that treatment is relatively unimportant and that interactions between the two are
uncommon. (Smart, 1978, p. 74)

Gibbs and Flanagan (1977) conducted a meta-analysis on a much larger body of data than that considered by Smart and reached quite different conclusions. They analyzed 45 predictive studies published between 1937 and 1974 involving 55 different treatment groups including some 11,350 subjects in an attempt to isolate personal characteristics of alcoholics associated with prognosis. They concluded that while some

patient characteristics are of greater general predictive value than others, stable predictors across studies were not apparent. One






41


possible reason for the absence of predictor stability, they held, is that certain treatments may be more effective with certain alcoholics.

Pattison (1979) set forth a more positive position than either Smart or Gibbs and Flanagan, noting that persons with certain

psychosocial profiles are apt to affiliate successfully with AA (p. 154). Pattison advocated the notion that the interaction of specific client variables and certain treatment variables is at least partially determinative for outcomes. Convinced of this, he argued for the need for matching client and treatment for maximum positive treatment outcomes. He maintained that the most powerful predictor of successful treatment outcomes is the matching of therapist-client values and goals (p. 169) and provided in vignette form illustrations of effective matching in four quite different treatment settings: the

aversion-conditioning hospital, the alcoholism outpatient clinic, the alcoholism halfway house, and the police farm work center. He found

that each of the four facilities tended to draw clients from different subpopulations of the overall alcoholic population and that each of those subpopulations held distinguishably different definitions of alcoholism, self-defined their treatment goals differently, and received different treatment from different personnel.

It appears, then, that the development of schemata for matching client and treatment would be a useful enterprise. Even although he was

pessimistic about the possibility of discovering a "best" treatment for alcoholism and asserted that technique variables are not likely powerful

determinants of long-term outcomes of treatment, Emrick (1975, p. 95) suggested that treatment personnel should seek to match each alcoholic with the treatment setting and approach which would provide the best fit






42


with that individual's own perspectives on the nature, etiology, and treatment of alcoholism.

Pattison (1979) was more sanguine and forceful than Emrick on the subject of matching. He held the matching of client, facility,

personnel, and treatment to be a goal worthy of expenditure of great effort. He pointed out that attaining the goal would be no small order, since most reviewers of treatment methods hold that no clear, large-scale indicators for client-treatment matching are easily derivable from the data. He added, however, that such a conclusion was

not surprising in light of the fact that treatment methods have been rather indiscriminantly applied, so that clear matching criteria could scarcely be derived from global reviews.

Pattison (1979, p. 132) cited examples, however, of smaller-scale discrete research projects which demonstrated the usefulness of constructing means for matching client and treatment (Kissin, Platz, & Su, 1970; McLachlan, 1974; Pattison, Coe, & Rhodes, 1969; Pattison, Coe, & Doerr, 1973; Trice, Roman, & Belasco, 1969). Pattison's (1979)

summaries of their salient points follow.

The finding of Pattison et al. (1969, 1973) that alcoholics of different types enter different treatment centers, get different treatment, and manifest different treatment results suggests that some unnoticed matching by self-selection occurs naturally. Similar

discoveries in other research projects have allowed the formulation of some predictors for matching (Kissin et al., 1970; Trice et al., 1969). McLachlan (1974) provided a clear illustration of the potential utility of matching efforts in his report that 77% of the patients matched to both the therapy and aftercare environments met recovery criteria, while






43


those matched only to either the therapy or aftercare environment had recovery rates of 65% and 61%, and of those mismatched to both environments, only 38% were classified as recovered.

Pattison et al. (1969), reporting on clientele of three different alcoholism treatment facilities, found that the treatment-successful cases at each facility differed substantially from one another. They constituted unique, differentiated subpopulations. The subpopulation of

each facility improved, but the improvement was in different patterns. In a later study of the characteristics of populations presenting prior to treatment at four different treatment centers, Pattison et al. (1973)

found that the subpopulations were not merely the results of the treatment undergone, but preexisted as distinct subpopulations upon their entering the various facilities.

Of the writers reviewed by the present author, E.M. Pattison is the

strongest advocate of the opinion that the need for workable ways of matching client and treatment is sufficiently great to merit investment of time and energy. Citing as evidence large-scale studies by Bromet, Moos, Bliss, and Wuthman (1977) and Cronkite and Moos (1978), Pattison

(1979) concluded that the selection of treatment method is more crucial to treatment success than had previously been realized. The following quotation from Pattison's 1979 article on selection of treatment modalities can serve well as a summary and conclusion of this portion of the current writing:

Our current state of knowledge about treatment is still too global and imprecise to formulate exact treatment guidelines. Our measurement and evaluation methods are too crude to assess our methods accurately. Nevertheless, the
accumulation of research to date does strongly suggest the value of matching subpopulations of alcoholics with the most
appropriate facilities, methods, and treatment personnel.
(p. 205)





-+4


Need to Identify Pragmatically Tested Theories and Practices For Inclusion In Training


The present author was unable to find in the literature reviewed explicit statements to the effect that there is a need to identify, for inclusion in the training of alcoholism workers, those theories and practices which have been demonstrated to be most effective. Hence, the following paragraphs represent inferred support for the proposition that such a need exists.

There can be no doubt that persons are being trained in the diagnosis and treatment of alcoholism. The August 1981 current

literature issue of Journal of Studies on Alcohol lists no fewer than thirteen recent titles related to the training of professionals or other alcoholism workers ("Current literature," pp. 565-567).

The NIAAA 1981 report to Congress (DeLuca) states: "Persons

involved in treating alcoholics need specific training, education, and experience for their jobs. This requirement applies to every level of the work force, and should be enforced through competency-based evaluation systems" (p. 185). The same report stresses the unmet need

for adequate training and notes the development of Project Cork, an alcoholism-specific curriculum initiated by Dartmouth Medical School, mentioning with apparent approval that "Project leaders expect that upon graduation students will not only know about alcoholism but will be able to do something about it" (p. 186). The present writer contends that

for students of alcoholism to be able to "do something about it" it is necessary that they be given the best possible information about what seems to work.






45


Camp and Kurtz (in press) are cited in the NIAAA report as having defined six areas of job competency for alcoholism counselors, including "(1) communication; (2) knowledge of alcohol use, prevention of alcoholism, treatment, and rehabilitation; (3) evaluation and assessment; (4) planning; (5) information and referral; and (6) counseling and treatment" (p. 187). It is important to note that the latter five of the six imply a need for training with incorporates the knowledge which the present author maintains is needed.

The attitudes of alcoholism workers are vitally important, and those attitudes are in part determined by the beliefs workers hold about alcoholism and alcoholics. Wolf, Chafetz, Blane, and Hill (1970)

studied the attitudes of 15 physicians toward their alcoholic patients and found that the doctors tended to view alcoholism as a disorder of derelicts and to be much more hesitant to diagnose alcoholism in socially intact persons as compared to derelict patients (p. 132). The researchers also held that their physician sample preferred a strictly medical diagnosis of alcoholism to one that included dysfunction in the psychosocial sphere (p. 134). Given that both attitudes are substantially at odds with widely accepted and fairly well supported views of alcoholism, it is apparent that this sample of physicians held attitudes inappropriate to alcoholism workers. The need for inculcating

more realistic and appropriate attitudes is apparent. The present

author submits that if strongly supportive evidence of the real-world effectiveness of identified theories and practices were included in the training of scientist-practitioners such as physicians, it would result in attitudes and behaviors more conducive to accurate diagnosis and efficacious treatment.






46


In the earlier discussion of Pattison's strong argument for careful

attention to client-treatment matching (1979), it was apparent that in order for such matching to be done well, the "matcher" would need to possess considerable knowledge about which combinations would be likely to produce what results. There is again a clear inference of the

reality of the training need here under consideration.

The present writer examined the indices and tables of contents and extensively sampled the texts of four recent works containing materials intended for use in training alcoholism workers (Johnson, 1973; Poley, Lea, & Vibe, 1979; Schuckitt, 1979; Zimberg, Wallace, & Blume, 1978). While each work discussed theories and practices of alcoholism and alcoholism treatment, none was found to include information concerning the broad-based support or lack of support for the relative worth, in terms of treatment outcomes, of the theories and techniques discussed. This writer assumes that this lack does not represent oversight on the part of the several authors and editors of these books, but rather points to the fact that the needed knowledge was not available to them.

In light of the present author's earlier argument that there is no widely acceptable synthesis of knowledge about the interrelationships of

alcoholism clients, treatments (including theoretical stances thereof), and outcomes, the immediately preceding paragraphs seem almost superfluous. Since the knowledge in question has not been synthesized, it is not being included in training although it is clearly needed.

Support for the Approach of the Study

The approach of the current study is descriptive with an emphasis on pragmatic utility of findings. Isaac and Michael (1971, p. 18) state

that the purpose of descriptive research is "to describe systematically






47


the facts and characteristics of a given population or area of interest,

factually and accurately." This statement fits well with the current author's intent to fulfill the need for 'a factual and accurate descriptive synthesis of what has been learned in research on outcomes of psychologically oriented alcoholism treatment.

Somewhat similar descriptive studies in the same general subject area have been undertaken by other authors (Baekeland, 1977; Costello, 1975 a, b; Emrick, 1973, 1974, 1975, 1979; Hill & Blane, 1967). In each

case it was apparent or plainly stated (e.g., Emrick, 1973, pp. 13-14; Hill & Blane, p. 76) that the number and diversity of research reports practically dictated that descriptive, synthetic studies should be undertaken. Baekeland (p. 386) saw that firm conclusions about the indications for and relative efficacy of alcoholism treatment methods were sorely needed, and part of his response to that need was to produce a critical review of the relevant literature.

Further support for the appropriateness of descriptive, synthetic studies such as this current one is found in a discussion by G.V. Glass (1976) of the worth of what he termed "meta-analysis." He used the term

to refer to "the statistical analysis of a large collection of analysis results from individual studies for the purpose of integrating the findings" (p. 3). While Glass's frame of reference was the field of educational research, his argument for meta-analysis is applicable to alcoholism research as well. Glass asserted that in questions of

outcome research, since findings are apt to vary in such confusing fashion across context, populations, and unnumbered other factors that commonality of results is rare, it is especially important that efforts be made to organize findings across studies in a way that can bring





48


order out of the confusion. Glass contradistinguished information and knowledge, characterizing the latter as organized, integrated, and, therefore, useful information. He wrote:

Our problem is to find the knowledge in the information. We need methods for the orderly summarization of studies so that knowledge can be extracted from the myriad individual
researches. (p. 4)


As has been mentioned above, the approach of the current study includes an element of pragmatism, a primary interest in the possible and the workable. Hill and Blane (1967, p. 94) in critiquing the methodology of alcoholism psychotherapy research noted that in some instances research was weakened by the researchers' having undertaken research tasks beyond their resources, financial, temporal, or human, and that those researchers would have done better to design their studies to fit within their limitations. Hence, while the current

author recognizes the need for a much more exhaustive study than the present one, he accepts the limits of practicality that Hill and Blane imply.

A second facet of the pragmatic aspect of the current approach has to do with the quest for accurate information presented in a manner that

will be understandable to those who may put it to use. "Those who may put it to use" refers to alcoholism workers, most of whom do not hold graduate degrees (DeLuca, 1981, p. 184; Pattison, 1979, p. 165). The implication is clear that for the knowledge sought in this study to be utilized by alcoholism workers it must be presented in a manner understandable to an audience not familiar with sophisticated statistical analyses.






49



Summary

The preceding review of literature has shown that there is support in the literature for the problem to which the current study is addressed, the needs which constitute that problem, and the approach of the study to fulfilling those needs. The support found in the

literature was sometimes implicit and sometimes direct.













CHAPTER III
METHODOLOGY


Overview of the Study

In the context of a growing body of empirical research on alcoholism treatment here appears to be a need for synthesis of those findings which have emerged. The need for a summary seems apparent from

the facts that while substantial human and monetary resources are being expended on the psychological treatment of large numbers of alcoholics, generalizable conclusions about the outcomes of that treatment are hard to come by. Hence, the purposes of this study are to identify promising

approaches to and models for outcome research as well as to describe criteria for selecting theories related to positive treatment outcomes, criteria for matching client and treatment, and criteria for selection of theories and practices for inclusion in training of alcoholism workers.

The rationale for the descriptive approach taken herein is that such an approach appears to be the necessary one in order to make sense of a bewildering mass of empirical findings.

The present study seeks answers to one general research question and six more specific ones. The general question is: "How have the

outcomes of psychologically oriented alcoholism treatment been studied and what has been learned?" The specific research questions are: "What

client, treatment, and outcome variables have been studied?" "To what extent have theoretical stances of treatment programs been taken into account as variables contributing to outcome?" "What are the ways in


50






51


which client, treatment, and outcome variables have been measured?" "What research models have been used?" "What research models have

yielded significant results?" "What interactions of client, treatment, and outcome variables have been observed?"

Delineation of Variables

The variables under consideration here fall into three basic categories: subject, treatment, and outcome.

The category "subject variables" is composed of the set of facts reported for the individuals described as clients, patients, or subjects in the various research reports considered herein. Generally, "subject

variable" herein refers to some fact about the subjects which pertained at or prior to their admissions to treatment. For convenience in data analyses, subject variables have been clustered in clusters parallel to those used in grouping outcome variables. Thus, in the following

discussion of outcome variables, the insertion or substitution of the notion "at or prior to admission to treatment" as appropriate will translate the respective outcome variables into subject variables.

Emrick (1973, pp. 31-37) found that the wide variety of outcome criteria in treatment outcome studies could logically be collapsed into 19 categories. Emrick's 19 criteria clusters are used here to delineate outcome variables, and, mutatis mutandis, subject variables. 1 The

following list, taken directly from Emrick's use of outcome criteria table, first names the outcome variable in question and then specifies the content of that variable.



1The author gratefully acknowledges Emrick's contribution in developing these clusters.






52


1. Drinking amount or frequency: amount of drinking,
frequency of drinking.

2. Affective-cognitive: psychological test results,
thoughts about treatment, cognitive-emotional
responses to drugs such as LSD and tranquilizers, clinical ratings about patients' emotional
responses and cognitive functioning.

3. Work situation: work status, work adjustment.

4. Further inpatient treatment: rehospitalization or
hospitalization after outpatient treatment, the hospitalization being for consequences directly
related to excessive drinking.

5. Home situation: relationships: marital status,
marital adjustment, adjustment to family life.

6. Physical: hospitalization and outpatient treatment
for physical problems, physical reactions to drug treatment such as LSD, disulfiram and scoline
(a muscle paralyzer), physical reactions to alcohol after treatment (e.g., blackouts, hangovers, DT's).

7. Arrests and other legal problems: arrests, illegal
activities, imprisonment.

8. Mixed: a mixture of two or more of the above criteria,
the mixture not intended to cover all areas of functioning.

9. Social situation: interpersonal adjustment not specifically related to home or work.

10. Further treatment: outpatient: AA: AA attendance
occurring after treatment and not during aftercare for more than 50% of the patients, assuming of course that AA was not the only mode of treatment
evaluated.

11. Financial situation: amount of income, degree of
responsibility for financial affairs, welfare status.

12. Global adjustment: overall functioning with no one
aspect being of paramount importance and with no
aspect being purposefully excluded.

13. Miscellaneous: criteria fitting none of the other
clusters, such as "day-to-day functioning" and
"cigarette consumption."






53


14. Drinking behavior other than amount or frequency:
change in drinking habits such as locus of drinking, time of day or week drinking is done, the type of beverage drunk, and amount of time between sips of
alcohol.

15. Home situation: residence: residential mobility,
quality of accommodations.

16. Further treatment: outpatient: other: outpatient
treatment during aftercare for less than 50% of the patients, the treatment taking place in another
setting with different personnel.

17. Use of leisure time: quality of leisure-time activities.

18. Religious life: quality and quantity of religious life.

19. Further treatment: locus unclear: vague references to
"other treatment." (Emrick, 1973, pp. 31-33)


The category "treatment variables" consists of the set of relevant facts reported in the several studies examined herein as to the length of treatment and the setting, procedures, personnel and/or theoretical stance of the facilities to which subjects were admitted.

The smaller category of treatment variables designated "setting" includes the geographic location of the facility, the locus of treatment

(inpatient versus outpatient,) and the type of facility. Some "types" of facilities are VA hospital, public outpatient clinic, private

aversive-conditioning hospital, halfway house, and state psychiatric hospital.

A second subset of treatment variables has to do with treatment procedures. Treatment procedures described in the reports under consideration here are widely varied and include, but are not restricted

to, self-help groups, family counseling, relaxation therapy, vocational counseling, assertiveness training, aversive conditioning, milieu therapy, occupational therapy, and psychodrama.





54


Inclusion of "personnel" as an aspect of treatment variables allows

the discussion of the training, attitudes, and behavior of the persons administering treatment as variables which' may contribute to outcomes.

Since "theoretical stance" is not easily included under either setting or personnel without ambiguity, it is considered to be a separable variable including explicit or implicit assumptions and

beliefs concerning alcoholism and alcoholic persons insofar as those assumptions and beliefs contribute to the goals, procedures, and outcomes of alcoholism treatment.

Description of the Population

The population of interest for this study is the universe composed of English-language articles on alcoholism published between 1973 and late 1981. For all languages, the population appears to be distributed through more than 1,000 periodicals, with approximately 50% of the articles appearing in English (Moll & Narin, 1977, p. 2177). In 1973 the population numbered more than 2,300 publications for that year (Moll & Narin, p. 2167). The articles which composed this population exhibit a broad range of interests and approaches. Moll and Narin characterized the foci of interest as biomedical, biosocial, and psychosocial. The approaches taken range from hortatory essay to carefully controlled experimental study.

Sampling Procedures

Five criteria were chosen for including publications in this study.

Firstly, publications must have been published research reports on psychologically oriented alcoholism treatment. This criterion was used to exclude works not germane to the intent of the study and, by the requirement "published," to exclude unpublished works which would have






55


been impossible to obtain or not sufficiently well done to merit publication. Secondly, the publications must have examined the outcomes of such treatment. This criterion was used to preserve the pragmatic, i.e., outcome-oriented, focus of the present study.

Thirdly, the reports must have considered no fewer than ten subjects. This criterion was used in order to have the data pool include only studies with relatively more generalizable findings. Fourthly,

reports must have reported on outcome measures taken no less than 12 months after treatment. This criterion was used to include only studies with greater reliability and validity. Other researchers (e.g., Costello, 1975 a, b; Polich et al., 1981) have maintained that

alcoholism treatment outcomes measured at less than one year after treatment are substantially less reliable than those measured at one year or more after treatment. Fifthly, the studies must have been in English and available to the author. This criterion was used in order to keep the present study within the bounds of realistic limitations.

The first step in selecting the sample was to procure

computer-assisted literature search printouts from NIAAA, NIMH, and Psych Info, using the nearest available approximations in each system to the descriptor-set: "the intersection of the sets 'alcoholism treatment, psychological, outcome, English language, 1973-1981.'" The

computer-assisted literature searches resulted in a pool of approximately 430 titles. The titles and abstracts in each printout were examined and articles whch did not meet the inclusion criteria were excluded from consideration.

Exclusion per the criteria listed above reduced the pool of documents appearing to be appropriate for inclusion herein to 21. No actual





56


count of articles failing to meet each criterion was taken, but the author's impression is that the rank order of frequency of reason for exclusion, from most frequent to least frequent, was as follows: were not research reports on psychologically oriented alcoholism treatment, were not outcome studies, had followup of less than one year, had fewer than ten subjects, were not available in English to the present author. The vast majority of documents excluded were rejected by the application of the first three criteria as rank-ordered above. No single-subject intensive design studies were found in either the computer-assisted literature searches or the manual literature search. Fewer than ten

studies were rejected for having fewer than ten subjects.

Original or photostatic copies of the articles appearing to meet the inclusion criteria were obtained and examined, and those which failed, on close inspection, to meet the criteria were excluded.

Further manual literature search included inspection of the reference lists in articles examined and inspection of the indices of Journal of Studies on Alcohol. This search increased the number of documents to be included to 34.

Description of Sample

The resultant sample consisted of 34 English-language articles reporting on psychologically oriented alcoholism treatment outcome research with outcome measures taken a minimum of 12 months posttreatment on a minimum of ten subjects.

Compared to the population from which it was drawn, the sample is smaller, restricted to the psychosocial sector, and further restricted by its being circumscribed by the limiting criteria mentioned above.






57


Description of Research Design

The design of this study is analagous to that of a descriptive study with human subjects. In the case at hand, the articles comprising the sample are analagous to subjects. The method of obtaining data is a set of questions asked about each article. The answers to those

questions constituted the data to be analyzed. The data were inspected and analyzed to ascertain commonalities and differences among and between studies. Conclusions were then drawn regarding the answerability of and answers to the research questions.

Description of Research Procedures

The above-mentioned set of questions was developed by reading sample articles and listing the subject, treatment and outcome variables mentioned in each. The lists so produced were then combined into lists of categories subsuming the specific variables mentioned into larger classes characterized by face validity and apparent manageability. The larger classes in turn provided the categories of questions to be asked of the articles in the sample.

The author and another judge, a Ph.D. psychologist with published research, administered the resulting data-gathering protocol independently to the same five articles from the sample. Their recorded

answers were compared and yielded an inter-judge agreement rate of 90.5%.2

Description of Data Analysis Procedures

When the data had been gathered by means of the data-gathering protocol, they were arranged in 13 tables. The tables were devised so


1
2See Appendix A.
See Appendix B.






58


that the research

The


data were arranged in ways designed to provide answers to the questions.

13 tables are as follow:

Table 2: Descriptions of Selected Outcome Studies Published

1973-1981;

Table 3: Numbers of Studies Reporting on Various Subject Variables;

Table 4: Numbers of Studies Reporting on Various Outcome Variables;

Table 5: Numbers of Studies Reporting on Various Treatment Variables;

Table 6: Major Findings with Statistical Significance of

p--,05 or Better in Studies Utilizing Correlational

Research Models;

Table 7: Major Findings with Statistical Significance of pc.05 or Better in Studies Utilizing Experimental

Research Models;

Table 8: Major Findings with Statistical Significance of P<.05 or Better in Studies Utilizing

Quasi-experimental Research Models;

Table 9: Major Findings with Practical but not Statistical Significance in Studies Utilizing Correlational

Research Models;

Table 10: Major Findings with Practical but not Statistical Significance in Studies Utilizing Descriptive

Research Models;






59


Table 11: Major Findings with Practical but not Statistical Significance in Studies Utilizing Experimental Research Models;

Table 12: Major Findings with Practical but not Statistical Significance in Studies Utilizing Quasi-experimental Research Models;

Table 13: Major Research Questions of Outcome Studies of Psychologically Oriented Treatment; Table 14: Major Interactions Observed.

The data were then analyzed by inspection to ascertain and describe patterns and trends.

Methodological Limitations

The present study is limited by the inclusion criteria to a small sample of the literature on alcoholism treatment. The longitudinal

limitation is circumscribed by the years 1973-1981, a limit chosen because somewhat similar meta-investigations had already been conducted on the relevant literature prior to 1973. The latitudinal limitations are those implied in the inclusion-exclusion criteria set forth above.

The fact that the data under consideration herein are published studies constitutes a methodological limitation. Published studies must meet the restrictions of space of the publications in which they appear.

It is likely the case that explanations or data which might be of value to the present undertaking have been excluded from published studies in order to keep within the prescribed limits of length. An example of

that limitation is that few of the studies examined in the present writing set out data for individual subjects; most of the reporting is in terms of group means or other per-group data. Thus, the method of





60



the present work does not allow for the analysis of similarities or differences on an individualized basis.

Other methodological limitations are-discussed in Chapter I above under the rubric "Rationale for the Approach."












CHAPTER IV
RESULTS



This chapter presents data drawn from 34 studies selected for inclusion. The data are set forth in tables with accompanying explanatory text.

Description of Studies

Table 2, "Description of Selected Outcome Studies Published 1973-1981," should be read with the following explanatory statements in mind.

The data grouped under the heading "Drinking outcome" are organized

to indicate which groups of subjects were categorized under drinking outcome classes established by Emrick (1974) and whether outcomes reported were uncommon by Emrick's guidelines.

Emrick's criteria for evaluating the uncommonness of results were established by his calculating the mean and median estimates of outcomes

for a large number of studies. He selected the points one standard deviation above and one standard deviation below the mean as the points outside which outcome percentages might be judged atypical. This

produced the following guidelines:

Class 1 (abstinent) 10.5%-53.3%
Class 2 (abstinent-or-controlled) 19.8%-67.0%
Class 3 (much-improved) 4.7%-26.3%
Class 4 (somewhat-improved) 0.9%-34.9%
Class 5 (much-or-somewhat-improved) 14.4%-44.4%
Class 6 (total improved) 47.8%-84.2%
Class 7 (total unimproved) 15.8%-52.2%
Class 8 (deteriorated) 0.2%-20.6%


Emrick noted that the range of one standard deviation on either side of the mean appeared to be a reasonable range of typicality for all 61





62


outcome categories since, except for the controlled-drinking outcome, estimates appeared to be normally distributed within groups. The

non-normal distribution within groups of controlled-drinking outcomes appeared to make it impossible for him to derive a reasonable range of typicality for that outcome. Hence, controlled drinking is not listed as a separate outcome class, although it is included in Class 2 (abstinent-or-controlled).

In interpreting Table 2 entries under "Drinking outcome" the reader

should note that the "group" column contains Roman numerals to indicate to which of the treatment groups described in the "Treatment type" column reference is made. In the column headed "class/%" the first Arabic numeral indicates into which of Emrick's drinking outcome classes the reported outcome was categorized. The second number is the within-group percentage of subjects evaluated who fell into the drinking outcome class in question. The column farthest right contains a plus-, minus-, or zero-sign to indicate whether the drinking outcome in question was, respectively, uncommonly good, uncommonly poor, or not uncommon per Emrick's guidelines.

With the preceding clarification in mind, then, one would read the first entry in the "Drinking outcome" columns of Table 2 as meaning that

25% of the subjects who had received VA hospital care only (group I) were included in the abstinence drinking outcome category (class 1), and that the outcome was neither uncommonly good nor uncommonly poor.

Three abbreviations in Table 2 may require explanation. "BAC"

refers to blood alcohol content, a measure of the ratio of ethanol to blood in the body. "NR" means "not reported" and that the relevant data were not reported. When "NR" appears in the "Drinking outcomes" columns






63



it means either that drinking outcomes were not reported or that within-group percentages could not be derived from the data reported. "Ss" is an abbreviation for subjects.














Table 2
Description of Selected Outcane Studies Published 1973-1981


Mean Drinking outcome
treatment Time of N unduration evalua- Sanple size evalu- ccaMonStudy Date Treatment type (if given) Outcome criteria tion and cimosition ated group class/% ness


1973 VA hospital care:
I. Standard treatment
II. Same as I plus group
aftercare


1973 State hospital, inpatient:
I. Group therapy; chenotherapy; AA; with abstinence as goal
II. Same as I plus individualized behavior
therapy
III. Same as I but with
controlled drinking as goal
IV. Same as II but with
controlled drinking
as goal


60 days inpatient
plus
aftercare


Drinking behavior and its consequences; family, social, and work adjustments


40 days Daily drinking
(inferred) disposition,
general emotional adjustment, vocational satisfaction, occupational
status, residential status/
stability


12 months posthospital



12 months posthospital


156 male alcoho- 122 lics


70 male alcoholics who volunteered for experimental treatment in a state hospital; all Gammatypea


69


aGanna-type alcoholism is "alcohol addition" characterized by acquired increased tissue tolerance to alcohol, adaptive cell metabolism, physical dependence, and loss of control (Jellinek, 1960, p. 37).


Pokorny, Miller, Kanas, &
Valles


Sobell & Sobell


I 1 25
II 1 53


0
+


0
+

+


I
II III
IV


2
2
2
2


27 87
32 85











Table 2-continued.


Willens, Letenendia, & Arroyave


1973 Hospital, inpatient:
problem solving groups,
AA:
I. Short stay II. Long stay


Tcmsovic 1974 VA hospital, inpatient : detoxication; educational lectures, tapes, and
movies; group therapy;
relaxation therapy;
physical fitness training; same individual
counseling; AA


1975 State hospital:
I. Detoxication and
aftercare planning
II. Detoxication plus
25 days intensive,
inpatient, psychosocial treatment,
milieu therapy including group psychotherapy, AA, discussion groups, recreational therapy,
and didactic lectures


I. 20 days Drinking behavior: II. 80 days abstinent, improved but drinking, unimproved


82 days Drinking behavior and
social adjustment


I. 9 days Drinking behavior; social
relationships;
II. 30 financial status;
days legal involvement;
use of ccumunity
resources


12 & 24 69 male, British, 62 months Gamna-type alcopost-dis- holics, mostly charge middle class (Drinking
outcomes listed here are at 24
months)


12 months post
discharge


12 months postdischarge


381 male alcoho- 179 lics, mean age 46, mean years education 11.5, all of average intelligence. Group I were
"binge" drinkers. Group II were continuous drinkers.


58 male alcoholics from one Wisconsin county, mean age 42, mean
years education 11, with less than five previous alcoholism admissions


I 1 39 o
II 154 +

I 6 68 o
II 6 67 o


I 1 28 o
II 1 31 o


I 337 II 3 27


58 I
II


+
+


1 37 o
1 34 o


I 2 37 o
II 2 41 o


I and II canbined


1 36 2 39


0
0


Stein, Newton,
& Bowman


ul
Mr













Table 2-continued.



Mean Drinking outcome
treatment Time of N unduration evalua- Sanple size evaTu- ocnTnStudy Date Treabtent type (if given) outcome criteria - tion and composition ated group class/% ness


Vogler, Capton, &
Weissbach


Vogler, Ferstl, Kraemer, & Brengelmann


1975 State hospital, inpatient:
I. Integrated behavior
change therapy including behavior counseling with followup
and booster sessions
II. Same as I with the
addition of baseline
alcohol drinking,
videotape playback of
drunken comportment,
blood alcohol content (BAC) discrimination
training, electric shock aversive conditioning, and avoidance practice


1975 German hospital, inpatient:
I. Standard hospital
treatment:
detoxication; group
therapy; improved diet; physical exercise
II. Sane as I plus
drinking-contingent electric shock conditioninq


Mean NR median 30 days for all Ss; median for group I: 22 days; median for group II: 45 days


Standard hospital
treatment duration was not reported


Drinking behavior: abstinence; controlled drinking; relapse; choice of beverage, drinking companions, and drinking environment


12 months
posthospital


Drinking behavior: 12 months abstinence posthospital


42 (37 male) hos- 42 pitalized chronic alcoholics, 21-55
years old, in sufficiently good health to ingest up to 16 oz. of
86 proof alcohol without innediate threat to health,
not diagnosed as psychotic or having organic brain
syndrome, no history of felonious assault, not taking
contraindicated medications

Experimental sub- 59 jects were 67 German chronic alcoholics who volunteered to participate in the experiment


I 1 36 o
II 1 30 0


I 2 57 o
II 2 64 o


I and II canbined


2 62 o


I 1 7


II 1 21


0











Table 2-continued.


III. Same as I plus random (non-responsecontingent) electric shock conditioning
IV. Sane as I plus
mixed (sane radan and some drinkingcontingent) electric
shock conditioning
1976 Outpatient, camunity reinforcement:
I. Behavioral job and
marital counseling, resocialization procedures, problen-preven-tion rehearsal, daily reports to counselor,
disulfiram, group
counseling, "buddy"
procedure, behavioral
contracts
II. (control group) instruction regarding
alcoholism and its dangers, individual
and group counseling, advice to take disulfiram, encouragement to join AA

1976 Outpatient behavior modification, individual
therapy:
I. Self-regulation training including BAC
feedback training, discriminated aversive electric shock conditioning, education,
and psychotherapy


III


Groups IIIV recceived 89 hours aversive conditioning


1 38


IV 1 40


I, II,
and III can
bined


24 months Time spent drinking, uneAployed, away fran hae,
and institutionalized


At the end 20 men admitted of 24 to a state hosmonths of pital for alcotreatment holism, ages 2060; all agreed to participate
in the study


1 31


18 I 1 98


0


0

0




+


II 1 45 o


Ten therapy sessions of 2 hours each


Controlled drinking


12 months after treatment


60 (49 male) Australian outpatients, mean age 44 years, majority having an 8 year history of alcoholism


37 I 2 38 0


II 2 30 o


III 2 10


I,11, 2 27 0


Azrin


Caddy & Invibond


-n











00~


Table 2-continued.



Mean Drinking outcome
treatment Time of N unduration evalua- Sauple size evalu- ocmonStudy Date Treatment type (if given) Outome criteria tion and composition ated group class/% ness

II. Same as I except and III


electric shock was
not used
III. Sane as I except
that no emphasis
was given to selfregulatory process


1976 Residential, public funded
agency: the treatment discussed is group training
in assertiveness with
videotaped feedback vs.
with verbal feedback


1976 State hospital, inpatient:
I. Group therapy, chemotherapy, AA, abstinence as goal
II. Same as I plus individualized behavior
therapy
III. Same as I but with
controlled drinking
as goal
IV. Same as II but with
controlled drinking
as goal


combined

I II
III I, II, and III combined


Residential
treatment: 21 days; assertiveness training: 12 hours


Degree of underor over-assertion


40 days Daily drinking
(inferred) disposition,
general emotional adjustment,
vocational satisfaction, occupational status, driving status,
residential
status/stability,
general health


12 months after
treatment


24 months posthospital


137 lower to mid- 60 dle SES Canadian alcoholics, imean age 44.2 years


70 male alcoholics who volunteered for experimental treatment in a state hospital


4
4
4
4


38
36
40 38


+
+
+
+


NR NR


69 I 2 42


0


II 2 64 o


III 2 42 o


IV 2 85


+


Scherer &
Freedberg


Sobell & Sobell











Table 2-continued.


1976 Private aversion conditioning hospital:
pharmacological (emetic)
counterconditioning: 5
sessions in 8 days after
withdrawal plus 5 reconditioning sessions over
the following year


13 days in hospital (range: 8-32
days)


Abstinence


12 months after admission


261 alcoholics 261 (80% males) who underwent treatment voluntarily, mean age 48.5, mean years of education 13.8,
most married and middle SES or above


Cutter, 1977 VA hospital, inpatient:
Boyatzis, I. Standard treatment:
& Clancy group therapy; individual counseling; exposure to AA, disulfiram
II. Same as I plus "power motivation training" (PMr); instruction in identifying links between power concerns and excessive drinking; teaching methods for identifying feelings of powerlessness; guidance in identifying and practicing alternative ways of feeling powerful McClelland 1977 Same as above


MdWilliams & Brown


1977 State psychiatric hospital, inpatient: small group therapy; individual therapy and counseling; family therapy;


Range of hospital
stay 30180 days, mean not reported; PH! group received 35 hours PMr in 10 days


Days intoxicated and weeks worked


Same as Drunk less than
above once a month and
employed 5/6 time


Range 4256 days,
mean not reported


Freguency of readmission for detoxication; MPI scores


12 months posttreatment


















12 months posttreatment

1, 6, 12, and 18 months posthospital


100 (98 male) alcoholics, trwan years of alcoholism 10, considerable prior treatment, at least one physical indicator of alcoholism












Same as above


120 men (94% voluntary) admitted to a state hospital alcoholism re-


74 NR NR




















93 I 2 24
II 2 46


III NR NR


Wiens, Montague, Manaugh, & English


All Ss 1 63


+


0
0










__J C)


Table 2-continued.



Mean Drinking outcxne
treatment Time of N unduration evalua- Sanple size evalu- camnStxdy Date Treatment type (if given) Outcome criteria tion and opposition ated group class/% ness


vocational rehabilitation services; AA meetings; physical exercise; relaxation therapy; disulfiram for 26% of
patients

1978 State hospital, inpatient:
I. Group therapy; chemotherapy; AA; with abstinence as goal
II. Same as I plus individualized behavior
therapy
III. Same as I but with
controlled drinking
as goal
IV. Same as II but with
controlled drinking
as goal
1978 VA hospital inpatient and
day treatment:
I. Group therapy and didactic presentations
II. Same as I plus skill
training: group instruction in problemsolving processes and
guided generation of and rehearsal of alternative responses in situations associated with drinking
behavior


40 days Daily drinking 36 months
(inferred) disposition; gen- posteral entional ad- hospital
justment; vocational satisfaction;
occupational status;
residential status/
stability


28 days inpatient
followed by 20 days day treatment.


Drinking behavior, 1, 3, 6, employment, hospi- and 12 talizations, weeks months in aftercare postday
treatment


habilitation ward, none diagnosed as psychotic or OBS


70 male alcoholics who volunteered for experimental treatment in a state hospital; all Ganma-type


40 male alcoholics who volunteered to participate in the
study; not actively psychotic or severely organically impaired; low middle SES; mean years
problem drinking 17; mean age 45.6 years


53 I 2 67 o



II 2 81 +


37


III 2 75


IV 2 95


NR NR


+


+


Caddy, Addington,
& Perkins











Chaney,
O'Leary, & Marlatt










Table 2--continued.


Jackson &
Snith


1978 Aversion conditioning hospital, inpatient:
I. Pentothal interviews,
nursing care; medical services; psychiatric consultation; alcohol
education lectures;
access to counseling;
chemical (enetic)
aversive counterconditioning
II. Same-as I except
aversive conditioning is electric
shock paired with
smelling and tasting an alcoholic beverage


Miller 1978 Outpatient, behavioral,
with controlled drinking as goal:
I. Behavioral self-control training: daily selfnonitoring and recordkeeping; training to identify stimulus antecedents of heavy drinking; instruction in rate reduction; instruction in selfcontrol
II. Aversive counterconditioning: electric shock paired with sniffing an alcoholic beverage and imagining tasting it, 40-50 trials per weekly session; self-mnnitoring and recordkeeping; no self-control instruction


5 aversion sessions,
usually 12 days in hospital


1 outpatient session weekly for 10 weeks


Abstinence at the
tixe of followup


Drinking frequency and amount


24 months post
treatment


12 months after treatment


344 (approximat- 192 ely 70% male)
aversion therapy hospital patients; 99% were white;
nost were employed and married; 287 received
chemical aversion therapy; 57 received electrical shock aversion therapy


46 (32 male) problem drinkers; mean age 36; mean education 13 years; low to middle income; 36%
court-referred; all approved by
alcoholism treatment center personnel for controlled drinking


I 1 77 +


II 1 85


35 I 1 7


II 1 20


III

I, II,
and III comtined


18

1 14 0


__j
F1


+


0













'able 2-continued.



Mean Drinking outome
treatment Time of N unduration evalua- Sample size evalu- coxutnStudy Date Treatment type (if given) Outcome criteria tion and composition ated group class/% ness

III. Controlled drinking I / 3V U


cxposite: Same instruction and selfmonitoring as I included in 10 sessions
of 120-150 minutes
each, which also included electric shocks
aversive counterconditioning with shock
paired to actual drinking beyond a
specified BAC

1979 Hospital, inpatient: irdi- 90 days
vidual and group behavioral counseling focused on identifying discriminative cues for drinking and generating and rehearsing nondrinking
responses


1979 VA hospital: inpatient:
milieu therapy; involvement of significant
others; disulfiram; extensive, aggressive
aftercare and followup


II 2 57


III

I, II, and III combined


Abstinence


Drinking status; residential stability; interpersonal relationships; employment status; leisure activities; general health


More than 13 months postdischarge;
mean: 20 months


24 months
posthospital


60 West Gernan alcoholic wcmen, mean age 35.3 years; mean years alcohol abuse 6.5; never diagnosed as endogenous depression; actively seeking help

75 (72 male) alcoholics (47 Caucasian, 26 Mexican-American, 2 Black); mean age 45.8 years; low middle SES; low
marital stability


58 All Ss
All Ss


65 All Ss


2 46 2 66







1 40 2 52


6 39 o


Cohen, Appelt, Olbrich, & Watzl


Costello, Baillargeon, Biever, & Bennett


NR












Alford


Bowen & Tweemlow


1980 Inpatient chemical dependency unit adjacent to a metropolitan general hospital, Moriented: individual,
group, and family
counseling; didactic
presentations; goal to
carplete the first five of the twelve
steps of AA while in
treatment

1980 No treatment: a study
of one-year-later
status of persons who
applied for alcoholism treatment at a VA hospital but did not
appear for treatment


38 days;
range: 35-77
days


Costello 1980 VA hospital, inpatient 42
therapeutic ccrmunity inp
plus outpatient aftercare and disulfiram






Cronkite 1980 Salvation Army; public hos- NR & moos pital alcoholism unit;
county funded halfway
house; private milieu therapy facility; aversion hospital. No detailed description of treatment
is provided.


Table 2-continued.

"AA criteria": 6, 12, &
abstinent; en- 24 months
played; socially poststable discharge
"General criteria": abstinent or light, nonabusive drinking


Abstinence; enployment; hospitalizations; arrests; living situation


days Drinking status;
patient residential stability; interpersonal relationships; leisure activities; employment status; general
health

Alcohol consunption; abstinence; depression; occupational functioning


12 months after application for
treatment



24 months postdischarge







A~proximately 24 months post-. treatment


56 (27 male) al- 43 coholic patients; mean age 49 years; mean education
12.5 years; mean years of problem drinking 15; middle SES


77 men who applied for alcoholism treatment in a VA hospital and did not appear for treatment


41 Anglo-American 37 men treated in a
VA hospital alcoholism unit


124 (94 male) alcoholic patients; mostly middle-aged,
white, and educated beyond high school. Distributed by treatment facility thus: Salvation Army


All Ss All Ss


77 All Ss


NR


120


1 51 o

2 66 o


1 24


0


NR


NR


-.2
C..J













Table 2-continued.



Mean Drinking outccnie
treatment Time of N unduration evalua- Sanple size evaluStudy Date Treatment type (if given) Outocrme criteria tion and crnposition ated group class/% ness

8%; public hospital 13%; halfway house 4%; private
facility 48%; aversion hospital 27%


1980 Private, AA-oriented,
residential treatment facility


1980 Free-of-charge outpatient behavior modification program for
problem drinkers desiring to moderate
their use of alcohol
by behavioral self
control training (BSCr): I. Bibliotherapy group:
intake interview;
provision of a behavioral self control
training self-help
manual and selfnonitoring record


33 days Abstinence; moderate drinking; other drug use;
physical health; self image; professional performance; personal
adjustment

I: intake Drinking amount only and frequency;
II and III: mean BAC 10 onehour sessions in 10 weeks
IV: 10
90-minute sessions in
10 weeks


12 months postdischarge


12 months after treatment


85 male physicians, typically middle-aged and married with a history
of 14 years of problem drinking


41 (25 male) al- 38 coholics by mean score on Michigan Alcoholism Screening Test;
mean age 45.4 years; mean years of education 15.8; mean annual incaxe $23,732.00; all voluntary clients


67 All Ss
All Ss


I
II
III
IV
I-IV
ccnbined


Kliner, Spicer, & Barnett


Miller & Taylor


1 76 +
2 92 +


2
2
2
2
2


33
27 38 60
42


0
0
0
0
0










Table 2--continued.


cards; prompting by
telephone to mail in
record cards
II. Same as I with the
addition of 10 individual one-hour BSCr
sessions following
the manual
III. Sane as II except
that the individual
sessions included
training in progressive deep muscle
relaxation
IV. Intended to be identical in content to
III; sessions were
in group format and
90 minutes long

1980 Individualized behavior
therapy, locus and specifics not reported in this
article

1981 Hospital inpatient: AAoriented group therapy; transactional analysis; grief groups;abstinence
viewed as the only viable treatment goal


NR




56-70
days; mean NR


Days abstinent; days of controlled drinking


Minimum number of weeks fran leaving the hospital to first drink; abstinence; controlled drinking; relapse


24 months following treatment


12 months postdischarge


70 male alcoholics



106 (74 male) alcoholics; nean age 44 years; 5 Maori; 101 of European descent; most saniskilled
workers; mean years prior alcoholism 10.4


69 NR NR




100 All Ss 1 48
All Ss 2 62 Ali Ss 7 38


1981 Private, residential, AAoriented. (See "Sample"
column for subject groupings.)


Abstinent at
followup; level of psychosocial functioning


12 months posttreatment


158 male alco- 158 holics; predominately middle class; 42% had attended college; median age 42; grouped by MMI


Maisto, Sobell, & Sobell


Abbott & Gregson


Conley


0
0
0


I
II III
IV


1 48 1 66 1 70 1 57


0
+
+
+


-J










-4


Table 2-continued.



Mean Drinking outcome
treatment Tine of N unduration evalua- Sanple size evaTu- nStudy Date Treatment type (if given) Outcome criteria tion and composition ated group class/% ness


profiles thus: I. Neurotic II. Classic alcoholic
III. Psychopathic
IV. Psychotic


1981 Private, residential alco- 28 days
holism treatment center:
insight-oriented individual, group, and family
therapy; AA as a major
support


Drinking behavior; adaptive behavior


12 months postdischarge


66 alcoholism treatment center patients who aged 18-65; did
not have organic brain syndrwe; were not psychotic on admission. The 38 (31 male) subjects studied
were mostly married and employed; mean age 45.5 years; upper middle class


1981 Private aversion conditioning hospital: mean
of five chemical (emetic)
aversive counterconditioning sessions; irdividual counseling; relaxation training including
biofeedback; discussion
groups; orientation to AA;


11-14 day detoxication; 10 days (5 sessions) counterconditioning; 7 aftercare


Abstinence


12 months after admission


908 alcoholics 843 (76% male); mean age 48.5 years; mean years of education 12.3; 58% married; 44% employed; mean
years uncontrolled drinking 7.8


Jones & Canyon


38 All Ss
All Ss


1 52 2 78


0
+


Neuberger,
Hasha, Matarazzo, Schmitz, & Pratt


All Ss


1 53 0










Table 2-continued.


family workshops; unlimited aftercare visits if abstinent


overnight visits in one year


1981 Court-ordered, outpatient: I
I. Power motivation train- h
ing: group experiential 4 exercises in risk- I
taking, goal-setting, h and interpersonal ccn- 6
munication during
stressful situations
II. EII group therapy workshops: emphasis on identifying problems and
giving information about
alcohol
III. Hcnv-study course:
doing written exercises from an alcohol education guide


1981 Hospital alcoholism treatment unit, inpatient: individual and group counseling; alcohol education lectures; AA metings, psychological evaluation; recreational therapy


: 32
iours in sessions 1: 15 ours in sessionsIII: NR


12 days; range: 139 days


Amount of frequency of drinking; personal adjustment


Drinking behavior; personal, familial, vocational, and social adjustment


18 mnths after treatment


6, 12, 18, and 24 months after discharge


351 male "mid- 218 range problem drinkers" arrested for IKI; mostly white; mean age 31.4 years; mean years of education 11.7;
mostly low middle incae


247 alcoholic first-time admissions; mostly white males; mean age 44 years; mean years of education 12;
about 66% were married; about 50% reported a family history of alcoholism


247


Swenson, StruckmanJohnson, Ellingstad, Clay, & Nichols


Valle


NR NR


NR NR






78


Idiosyncrasies in Table 2

Some of the studies summarized in Table 2 reported outcomes at one or two years after admission rather than after the termination of treatment. Azrin (1976) reported on outcomes at the end of two years of treatment. Three studies on aversion-conditioning hospital treatment (Jackson & Smith, 1978; Neubuerger et al., 1981; Wiens et al., 1976) reported outcomes at one year after admission, which year included, for many of their subjects, return visits to the hospital. The outcomes

reported in these studies should not be understood to be outcomes measured after all treatment had ceased for, respectively, two or one years.

The study by Bowen and Tweemlow (1980) is unique among those included here in that it focuses on persons who applied for alcoholism treatment but did not appear for treatment when scheduled.

For nine studies, the abbreviation "NR" (not reported) appears in the "drinking outcomes" portion of Table 2. Scherer and Freedberg

(1976) did not report drinking outcomes; their outcome focus was on the dimension of assertiveness, not drinking. The other eight studies did report drinking outcomes in some form or other, but within-group percentages could not be derived from their drinking outcome data as reported.

Patterns in Table 2

Of the 25 studies for which an "uncommonness of outcome" entry could be made, three are coded as reporting uncommonly poor drinking outcomes, 12 as reporting uncommonly good drinking outcomes, and ten as reporting not uncommon drinking outcomes.






79


In each instance in which an uncommonly poor drinking outcome was reported, the poor outcome pertained to only one group of subjects within a larger cohort. In the Vogler, Ferstl, Kraemer, and Bengelmann

(1975) study the poor-outcome group was a control group receiving only standard hospital treatment. The poor-outcome group, with an abstinent-or-controlled rate of 10%, in the report by Caddy and Lovibond (1976) was just below the lower boundary (10.5%) for uncommonness. In the Miller (1978) study, which reported two groups with uncommonly poor abstinence outcomes (7% and 8%), the treatment goal was controlled drinking. In all three cases the elapsed time between the end of treatment and evaluation was one year.

The twelve papers reporting uncommonly good outcomes were reports on only ten programs. The outcomes at one, two, and three years for one

program were reported in three separate papers (Caddy et al., 1978; Sobell & Sobell, 1973, 1976). The apparent uncommonly good drinking outcome of another program (Jones & Lanyon, 1981) should, according to the authors of the report, not be considered valid for program evaluation because too many subjects were lost to followup. Thus, only nine programs are discussed here.

The distribution of types of treatment facility among the nine is unremarkable with one exception. There are two outpatient programs, one state hospital inpatient program, two private general hospital programs,

two private aversive conditioning hospital programs, and two private residential treatment center programs. The last two, together, provide the aforementioned remarkable exception. The studies by Kliner et al. (1980) and Conley (1981) are both on different samples of the treated population of the same residential treatment center, which has a good






80


reputation, a long history and a psychodynamic AA-oriented approach. The Kliner et al. study showed uncommonly good results in both the abstinent and the abstinent-or-controlled classes. The Conley study

reported uncommonly good results in the abstinent class, with a highly selected sample (physicians).

There appears to be a pattern with regard to treatment modality and uncommonly good outcome. Seven of the nine programs were inpatient and two were outpatient. Of the seven inpatient programs one was

behaviorally oriented, four were dynamically oriented, and two offered dynamically oriented counseling as an adjunct to aversive counterconditioning. The two outpatient programs were behaviorally oriented. Thus it would appear that unusually successful inpatient programs share some species of dynamic orientation, and unusually successful outpatient programs share a behavioral orientation.

There was a wide range of treatment duration. Among inpatient

programs, mean treatment duration ranged from 12 days for aversive conditioning hospitals to 82 days for a general hospital. Outpatient treatment duration was ten sessions (Caddy & Lovibond, 1976) and two years (Azrin, 1976). While no general pattern of relationship between

treatment duration and uncommonly good outcome seems to emerge, it is interesting to note that Willems et al. (1973) reported a long-stay treatment group reached a 54% abstinence level, which is in the low end of the uncommonly good outcome range, but concluded that there are no important outcome differences between groups treated 20 days and 80 days, respectively. It is noted, however, that Willems et al. did not use Emrick's uncommonality of outcome classes in evaluating their data.






81


There appears to be a relationship between treatment goals and uncommonly good treatment outcomes. Six of the programs under discussion appeared to have abstinence as a sole drinking outcome goal (Azrin, 1976; Conley, 1981; Jackson & Smith, 1978; Kliner et al., 1980; Wiens et al., 1976; Willems et al., 1973). The program first reported on by Sobell and Sobell (1973) and then at two other intervals had abstinence as a treatment goal for two of four subgroups. Among the seven programs in which there was a goal of abstinence for at least one treatment group, six studies reported uncommonly good abstinence-class drinking outcomes for at least one group.

Two programs (Caddy & Lovibond, 1976; Tomsovic, 1974) had reduced or controlled drinking as outcome goals. The Sobell and Sobell program had controlled drinking as a goal for two of four subgroups. Among the three programs in which there was a controlled drinking outcome goal for at least one group, Caddy and Lovibond reported uncommonly good outcomes in the somewhat-improved drinking outcome class, while Tomsovic reported

uncommonly good outcomes in the much-improved category. The Sobell and Sobell program reports indicated uncommonly good abstinent-or-improved category outcomes for two groups at one year, for one group at two years, and for two groups at three years.

It is possible that the time of evaluation was important in two cases. Weins et al. (1976) reported an uncommonly good abstinence rate for all subjects at one year after admission to an aversive conditioning

hospital, during which year some patients had returned for additional treatment. Azrin (1976) recorded an uncommonly good abstinence rate (98%) for experimental subjects two years after they began a community reinforcement behavior modification program.





82


All of the subjects in all of the programs reporting uncommonly good treatment outcomes were classified as alcoholics, mostly male, mostly middle class. The one instance in which the uniqueness of the subject sample may have made an important difference is the study by Kliner et al. (1980) in which the sample was composed entirely of physicians. At one year after discharge, Kliner et al. reported 76% abstinent and 92% abstinent-or-controlled.

Subject Variables

Table 3 summarizes the subject variables alluded to in the studies under consideration. "Subject variables," in this context, refers to facts pertaining to the clients previous to or at the beginning of treatment. In Table 3 subject variables are categorized by the methods by which they were measured. "Standardized methods" refers to the use of questionnaires, inventories, scales, or psychological tests which had

been standardized on some other population before they were used in the study in question. "Nonstandardized methods" refers to the use of structured interviews, unstructured interviews, or instruments which had not been standardized on some other population before they were used in the study in question.

The numbers in the columns of Table 3 indicate the number of studies alluding to the subject variable in question, and whether the method of measurement was standardized or not. Thus, one would

understand by reading the first line of Table 3 that for the subject variable "drinking amount or frequency" the variable was alluded to in 11 studies in which it had been measured by standardized methods and in 12 studies in which it had been measured by nonstandardized methods, so that it was alluded to in a total of 23 studies.






83





Table 3
Numbers of Studies Reporting on
Various Subject Variables


By standardized By nonstandardized Variable methods methods Row totals


Drinking amount
or frequency 11 12 23

Affectivecognitive 14 6 20

Work situation 5 10 15

Home situation:
relationships 3 8 11

Previous inpatient treatment 0 14 14

Physical 5 12 17

Legal problems 2 6 8

Mixed 0 0 0

Social situation 3 3 6

Previous A.A.
involvement 0 7 7

Finances 1 7 8

Global adjustment 1 1 2

Miscellaneous 3 1 4

Other drinking
behavior 1 3 4

Home situation:
residence 2 0 2

Previous outpatient
treatment 0 2 2

Leisure time 0 0 0

Religious life 0 2 2





84


Table 3--continued.


By standardized By nonstandardized Variable methods methods Row totals


Previous treatment:
locus unclear 0 1 1

Column totals 51 95

Patterns in Table 3

Subject variables were measured 95 times by nonstandardized methods and 51 times by standardized methods. Although some variables, e.g., those pertaining to history of previous treatment, are not amenable to

measurement by standardized instruments, it appears that the use of nonstandardized or "homegrown" methods of measuring subject variables predominates. The sole exception to this pattern is the use of standardized methods to measure variables in the affective and cognitive domains, in which standardized methods were used 14 times and nonstandardized methods only six times.

Outcome Variables

Table 4 summarizes the outcome variables alluded to in the studies under consideration. "Outcome variables," in this context, refers to facts pertaining to the clients after treatment. In Table 4 outcome

variables are categorized by the methods by which they were measured. "Standardized methods" and "nonstandardized methods" have the same meanings as in Table 3.

The numbers in the columns of Table 4 indicate the numbers of studies alluding to the respective outcome variables, and the method of measuring those variables.






85


Table 4
Numbers of Studies Reporting on
Various Outcome Variables


Variable Drinking amount or frequency Affectivecognitive Work situation Home situation: relationships Subsequent inpatient treatment Physical Legal problems Mixed

Social situation Subsequent A.A. involvement Finances Global adjustment Miscellaneous Other drinking behavior Home situation: residence Subsequent outpatient treatment Leisure time Religious life


By standardized By nonstandardized
methods methods Row totals


4


29


8 15 10 10

8 10

4

7


3

4


0


0

4

2

0

4


0

1

3

0


0


2


0

1

0


33 11 19 10 10

12 12

4 11


9

4

8

8


5


9

3

5

8


5


5


3

5

2


7


3

6

2





86


Table 4--continued.


By standardized By nonstandardized Variable methods methods Row totals


Subsequent treatment: locus unclear 0 2 2

Column totals 28 148

Patterns in Table 4

Outcome variables were measured 148 times by nonstandardized

methods and 28 times by standardized methods. Of the 34 studies under consideration all but one reported on amount or frequency of drinking as an outcome variable. The outcome variable second most frequently

alluded to was "work situation," which was referred to in 19 studies. Physical health and legal problems were each alluded to in 12 studies, affective-cognitive variables and social situation in 11 studies each, and at-home relationships and subsequent in-patient treatment in ten studies each. Subsequent AA involvement, global adjustment, and the miscellaneous variable were respectively alluded to nine, eight, and eight times each.

There appears to be a pattern of almost ubiquitous interest in drinking as an outcome variable, while vocational adjustment holds second place in terms of interest. Trailing at considerable distance are physical health, legal problems, affective-cognitive status, social situation, family relationships, subsequent treatment, AA involvement, and global adjustment.

Treatment Variables

Table 5 summarizes the treatment variables alluded to in the studies considered here. The numbers in the columns of Table 5 indicate





87


the numbers of studies alluding to the respective treatment variables and whether the respective variables were reported directly or inferred from the study. The following paragraphs clarify the structure and content of Table 5.

In three studies length of treatment was not specifically reported but could be inferentially estimated from the description of the treatment regimen. In five studies length of treatment was neither reported nor inferrable.

The "locus of treatment" category classifies treatment as inpatient

or outpatient as well as referring to type of treatment facility, i.e., state, VA, or private hospital; private residential facility; halfway house; or outpatient counseling center. For one study the locus of

treatment was neither reported nor inferrable.

"Treatment methods" refers to both the type of treatment (e.g., behavioral vs. psychodynamic) and specific treatment procedures (e.g., electric shock aversive conditioning, assertiveness training). In three studies treatment methods were neither specifically reported nor inferrable.

"Theoretical stance" refers to the position held by a treatment program on the nature of alcoholism and/or the nature and course of therapeutic practice in the treatment of alcoholics. Two studies made brief, direct reference to theoretical stance. For 18 studies some

rough and faint approximation of theoretical stance could be inferred from the stated treatment methods of choice, the statement that the program was AA oriented, or the selection of treatment goals.

Some studies made direct reference to the definition of alcoholism used, e.g., stated that World Health Organization criteria for the










diagnosis of alcoholism were used. In some other studies the definition

of alcoholism used was inferred from the authors' use of such terms as "gamma-type alcoholics." For 23 studies the definition of alcoholism was neither directly reported nor inferrable.

"Characteristics of treatment personnel" refers to such items as the training, professional or paraprofessional group and level, experience, and personality characteristics of treatment personnel.


Table 5
Numbers of Studies Reporting on
Various Treatment Variables



Variable Reported directly Inferred Row totals


Length of treatment 26 3 29

Locus of treatment 33 0 33

Treatment methods 31 0 31

Theoretical stance 2 18 20

Definition of alcoholism 3 8 11

Characteristics of
treatment personnel 4 0 4

Patterns in Table 5

Length, locus, and methods of treatment were reported in almost all studies. Theoretical stance was inferred for 18 studies but directly referred to in only two. The definition of alcoholism used was reported or inferrable in less than one-third of the studies. Characteristics of treatment personnel were reported in four studies and were not inferrable from the other 29.


88





C 0 U)


Research Models and Statistical Significance

Tables 6, 7, and 8 summarize the major findings with statistical significance of p-.05 or better for those studies whose major findings were statistically significant. The findings are organized by research

model, with the studies listed alphabetically within the respective years of publication.

The categories of research models were drawn from the descriptions of basic methods of research set forth by Isaac and Michael (1971). The

research models used in the studies under consideration here were subsumed under the categories of correlational, descriptive, experimental, or quasi-experimental methods.

Of the studies which included statistically significant findings, one used a descriptive research model. Alford (1980) noted two major findings significant at the .05 level or better. He found that for

alcoholics treated in an AA-oriented inpatient program, at two years posttreatment:

(a) Ninety-six % of then-abstinent alcoholics were socially stable

as compared with 50% of those who were drinking moderately and 14%

of those who were drinking heavily;

(b) Pre-treatment alcohol intake is a predictor for treatment outcomes; subjects whose reported pre-treatment drinking was in the 100-200 oz. per week range had a higher success rate than did either those in the over 200 oz. per week range or those in the under 100 oz. per week range, although the third group fared better

than the second.






90


Table 6
Major Findings with Statistical Significance of p.-.05 or Better in Studies Utilizing Correlational Research Models


Study


Findings


Pokorny et al., 1973







McWilliams & Brown, 1977





Bowen & Tweemlow, 1980




Cronkite & Moos, 1980









Maisto et al., 1980


1. Patients with histories of higher stability in
marriage, social relationships, and employment are more likely to participate in aftercare. 2. Male alcoholics treated by 60 days of inpatient care followed by outpatient group
therapy show the same rate of improvement as those who had 90 days of hospitalization
without aftercare.

1. MMPI data are of little use in identifying
alcoholics with good prognoses.
2. The success of alcoholism treatment cannot
be evaluated from MMPI change scores at six weeks of treatment compared to the second week
of treatment.

One year later, there were no significant differences in rates of abstinence, hospitalizations, or police arrests between men who applied for alcoholism treatment but failed to appear for treatment and men who applied for and underwent alcoholism treatment.

1. Posttreatment stresses and coping responses
have strong effects on outcome.
2. Effects of both pre-treatment subject variables and treatment-related factors are
frequently shared with the effects of posttreatment factors.
3. The relative strength of pre-treatment,
treatment, and posttreatment variables as
predictors of outcome will vary with the kinds
of outcome criteria considered.

1. Individualized behavior therapy (IBT) with a
controlled-drinking goal has a strong positive relationship to controlled drinking as an
outcome.
2. IBT with an abstinence goal does not influence
days abstinent as an outcome.
3. There is a negative relationship between the
number of pre-treatment alcohol-related hospitalizatons and controlled drinking outcomes.





91


Table 6--continued.


Findings


Abbott & Gregson, 1981 Conley, 1981






Jones & Lanyon, 1981 Patterns in Table 6


The level of cognitive dysfunction at intake
predicted the one-year outcome for 58% of subjects.

Some MMPI types are predictive for differential drinking outcomes; at one year posttreatment the high-to-low rank ordering of abstinence rates
between groups of subjects typed by MMPI profile was: psychopathic; classic alcoholic; psychotic; neurotic.

At one year after inpatient treatment subjects' positive scores on an instrument measuring adaptive skills correlated positively with positive treatment outcome.


Of the 14 findings listed in Table 6, seven had to do with subject variables as predictors of treatment outcomes. Four findings had to do with the relationships of treatment and outcome. Two findings had to do


with the relationships of posttreatment factors to outcomes.


One


finding had to do with how the kinds of outcome criteria chosen will cause apparent variation in the predictive power of other variables.


Table 7
Major Findings with Statistical Significance of p.-<05 or
Better in Studies Utilizing Experimental Research Models


Study


Findings


Sobell & Sobell, 1973




Vogler, Compton, & Weissbach, 1975


Patients receiving individualized behavior therapy functioned better at 12 months than control groups receiving conventional therapy, regardless of whether the treatment goal was abstinence or controlled drinking.

Higher levels of pre-treatment alcohol ingestion are strong predictors for higher levels of posttreatment alcohol drinking.


Study






92


Table 7--continued.


Findings


Vogler, Ferstl, Kraemer, & Brengelmann, 1975



Azrin, 1976




Sobell & Sobell, 1976










Caddy et al., 1978














Chaney et al., 1978


There were significant between-groups differences on drinking outcome measures for groups receiving, respectively, standard hospital treatment, drinkingcontingent shocks, random shocks while drinking, and mixed drinking-contingent and random shocks.

Compared with matched controls not receiving community-reinforcement treatment, alcoholics in a community-reinforcement treatment program did better on all outcome measures.

1. At 24 months posttreatment, clients who had
received individualized behavior therapy (IBT) with controlled drinking as a goal
functioned better in terms of both drinking outcomes and general adjustment outcomes than did a control group who received standard
hospital treatment for alcoholism.
2. At 24 months posttreatment, there was no
significant difference between IBT clients with
a nondrinking treatment goal and their controls.

1. At three years posttreatment, clients in the
individualized behavior therapy groups, regardless of whether the treatment goal was abstinence or controlled drinking, did better in terms of overall improvement than their
controls.
2. At three years posttreatment, there were no
significant differences in drinking outcomes between IBT clients with abstinence as a treatment goal and their conventionally treated
controls with the same treatment goal.
3. In the controlled-drinking-as-treatmentgoal condition, IBT clients drank less at three years than their conventionally treated
controls in the same goal condition.

At 12 months posttreatment, alcoholics who had
received skill training specific to the avoidance of alcohol abuse had an average number of days drunk
equal to one-sixth that of pooled control groups, drank one-fourth as much, and had an average drinking period less than one-eighth as long. The difference in days abstinent was not significant.


Study




Full Text

PAGE 1

A SYNTHESIS OF PSYCHOLOGICALLY ORIENTED ALCOHOLISM TREATMENT OUTCOME RESEARCH BY JAMES DAVID POAGE A DISSERTATION PRESENTED TO THE GRADUATE COUNCIL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 1982

PAGE 2

This work is affectionately dedicated to Peg, a warmhearted humanist, who nevertheless has a real understanding of carrots and sticks; Maureen, who knows why; a few dozen alcoholics, who are my respected teachers; and my mother and father, who between them showed me how to be gentle, muleheaded, and logical.

PAGE 3

ACKNOWLEDGMENTS There are several persons who have helped me in the writing of this dissertation. Larry C. Loesch, Ph.D., who served as my committee chairman, provided me with good-humored encouragement and guidance dispensed with a kindly hand. The other members of my committee, Paul W. Fitzgerald, Ed.D., and Franz R. Epting, Ph.D., patiently stuck with my meandering process for over five years. All have earned my gratitude. Mary Kay Hartung, reference librarian at the University of South Florida, was of invaluable assistance in helping me talk to a computer so that it could understand what I wanted in the way of a literature search. Linda Mazzone typed and retyped the manuscript with patience and care. Margaret Scripps Buzzelli, owner and President of Anabasis, Inc., urged and permitted me to spend company time on the project. She knew when to say, "You can do it!" and she did. My sons, John, Bob, and David Poage, alternately harrassed and loved me toward finishing the work. Katherine Bailey, my stepdaughter, an accomplished applied psychologist at age eleven, administered her smile as a selective reinforcement when needed. Finally, firstly, and in between, my wife, Maureen Ann O'Harra, Ph.D., gave me loving support without stint, served as second reader and first-rate editorial consultant, and was optimistic enough for both of us . iii

PAGE 4

TABLE OF CONTENTS PAGE ACKNOWLEDGMENTS iii LIST OF TABLES vi ABSTRACT vii CHAPTER PAGE I INTRODUCTION 1 Context 1 General Assumptions 2 Statement of the Problem 2 Need for the Study 3 Purpose of the Study 4 Rationale for the Approach 5 Research Questions 11 Definition of Terms 11 Overview of the Remainder of the Study 13 II SURVEY OF RELATED LITERATURE 15 Support for the Delineated Problem 15 Support for the Need for the Study 24 Support for the Approach of the Study 46 Summary 49 III METHODOLOGY 50 Overview of the Study 50 Delineation of Variables 51 Description of the Population 54 Sampling Procedures 55 Description of Sample 57 Description of Research Design 57 Description of Research Procedures 57 Description of Data Analysis Procedures 58 Methodological Limitations 59 IV RESULTS 61 Description of Studies 61 Subject Variables 81 Outcome Variables 83 Treatment Variables 85 Research Models and Statistical Significance 87 Research Models and Practical Significance 94 iv

PAGE 5

Research Questions 99 Interactions 103 V DISCUSSION 109 Generalizability Limitations 109 Answers to Research Questions 110 Conclusions and Interpretations 116 Implications for Theory, Research, Training, and Practice 120 Recommendations 122 Summation 127 LIST OF REFERENCES 129 APPENDICES A DATAGATHERING PROTOCOL 132 B INTERJUDGE AGREEMENT 134 C REFERENCE LIST OF STUDIES ANALYZED 135 BIOGRAPHICAL SKETCH ' 138 V

PAGE 6

LIST OF TABLES TABLE PAGE Table 1 Alcoholism Treatment Articles in Psych Info 18 Table 2 Description of Selected Outcome Studies Published 1973-1981 64 Table 3 Numbers of Studies Reporting on Various Subject Variables ' '. . .82 Table 4 Numbers of Studies Reporting on Various Outcome Variabes 84 Table 5 Numbers of Studies Reporting on Various Treatment Variables 87 Table 6 Major Findings with Statistical Significance of £<. 05 or Better in Studies Utilizing Correlational Research Models 88 Table 7 Major Findings with Statistical Significance of £<. 05 or Better in Studies Utilizing Experimental Research Models 90 Table 8 Major Findings with Statistical Significance of £<.05 or Better in Studies Utilizing Quasi-experimental Research Models 92 Table 9 Major Findings with Practical but not Statistical Significance in Studies Utilizing Correlational Research Models 94 Table 10 Major Findings with Practical but not Statistical Significance in Studies Utilizing Descriptive Research Models 95 Table 11 Major Findings with Practical but not Statistical Significance in Studies Utilizing Experimental Research Models 96 Table 12 Major Findings with Practical but not Statistical Significance in Studies Utilizing Quasi-experimental Research Models 97 Table 13 Major Research Questions of Outcome Studies of Psychologically Oriented Alcoholism Treatment 99 Table 14 Major Interactions Observed 104 vi ; . i

PAGE 7

Abstract of Dissertation Presented to the Graduate Council of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy A SYNTHESIS OF PSYCHOLOGICALLY ORIENTED ALCOHOLISM TREATMENT OUTCOME RESEARCH By James David Poage May, 1982 Chairman: Larry C. Loesch, Ph.D. Major Department: Department of Counselor Education Thirty-four studies published from 1973 through 1981, and evaluating outcomes of psychological treatment for ten or more alcoholics at 12 months or more after treatment, were reviewed. All studies included data on subject, treatment, and outcome variables. Subject variables were reported in 17 categories, with previous drinking history reported in all studies. Outcome variables were reported in 19 categories, with drinking behavior as an outcome variable considered most frequently (in 33 studies). Treatment variables were reported in six categories, with treatment method reported in all studies. No study treated the theoretical stance of the treatment program as a variable of focal importance. Subject variables were measured 95 times by nonstandardized methods and 51 times by standardized methods. Outcome variables were measured 148 times by nonstandardized methods and 28 times by standardized methods . vii

PAGE 8

Correlational research models were used in nine studies, descriptive research models in six, experimental research models in ten, and quasi-experimental research models in nine. Every study in which an effort was made to compute statistical significance reported at least one finding significant at the .05 level or better. Findings deemed practically significant but not statistically significant were identified in 21 studies. Two studies reported subjecttreatment interactions. Subjectoutcome interactions were reported in 13 studies. Seventeen studies reported treatment-outcome interactions. One study reported on subject-treatment-outcome interactions . Implications for theory, research, training, and practice were described. Recoinmendations were made for further developments in theory, research, training, and practice. viii

PAGE 9

CHAPTER I INTRODUCTION The often stodgy and ingrown world of research owes a debt of gratitude to Charles Shultz for a mildly iconoclastic Snoopy cartoon he drew in the 1960's. In the first panel Snoopy is looking at a huge and untidy pile of bones. In succeeding panels he tosses the bones one by one over his shoulder, so that in the last picture he has a large and untidy pile of bones on the other side of the yard from their original location. Snoopy grins: "We call it research." The present dissertation intends to modify Snoopy' s model by the addition of considerable order and refinement. Research in alcoholism does in some ways resemble an untidy bone-pile that needs not just reshuffling, but purposeful sorting. The specific bone-pile of data to be sorted in this study is composed of selected research reports on the outcomes of psychologically oriented treatment of alcoholism published in English in the years 1973-1981. Context There is in the field of alcoholism an information explosion. The literature doubled in the seven years, 1967-1973 (Moll & Narin, 1977). It continues to grow rapidly in various directions, making any unified comprehension more difficult with each new addition. The present study is responsive to the notion that it is more helpful to make some sense of and suggest new direction for the existing corpus of research than to add one more set of empirical findings to an already less than comprehensible pile of such data. 1

PAGE 10

2 General Assumptions The present author has personal reasons for undertaking this particular study in this particular way. His work is the treatment of alcoholics, and he deems it important to learn what works, with whom, and, if possible, why. To undertake the "treatment" of "alcoholics" is to embrace, implicitly at least, two assumptions. The first is that there are persons to whom the term "alcoholic" is appropriately applied, and that there is something about the condition of an "alcoholic" which is maladaptive and undesirable. The second assumption is that "treatment" may ameliorate that which is maladaptive, undesirable, and specific to "alcoholics." A third assumption is also important to the present study. It is that something of pragmatic and theoretical value may be learned from the analysis of a relatively thin slice of the research literature on outcomes of psychologically oriented treatment of alcoholism. Statement of the Problem Succinctly put, the problem to which this study is addressed may be stated in three propositions: (a) There are several hundred thousand alcoholics in the United States; (b) Substantial resources, both human and material, are expended on the treatment of these alcoholics; (c) There exists a huge body of research on the treatment of alcoholics, but there is no definitive synthesis on that research which has yielded a set of conclusions which serve as reliable guides to the most effective research, theory, practice, and training for the treatment of alcoholics .

PAGE 11

3 Need for the Study There are more than 9,000 alcoholism treatment programs in the continental United States (NIAAA, 1979). In 2,821 of these programs for which data are available, approximately 28,000 treatment personnel provide some form of treatment for an estimated 1,712,000 problem drinkers and alcoholics annually at an annual expenditure of approximately $795,183,000 (DeLuca, 1981). Unfortunately, no one seems to know what is being done or how well. The preceding sentence is, of course, hyperbole. Literally hundreds of published studies provide data on what is done in alcoholism treatment and what the apparent outcomes are. However, no single, clear picture emerges from this embarrassment du richesse . While the attainment of such a single, clear picture seems unlikely at this time, it is also apparent that there are needs which might be met by making a start on the tasks of summarizing and clarifying. Those needs ar^ delineated in the following statements of contingency. If researchers knew how alcoholism treatment outcomes have been studied, then researchers might identify the more (and less) fruitful avenues of approach and be guided by those identifications in future research. If researchers knew which models of alcoholism treatment outcome studies have yielded the clearest and most significant results (and the contrary), researchers could then be more usefully selective in choosing models for further research. If practitioners and treatment planners knew which theoretical stances are positively associated with the most favorable treatment

PAGE 12

4 outcomes, then they could be more pragmatic in selecting the best theoretical bases for treatment. If they knew the treatment variables which are the best predictors for the best prognosis when combined with given client variables, practitioners and treatment planners could match the treatment to the client for maximum effectiveness and efficiency. If teachers and trainers had a clear and valid understanding of the issues mentioned above, then they could train alcoholism counselors and other alcoholism workers with greater confidence and precision in pragmatically tested theory, diagnosis, treatment planning, and practice. Purpose of the Study This study has several purposes. Of necessity, the first purpose is to determine whether the research considered herein can provide answers to the research questions of this study. If it cannot provide those answers, a second purpose will be to suggest ways of research that might be more effective in supplying the needed information. The other purposes of this study, all with reference to psychologically oriented treatment of alcoholics, are as follow: (a) To identify the more (and less) fruitful approaches to outcome research; (b) To describe criteria for selecting useful models for such outcome research; (c) To describe criteria for selecting theoretical models of alcoholism which are related to positive treatment outcomes; (d) To describe criteria for matching treatment to client; (e) To describe theories and practices appropriate for inclusion in the training of alcoholism counselors and other alcoholism workers.

PAGE 13

5 Rationale for the Approach The approach of this study is descriptive and, in a sense, historical. It is descriptive in that the emphasis is on delineating facts and characterics of a population of interest (Isaac & Michael, 1971), and historical in that the immediate population of interest is composed not of persons but of documentation and interpretation of events already past. The descriptive-historical approach has some disadvantages.^ Compared to experimental approaches, the data used are secondary in the sense that the author is not working directly with first-hand sensory data or such sensory data enhanced by instrumentation. Hence, descriptive-historical studies may be seen as less empirical than experimental studies. The conclusions of such descriptive-historical studies may be viewed as more subjective and less verifiable than those reached via "brass instrument" approaches. The corresponding advantage of the descriptive-historical approach in the case at hand lies precisely in the fact that so many empirical, if not purely experimental, data exist without organization into meaningful Gestalts . Information is not the same as knowledge, and knowledge is sorely needed in the field of alcoholism. Human beings gather data; computers store but do not integrate data. For data to become useful, it is necessary that they become integrated, organized under categories relevant to human interests, which interests in the final analysis determine what is "useful." Some "subjectivity," in the sense of value-informed The remainder of this section, "Rationale for the Approach," represents the current writer's own thoughts.

PAGE 14

6 choices which constitute organizing principals for the integration of information, is a necessity if information is to become useful. Thus, while the conclusions of a descriptive-historical study are more "subjective" than those of other approaches, that very "subjectivity," as described above, renders the work useful, albeit a price is extracted in terms of the lowered verif lability by independent observers. Whatever is lost in empiricism may well be regained in pragmatism. A second disadvantage of the descriptive-historical approach versus a more empirical approach is that the former has less formidable safeguards against author bias. In the present instance, one of the author's major biases is a preference for the practical; he is in search of that which works. This interest is not an unworthy one. Why, after all, is research undertaken if not finally to lead to workable solutions to significant problems? Perhaps the most weighty disadvantage of the descriptive-historical approach to the problem at hand lies in the nature of the problem. It is possible that the materials to be studied are so diverse in method, focus, assumptions, and treatment that they simply are not amenable to summarization and integration in a manner that is both meaningful and logically valid. The counter to this disadvantage is that the need is so great that some attempt to address the problem is justified, even if ultimately it should fall far short of desired success. A descriptive-historical method of studying research on alcoholism treatment outcomes has the advantage of being responsive to the need for a sort of Baedeker's Guide to Researchland. There is a finite but huge number of paths that might be taken (or hacked out) in the wilderness of potential alcoholism research. If the present writing can identify only

PAGE 15

a few demoQstrably dead-end trails and one or two routes that may lead to the useful, the project will have been worthwhile. While the "Baedeker Factor" points to a clear advantage of the descriptive-historical approach, the approach has also an inherent disadvantage for such purposes. The author might through bias, ineptitude, or ignorance miss the mark and encourage ultimately fruitless lines of exploration or inappropriately dismiss as not worth the effort approaches which could in fact yield valuable results. The descriptive-historical approach has fewer inherent safeguards against error due to bias than more rigorously empirical methods--although no known methodology is proof against determined ineptitude or ignorance. For all of that, these risks are counterbalanced by the facts that the author's conclusions will be based not only on bias but also on data in the form of published research, that alcoholism researchers are not famous for being easily discouraged from following their own bents, and that a Baedeker is needed and will be subject to critical review which could reveal its errors. The current study makes minimal use of statistical analysis. The decision to work in this way has the disadvantages of rendering the findings less precise than they might otherwise be and of overlooking subtle differences and relationships visible only through more sophisticated statistical analysis. However, the primarily non-statistical summary and integration undertaken here is more compatible with the scope of the author's intent and capabilities. The reader should keep in mind that this study is intended to be of practical value for front-line alcoholism workers, many of whom do not hold college degrees and have had no exposure to more than the simplest statistics (e.g., averages).

PAGE 16

8 The commitment to practicality brings with it a limitation; the current study is not apt greatly to advance alcoholism theory per se . The advantage of the pragmatically oriented approach is that it lends itself well to improvement in research, practice, and training. Moreover, it is hoped that the observation of correlations of programs' theoretical stances with treatment outcomes may lead toward reality-based theory selection in program design and training. The current project is not exhaustive in its scope. The vast volume of published research on alcoholism and the span of human life preclude the practical possibility of any one person even reading all the material, much less synthesizing it. The focus must be narrowed in both longitude and latitude. The longitudinal limitation of this study is roughly demarcated by the years 1973-1981. One reason for this is that to examine a nine-year slice of the literature is obviously more manageable than to survey references to alcohol problems from the early Babylonion epoch to the present. Also, other writers (Baekeland, 1977, Costello, 1975 a, b, Emrick, 1973) have carried out somewhat similar studies of the relevant literature into the early 1970' s. The selection of the more recent literature also reflects the assumption that each generation of researchers builds on the work of its predecessors, screening out gross error and retaining proven fact and useful knowledge, so that more recent productions are a kind of distillate and collectively constitute at least the beginnings of integration and synthesis of what the research community has learned. The disadvantage of a nine-year longitudinal slice of the literature is the possibility of missing longer-term trends in research,

PAGE 17

9 but, since it is not a central emphasis of this paper to analyze such trends, and since literature reviewing older literature is considered herein, this is not a telling disadvantage. ' Three latitudinal limitations of note have been chosen for the current study. Firstly, it is limited to published work in English. This carries with it the possibility of missing some important reports in other languages. However, a computer-assisted literature search, per the descriptors used, turned up only five non-English language reports which appeared to be of even remote interest and which on the basis of their abstracts could be excluded as not of interest. Secondly, the current study is restricted to reports dealing with psychologically oriented treatments of alcoholism. This limitation of focus is disadvantageous in that alcoholism has biological complications, may well have biological causal components, and has often been treated as a physical illness. The literature on medical and especially pharmaceutical treatment is voluminous, highly technical, intriguing, and beyond the scope of both the author's capabilities and his primary interest. The choice to exclude nonpsychological treatment of alcoholics is predicated on four facts: (a) Beyond detoxification and the use of chemical antagonists of alcohol (e.g., disulfirara), alcoholism per se is usually regarded as requiring primarily psychological treatment, although concomitant conditions such as cirrhosis of the liver or alcoholic gastritis of course call for medical treatment; (b) Enormous amounts of money and energy are expended on psychologically oriented treatment, and present knowledge about the efficacy of such treatment is insufficient; (c) The literature on alcoholism treatment regardless of treatment orientation is too great in

PAGE 18

10 size to be encompassed in the present study; (d) The context of this dissertation is psychological, not medical, penal or religious. The mention of the words "penal" and "religious" above leads into the third latitudinal limit of this dissertation. While punitive measures have been evoked as a means of social control or "treatment" of alcoholics, to consider this aspect of response to alcoholism would go beyond the intended scope of this study, even though such measures surely have psychological dimensions. In a similar vein, there is a time-honored tradition of viewing alcoholism as a spiritual-religious problem to which the spiritually oriented approaches have been myriad and powerful. Foremost among those spiritually oriented approaches, certainly in the twentieth century, has been the fellowship of Alcoholics Anonymous. There is no doubt that AA may properly be described as a psychological treatment (broadly defined) of alcoholism. However, Alcoholics Anonymous as a treatment will not receive major attention in the present work for three reasons. (a) AA is infrequently treated as a sole treatment modality in the research literature, partially because a group whose cornerstone is anonymity is very difficult to study scientifically, (b) While many treatment programs encourage client exploration of and involvement with the principles and program of AA, well defined AA tradition prohibits that there should be a formal, institutional AA treatment program extant anywhere save in the voluntary and anonymous fellowship of AA. (However, in instances in which formal, institutional treatment program researchers reviewed herein have reported AA attendance as a client or treatment variable, such data are

PAGE 19

11 included in this study); (c) A study of the treatment outcomes of Alcoholics Anonymous would constitute in and of itself a monumental undertaking far beyond the resources and primary interest of the author. Hence, while AA is seen as a potent force in the psychologically oriented treatment of alcoholism, AA per se must be by and large excluded from consideration herein. Research Questions The general research question addressed in this study is: "How have the outcomes of psychologically oriented alcoholism treatment programs been studied and what has been learned?" The more specific research questions are six in number. (a) What client, treatment, and outcome variables have been studied; (b) To what extent have theoretical stances of treatment programs been taken into account as variables contributing to outcomes; (c) What are the ways in which client, treatment, and outcome variables have been measured; (d) What research models have been used; (e) What research models have yielded significant results; (f) What interactions of client, treatment, and outcome variables have been observed? Definition of Terms The history of attempts to define alcoholism has been fraught with ambiguity, frustration, conflict, and some might say, failure. Definitions abound, their number sometimes seeming closely to approach the number of persons using the term, but no one definition has ever captured the loyalty of all concerned parties, and perhaps not even of a majority. The lack of a firm, universally accepted definition is no minor inconvenience but has at times constituted a serious impediment to understanding and treating the condition (Keller, 1960). In a

PAGE 20

popularized discussion of alcoholism, Johnson (1973, p. 1) characterizes alcoholism as a disease which "involves the whole man: physically, mentally, psychologically, and spiritually" and which is "primary, progressive, chronic, and fatal." It should be noted that Johnson founded and guided the development of The Johnson Institute in Minneapolis, one of the more prestigious alcoholism treatment centers in the United States, and that his definition approximates a widely held conception of alcoholism. ' A more formal definition is offered by a source of equal or greater prestige, the National Council on Alcoholism/ American Medical Society on Alcoholism: Alcoholism is a chronic, progressive and potentially fatal disease. It is characterized by tolerance and physical dependency or pathological organ changes, or both — all the direct or indirect consequences of the alcohol ingested (DeLuca, 1981, p. 36). The most recent (1977) World Health Organization definition of the syndrome of alcohol dependency stresses the "compulsion to take alcohol on a continuous or periodic basis in order to experience its psychic effects, and sometimes to avoid the discomfort of its absence" (DeLuca, 1981, p. 36). * . There exists a wide range of definitions-by-example of alcoholism, from any use of ethanol as a beverage (Gordon, 1958) to the chronic intoxication of the stereotypical skid-row wino. Since personnel who make the diagnostic judgments which determine the inclusion or exclusion of subjects for research reports range from non-degreed para-professional counselors to rigorously trained research scientists, it is reasonable to believe that research reports span a correspondingly wide range of definitions of alcoholism. Moreover, Keller (1960) argued

PAGE 21

13 convincingly that there is no escape from the use of the diagnostician's subjective judgment in diagnosing alcoholism. For purposes of the present writing, it is necessary to recognize that the literature surveyed neither uses a single definition nor restricts itself to a single set of diagnostic criteria. Therefore, however circular and semantically unsatisfactory it may be, in this study, the word "alcoholism" refers to those conditions believed by the various diagnosticians represented in the several studies to be present in persons treated for alcoholism. "Alcoholic," herein, refers to a person who "has alcoholism." More simply, in the present writing, "alcoholism" is whatever is so named by an investigator and an "alcoholic" is whomever an investigator designates as such. Whenever investigators have indicated which definition of alcoholism was used for diagnosing subjects, that fact will be noted. As used herein "psychologically oriented treatment" refers to actions undertaken with the intentions that the client: (a) cease and desist from, or exhibit a lower frequency and/or intensity of, maladaptive behaviors associated with alcoholism and (b) exhibit a higher frequency and/or intensity of adaptive behaviors. "Behaviors" in this context is broadly construed to include such constructs as attitudes, thoughts, feelings, beliefs and values, whether or not manifest in "observable behavior." Excluded from this category are treatments which have no specific psychological orientation, such as incarceration, general medical treatment, or the administration of non-psychoactive pharmacological preparations.

PAGE 22

14 Overview of the Remainder of the Study The remainder of this study is organized in four chapters. Chapter II consists of a review of the related literature on reviewing and analyzing alcoholism treatment outcome literature. Chapter III sets forth the methodology used herein. Chapter IV contains the summary and analyses of the data examined. The final chapter is a discussion of the results described in Chapter IV, including answers to the research questions, conclusions and implications and recommendations for theory, research, practice, and training.

PAGE 23

CHAPTER II SURVEY OF RELATED LITERATURE Support for the Delineated Problem The problem Co which this study is addressed has been outlined in the preceding chapter in three propositions, which are restated and supported below from relevant literature. Size of Alcoholic Population First, there are many alcoholics in the United States — and, of course, in other nations, but the present study is focused primarily on this country. A highly regarded resource for information on the prevalence of drinking problems in the United States is the series of special reports to the U.S. Congress on alcohol and health from the Secretary of Health and Human Services. The latest such report (DeLuca, 1981) describes a pattern of increasing apparent alcohol consumption in the U.S. from two gallons per capita per year in the late 1950' s to 2.5 gallons by 1970. Thus, in a single decade there was a 25% increase in apparent per capita consumption. It should be noted that these figures are in terms of ethanol , which in the report is used synonymously with the term absolute alcohol or pure ethyl alcohol regardless of whether it was consumed in its more diluted forms (beer, wine) or in its less diluted form (distilled spirits). Apparent consumption rates were derived by dividing the number of gallons of ethanol sold in the form of alcoholic beverages by the number of persons in the population 14 years of age and older, with the assumption that what was sold as a beverage was consumed as a beverage. 15

PAGE 24

16 By 1978, apparent consumption stood at slightly more than 2.7 gallons per year per person aged 14 years or more. By 1981, apparent daily consumption averaged one ounce of ethanol (roughly two drinks) per person 14 years of age or older; the annual apparent consumption figure would be 2.85 gallons. However, one-third of the adult population reports not drinking alcohol at all, forcing the conclusion that the mean daily ingestion is higher (about 1.5 ounces or three drinks) for those who do drink. Even this average can be misleading, since a small segment of the adult population drinks much more than the average while the majority drink far less. The report offers an estimate that approximately 11% of the adult population consume about half of all alcoholic beverages sold. The estimate that 11% of the drinking age population consume one-half of the beverage alcohol sold is especially interesting in light of the conclusion in the same Fourth Special Report to Congress that approximately one out of ten adult American drinkers are likely to be either alcoholics or problem drinkers at some time in their lives (DeLuca, 1981). Taking as a conservative estimate that 100 million Americans drink alcohol, and using a low cutting point for problem drinking or alcoholism, it would be concluded that about twenty million fall into the alcoholism or problem drinking category, while a much higher cutting point would still yield an estimate of ten million alcoholics or problem drinkers in the United States. Even if one presumes that heavy consumption, problem drinking, and alcoholism are overreported, one is still led to the unavoidable conclusion that there are in the nation large numbers of persons described as alcoholic.

PAGE 25

17 Extensive Expenditure of Resources on Alcoholism Treatment The second proposition to be supported in this chapter is that there are vast resources spent in alcoholism treatment. The evidence that great amounts of human and financial resources are expended in the treatment of alcoholics is even clearer than the evidence supporting the existence of large numbers of alcoholics, and further supports the assertion that there are many alcoholics. Reference has already been made above (Chapter I) to the 1979 NIAAA Directory's listing of over 9,000 alcoholism treatment programs. In surveying that directory, the author found that the majority of private alcoholism treatment facilities personally known by him to be extant in 1979 or earlier were not listed. One must conclude that there are many more than 9,000 such facilities. It should be noted that the $795 million 1979 expenditure figure for alcoholism treatment mentioned in Chapter I was derived from only 4,073 alcoholism treatment units which reported funding information in the 1979 National Drug and Alcohol Treatment Survey (DeLuca, 1981). Since fewer than one-half of the known alcoholism treatment units reported, it is clearly the case that more than $800 million is being spent annually on alcoholism treatment in the United States. Paucity of Knowledge About Treatment Outcomes The third proposition stated in the delineation of the problem at hand is that too little is known about the effect of the efforts to treat alcoholism. The problem is not primarily that too little alcoholism research has been done and reported. Rather, the difficulty has to do with the fact that so much has been reported and so little synthesized.

PAGE 26

18 To determine the accuracy of his impression that the research literature on alcoholism is indeed increasing rapidly in volume, the current writer queried a computerized data retrieval system knovm as Psych Info , which accesses materials referenced in Psychological Abstracts. The question put to the system was: "By year, how many articles on alcoholism treatment are in the Psych Info data pool?" The printout received included all articles so categorized from the 1967 inception of the system to the query date, September 21, 1981. The information yielded, plus derived percentages, is incorporated in Table 1. Table 1 Alcoholism Treatment Articles in Psych Info Projected no. No. alcoholism % Citations on alcoholism treatment Total alcoholism treatment Year citations citations treatment articles 1981 6 1,518 .4 (195) 1980 48 16,092 .3 (195) 1979 98 23,447 .4 (195) 1978 195 30,366 .6 1977 161 30,029 .5 1976 147 28,258 .5 1975 144 27,546 .5 1974 159 28,350 .6 1973 148 25,753 .6 1972 107 22,658 .5 1971 99 22,141 .4 1970 76 20,034 .4 1969 89 21,142 .4 1968 109 21,032 .5 1967 81 18,115 A (TOTALS) 1,714 336,481 (mean %=.5) Average number alcoholism treatment articles per year = 114 Average total citations per year = 22,432

PAGE 27

19 The increase in the number of articles on alcoholism treatment over the 12 year period 1967-1978 is impressive. From 81 articles in 1967 the number grew to 195 articles in 1978, an increase of 240%. Not only has there been an increase in the number of such articles, but also the proportion of alcoholism treatment articles to the total number of citations in Psychological Abstracts has increased from .45% in 1967 to .64% in 1978. It is inferred from the total citation figures in Table 1 that for the years 1979-1981 not all of the citations in Psychological Abstracts had been entered in Psych Info , so that the listed percentages of alcoholism treatment articles for those three years ought not to be regarded as trustworthy. Therefore, both the total number of citations (30,366) and the percent of alcoholism treatment articles (.64%) for 1978 were chosen as constants to yield a projected 195 alcoholism treatment articles per year for the years 1979-1981. By conducting some manual literature search and accessing the computer data pool of NIAAA via a computer search, the author discovered that in regard to alcoholism research Psych Info is exhaustive neither in its descriptor system nor in its scope of publications. It is reasonably concluded, then, that the literature on alcoholism treatment since 1967 far exceeds the 1,714 citations turned up by the single query of Psych Info . Further corroboration of the fact that the literature on alcoholism research is voluminous and expanding is found in Moll and Narin's (1977) effort to characterize that body of literature. They found that alcohol research publications doubled in the seven year period (1967-1973)

PAGE 28

20 studied. In 1967, 1,134 articles were noted. By 1970 the number was 1,220, and by 1973, 2,348, with the publications scattered through more than 1,000 different journals. Moll and Narin categorized the literature in three sectors: biomedical, biosocial, and psychosocial. Their psychosocial sector would seem to include the majority of works which in the present study would be classed as dealing with psychologically oriented treatment of alcoholism. While that sector occupied some 14% of the literature found in their search, the category was not exclusively restricted to treatment articles, so that it would seem inappropriate to conclude that the scope of the literature of interest here was as large as 14% of their totals, i.e., 160 for 1967, 170 for 1970, and 328 for 1973. Nevertheless, it is reasonable to hold that the volume of alcoholism treatment literature is even greater than that implied by Table 1 above. Thus far it has been demonstrated that there are a significant number of alcoholics, that they are treated at great expense, and that there is a large and rapidly growing body of research on that treatment. The remaining assertion in the delineation of the problem is not so easily demonstrable, because it is the negative proposition that there is no definitive synthesis of the research on alcoholism treatment. "Empirical proof" of a negative proposition is, in the strictest sense, impossible, since it would require an exhaustive inspection of all of the relevant data along with proof that the inspection was exhaustive. Hence, it must serve to state here that among the literature surveyed by the present author, there have been found only four recent major reviews of treatment outcome reports, and no one of them appears to have been accepted by the alcoholism research community as a definitive synthesis.

PAGE 29

21 The earliest of the four reviews is that of Hill and Blane (1967), in which they examined 49 studies published in the United States and Canada from 1952 through 1963. The focus of interest in their survey was on the methodological aspects of the studies reviewed, with attention primarily to the designs of the studies and the methods of reporting findings. They found that the various authors represented in their sample generally failed to meet one or more basic requirements of scientific evaluative research and that many reported their findings in ways that were vague, incomplete, or confusing. Hill and Blane did not in their article make statements attempting to set forth a synthesis of the findings of the research evaluating psychotherapy with alcoholics. Indeed, given the breadth and cogency of their methodological criticisms, one would feel safe in deducing that they would maintain that no nontrivial synthesis could be made of those findings. Emrick (1974, 1975) reviewed studies published in English from 1952 through 1973 which reported patient outcomes of psychologically oriented alcoholism treatment. He located 384 such studies published over the 22 year period. With regard to outcomes following treatment he concluded (Emrick, 1974) that approximately 36% of patients were abstinent during follow-up and about 5% were controlled drinkers, so that the abstinent or controlled drinking segment constituted roughly 40% of the sample. Using other categories, he found that while less than 20% were totally abstinent, about one-third were improved although not necessarily totally abstinent or controlled; two-thirds were at least somewhat improved; one-third were unimproved, and between one-tenth and one-twentieth were in worse condition. While these findings may appear to be so general as to be trivial, something

PAGE 30

22 else that Emrick noted in process of searching them out is of significance. Emrick (1974, p. 533) established statistical guidelines for evaluating the commonness of results, specifically, that for rates of abstinence, abstinence-or-controlled-drinking, muchimproved, somewhat-improved, much-or-somewhat-irap roved, total-improved, total-unimproved, and deteriorated, specific percentage ranges can be used to judge with objectivity whether the results of a given study are common or are especially noteworthy as being unusually high or low. Later, Emrick (1979) analyzed 90 studies of randomized controlled trials of alcoholism treatment published between 1952-1978, seeking evidence for the relative effectiveness of various treatment approaches. He stated that his most salient findings were (a) when nonbehavioral treatment is applied to heterogenous groups of alcohol abusers, its effectiveness is not increased by rendering more than brief care and (b) certain behavioral approaches have been shown to be relatively effective in diminishing problem drinking. While Emrick' s work is useful in that it explicates a number of issues vital to research on alcoholism treatment and provides some clear guidelines for future research, it does not constitute a definitive or universally acceptable synthesis. Costello (1975 a) collated 58 research reports published between 1951-1972 on outcomes of alcoholism treatment over a one-year followup period. He categorized outcomes of each study in terms of case fatality rate, problem-drinking rate, success rate, and followup unavailability rate. By a hierarchical grouping method he clustered the studies in groups of maximum multivariate similarity. He noted some similarities

PAGE 31

23 within clusters of studies characterized as having good outcome profiles and some consistencies within those groups of studies described as having poor outcome profiles. Following the same procedure, Costello (1975 b) collated the findings of 23 studies with two-year followup published between 1952-1972. His results were similar to those of his earlier effort (1975 a) and suggested to him some general "baselines" for measuring treatment effectiveness. He proposed that overall treatment program goals should be relative to the expected outcomes for various prognostic subgroups of patients as those prognoses may be expected to interact with the varieties of treatments which each treatment program is able to offer. Baekeland (1977) undertook a more general approach in his critical review of English-language literature for the years 1953-1972 on evaluation of treatment methods in chronic alcoholism. (The present writer has found no way to account for the rather late publication of this survey.) Having reviewed a 20 year segment of the literature, Baekeland concluded that (a) multifactorial outcome measures are superior to abstinence alone as treatment success criteria; (b) patient, rather than treatment, factors play a dominant role in outcomes; (c) no great differences appear in the effectiveness of different kinds of treatment regimens. Baekeland 's survey and conclusions are helpful, and all three conclusions mentioned above are frequently reflected in research carried out since his publication. However, there is a need to discover and synthesize what has been happening in the research literature more recently and what it means for the future.

PAGE 32

24 The present writer judges that while the reviews by Hill and Blane (1967), Emrick (1973, 1974, 1975), Costello (1975 a, b) , and Baekeland (1977) are both admirable and useful; they do not constitute, individually or collectively, a definitive or universally accepted synthesis of the findings on outcomes of alcoholism treatment. Furthermore, since he has found in surveying the literature no reference to any other work that seems to be so regarded, he concludes that there is extant no definitive synthesis. In summary, the literature cited in this section supports the propositions which bound the problem to be addressed herein. Support for the Need for the Study The need for the study was delineated in the preceding chapter in five statements of contingency. Support from the literature is provided for those statements below. Need to Identify Fruitful Avenues of Approach It was stated in the first chapter that there is a need to identify the more (and less) fruitful avenues of approach to the study of alcoholism treatment. More explicitly, in order for research on alcoholism treatment outcomes to go forward in useful fashion, it is important to know what questions to ask of the phenomena being studied. "What are the fruitful avenues of approach?" may be reduced to "What sorts of questions are most likely to yield significant and useful information?" When one begins to probe for an understanding of whence research questions arise, it soon seems obvious that scientific objectivity is somehow inextricably related to subjectivity of what could be termed an artistic nature. In alcoholism research, as in other human science disciplines, there is an artistic struggle to discern the

PAGE 33

25 concrete and human in the intangible and abstract statistic (Keeley, 1979). There is the quest for that knowledge which will help , and thus the valid, primitive reason for research both gives research its direction and sows the seeds of potential error. The human interest of the researcher gives limits to the research questions and thus at the same time renders the endeavor possible (because one cannot ask everything and get anywhere) and biased (because if one asks question A, one may miss the valuable information that would be gained by asking question B) . The paradoxical necessity of researcher bias in choosing avenues of approach is especially relevant to alcoholism research. Pattison, Sobell, and Sobell (1977 , p. 36) have claimed that the traditional concepts of alcoholism distort treatment outcome evaluation. They assert that the traditional concept is that alcoholism is a unitary phenomenon, discontinuous with normal drinking patterns, frequently marked by an apparently irresistible craving for ethanol and loss of control of drinking, representing a permanent and irreversible condition, and constituting a progressive disease which proceeds on an inexorable course through a distinct series of phases. Such a conception, of course, dictates to a significant degree what questions the researcher will ask about alcoholism treatment. It may lead to an oversiraple classification of treatment outcomes into the dichotomous grouping of "drunk" or "sober," with no attention given to intermediate drinking outcomes. Fortunately, many researchers have gone beyond the simple dichotomy; unfortunately, they have often used vague and arbitrary categories such as "occasional slips" or "improved drinking" rather than explicitly defined and quantified classes of outcomes. One

PAGE 34

26 result is that direct comparisons between outcome studies are difficult and of doubtful worth (Pattison et al., p. 36). Moreover, Jacobson (1976, pp. 15-16) has pointed out that there is evidence that alcoholism is not a unitary clinical phenomenon but a complex multidimensional problem, and that to attempt to measure the problem in terms of the presence or absence of a single common symptom may lead to faulty understanding. He advocates the use of the concept of "alcoholisms" which might be identified and amenable to treatments specific to each. Whether or not his notion will prove useful, it is clear that he understands and takes serious account of the fact that the researcher's conception of alcoholism is a crucial determinant of the questions to be put to the phenomenon and its data. While much research on the effectiveness of alcoholism treatment has been published, reviews of the literature (Baekeland, 1977; Emrick, 1973, 1974, 1975; Hill & Blane, 1967) afford no clear consensus with regard to expected rates of relapse or successful treatment (Gottheil, Thornton, Skoloda, & Alterman, 1979, p. 91). Gottheil et al. (p. 92) note that the reported results of treatment vary greatly between studies and may generally be adjudged good or bad depending whether one's criteria are stringent, such as absolute and lasting abstinence, or more lenient, such as decreased drinking or improved psychosocial functioning. They stress the fact that while the varying stringencies of outcome judgment criteria affect statistics, they do not really alter the outcome for the patient, and that, nevertheless, the controversy over abstinence versus moderate drinking as treatment goals is no mere semantic debate. The opposed stances are representative of theoretical postures based on real world experiences and having potent implications

PAGE 35

27 not only for treatment goals and methods but for the kind of research that should be undertaken. The present author would add that the theoretical position of the researcher also implies or even dictates the research approach that will be taken. While it is true that the traditional conception of alcoholism holds sway over most alcoholism treatment personnel and many researchers (Pattison at al., 1977, pp. 37-39), new conceptions are being developed (Jacobson, 1976, Pattison et al., 1977), and each tends to generate approaches to research. Reviewers cited above (Hill & Blane, 1967; Emrick, 1973, 1974, 1975; Baekeland, 1977) have stated or implied in one way or another that research reports on alcoholism treatment outcomes are so varied and often inexplicit in their approaches that cross-study comparisons are difficult to achieve at best, and that the results of such comparative efforts are problematic. Hence, the present author concludes that the literature reviewed herein supports the contention that there is a real need for clarification of research approaches taken and for identification of the more and less fruitful avenues of approach to research on alcoholism treatment outcomes. Need for Identification of Productive Research Models That productive research models are needed is a proposition with which few would contend. However, it is important that the question, "Productive of what?" be addressed. By "a productive model for research" in the present context is meant a way of studying and reporting on the phenomena of interest which yields clear and significant results. "Clear" in this context is not far from the ordinary dictionary meaning of the term. Clear results are those which are relatively unambiguous and understandable. "Significant" here is

PAGE 36

28 used with a pragmatic connotation to refer to those results which are useful for guiding and improving the processes of treatment. "Statistical significance" is included in the notion of significance here insofar as statistical significance constitutes a trustworthy indicator of real world validity and reliability. Of the four reviews of the literature on psychologically oriented treatment of alcoholics referred to thus far, the work by Hill and Blane (1967) is the most attentive to the issue of choice of research model. Their contention was that any effort in evaluative research should attempt to meet some specific basic requirements of scientific behavior. The basic requirements which they set forth include the use of controls, either matched control and treatment groups or random assignment of subjects to treatment or control conditions, careful selection and definition of the behavior to be evaluated, the use of reliable measurement instruments and methods, the establishment of pretreatment baselines, and the application of the same measures before and after treatment (Hill & Blane, 1967, p. 77). These standards apply to conducting an evaluative study. Hill and Blane lay out other requirements for the reporting of such studies. They maintain that the setting in which the treatment took place should be described clearly; the type of therapy should be specified; descriptions of the population, sampling procedures, and characteristics of the samples should be included, along with the facts concerning "lost" subjects; the instruments of measurement used, their reliability, and the timing and manner of their administration should be specified; findings and attendant statistical applications should be presented in enough detail to permit readers to make their own interpretations of the data (p. 78).

PAGE 37

29 The fact that Hill and Blane were so careful to describe requirements for conducting and reporting evaluative research on psychotherapeutic treatment of alcoholics indicates their belief that the issue of research model is a vital one. What they discovered about the selection and use of research models was, by their standards, disappointing. In reviewing 49 studies published 1952 through 1963 in the United States and Canada, they found only two which were prospective (Hill & Blane, p. 78) and none which perfectly met their criteria for good scientific research. In their sample 43 articles used the patient as the patient's own control, most of them inadequately. None of the studies made use of a nontreatment control group. None controlled for motivation, although Hill and Blane (1967) regarded it as the one single factor most necessary to control. No study reviewed by them claimed to have used a representative sample of alcoholics-in-general. Sampling procedures were seldom clearly described and often not reported at all. About one-half of the studies failed to report even the gender, or number of subjects of each sex, in their sample. Only three studies used sampling procedures developed prior to treatment; in the other 46 random selection and representative sampling were precluded by the use of after-the-fact sampling procedures. Hill and Blane found serious faults in the selection and definition of criterion variables. All of the studies used drinking behavior as a major criterion; almost half used it as the sole criterion for improvement. This happened despite the fact that the notion that "a one-to-one correspondence exists between the amount of alcohol consumed and severity of problems" was deemed by Hill and Blane (p. 87) to

PAGE 38

30 represent a position then known to be untenable. The few studies reviewed which did undertake to use multiple outcome criteria were inadequate in their methodological management of those criteria. Almost all of the studies reviewed were weak on the point of reliability. The most frequently-occurring source of data was the interview, and in many instances the basic data sources were clinical interviews which had not been designed with evaluation in mind. Specific instruments such as attitude scales or other psychological tests were seldom used. On the whole, the studies were found lacking in attempts to deal with the reliability issue. For example, although all of the studies involved judgments on "improvement," only two reported interjudge reliability coefficients and all others were vague in reporting how ratings were conducted or the extent of agreement between rating sources (p. 92). On the whole, the articles reviewed failed the requirements for measurement before and after treatment. Hill and Blane found an acceptable use of pre-post measurement comparisons in only three instances. They held that because of the various authors' failure to provide sound or sufficient pretreatment information, the results of most of the studies could not easily be interpreted. While posttreatment data collection was more systematic, it frequently resulted in biases related to inadequate follow-up procedures and failures either to control for or report temporal relationships between date of admission, duration of therapy, and the interval between discharge and evaluation (p. 97). Hill and Blane found their 49 studies to be inadequate in regard to the requirements for reporting evaluative research almost as often as

PAGE 39

31 they were remiss in meeting criteria for conducting research. Most of the articles did include clear descriptions of the setting of the work described, but several made only vague reference, e.g. "psychiatric treatment," to the type of treatment. Generally their descriptions of the population and sample characteristics were less than optimum, and shrinkages of as much as 30% of the sample were frequently left unexplained and without discussion. Rarely did the studies report clearly on the definition of the evaluation criterion behavior or its measurement. Because of the above-noted inadequacies, along with the wide variations in settings and types of treatment conducted. Hill and Blane did not even attempt to make a summary regarding "improvement" (p. 100). Their decision underlines and emphatically supports the need for the identification (and use) of research models that will produce clear and significant results. Baekeland (1977) reviewed the English-language literature for the years 1953-1973 on methods for the treatment of chronic alcoholism. His conclusions were in some ways similar to those of Hill and Blane, although his emphasis of inquiry was somewhat different. Laying less stress on criticism of the studies themselves and more on the evaluation of the actual treatment reflected in the reports, Baekeland went in search of solid conclusions about the indications for various methods of treatment for alcoholism and the relative efficacy of such treatment methods. He chose to look at alcoholism treatment programmatically , in terms of inpatient versus outpatient and in terms of three currently popular single mode approaches, i.e.. Alcoholics Anonymous, pharmacological treatment, and behavioral psychotherapy. Baekeland reached eight major conclusions as follow:

PAGE 40

32 (a) Multifactorial outcome measures are superior to abstinence alone as success criteria; (b) The absolute minimum acceptablefollow-up interval is six months ; (c) Patient variables rather than treatment factors play a dominant role in treatment outcomes ; (d) Outpatient programs had higher improvement rates than inpatient programs; (e) The population served by Alcoholics Anonymous is markedly different from that served by hospitals and clinics, and when population differences are taken into account, it may be that clinic treatment has a higher success rate than AA; (f) Behavioral approaches appear to produce about the same results as other treatment methods, if account is taken of the fact that they are usually applied to carefully selected volunteers; (g) Because of high dropout rates and uncontrolled factors in the studies conducted, the usefulness of antidepressants or tranquilizing drugs in the treatment of alcoholism has not been fairly tested; (h) "Patients who do well on drugs, psychotherapy, or rehabilitation programs seem to have different characteristics and success rates go up with the number of treatment options given the patient" (p. 428). Support for the present writer's contention that there is a need for the identification of productive models for alcoholism research is to be found in the first three of Baekeland's conclusions as listed.

PAGE 41

33 Moreover, Baekeland's emphasis throughout on the dominance of patient variables over treatment factors in predicting outcomes leads clearly to the inference that there is a need for research models which incorporate patient variable data and demonstrate interactions of patient, treatment, and outcome variables. The intent of Emrick's 1974 article is similar to that of Baekeland's. Emrick reviewed 271 evaluative studies on outcome of psychologically oriented treatment of alcoholics published in English from 1952 through 1971 with the stated goal of "collecting data relevant to the nature and value of alcoholism treatment" (p. 523). He focused on three areas: (a) outcome criteria used to evaluate treatment; (b) how those criteria were related to one another; (c) the effect that treatment had on drinking behavior. Emrick found that a great number of outcome measures had been employed. He combined those measures into 19 criteria clusters, one of which was frequency or amount of alcohol ingestion. Dropping from consideration six studies which were single-case reports, he analyzed the remaining 265 in terms of the relationships between outcome criteria. (It is a fact worthy of note that only one study of the 265 reported data on such interrelationships.) He found that for more than two-thirds of the reports, drinking outcome related positively with other outcome criteria. He tested those relationships by applying the binomial test and found that the null hypothesis could be rejected at the .05 level or better. In short, although Emrick's primary concern was not with research methodology per se , he appears to have found the state of the art in better shape than did Hill and Blane some seven years previous. They

PAGE 42

34 expressed concern about researchers' using total abstinence as the sole criterion on the drinking dimension (Hill & Blane, 1967, p. 85). Finding as he did that many evaluators used more than the abstinent-not abstinent dichotomy in measuring drinking outcome, Emrick suggested that the concern of Hill and Blane, and, by extension, that same concern expressed later by Baekeland (1977) seemed more theoretical than actual (Emrick, 1974, p. 534). From a methodological standpoint, Emrick' s most important contribution coming out of his question about the impact of treatment on drinking behavior was his devising a means (p. 533) for judging objectively whether the drinking outcome results of a new study are normal or especially worthy of note as being uncommonly high or low. That Emrick was able to make sense of a highly varied mass of data contained in 265 reports is a tribute to his ingenious and patient re-analysis and meta-analysis of the reports, not to the methodological excellence or homogeneity of the reports themselves. The fact that his analysis had to be undertaken almost from the ground up is further testimony to the need for alcoholism researchers to be able to identify and use research models which yield easily understood results. The following quotation from the introduction to a careful, exhaustive, and controversial research report, The Course of Alcoholism (Polich, Armor, & Braiker, 1981), popularly known as "the second Rand Report," can serve as a summary of the support reviewed literature provides for the need for productive models of research. A great deal of the existing evidence on the extent of change in alcoholism comes from follow-up studies of clinical populations. This literature, which is now so extensive that the careful review of it has become almost a profession in itself, does not readily lend itself to generalization. . . .Comparison across studies is perilous

PAGE 43

35 because of the lack of explicit measurement procedures and the frequent use of broad clinical judgments of what is "improved," "unimproved," etc. Added to this are the numerous methodological deficiencies to which applied research is usually subject, such as the use of convenient samples rather than random ones, failure to measure all sample members at follow-up, and lack of randomization or statistical control in analysis, (pp. 7-8) Need to Identify Theoretical Stances Associated with Specific Treatment Outcomes It is the objective of this section to demonstrate that there is support in the literature for the notion that there is a need to identify those theoretical positions which appeared to contribute to desirable or undesirable treatment outcomes. In order to accomplish this, an attempt is made below to show that the literature supports the premises that theoretical stance is an important variable in determining treatment outcome and that too little is known about the relationship between the theoretical positions of treatment agents and the outcomes of treatment. From the bare fact of their having published a lengthy book entitled Emerging Concepts of Alcohol Dependence , Pattison, Sobell, and Sobell (1977) imply their strongly held belief that, in the field of alcoholism treatment, theory makes a difference. They specifically assert that the traditional model of alcoholism has dictated much of the design of alcoholism treatment services, determining that treatment should make alcoholics "aware of their permanent physiological abnormality and the necessity for them to be permanently abstinent" (p. 3). Furthermore, they claim, the traditional theory contains the implication that treatment should have as its main concern those problems which result from learning to live without alcohol.

PAGE 44

36 Linsky (1972) states that in the case of deviant behavior such as alcoholism the theories of etiology will govern the strategies of intervention. This viewpoint is somewhat' parallel to the concern of Pattison (1979) that while progress in therapeutic methods depends upon the development of new models of the illness, alcoholism models have been loosely structured and their associated epidemiological definitions ill-formed. While it certainly appears to be the case that theory may well be expected to have a potent influence on the processes and results of treatment, the present author has not found the theoretical stance of treatment personnel considered as a weighty variable in the reviews of Baekeland (1977), Costello (1975 a, 1975 b) , Emrick (1973, 1974, 1975, 1979), or Hill and Blane (1967) cited earlier as major surveys of the literature on treatment outcomes. This may be because the studies examined did not report clearly on this aspect. In any case, the absence of systematic consideration of theoretical posture as a variable suggests the need for such consideration. Need to Identify Combinations of Client and Treatment Variables Predictive of Positive Outcomes In this section, the author will set forth support from the literature for five propositions. First, alcoholism clients vary on several dimensions. Second, treatments for alcoholics vary in a number of ways. Third, there is variation in treatment outcomes. Fourth, some kinds of clients do better in some kinds of treatment than in others. Fifth, to work toward means of matching client and treatment is reasonable and needful. Emrick (1973, p. 23) shows that researchers have reported data on at least the following list of patient variables: age, term of problem

PAGE 45

37 drinking, gender, marital status, employment status, level of vocational skill, socioeconomic status, arrest record, prior treatment, type of alcohol problem, and psychiatric diagnosis. All of these items were noted by Emrick to have been thought by at least one researcher to relate in some way to alcoholism treatment outcome. Polich et al. (1981, p. 132) examined relationships between outcomes and the following classes of subject characteristics at admission: level of alcohol dependence symptoms, age, social stability, socioeconomic status, previous alcoholism treatment, and ethnicity. Pattison (1979, pp. 137-139) asserted that alcoholic populations may vary in personality structure, social class, gender, and ethnicity, all of which variables should be examined for interactions with treatment variables and outcome variables. He also noted (p. 205) that alcoholics entering treatment vary in terms of areas of disability, degrees of impairment, potentialities for change, and individual preferences regarding goals and methods of treatment. There is variability in the alcoholism syndrome itself, and hence in the manifestations of that syndrome in the individuals who present for treatment. Pattison (1979, pp. 134-137) points out that there appear to be multiple subtypes of alcoholism composed of complex sets of drinking, social, and personality variables. Moreover, the course of the syndrome is variable. Individuals may move in and out of symptomatic drinking or may exhibit a linear, progressive worsening of problem drinking. It is possible for the severity of problematic drinking to remain constant and nonprogressive. The processes of remission or progressive deterioration may significantly vary with time, place, and circumstance. Pattison held that for optimal matching of

PAGE 46

38 treatment to client, some or all of these complex sets of variables must be taken into account. Given that alcoholism is a highly variable phenomenon and that alcoholics vary widely on several dimensions, it should come as no surprise that treatment for alcoholism is also varied. Polich et al. (1981, p. 135) point out that whereas "treatment" in an experimental study may be expected to be well-defined, unitary, and explicitly bounded in time, actual clinical treatment in its naturalistic environment has none of these characteristics. Treatment varies as to duration; Emrick (1973) recorded outpatient treatment ranging from four sessions to ten years and inpatient treatment varying between one day and three months. There are, of course, inpatient versus outpatient treatment settings to be taken into account (Baekeland, 1977; Pattison, 1979). Treatment varies in terms of coerced versus voluntary, the stage of alcoholism at which clinical intervention is made, and the behaviors targeted as treatment goals (Pattison, 1979, pp. 155-156). There are variations in the theoretical postures and ideologies held by treatment programs and differences in treatment personnel as to degree and kind of preparation, ideological position, preferred techniques, and prior personal experience of alcoholism (Pattison, 1979, pp. 164-169). Individual differences among personnel contribute to differences in definition of treatment goals including predilections for abstinence, "social drinking," "attenuated drinking," "controlled drinking," and "normal drinking" (Pattison, pp. 194-202). Techniques for the treatment of alcoholism are numerous. Pattison (1979, pp. 156-164) discusses specific and nonspecific individual psychotherapies, group psychotherapies , family therapies, behavioral

PAGE 47

3^ therapies, and four categories of drug therapies. Erarick (1973, p. 17) found in use "a staggering array" of therapies including but not exhausted by "psychodrama; self-confrontation through videotape; conditioned aversion by electric shock, emetics and muscular paralysis; systematic desensitization; antidepressants, tranquilizers and antipsychotic drugs used in conjunction with psychotherapy; deterrent drugs (disulfirara, metronidazole, calcium cyanamide); hallucinogenic drugs; insulin shock; and individual and group insight-oriented analysis." He noted (p. 19) that these approaches were used, singly and in combination, and for varying terms, in inpatient, outpatient, halfway house, industrial, and prison settings. Since alcoholism patients and treatments vary, it is to be expected that the outcomes of treatment vary as well, and this is the case. As Pattison (1979, p. 132) has observed, "different types of alcoholics present themselves at different facilities, receive distinctly different treatments, and achieve different treatment outcomes." Polich et al. (1981, p. 101) studied treatment outcomes in terms of eight separate psychosocial variables and observed (p. 8) that the outcomes of treatment vary significantly along several continua for judging improvement. So varied are the outcomes of treatment, or at least the variables examined to measure outcome, that Emrick (1974, p. 526) found that his best effort to collapse variables into categories that might be at once parsimonious and inclusive resulted in no less than 19 clusters of outcome variables. Drinking behavior was the rubric for only one of these criterion clusters, and Emrick, in order to analyze data produced under varying classifications of drinking outcomes, formed nine subcategories under that major heading. Emrick (1973, pp. 80-81) found a

PAGE 48

40 number of statistically significant differences ( £. 05 or better) on the drinking outcome variables in studies reviewed by him. The profusion of therapeutic techniques and approaches implies a widely held belief that some treatments are better than others, or, put more cautiously, that some alcoholics do better in some kinds of treatment than in others. Emrick (1975) reviewed 384 studies of psychologically oriented treatment of alcoholics and judged that it was possible that differences in treatment outcome may have been due to a specific treatment having unusually beneficial and perdurable effects (p. 94), although the possibility could not be documented because the studies reviewed did not control for the negative experiences of patients in treatment. Smart (1978) studied 1091 alcoholics from seven treatment facilities, for each of whom full intake, treatment, and followup information was available. His purpose was to discover whether some alcoholics do better in some types of treatment than others. He concuded that: The data indicate that patient characteristics are most important in predicting outcome, that treatment is relatively unimportant and that interactions between the two are uncommon. (Smart, 1978, p. 74) Gibbs and Flanagan (1977) conducted a meta-analysis on a much larger body of data than that considered by Smart and reached quite different conclusions. They analyzed 45 predictive studies published between 1937 and 1974 involving 55 different treatment groups including some 11,350 subjects in an attempt to isolate personal characteristics of alcoholics associated with prognosis. They concluded that while some patient characteristics are of greater general predictive value than others, stable predictors across studies were not apparent. One

PAGE 49

41 possible reason for the absence of predictor stability, they held, is that certain treatments may be more effective with certain alcoholics. Pattison (1979) set forth a more positive position than either Smart or Gibbs and Flanagan, noting that persons with certain psychosocial profiles are apt to affiliate successfully with AA (p. 154). Pattison advocated the notion that the interaction of specific client variables and certain treatment variables is at least partially determinative for outcomes. Convinced of this, he argued for the need for matching client and treatment for maximum positive treatment outcomes. He maintained that the most powerful predictor of successful treatment outcomes is the matching of therapist-client values and goals (p. 169) and provided in vignette form illustrations of effective matching in four quite different treatment settings: the aversion-conditioning hospital, the alcoholism outpatient clinic, the alcoholism halfway house, and the police farm work center. He found that each of the four facilities tended to draw clients from different subpopulations of the overall alcoholic population and that each of those subpopulations held distinguishably different definitions of alcoholism, self-defined their treatment goals differently, and received different treatment from different personnel. It appears, then, that the development of schemata for matching client and treatment would be a useful enterprise. Even although he was pessimistic about the possibility of discovering a "best" treatment for alcoholism and asserted that technique variables are not likely powerful determinants of long-term outcomes of treatment, Emrick (1975, p. 95) suggested that treatment personnel should seek to match each alcoholic with the treatment setting and approach which would provide the best fit

PAGE 50

42 with that individual's own perspectives on the nature, etiology, and treatment of alcoholism. Pattison (1979) was more sanguine and forceful than Emrick on the subject of matching. He held the matching of client, facility, personnel, and treatment to be a goal worthy of expenditure of great effort. He pointed out that attaining the goal would be no small order, since most reviewers of treatment methods hold that no clear, large-scale indicators for client-treatment matching are easily derivable from the data. He added, however, that such a conclusion was not surprising in light of the fact that treatment methods have been rather indiscrirainantly applied, so that clear matching criteria could scarcely be derived from global reviews. Pattison (1979, p. 132) cited examples, however, of smaller-scale discrete research projects which demonstrated the usefulness of constructing means for matching client and treatment (Kissin, Platz, & Su, 1970; McLachlan, 1974; Pattison, Coe, & Rhodes, 1969; Pattison, Coe, & Doerr, 1973; Trice, Roman, & Belasco, 1969). Pattison's (1979) summaries of their salient points follow. The finding of Pattison et al. (1969, 1973) that alcoholics of different types enter different treatment centers, get different treatment, and manifest different treatment results suggests that some unnoticed matching by self-selection occurs naturally. Similar discoveries in other research projects have allowed the formulation of some predictors for matching (Kissin et al., 1970; Trice et al., 1969). McLachlan (1974) provided a clear illustration of the potential utility of matching efforts in his report that 77% of the patients matched to both the therapy and aftercare environments met recovery criteria, while

PAGE 51

43 those matched only to either the therapy or aftercare environment had recovery rates of 65% and 61%, and of those mismatched to both environments, only 38% were classified as recovered. Pattison et al. (1969), reporting on clientele of three different alcoholism treatment facilities, found that the treatment-successful cases at each facility differed substantially from one another. They constituted unique, differentiated subpopulations . The subpopulation of each facility improved, but the improvement was in different patterns. In a later study of the characteristics of populations presenting prior to treatment at four different treatment centers, Pattison et al. (1973) found that the subpopulations were not merely the results of the treatment undergone, but preexisted as distinct subpopulations upon their entering the various facilities. Of the writers reviewed by the present author, E.M. Pattison is the strongest advocate of the opinion that the need for workable ways of matching client and treatment is sufficiently great to merit investment of time and energy. Citing as evidence large-scale studies by Bromet, Moos, Bliss, and Wuthman (1977) and Cronkite and Moos (1978), Pattison (1979) concluded that the selection of treatment method is more crucial to treatment success than had previously been realized. The following quotation from Pattison' s 1979 article on selection of treatment modalities can serve well as a summary and conclusion of this portion of the current writing: Our current state of knowledge about treatment is still too global and imprecise to formulate exact treatment guidelines. Our measurement and evaluation methods are too crude to assess our methods accurately. Nevertheless, the accumulation of research to date does strongly suggest the value of matching subpopulations of alcoholics with the most appropriate facilities, methods, and treatment personnel, (p. 205)

PAGE 52

44 Need to Identify Pragmatically Tested Theories and Practices For Inclusion In Training The present author was unable to find in the literature reviewed explicit statements to the effect that there is a need to identify, for inclusion in the training of alcoholism workers, those theories and practices which have been demonstrated to be most effective. Hence, the following paragraphs represent inferred support for the proposition that such a need exists. There can be no doubt that persons are being trained in the diagnosis and treatment of alcoholism. The August 1981 current literature issue of Journal of Studies on Alcohol lists no fewer than thirteen recent titles related to the training of professionals or other alcoholism workers ("Current literature," pp. 565-567). The NIAAA 1981 report to Congress (DeLuca) states: "Persons involved in treating alcoholics need specific training, education, and experience for their jobs. This requirement applies to every level of the work force, and should be enforced through competency-based evaluation systems" (p. 185). The same report stresses the unmet need for adequate training and notes the development of Project Cork, an alcoholism-specific curriculum initiated by Dartmouth Medical School, mentioning with apparent approval that "Project leaders expect that upon graduation students will not only know about alcoholism but will be able to do something about it" (p. 186). The present writer contends that for students of alcoholism to be able to "do something about it" it is necessary that they be given the best possible information about what seems to work.

PAGE 53

45 Carap and Kurtz (in press) are cited in the NIAAA report as having defined six areas of job competency for alcoholism counselors, including "(1) communication; (2) knowledge of alcohol use, prevention of alcoholism, treatment, and rehabilitation; (3) evaluation and assessment; (4) planning; (5) information and referral; and (6) counseling and treatment" (p. 187). It is important to note that the latter five of the six imply a need for training with incorporates the knowledge which the present author maintains is needed. The attitudes of alcoholism workers are vitally important, and those attitudes are in part determined by the beliefs workers hold about alcoholism and alcoholics. Wolf, Chafetz, Blane, and Hill (1970) studied the attitudes of 15 physicians toward their alcoholic patients and found that the doctors tended to view alcoholism as a disorder of derelicts and to be much more hesitant to diagnose alcoholism in socially intact persons as compared to derelict patients (p. 132). The researchers also held that their physician sample preferred a strictly medical diagnosis of alcoholism to one that included dysfunction in the psychosocial sphere (p. 134). Given that both attitudes are substantially at odds with widely accepted and fairly well supported views of alcoholism, it is apparent that this sample of physicians held attitudes inappropriate to alcoholism workers. The need for inculcating more realistic and appropriate attitudes is apparent. The present author submits that if strongly supportive evidence of the real-world effectiveness of identified theories and practices were included in the training of scientist-practitioners such as physicians, it would result in attitudes and behaviors more conducive to accurate diagnosis and efficacious treatment.

PAGE 54

46 In the earlier discussion of Pattison's strong argument for careful attention to client-treatment matching (1979), it was apparent that in order for such matching to be done well, the "matcher" would need to possess considerable knowledge about which combinations would be likely to produce what results. There is again a clear inference of the reality of the training need here under consideration. The present writer examined the indices and tables of contents and extensively sampled the texts of four recent works containing materials intended for use in training alcoholism workers (Johnson, 1973; Foley, Lea, & Vibe, 1979; Schuckitt, 1979; Zimberg, Wallace, & Blume, 1978). While each work discussed theories and practices of alcoholism and alcoholism treatment, none was found to include information concerning the broad-based support or lack of support for the relative worth, in terms of treatment outcomes, of the theories and techniques discussed. This writer assumes that this lack does not represent oversight on the part of the several authors and editors of these books, but rather points to the fact that the needed knowledge was not available to them. In light of the present author's earlier argument that there is no widely acceptable synthesis of knowledge about the interrelationships of alcoholism clients, treatments (including theoretical stances thereof), and outcomes, the immediately preceding paragraphs seem almost superfluous. Since the knowledge in question has not been synthesized, it is not being included in training although it is clearly needed. Support for the Approach of the Study The approach of the current study is descriptive with an emphasis on pragmatic utility of findings. Isaac and Michael (1971, p. 18) state that the purpose of descriptive research is "to describe systematically

PAGE 55

47 the facts and characteristics of a given population or area of interest, factually and accurately." This statement fits well with the current author's intent to fulfill the need for a factual and accurate descriptive synthesis of what has been learned in research on outcomes of psychologically oriented alcoholism treatment. Somewhat similar descriptive studies in the same general subject area have been undertaken by other authors (Baekeland, 1977; Costello, 1975 a, b; Emrick, 1973, 1974, 1975, 1979; Hill & Blane, 1967). In each case it was apparent or plainly stated (e.g., Emrick, 1973, pp. 13-14; Hill & Blane, p. 76) that the number and diversity of research reports practically dictated that descriptive, synthetic studies should be undertaken. Baekeland (p. 386) saw that firm conclusions about the indications for and relative efficacy of alcoholism treatment methods were sorely needed, and part of his response to that need was to produce a critical review of the relevant literature. Further support for the appropriateness of descriptive, synthetic studies such as this current one is found in a discussion by G.V. Glass (1976) of the worth of what he termed "meta-analysis." He used the term to refer to "the statistical analysis of a large collection of analysis results from individual studies for the purpose of integrating the findings" (p. 3). While Glass's frame of reference was the field of educational research, his argument for meta-analysis is applicable to alcoholism research as well. Glass asserted that in questions of outcome research, since findings are apt to vary in such confusing fashion across context, populations, and unnumbered other factors that commonality of results is rare, it is especially important that efforts be made to organize findings across studies in a way that can bring

PAGE 56

48 order out of the confusion. Glass contradistinguished information and knowledge, characterizing the latter as organized, integrated, and, therefore, useful information. He wrote: Our problem is to find the knowledge in the information. We need methods for the orderly summarization of studies so that knowledge can be extracted from the myriad individual researches, (p. 4) As has been mentioned above, the approach of the current study includes an element of pragmatism, a primary interest in the possible and the workable. Hill and Blane (1967, p. 94) in critiquing the methodology of alcoholism psychotherapy research noted that in some instances research was weakened by the researchers' having undertaken research tasks beyond their resources, financial, temporal, or human, and that those researchers would have done better to design their studies to fit within their limitations. Hence, while the current author recognizes the need for a much more exhaustive study than the present one, he accepts the limits of practicality that Hill and Blane imply. A second facet of the pragmatic aspect of the current approach has to do with the quest for accurate information presented in a manner that will be understandable to those who may put it to use. "Those who may put it to use" refers to alcoholism workers, most of whom do not hold graduate degrees (DeLuca , 1981, p. 184; Pattison, 1979, p. 165). The implication is clear that for the knowledge sought in this study to be utilized by alcoholism workers it must be presented in a manner understandable to an audience not familiar with sophisticated statistical analyses.

PAGE 57

49 Summary The preceding review of literature has shown that there is support in the literature for the problem to which the current study is addressed, the needs which constitute that problem, and the approach of the study to fulfilling those needs. The support found in the literature was sometimes implicit and sometimes direct.

PAGE 58

CHAPTER III METHODOLOGY Overview of the Study In the context of a growing body of empirical research on alcoholism treatment there appears to be a need for synthesis of those findings which have emerged. The need for a summary seems apparent from the facts that while substantial human and monetary resources are being expended on the psychological treatment of large numbers of alcoholics, generalizable conclusions about the outcomes of that treatment are hard to come by. Hence, the purposes of this study are to identify promising approaches to and models for outcome research as well as to describe criteria for selecting theories related to positive treatment outcomes, criteria for matching client and treatment, and criteria for selection of theories and practices for inclusion in training of alcoholism workers . The rationale for the descriptive approach taken herein is that such an approach appears to be the necessary one in order to make sense of a bewildering mass of empirical findings. The present study seeks answers to one general research question and six more specific ones. The general question is: "How have the outcomes of psychologically oriented alcoholism treatment been studied and what has been learned?" The specific research questions are: "What client, treatment, and outcome variables have been studied?" "To what extent have theoretical stances of treatment programs been taken into account as variables contributing to outcome?" "What are the ways in ' • 50 '

PAGE 59

51 which client, treatment, and outcome variables have been measured?" "What research models have been used?" "What research models have yielded significant results?" "What interactions of client, treatment, and outcome variables have been observed?" Delineation of Variables The variables under consideration here fall into three basic categories: subject, treatment, and outcome. The category "subject variables" is composed of the set of facts reported for the individuals described as clients, patients, or subjects in the various research reports considered herein. Generally, "subject variable" herein refers to some fact about the subjects which pertained at or prior to their admissions to treatment. For convenience in data analyses, subject variables have been clustered in clusters parallel to those used in grouping outcome variables. Thus, in the following discussion of outcome variables, the insertion or substitution of the notion "at or prior to admission to treatment" as appropriate will translate the respective outcome variables into subject variables. Emrick (1973, pp. 31-37) found that the wide variety of outcome criteria in treatment outcome studies could logically be collapsed into 19 categories. Emrick' s 19 criteria clusters are used here to delineate outcome variables, and, mutatis mutandis , subject variables.^ The following list, taken directly from Emrick' s use of outcome criteria table, first names the outcome variable in question and then specifies the content of that variable. The author gratefully acknowledges Emrick' s contribution in developing these clusters.

PAGE 60

52', , 1 1. Drinking amount or frequency: amount of drinking, frequency of drinking. 2. Affective-cognitive: psychological test results, thoughts about treatment, cognitive-emotional responses to drugs such as LSD and tranquilizers, clinical ratings about patients' emotional responses and cognitive functioning. 3. Work situation: work status, work adjustment. 4. Further inpatient treatment: rehospitalization or hospitalization after outpatient treatment, the hospitalization being for consequences directly related to excessive drinking. 5. Home situation: relationships: marital status, marital adjustment, adjustment to family life. 6. Physical: hospitalization and outpatient treatment for physical problems, physical reactions to drug treatment such as LSD, disulfiram and scoline (a muscle paralyzer), physical reactions to alcohol after treatment (e.g., blackouts, hangovers, DT's). 7. Arrests and other legal problems: arrests, illegal activities, imprisonment. 8. Mixed: a mixture of two or more of the above criteria, the mixture not intended to cover all areas of functioning. 9. Social situation: interpersonal adjustment not specifically related to home or work. 10. Further treatment: outpatient: AA: AA attendance occurring after treatment and not during aftercare for more than 50% of the patients, assuming of course that AA was not the only mode of treatment evaluated. 11. Financial situation: amount of income, degree of responsibility for financial affairs, welfare status. 12. Global adjustment: overall functioning with no one aspect being of paramount importance and with no aspect being purposefully excluded. 13. Miscellaneous: criteria fitting none of the other clusters, such as "day-to-day functioning" and "cigarette consumption." J

PAGE 61

53 14. Drinking behavior other than amount or frequency: change in drinking habits such as locus of drinking, time of day or week drinking is done, the type of beverage drunk, and amount of time between sips of alcohol. 15. Home situation: residence: residential mobility, quality of accommodations. 16. Further treatment: outpatient: other: outpatient treatment during aftercare for less than 50% of the patients, the treatment taking place in another setting with different personnel. 17. Use of leisure time: quality of leisure-time activities . 18. Religious life: quality and quantity of religious life. 19. Further treatment: locus unclear: vague references to "other treatment." (Emrick, 1973, pp. 31-33) The category "treatment variables" consists of the set of relevant facts reported in the several studies examined herein as to the length of treatment and the setting, procedures, personnel and/or theoretical stance of the facilities to which subjects were admitted. The smaller category of treatment variables designated "setting" includes the geographic location of the facility, the locus of treatment (inpatient versus outpatient,) and the type of facility. Some "types" of facilities are VA hospital, public outpatient clinic, private aversive-conditioning hospital, halfway house, and state psychiatric hospital. A second subset of treatment variables has to do with treatment procedures. Treatment procedures described in the reports under consideration here are widely varied and include, but are not restricted to, self-help groups, family counseling, relaxation therapy, vocational counseling, assertiveness training, aversive conditioning, milieu therapy, occupational therapy, and psychodraraa.

PAGE 62

54 Inclusion of "personnel" as an aspect of treatment variables allows the discussion of the training, attitudes, and behavior of the persons administering treatment as variables which' may contribute to outcomes. Since "theoretical stance" is not easily included under either setting or personnel without ambiguity, it is considered to be a separable variable including explicit or implicit assumptions and beliefs concerning alcoholism and alcoholic persons insofar as those assumptions and beliefs contribute to the goals, procedures, and outcomes of alcoholism treatment. Description of the Population The population of interest for this study is the universe composed of English-language articles on alcoholism published between 1973 and late 1981. For all languages, the population appears to be distributed through more than 1,000 periodicals, with approximately 50% of the articles appearing in English (Moll & Narin, 1977, p. 2177). In 1973 the population numbered more than 2,300 publications for that year (Moll & Narin, p. 2167). The articles which composed this population exhibit a broad range of interests and approaches. Moll and Narin characterized the foci of interest as biomedical, biosocial, and psychosocial. The approaches taken range from hortatory essay to carefully controlled experimental study. Sampling Procedures Five criteria were chosen for including publications in this study. Firstly, publications must have been published research reports on psychologically oriented alcoholism treatment. This criterion was used to exclude works not germane to the intent of the study and, by the requirement "published," to exclude unpublished works which would have

PAGE 63

55 been impossible to obtain or not sufficiently well done to merit publication. Secondly, the publications must have examined the outcomes of such treatment. This criterion was used to preserve the pragmatic, i.e., outcome-oriented, focus of the present study. Thirdly, the reports must have considered no fewer than ten subjects. This criterion was used in order to have the data pool include only studies with relatively more generalizable findings. Fourthly, reports must have reported on outcome measures taken no less than 12 months after treatment. This criterion was used to include only studies with greater reliability and validity. Other researchers (e.g., Costello, 1975 a, b; Polich et al., 1981) have maintained that alcoholism treatment outcomes measured at less than one year after treatment are substantially less reliable than those measured at one year or more after treatment. Fifthly, the studies must have been in English and available to the author. This criterion was used in order to keep the present study within the bounds of realistic limitations. The first step in selecting the sample was to procure computer-assisted literature search printouts from NIAAA, NIMH, and Psych Info, using the nearest available approximations in each system to the descriptor-set: "the intersection of the sets 'alcoholism treatment, psychological, outcome, English language, 1973-1981."' The computer-assisted literature searches resulted in a pool of approximately 430 titles. The titles and abstracts in each printout were examined and articles whch did not meet the inclusion criteria were excluded from consideration. Exclusion per the criteria listed above reduced the pool of documents appearing to be appropriate for inclusion herein to 21. No actual

PAGE 64

56 count of articles failing to meet each criterion was taken, but the author's impression is that the rank order of frequency of reason for exclusion, from most frequent to least frequent, was as follows: were not research reports on psychologically oriented alcoholism treatment, were not outcome studies, had followup of less than one year, had fewer than ten subjects, were not available in English to the present author. The vast majority of documents excluded were rejected by the application of the first three criteria as rank-ordered above. No single-subject intensive design studies were found in either the computer-assisted literature searches or the manual literature search. Fewer than ten studies were rejected for having fewer than ten subjects. .. ''^ Original or photostatic copies of the articles appearing to meet the inclusion criteria were obtained and examined, and those which failed, on close inspection, to meet the criteria were excluded. Further manual literature search included inspection of the reference lists in articles examined and inspection of the indices of Journal of Studies on Alcohol . This search increased the number of documents to be included to 34. Description of Sample The resultant sample consisted of 34 English-language articles reporting on psychologically oriented alcoholism treatment outcome research with outcome measures taken a minimum of 12 months posttreatment on a minimum of ten subjects. Compared to the population from which it was drawn, the sample is smaller, restricted to the psychosocial sector, and further restricted by its being circumscribed by the limiting criteria mentioned above.

PAGE 65

57 Description of Research Design The design of this study is analagous to that of a descriptive study with human subjects. In the case at hand, the articles comprising the sample are analagous to subjects. The method of obtaining data is a set of questions asked about each article. The answers to those questions constituted the data to be analyzed. The data were inspected and analyzed to ascertain commonalities and differences among and between studies. Conclusions were then drawn regarding the answerability of and answers to the research questions. Description of Research Procedures The above-mentioned set of questions was developed by reading sample articles and listing the subject, treatment and outcome variables mentioned in each. The lists so produced were then combined into lists of categories subsuming the specific variables mentioned into larger classes characterized by face validity and apparent manageability. The larger classes in turn provided the categories of questions to be asked of the articles in the sample. The author and another judge, a Ph.D. psychologist with published research, administered the resulting data-gathering protocol^ independently to the same five articles from the sample. Their recorded answers were compared and yielded an interjudge agreement rate of 90.5%.^ Description of Data Analysis Procedures When the data had been gathered by means of the data-gathering protocol, they were arranged in 13 tables. The tables were devised so See Appendix A. See Appendix B.

PAGE 66

58 that the data were arranged in ways designed to provide answers to the research questions. The 13 tables are as follow: Table 2: Descriptions of Selected Outcome Studies Published 1973-1981; Table 3: Numbers of Studies Reporting on Various Subject Variables ; Table 4: Numbers of Studies Reporting on Various Outcome Variables ; Table 5: Numbers of Studies Reporting on Various Treatment Variables ; Table 6: Major Findings with Statistical Significance of £-^05 or Better in Studies Utilizing Correlational Research Models; Table 7: Major Findings with Statistical Significance of £<.05 or Better in Studies Utilizing Experimental Research Models; Table 8: Major Findings with Statistical Significance of £<.05 or Better in Studies Utilizing Quasi-experimental Research Models; Table 9: Major Findings with Practical but not Statistical Significance in Studies Utilizing Correlational Research Models; Table 10: Major Findings with Practical but not Statistical Significance in Studies Utilizing Descriptive Research Models;

PAGE 67

59 Table 11: Major Findings with Practical but not Statistical Significance in Studies Utilizing Experimental Research Models; Table 12: Major Findings with Practical but not Statistical Significance in Studies Utilizing Quasi-experimental Research Models; Table 13: Major Research Questions of Outcome Studies of Psychologically Oriented Treatment; Table lA: Major Interactions Observed. The data were then analyzed by inspection to ascertain and describe patterns and trends. Methodological Limitations The present study is limited by the inclusion criteria to a small sample of the literature on alcoholism treatment. The longitudinal limitation is circumscribed by the years 1973-1981, a limit chosen because somewhat similar meta-investigations had already been conducted on the relevant literature prior to 1973. The latitudinal limitations are those implied in the inclusion-exclusion criteria set forth above. The fact that the data under consideration herein are published studies constitutes a methodological limitation. Published studies must meet the restrictions of space of the publications in which they appear. It is likely the case that explanations or data which might be of value to the present undertaking have been excluded from published studies in order to keep within the prescribed limits of length. An example of that limitation is that few of the studies examined in the present writing set out data for individual subjects; most of the reporting is in terms of group means or other per-group data. Thus, the method of

PAGE 68

60 the present work does not allow for the analysis of similarities or differences on an individualized basis. Other methodological limitations are discussed in Chapter I above under the rubric "Rationale for the Approach."

PAGE 69

CHAPTER IV RESULTS This chapter presents data drawn from 34 studies selected for inclusion. The data are set forth in tables with accompanying explanatory text. Table 2, "Description of Selected Outcome Studies Published 1973-1981," should be read with the following explanatory statements ia mind. The data grouped under the heading "Drinking outcome" are organized to indicate which groups of subjects were categorized under drinking outcome classes established by Emrick (1974) and whether outcomes reported were uncommon by Emrick' s guidelines. Emrick' s criteria for evaluating the uncomraonness of results were established by his calculating the mean and median estimates of outcomes for a large number of studies. He selected the points one standard deviation above and one standard deviation below the mean as the points outside which outcome percentages might be judged atypical. This produced the following guidelines: Emrick noted that the range of one standard deviation on either Description of Studies Class 1 (abstinent) Class 2 (abstinent-or-controlled) Class 3 (much-improved) Class 4 (somewhat-improved) Class 5 (much-or-somewhat-improved) Class 6 (total improved) Class 7 (total unimproved) Class 8 (deteriorated) 10.5%-53.3% 19.8%-67.0% 4.7%-26.3% 0.9%-34.9% 14.4%44. 4% 47.8%-84.2% 15.8%-52.2% 0.2%20. 6% side of the mean appeared to be a reasonable range of typicality for all 61

PAGE 70

62 outcome categories since, except for the controlled-drinking outcome, estimates appeared to be normally distributed within groups. The non-normal distribution within groups of controlled-drinking outcomes appeared to make it impossible for him to derive a reasonable range of typicality for that outcome. Hence, controlled drinking is not listed as a separate outcome class, although it is included in Class 2 (abstinent-or-controlled) . In interpreting Table 2 entries under "Drinking outcome" the reader should note that the "group" column contains Roman numerals to indicate to which of the treatment groups described in the "Treatment type" column reference is made. In the column headed "class/%" the first Arabic numeral indicates into which of Emrick's drinking outcome classes the reported outcome was categorized. The second number is the within-group percentage of subjects evaluated who fell into the drinking outcome class in question. The column farthest right contains a plus-, minus-, or zero-sign to indicate whether the drinking outcome in question was, respectively, uncommonly good, uncommonly poor, or not uncommon per Emrick's guidelines. With the preceding clarification in mind, then, one would read the first entry in the "Drinking outcome" columns of Table 2 as meaning that 25% of the subjects who had received VA hospital care only (group I) were included in the abstinence drinking outcome category (class 1), and that the outcome was neither uncommonly good nor uncommonly poor. Three abbreviations in Table 2 may require explanation. "BAG" refers to blood alcohol content, a measure of the ratio of ethanol to blood in the body. "M" means "not reported" and that the relevant data were not reported. When "NR" appears in the "Drinking outcomes" columns

PAGE 71

'J. 63 it means either that drinking outcomes were not reported or that within-group percentages could not be derived from the data reported. "Ss" is an abbreviation for subjects.

PAGE 73

65 0 + 0 0 0 0 + + 0 0 0 0 0 0 o\ 00 r~00 t-VO a\ n in vo tN m ro m ro ro ro n iH i-H vo vo i-l m ro iH f-H r>i OJ iH rs H H M M M M M M M M M and H H H I-l H M M Mia 1 I tM •H (0 4J M ^ V B IS £ S 5 'g d, I cn a! to C 01 J3 CM ^ 5 i, B-l 8 ? « 5 3 g alS8--:{l3 8 0) I H (8 (0 01 -i-l y !" > US 01 _ a) — -tJ 10 0 . ^S^ g S-&.5 8 » & o 1 1 3 § c -H o I o c S 8 u-i C 8.7 01 E I 01 I _ c C o 0 -3 H JJ 01 n} P SI 1 ^ in

PAGE 75

67 o CO in 00 o (N 1 H s CM I C tn 6 fN "p 0) JJ 0 O ^ P £ m S M S-'s 5! ^^ to (0 u -u So -H W •0 5!^ C Ifl 3 M-l -P 0 c 5 S 5 (0 c o < 0 I JJ ^8 0^ (0 C ^ -H >. O 1-1 I I tn jJ 5 1 i-( (0 4J I 00 vo 5

PAGE 77

69 o o n CO 01 5 ^ -i J 00 I I >i TJ IJ ffi ra O l-i lh >u u £j ra • o 00 o J J 00 •35 z ^ 03 rn th c .-I flj 3 _ fi W a p tT> 0) c a S > (0 >, O E iij_ o\ 0 o a £ Q g — 'j::tn>-i 4J''^''^^ .H (0 >i-H (B E O 'u j: O rH JJ « 0 CM O -d "H > E 0) m Si 13 (0 I ^1 C -H M I ft CM to in >, in -> 1 . B E ra 11 P If in 2 § <4-l rH I Ul . O <0 O >1 &> IX >. fi tn (0 o _ „ in O -i^ 00 SI) 3)^ K £ in -H S ij a TJ D "0 in P $ bi u p > 3 &1-H Q •ri in JJ o 0 --I 0 c tn 9-5 O -H ^ 0*0 la 5) 0 x» 0 CP i d 3 i I's I >H C -H 4J w in o g g> in c m (U •H U > •H •-( CLi-H Ot CP >, U 0) jJ x: 4J B id H ' 8 01 -H cr>'<-' c 4J -H o &>->H nJ >»-l 8 o K e •H -H .t! -U (u a c •U 3 r-l -H (B ^ O 10 <-l to *J D<'6 3) c ^ 10 ^ 5

PAGE 78

00 in in 00 00

PAGE 79

71 I

PAGE 80

0 3 O O 0 D vo n in ^ \o N (N CM ^ HI 0 0 rH i-l — vo 0 gCM -H O » 53 CM -H C rO o 2 VI , (U 10 15 00 -^ i « -1 1^ 5 rH ii rH ..^ •iH -H rH &,H a rH gi. ..I 1. Q 03 !§ ' 3 '"^ ""^ S ! iJ T! •• ' Ql O "S ^ >-l JJ O -H -H iH jd 0) --I -H -H M IS 001-P3't3O T3 O Q'-t'H "S r-i UOUlSoOOJO •H C T) 14-1 0) O g ^ -H ^ J) x! -3 -H ^ I •H > -H 3 (u 1-1 fl 43 rH C > S Ul -H Q (U C pa J) -H a

PAGE 81

73 fM n m CO ra 3 3 H c 8 I 51 § >, o Q) 4I CTi J — p in "-I 5 ^ (0 tn Q<'-i w Q) 3 (6 >M 5> w U tT"rJ 0) o C e en *0 (N CWO "(0 ^ -H •PI H C T! 0) c 3 rH B 'a ii _ (0 ffl tn 4J 2 , II) o Q w i: I u c 43 10 O 0! S a! '.H c (0 4J &, O 0) •H £ N ui g in 0 ^ 1-1 JJ i§ 0143 (3 CO •H C ^ M -H (0 JJ 4J u] u m 0) -H m m i-H il 4J .-H 43 0) T-l 0) (B C M > ^ 2,1! C -H r< CO •H If) M 0) -H Q M XI ^ ^ 0) -I (0 M (0 5^ o -C 4J " o •H ^ £ r-1 H) Q .oi'DTd'otTimii.-ir-i •0 "o 5 p c u . o 0 ajj en rj 0) « C < 4J -o a-3 V) •a girto 00 i3 in m •do). -H (8 „ I, 3 i> "3 ^ 5 ^ .-^ o 03 o 4J ••3 a'^^l c II -^ip 5.-3 5^ I JJWTJfl.HIBQ) U-lC g ^ 05 43 IH (fl 4J 10 iH 0) 3 n S o 00 s 9 iig:^ fel-H 0) . IW E ^ T) 43 ^ -4 0 S i ? e b R 5 "a 8 o GO 5

PAGE 82

+ + O 0 0 o o I •H 0) -a so m OP tn ro 3 (1) CD O O >1 C tin >,.H 4J 0 OP •a 10 n -3 rH ^ Q, ^ in CN a g _g _g * .« . . . (0 1 1 i; i 8 > 8.: mm fo 0) (Tii-H W 10 £ >irH m +3 x: (u (u O Ol 01 M-l 1 (0 13 , S 01 C o CO aJ a) 0) m 00 o > • H H H H . 00 I 1-1 10 3 — . o CM i-i 0) £ a) I0 ui fl! w 4J CO ^ >i 'Is 11^ iJ f ^1 I -H -H a E Bi 3 _ r CO CMO 5-H0>0 (DO MOHi-I^ (Oi-HMOl OliH p S "S ^ in (0 0 0) 0) a 10 £< "2 Q ly lO a-H S o i-i 10 8? 2. 3 'I a in 4J >4

PAGE 85

77 z CO CM 0] CR I ' " •O (t> U (n D 01 ^ in 3 "p >. s y ^ i ^ 10 O Q) -H O 'O •H 10 10 4-1 r~u -H CM VW I — 4J 01 0) 0) O -H tn ^ m X) 3 (0 JJ 0) i-H •5 £ J: "8 0) >, 8 g«N OP IW iH VD cn O tf> <0 C in <*-! O i ^ 10 CO tM H 0) 01 » CO «T uj x! VO f-H CN (0 CJ ° ° 93 c y 7 w (3 O i8 re (0 4J C 10 0 > 0 w in m 04 0) 01 (0 "S '-d B* m 8i a) ^ 8 n > 4J •p , &-H cn C O 0 H l-l 0) O _ I 0) H (8 B J 'O _ -= p -H 13 S C § C c jj fo 01 E 4J •H 10 U -H 01 a u >w c o •H 3 . s s ^ (tl 01 •H o Q) ^ ^ 01 X o S ^ *j D >l4J l-l (D 'O fO TS =i fc! _ -d 171 <-( 01 & So 01 JJ C7> <0 01 01 01 (0 flj jg H 3 c 01 . i S Q, lo oT c O -5 "O r-l S rH to ^.^^ 8^ ^5 10 C Q M iH JJ 3 3 10 10 •H -o 5f! 0 a CO en w 1-3 w u z

PAGE 86

78 Idiosyncrasies in Table 2 Some of the studies summarized in Table 2 reported outcomes at one or two years after admission rather than after the termination of treatment. Azrin (1976) reported on outcomes at the end of two years of treatment. Three studies on aversion-conditioning hospital treatment (Jackson & Smith, 1978; Neubuerger et al. , 1981; Wiens et al., 1976) reported outcomes at one year after admission, which year included, for many of their subjects, return visits to the hospital. The outcomes reported in these studies should not be understood to be outcomes measured after all treatment had ceased for, respectively, two or one years . The study by Bowen and Tweemlow (1980) is unique among those included here in that it focuses on persons who applied for alcoholism treatment but did not appear for treatment when scheduled. For nine studies, the abbreviation "NR" (not reported) appears in the "drinking outcomes" portion of Table 2. Scherer and Freedberg (1976) did not report drinking outcomes; their outcome focus was on the dimension of assertiveness , not drinking. The other eight studies did report drinking outcomes in some form or other, but within-group percentages could not be derived from their drinking outcome data as reported. Patterns in Table 2 Of the 25 studies for which an "uncommonness of outcome" entry could be made, three are coded as reporting uncommonly poor drinking outcomes, 12 as reporting uncommonly good drinking outcomes, and ten as reporting not uncommon drinking outcomes.

PAGE 87

79 In each instance in which an uncommonly poor drinking outcome was reported, the poor outcome pertained to only one group of subjects within a larger cohort. In the Vogler, Ferstl, Kraemer, and Bengelmann (1975) study the poor-outcome group was a control group receiving only standard hospital treatment. The poor-outcome group, with an abstinent-or-controlled rate of 10%, in the report by Caddy and Lovibond (1976) was just below the lower boundary (10.5%) for uncommonness . In the Miller (1978) study, which reported two groups with uncommonly poor abstinence outcomes (7% and 8%), the treatment goal was controlled drinking. In all three cases the elapsed time between the end of treatment and evaluation was one year. The twelve papers reporting uncommonly good outcomes were reports on only ten programs. The outcomes at one, two, and three years for one program were reported in three separate papers (Caddy et al., 1978; Sobell & Sobell, 1973, 1976). The apparent uncommonly good drinking outcome of another program (Jones & Lanyon, 1981) should, according to the authors of the report, not be considered valid for program evaluation because too many subjects were lost to followup. Thus, only nine programs are discussed here. The distribution of types of treatment facility among the nine is unremarkable with one exception. There are two outpatient programs, one state hospital inpatient program, two private general hospital programs, two private aversive conditioning hospital programs, and two private residential treatment center programs. The last two, together, provide the aforementioned remarkable exception. The studies by Kliner et al. (1980) and Conley (1981) are both on different samples of the treated population of the same residential treatment center, which has a good

PAGE 88

80 ) • ' .... '"-^ reputation, a long history and a psychodynamic AA-oriented approach. The Kliner et al. study showed uncommonly good results in both the abstinent and the abstinent-or-controlled classes. The Conley study reported uncommonly good results in the abstinent class, with a highly selected sample (physicians). There appears to be a pattern with regard to treatment modality and uncommonly good outcome. Seven of the nine programs were inpatient and two were outpatient. Of the seven inpatient programs one was behaviorally oriented, four were dynamically oriented, and two offered dynamically oriented counseling as an adjunct to aversive counterconditioning. The two outpatient programs were behaviorally oriented. Thus it would appear that unusually successful inpatient programs share some species of dynamic orientation, and unusually successful outpatient programs share a behavioral orientation. There was a wide range of treatment duration. Among inpatient programs, mean treatment duration ranged from 12 days for aversive conditioning hospitals to 82 days for a general hospital. Outpatient treatment duration was ten sessions (Caddy & Lovibond, 1976) and two years (Azrin, 1976). While no general pattern of relationship between treatment duration and uncommonly good outcome seems to emerge, it is interesting to note that Willems et al. (1973) reported a long-stay treatment group reached a 54% abstinence level, which is in the low end of the uncommonly good outcome range, but concluded that there are no important outcome differences between groups treated 20 days and 80 days, respectively. It is noted, however, that Willems et al. did not use Emrick's uncommonality of outcome classes in evaluating their data.

PAGE 89

81 There appears to be a relationship between treatment goals and uncommonly good treatment outcomes. Six of the programs under discussion appeared to have abstinence as a sole drinking outcome goal (Azrin, 1976; Conley, 1981; Jackson & Smith, 1978; Kliner et al., 1980; Wiens et al. , 1976; Willems et al., 1973). The program first reported on by Sobell and Sobell (1973) and then at two other intervals had abstinence as a treatment goal for two of four subgroups. Among the seven programs in which there was a goal of abstinence for at least one treatment group, six studies reported uncommonly good abstinence-class drinking outcomes for at least one group. Two programs (Caddy & Lovibond, 1976; Tomsovic, 1974) had reduced or controlled drinking as outcome goals. The Sobell and Sobell program had controlled drinking as a goal for two of four subgroups. Among the three programs in which there was a controlled drinking outcome goal for at least one group, Caddy and Lovibond reported uncommonly good outcomes in the somewhat-improved drinking outcome class, while Tomsovic reported uncommonly good outcomes in the much-improved category. The Sobell and Sobell program reports indicated uncommonly good abstinent-or-improved category outcomes for two groups at one year, for one group at two years, and for two groups at three years. It is possible that the time of evaluation was important in two cases. Weins et al. (1976) reported an uncommonly good abstinence rate for all subjects at one year after admissi on to an aversive conditioning hospital, during which year some patients had returned for additional treatment. Azrin (1976) recorded an uncommonly good abstinence rate (98%) for experimental subjects two years after they began a community reinforcement behavior modification program.

PAGE 90

82 All of the subjects in all of the programs reporting uncommonly good treatment outcomes were classified as alcoholics, mostly male, mostly middle class. The one instance in which the uniqueness of the subject sample may have made an important difference is the study by Kliner et al. (1980) in which the sample was composed entirely of physicians. At one year after discharge, Kliner et ai. reported 76% abstinent and 92% abstinent-or-controlled. Subject Variables Table 3 summarizes the subject variables alluded to in the studies under consideration. "Subject variables," in this context, refers to facts pertaining to the clients previous to or at the beginning of treatment. In Table 3 subject variables are categorized by the methods by which they were measured. "Standardized methods" refers to the use of questionnaires, inventories, scales, or psychological tests which had been standardized on some other population before they were used in the study in question. "Nonstandardized methods" refers to the use of structured interviews, unstructured interviews, or instruments which had not been standardized on some other population before they were used in the study in question. The numbers in the columns of Table 3 indicate the number of studies alluding to the subject variable in question, and whether the method of measurement was standardized or not. Thus, one would understand by reading the first line of Table 3 that for the subject variable "drinking amount or frequency" the variable was alluded to in 11 studies in which it had been measured by standardized methods and in 12 studies in which it had been measured by nonstandardized methods, so that it was alluded to in a total of 23 studies.

PAGE 91

83 Table 3 Numbers of Studies Reporting on Various Subject Variables By standardized By nonstandardized Variable methods methods Row totals Drinking amount or frequency 11 12 23 Affectivecognitive 14 6 20 Work situation 5 10 15 Home situation: relationships 3 8 11 Previous inpatient treatment 0 14 14 Physical 5 12 17 Legal problems 2 6 8 Mixed 0 0 0 Social situation 3 3 6 Previous A. A. involvement 0 7 7 Finances 1 7 8 Global adjustment 1 1 2 Miscellaneous 3 1 4 Other drinking behavior 1 3 4 Home situation: residence 2 0 2 Previous outpatient treatment 0 2 2 Leisure time 0 0 0 Religious life 0 2 2

PAGE 92

84 Table 3 — continued. Variable By standardized methods By nonstandardized methods Row totals Previous treatment: locus unclear 0 1 1 Column totals 51 95 Patterns in Table 3 Subject variables were measured 95 times by nonstandardized methods and 51 times by standardized methods. Although some variables, e.g., those pertaining to history of previous treatment, are not amenable to measurement by standardized instruments, it appears that the use of nonstandardized or "homegrown" methods of measuring subject variables predominates. The sole exception to this pattern is the use of standardized methods to measure variables in the affective and cognitive domains, in which standardized methods were used 14 times and nonstandardized methods only six times. Table 4 summarizes the outcome variables alluded to in the studies under consideration. "Outcome variables," in this context, refers to facts pertaining to the clients after treatment. In Table 4 outcome variables are categorized by the methods by which they were measured. "Standardized methods" and "nonstandardized methods" have the same meanings as in Table 3. The numbers in the columns of Table 4 indicate the numbers of studies alluding to the respective outcome variables, and the method of measuring those variables. Outcome Variables

PAGE 93

Table 4 Numbers of Studies Reporting on Various Outcome Variables Variable By standardized methods By nonstandardized methods Row total Drinking amount or frequency 4 29 33 Affectivecognitive 3 8 11 Work situation 4 19 Home situation: relationships 0 10 10 Subsequent inpatient treatment 0 10 t : ' 10 Physical 4 8 12 Legal problems 2 10 12 Mixed 0 4 ; 4 Social situation 4 7 11 Subsequent A. A. involvement 0 9 9 Finances 1 3 4 Global adjustment 3 5 8 Miscellaneous 0 Z : \ 8 Other drinking behavior 0 e,. Home situation: residence 2 5 Subsequent outpatient treatment 0 3 3 Leisure time 1 5 6 Religious life 0 2 2

PAGE 94

86 Table 4 — continued. Variable By standardized methods By nonstandardized methods Row totals Subsequent treatment: locus unclear 0 2 2 Column totals 28 148 Patterns in Table 4 Outcome variables were measured 148 times by nonstandardized methods and 28 times by standardized methods. Of the 34 studies under consideration all but one reported on amount or frequency of drinking as an outcome variable. The outcome variable second most frequently alluded to was "work situation," which was referred to in 19 studies. Physical health and legal problems were each alluded to in 12 studies, affective-cognitive variables and social situation in 11 studies each, and at-home relationships and subsequent in-patient treatment in ten studies each. Subsequent AA involvement, global adjustment, and the miscellaneous variable were respectively alluded to nine, eight, and eight times each. There appears to be a pattern of almost ubiquitous interest in drinking as an outcome variable, while vocational adjustment holds second place in terms of interest. Trailing at considerable distance are physical health, legal problems, affective-cognitive status, social situation, family relationships, subsequent treatment, AA involvement, and global adjustment. Table 5 summarizes the treatment variables alluded to in the ies considered here. The numbers in the columns of Table 5 indicate Treatment Variables

PAGE 95

87 the numbers of studies alluding to the respective treatment variables and whether the respective variables were reported directly or inferred from the study. The following paragraphs clarify the structure and content of Table 5. In three studies length of treatment was not specifically reported but could be inferentially estimated from the description of the treatment regimen. In five studies length of treatment was neither reported nor inferrable. The "locus of treatment" category classifies treatment as inpatient or outpatient as well as referring to type of treatment facility, i.e., state, VA, or private hospital; private residential facility; halfway house; or outpatient counseling center. For one study the locus of treatment was neither reported nor inferrable. "Treatment methods" refers to both the type of treatment (e.g., behavioral vs. psychodynamic) and specific treatment procedures (e.g., electric shock aversive conditioning, assertiveness training). In three studies treatment methods were neither specifically reported nor inferrable. "Theoretical stance" refers to the position held by a treatment program on the nature of alcoholism and/or the nature and course of therapeutic practice in the treatment of alcoholics. Two studies made brief, direct reference to theoretical stance. For 18 studies some rough and faint approximation of theoretical stance could be inferred from the stated treatment methods of choice, the statement that the program was AA oriented, or the selection of treatment goals. Some studies made direct reference to the definition of alcoholism used, e.g., stated that World Health Organization criteria for the

PAGE 96

88 diagnosis of alcoholism were used. In some other studies the definition of alcoholism used was inferred from the authors' use of such terms as "gamma-type alcoholics." For 23 studies the definition of alcoholism was neither directly reported nor inferrable. "Characteristics of treatment personnel" refers to such items as the training, professional or paraprofessional group and level, experience, and personality characteristics of treatment personnel. Table 5 Numbers of Studies Reporting on Various Treatment Variables Variable Reported directly Inferred Row totals Length of treatment 26 3 29 Locus of treatment 33 0 33 Treatment methods 31 0 31 Theoretical stance 2 18 20 Definition of alcoholism 3 8 11 Characteristics of treatment personnel 4 0 4 Patterns in Table 5 Length, locus, and methods of treatment were reported in almost all studies. Theoretical stance was inferred for 18 studies but directly referred to in only two. The definition of alcoholism used was reported or inferrable in less than one-third of the studies. Characteristics of treatment personnel were reported in four studies and were not inferrable from the other 29.

PAGE 97

89 Research Models and Statistical Significance Tables 6, 7, and 8 summarize the major findings with statistical significance of £-=.05 or better for those studies whose major findings were statistically significant. The findings are organized by research model, with the studies listed alphabetically within the respective years of publication. The categories of research models were drawn from the descriptions of basic methods of research set forth by Isaac and Michael (1971). The research models used in the studies under consideration here were subsumed under the categories of correlational, descriptive, experimental, or quasi-experimental methods. Of the studies which included statistically significant findings, one used a descriptive research model. Alford (1980) noted two major findings significant at the .05 level or better. He found that for alcoholics treated in an AA-oriented inpatient program, at two years posttreatment: (a) Ninety-six % of then-abstinent alcoholics were socially stable as compared with 50% of those who were drinking moderately and 14% of those who were drinking heavily; (b) Pre-treatment alcohol intake is a predictor for treatment outcomes; subjects whose reported pre-treatment drinking was in the 100-200 oz. per week range had a higher success rate than did either those in the over 200 oz. per week range or those in the under 100 oz. per week range, although the third group fared better than the second.

PAGE 98

90 Table 6 Major Findings with Statistical Significance of ^.-t-OS or Better in Studies Utilizing Correlational Research Models Study Findings Pokorny et al. , 1973 McWilliams & Brown, 1977 Bowen & Tweemlow, 1980 Cronkite & Moos, 1980 Maisto et al, 1980 1. Patients with histories of higher stability in marriage, social relationships, and employment are more likely to participate in aftercare. 2. Male alcoholics treated by 60 days of inpatient care followed by outpatient group therapy show the same rate of improvement as those who had 90 days of hospitalization without aftercare. 1. MMPI data are of little use in identifying alcoholics with good prognoses. 2. The success of alcoholism treatment cannot be evaluated from MMPI change scores at six weeks of treatment compared to the second week of treatment. One year later, there were no significant differences in rates of abstinence, hospitalizations, or police arrests between men who applied for alcoholism treatment but failed to appear for treatment and men who applied for and underwent alcoholism treatment. 1. Posttreatraent stresses and coping responses have strong effects on outcome. 2. Effects of both pre-treatment subject variables and treatment-related factors are frequently shared with the effects of posttreatment factors. 3. The relative strength of pre-treatment, treatment, and posttreatraent variables as predictors of outcome will vary with the kinds of outcome criteria considered. 1. Individualized behavior therapy (IBT) with a controlled-drinking goal has a strong positive relationship to controlled drinking as an outcome. 2. IBT with an abstinence goal does not influence days abstinent as an outcome. 3. There is a negative relationship between the number of pre-treatment alcohol-related hospitalizatons and controlled drinking outcomes.

PAGE 99

91 Table 6 — continued. Study Findings Abbott & Gregson, 1981 Conley, 1981 Jones & Lanyon, 1981 Patterns in Table 6 The level of cognitive dysfunction at intake predicted the one-year outcome for 58% of subjects. Some MMPI types are predictive for differential drinking outcomes; at one year posttreatment the high-to-low rank ordering of abstinence rates between groups of subjects typed by MMPI profile was: psychopathic; classic alcoholic; psychotic; neurotic. At one year after inpatient treatment subjects' positive scores on an instrument measuring adaptive skills correlated positively with positive treatment outcome. Of the 14 findings listed in Table 6, seven had to do with subject variables as predictors of treatment outcomes. Four findings had to do with the relationships of treatment and outcome. Two findings had to do with the relationships of posttreatment factors to outcomes. One finding had to do with how the kinds of outcome criteria chosen will cause apparent variation in the predictive power of other variables. Table 7 Major Findings with Statistical Significance of £.-<:05 or Better in Studies Utilizing Experimental Research Models Study Sobell & Sobell, 1973 Vogler, Compton, & Weissbach, 1975 Findings Patients receiving individualized behavior therapy functioned better at 12 months than control groups receiving conventional therapy, regardless of whether the treatment goal was abstinence or controlled drinking. Higher levels of pre-treatment alcohol ingestion are strong predictors for higher levels of posttreatment alcohol drinking.

PAGE 100

92 Table 7 — continued. Study Findings Vogler, Ferstl, Kraemer, & Brengelmann, 1975 Azrin, 1976 Sobell & Sobell, 1976 Caddy et al. , 1978 Chaney et al., 1978 There were significant between-groups differences on drinking outcome measures for groups receiving, respectively, standard hospital treatment, drinkingcontingent shocks, random shocks while drinking, and mixed drinking-contingent and random shocks. Compared with matched controls not receiving community-reinforcement treatment, alcoholics in a community-reinforcement treatment program did better on all outcome measures. 1. At 24 months posttreatment, clients who had received individualized behavior therapy (IBT) with controlled drinking as a goal functioned better in terms of both drinking outcomes and general adjustment outcomes than did a control group who received standard hospital treatment for alcoholism. 2. At 24 months posttreatment, there was no significant difference between IBT clients with a nondrinking treatment goal and their controls. 1. At three years posttreatment, clients in the individualized behavior therapy groups, regardless of whether the treatment goal was abstinence or controlled drinking, did better in terms of overall improvement than their controls . 2. At three years posttreatment, there were no significant differences in drinking outcomes between IBT clients with abstinence as a treatment goal and their conventionally treated controls with the same treatment goal. 3. In the controlled-drinking-as-treatmentgoal condition, IBT clients drank less at three years than their conventionally treated controls in the same goal condition. At 12 months posttreatment, alcoholics who had received skill training specific to the avoidance of alcohol abuse had an average number of days drunk equal to one-sixth that of pooled control groups, drank one-fourth as much, and had an average drinking period less than one-eighth as long. The difference in days abstinent was not significant.

PAGE 101

93 Table 7 — continued. Study Findings Miller, 1978 1, Miller & Taylor, 1980 Swenson et al., 1981 1. 2. At 12 months posttreatraent, while all groups showed a significant decrease in drinking, there were no significant outcome differences between groups of outpatient problem drinkers respectively exposed to aversive counterconditioning, behavioral selfcontrol training, and a composite of the two. Clients who received and read a self-control manual showed a small but significant improvement in maintaining controlled drinking compared to those who did not receive and read the manual. At one year posttreatment, all alcoholics given outpatient behavioral self-control training had lowered their alcohol consumption and had improved their scores on a profile of mood states measuring tension/anxiety, depression/dejection, anger/ hostility, vigor, fatigue, and confusion. There were no significant differences between groups of subjects receiving behavioral self-control training in four differing modalities. At 18 months posttreatment there were no significant between-groups outcome differences for subjects receiving either power motivation training, group therapy or a home-study alcohol course as treatment. Patterns in Table 7 Of 15 findings listed, 14 had to do with the relationships of treatment variables to outcomes. One finding had to do with a subject variable as a predictor of outcome.

PAGE 102

94 Table 8 Major Findings with Statistical Significance of £-<.05 or Better in Studies Utilizing Quasi-experimental Research Models Study Findings Willems et al. , 1973 1. Patients treated in a hospital for a mean of 20 days had as good drinking outcomes at a 24 month followup as patients treated a mean of 82 days. 2. Lower social class, higher social maladjustment, and a history of delirium tremens correlate strongly with poor outcome. 3. Understanding oneself as an alcoholic and actively seeking sobriety at discharge from treatment are variables strongly associated with better outcomes. Caddy & Lovibond, 1976 Stein et al . , 1975 Scherer & Freedburg, 1976 McClelland, 1977 Costello et al. , 1979 Costello, 1980 The treatment response of clients receiving a combination of self-regulation training and electric shock aversive counterconditioning was better than the treatment response of clients receiving either self-regulation training only or aversive counterconditioning only. There was no significant difference between short treatment (mean: nine days) and long treatment (mean: 30 days) groups on any outcome measure at 12 months post-discharge. Videotaped feedback, as compared with verbal feedback, does not significantly enhance assertiveness skill acquisition and maintenance in alcoholics. At 12 months posttreatment, clients who had received power motivation training had a higher improvement rate than those who had received standard treatment. At two years posttreatment, the ethnicity of clients (Mexican-American vs. Anglo-American) was not significantly associated with treatment outcome. Greater participation in aftercare correlates positively with better outcomes; the relationship is stronger than the relationship of in-hospital adjustment to outcome; participation in aftercare contributes almost as much to outcome as does pre-treatment social stability.

PAGE 103

95 Table 8 — continued. Study Findings Valle, 1981 Subjects whose primary therapists were characterized by higher measured levels of accurate empathy, genuineness, concreteness , and respect were less likely to be drinking at two years posttreatment and likely to have fewer days of relapse than subjects whose primary therapists had lower levels of interpersonal functioning. Patterns in Table 8 Of ten findings listed, seven had to do with the relationships of treatment variables to outcomes. Three findings had to do with subject variables as predictors of outcomes. Summary of Patterns in Tables 6-8 A total of 39 findings were listed. Of that number, 11 findings had to do with subject variables as predictors of outcomes. Twenty-five findings had to do with the relationships of treatment variables to outcomes. Two findings had to do with the relationships of posttreatment factors to outcomes. One finding had to do with the ways that the kinds of outcome criteria chosen will cause apparent variation in the predictive power of other variables. Research Models and Practical Significance Tables 9, 10, 11, and 12 summarize major findings to which prac-tical, but not statistical, significance was attributed by the authors. The findings are organized by research model, with the studies listed alphabetically within the respective years of publication.

PAGE 104

96 Table 9 Major Findings with Practical but not Statistical Significance in Studies Utilizing Correlational Research Models Study Findings Bowen & Tweemlow, 1980 Maisto et al. , 1980 Abbott & Gregson, 1981 The fact that there were no significant betweengroup differences in one-year outcome measures for men who applied for alcoholism treatment but failed to appear for treatment and men who were treated may be evidence for the notion that alcoholism may be appropriately viewed as a way of life and that a decision to do something about problem drinking may be a pivotal factor. 1. Treatment goal orientation appears to be strongly related to controlled-drinking outcomes; individualized behavior therapy with a controlled-drinking goal tends to produce more favorable outcomes in terras of both controlled drinking and abstinence than does the same therapy with a goal of abstinence. 2. The best candidates for success with controlled-drinking treatment goals appear to be persons who have not experienced severe alcoholic deterioration. Alcoholics' self-rating at intake of the severity of their drinking problems is a strong predictor of outcome; the worse alcoholics perceive their drinking problems the worse the outcome tends to be. Patterns in Table 9 Of three findings listed, two had to do with subject variables as predictors of outcome and one had to do with the relationships of treatment variables to outcome.

PAGE 105

97 Table 10 Major Findings with Practical but not Statistical Significance in Studies Utilizing Descriptive Research Models Study Findings Tomsovic, 1974 Wiens et al . , 1976 Jackson & Smith, 1978 Cohen et al. , 1979 Alford, 1980 Kliner et al. , 1980 Neubuerger et al., 1981 "Binge" and continuous-drinking alcoholics showed about the same relative improvement at one year posttreatment. Of 261 alcoholics voluntarily undergoing pharmacological aversive counterconditioning in an aversive treatment hosital, 63% were reported abstinent at one year post-discharge. Electric shock aversion therapy appears to be a promising alternative to chemical aversion therapy in cases in which the latter is contraindicated. The outcomes of behaviorally oriented treatment of German women alcoholics did not correlate significantly with subjects' drinking history, social background, or personality factors. At two years posttreatment, a private, residential, AA-oriented program showed success rates comparable to those of other programs with different orientations . Compared with the general patient population of a large residential alcoholism treatment center, physicians treated there showed more favorable response to treatment when treatment response was measured by the criteria of abstinence and alcohol misuse at one year posttreatment. For alcoholics treated in a private aversive conditioning hospital four subject-status variables were noted to be prognostic indicators for abstinence at one year after admission. While the one year abstinence rate for all patients was 53%, those on Medicare with physical or psychiatric disability status had an abstinence rate of only 33%. Married persons had an abstinence rate of 62%. Employed persons had an abstinence rate of 65%, and those who were both married and employed had an abstinence rate of 73%.

PAGE 106

98 Patterns in Table 10 Of six findings listed, three had to do with subject variables as predictors of outcome and three had to do with the relationships of treatment variables to outcomes. Table 11 Major Findings with Practical but not Statistical Significance in Studies Utilizing Experimental Research Models Study Findings Sobell & Sobell, 1973 Vogler , Compton, & Weissbach, 1975 Sobell & Sobell, 1976 Caddy et al. , 1978 Chaney et al . , 1978 Miller, 1978 Some persons diagnosed as alcoholic did acquire and maintain controlled drinking behavior for one year posttreatment. When treated by in-treatment-session alcohol drinking followed by videotaped playback of drunken comportment, alcohol education, BAG discrimination training, electric shock aversion training, avoidance training, behavior counseling, followup, and booster sessions, alcoholics did no better on gross outcome measures than did their alcoholic controls who received only alcohol education, behavior counseling, followup, and booster sessions, even though the median treatment duration was twice as long for the former group as for control subjects. Some persons diagnosed as alcoholic did acquire and maintain controlled drinking behavior for two years posttreatment. Some persons diagnosed as alcoholic acquired and maintained controlled drinking behavior for three years posttreatment. Alcoholics trained to recognize "drink-temptation situations" and to generate alternative behaviors had less severe and shorter relapse episodes than alcoholics not so trained. 1. Behavioral training in functional analysis of drinking behavior, self -monitoring, and drinking rate reduction was equally effective with or without aversive counterconditioning and more cost-effective without it. 2. Neither a previous history of alcoholism nor an advanced degree appeared to be necessary for

PAGE 107

99 Table 11 — continued. Study Findings therapists effectively to carry out behavioral self-control training, aversive counterconditioning, or a composite of the two, with alcoholic clients . ^ 3. The higher frequency of abstinence among traditionally-treated alcoholics compared to alcoholics receiving behavioral training with a controlled-drinking goal suggested a possible relationship between treatment outcome and the ideology and expectations of the treatment agency. Pattern in Table 11 All eight findings listed had to do with the relationships of treatment variables to outcomes. Table 12 Major Findings with Practical but not Statistical Significance in Studies Utilizing Quasi-experimental Research Models Study Findings Cutter et al.. Persons characterized as rebellious and noncon1977 forming are more likely than persons characterized as conforming to benefit from therapy with an egalitarian approach and to participate in AA. "Conformers" tend to respond well to more directive therapeutic approaches. Costello et al. , Costello's Treatment Difficulty Scale appeared to 1979 be a valid predictor of outcome across ethnic groups. Pattern in Table 12 Both findings listed had to do with subject variables. One finding had to do with a subject-variable dimension in relation to therapeutic approach. One finding had to do with a subject variable as a predictor of outcome.

PAGE 108

100 Sununary of Patterns in Tables 9-12 A total of 23 findings were listed. Of that number, ten had to do with subject variables as predictors of outcome. Twelve findings had to do with relationships of treatment variables to outcomes. One finding had to do with the relationship of a subject-variable dimension to therapeutic approach. All eight of the findings listed from studies with experimental research approaches had to do with the relationships of treatment variables to outcomes. Summary of Patterns in Tables 6-12 Two findings, not given a table of their own because they were from the lone descriptive study with statistically significant findings, were mentioned in the text preceding Table 6. Those two findings are included in this general summary of patterns of major findings. A total of 64 findings were listed. Of that number, 22 findings had to do with subject variables as predictors of outcome. Thirty-seven findings had to do with the relationships of treatment variables to outcomes. Three findings had to do with the relationships of posttreatment factors to outcomes. One finding had to do with the ways that the kinds of outcome criteria chosen will cause apparent variation in the predictive power of other variables. One finding had to do with the relationship of a subject-variable dimension to therapeutic approach. Studies using correlational research models produced nine findings having to do with subject variables as outcome-predictors, five findings having to do with relationships of treatment variables to outcomes, and two findings having to do with the relationships of posttreatment factors to outcomes.

PAGE 109

101 Studies using descriptive research models produced four findings having to do with the subject variables as outcome-predictors and three findings having to do with relationships of treatment variables to outcomes. Studies using experimental research models produced one finding having to do with subject variables as outcome-predictors and 22 findings having to do with relationships of treatment variables to outcome. Studies using quasi-experimental research models produced four findings having to do with subject variables as outcome-predictors and seven findings having to do with relationships of treatment variables to outcomes . Research Questions Table 13 lists the major research questions implied or stated in the studies under consideration. One study (Cronkite & Moos, 1980) included a statement of research questions; those questions appear, verbatim, in quotation marks, in Table 13. All other research questions in the table were inferred from the studies. The studies are arranged alphabetically by author within their respective years of publication. Table 13 Major Research Questions of Outcome Studies of Psychologically Oriented Alcoholism Treatment ^^^<^y ' Research Questions Will male alcoholics who had 60 days of hospitalization followed by outpatient group therapy show the same rate of improvement as those who had 90 days of hospitalization without aftercare? Pokorny et al. , 1973

PAGE 110

102 Table 13--continued. Study Research Questions Sobell & Sobell, 1973 Willems et al . , 1973 Tomsovic, 1974 Stein et al. , 1975 Vogler , Compton & Weissbach, 1975 Vogler, Ferstl, Kraemer, & Brengelmann, 1975 Azrin, 1976 Caddy & Lovibond, 1976 Scherer & Freedberg, 1976 Sobell & Sobell, 1976 1. Will individualized behavior therapy result in better treatment outcomes at one year than conventional treatment? 2. Can some gamma-type alcoholics acquire and maintain controlled drinking behavior? Is inpatient treatment of more than one month's duration necessary for successful treatment of alcoholism? Will "binge" and "continuous" drinkers in the same treatment program have different outcomes? Does hospital stay beyond detoxication relate to outcome of treatment? 1. Is the "integrated behavior change package" effective in the treatment of alcoholics? 2. Is pre-treatment level of drinking related to successful outcome? 3. Can some chronic alcoholics be trained drink alcohol in moderation? to Is a mixed contingent-and random-shock schedule more effective in suppressing the drinking habits of chronic alcoholics, in the long run, than either an all-contingent or an all-random schedule? 1. 2. What effect will a communityreinforcement treatment program have on alcoholics? Is such a program cost-effective? Do controlled-drinking outcomes vary with the treatments self-regulation training, aversive conditioning plus self-regulation training, and aversive conditioning only, when discrimination training is held constant in all three conditions? Does videotaped feedback enhance assertiveness training with alcoholics? 1. Will individualized behavior therapy result in better treatment outcomes at two years than conventional treatment? 2. Is controlled drinking an appropriate treatment goal with some alcoholics?

PAGE 111

Table 13 — continued. 103 Study Research Questions Wiens et al . , 1976 Cutter et al. , 1977 McClelland, 1977 McWilliams & Brown, 1977 Caddy et al. , 1978 Chaney et al. , 1978 Jackson & Smith, 1978 Miller, 1978 Cohen et al. , 1979 Costello et al., 1979 What is the abstinence rate of alcoholics treated by chemical (emetic) aversive counterconditioning at one year? Will treatment efficacy be better served if clients' preferred relationship to authority is considered in matching treatment styles to clients? Will alcoholics receiving power motivation training in addition to standard treatment be more improved at one year posttreatment than those receiving standard treatment only? Can MMPI scores predict successful completion of alcoholism treatment? 1. Will individualized behavior therapy result in better treatment outcomes at three years than conventional treatment? 2. Is controlled drinking an appropriate treatment goal for some alcoholics? 1. Does skill training in recognizing situations constituting drinking cues, and generating alternative responses increase problem drinkers' effectiveness in responding to stressful situations? 2. Does an improvement in psychosocial problem solving skills have an effect on subsequent frequency of problem drinking? Does electric shock aversive conditioning therapy work as well as chemical (emetic) aversive conditioning for alcoholics? What are the outcome differences at one year between groups treated by: (a) outpatient group therapy plus disulfiram; (b) aversive counterconditioning; (c) behavioral self-control training; and (d) behavioral self-control training plus aversive counterconditioning? What are the 18 month followup results, in terms of abstinence, for a 90 day behaviorally oriented inpatient treatment program for alcoholic women? 1. Is the ethnicity (Mexican-American vs. AngloAmerican) of alcoholic men a factor in treatment outcome at two years?

PAGE 112

Table 13 — continued. Study Research Questions Alford, 1980 Bowen & Tweemlow, 1980 Costello, 1980 Cronkite & Moos, 1980 (p. 306) Kliner et al. , 1980 2. Is the Treatment Difficulty Scale a valid predictor of success across ethnic groups? What are the two year outcomes for AA oriented inpatient treatment? One year later, what is the status of persons who applied for alcoholism treatment but did not appear for treatment compared with those who applied and were treated? Do aftercare contacts relate to improved adjustment at one and two years posttreatment? 1. "What are the interrelationships among patient characteristics, treatment experiences, posttreatment factors, and outcome?" 2. "What is the relative importance of posttreatment factors compared with patient-related and treatment-related variables in predicting outcome?" 3. "How is treatment related to posttreatment factors, and in turn what are the mediating effects of posttreatment experiences between treatment and outcome or between patient characteristics and outcome?" At one year posttreatment, how did physicians treated for alcoholism at a private residential AA oriented facility fare in terms of drinking, vocational, general health, and personal adjustment status? Maisto et al. , 1980 Among pre-treatment, treatment variables, dictors of outcome? within-treatment, and postwhich are the stronger preMiller & Taylor, 1980 1. Are there significant between-groups differences, at one year, in mood scale scores and drinking outcomes for alcoholics undergoing behavioral self-control training in four differing modes of presentation? 2. Is bibliotherapy with a behavioral self-control manual an effective treatment for problem drinkers?

PAGE 113

105 Table 13 — continued. Study Research Questions Abbott & Gregson, 1981 Conley, 1980 Jones & Lanyon, 1981 Neubuerger et al., 1981 Swenson et al. , 1981 Valle, 1981 Is level of cognitive dysfunction measured at intake a predictor of treatment outcome at one year? 1. Are male alcoholics classifiable into subgroups by MMPI profiles? 2. Are there different treatment outcomes associated with four MMPI-prof ile-types of male alcoholics? Is successful outcome of alcoholism treatment associated with patients' ability to cope with stimuli that could trigger drinking? What is the abstinence rate at one year after admission for aversive conditioning hospital patients? At 18 months posttreatment will the drinking and personal adjustment patterns of DWI arrestees exposed to short-term rehabilitation be improved more than those of DWI arrestees exposed to a minimum of treatment? Does the level of interpersonal functioning of the primary therapist have a significant and important impact on treatment outcome at two years for alcoholics treated in a residential facility? Patterns of Research Questions There are 40 research questions listed in Table 13. Of that number, 25 questions have to do with the relationships of treatment variables to outcomes; four of those 25 questions have to do with length of treatment, and 21 have to do with treatment methods. Nine questions have to do with relationships of subject variables to outcomes. Two questions have to do with interactions of subject variables, treatment variables, posttreatment factors, and outcomes. Each of the following categories had one question each: posttreatment factors relevant to outcome; cost-effectiveness of treatment method; interaction of treatment variables, posttreatment factors, and

PAGE 114

106 outcome; interaction of counselor level of interpersonal functioning with outcome. Interactions The major interactions which were observed in the sample studies and which are relevant to the present writing are summarized in Table 14. The interactions are presented in four groups: interactions of client variables with treatments; interactions of client variables with outcomes; interactions of treatments with outcomes; and interactions of client variables with treatment and outcome. Table 14 Major Interactions Observed Study Interaction Clienttreatment Patients with pre-treatment histories of higher stability in marriage, social relationships and employment are more likely to participate in aftercare than those with poorer histories in the same dimensions. Rebellious, nonconforming personality types are more likely to participate in AA or egalitarian therapies, while clients whose personalities are characterized as more compliant and conformist respond well to firm medical management. Client-outcome Clients who come from lower socio-economic strata, have a history of delirium tremens, and have a history of social maladjustment are apt to have poor treatment outcomes. Pokorny et al. , 1973 Cutter et al. , 1977 Willems et al . , 1973

PAGE 115

107 Table l4--continued. Study Interaction Client-outcome Torasovic, 1974 Vogler, Compton, & Weissbach, 1975 Cohen et al . , 1979 Costello et al. , 1979 Alford, 1980 Bowen & Tweemlow, 1980 Costello, 1980 Kliner et al . , 1980 Maisto et al . , 1980 Although "binge" and continuous drinkers could be distinguished by health and social characteristics at entry, one year outcomes were about the same for the two groups. Higher levels of pre-treatment drinking predict poorer treatment outcomes. For German women treated in an inpatient behaviorally oriented program, treatment outcomes did not correlate significantly with drinking history, social background, or personality factors. Clients whose scale-measured difficulty of treatment scores indicated they would be difficult to treat had poorer treatment outcomes at two years posttreatment than those whose scores indicated they would be easier to treat. Ethnicity of clients (Mexican-American versus Anglo-American) was not shown to be a factor in treatment outcome at two years . Clients whose pre-treatment alcohol consumption was over 200 oz. per week and clients whose pretreatment alcohol consumption was less than 100 oz. per week had poorer success rates at two years posttreatment than clients whose pre-treatment alcohol intake was 100-200 oz. per week. One year later, persons who had applied for alcoholism treatment but did not appear for treatment showed no difference in drinking status compared with persons who had applied for and received treatment for alcoholism. Client social stability at intake is positively related to good treatment outcome at one year. Physicians had better treatment outcomes at one year than the general patient population of a large residential treatment facility for alcoholics. Clients with histories of more alcohol-related hospitalizations had poorer treatment outcomes at two years than did clients with fewer

PAGE 116

108 Table 14 — continued. Study Interaction Client-outcome Abbott & Gregson, 1981 Conley, 1981 Neubuerger et al., 1981 alcohol-related hospitalizations. The best candidates for success with a controlled-drinking goal were clients who had experienced less pre-treatment deterioration. Level of cognitive dysfunction predicted one-year outcome for 58% of alcoholic clients; the more cognitively impaired clients were, the poorer their outcomes were. Clients who, at intake, viewed their drinking problems as more severe had poorer one-year outcomes than those who viewed their drinking problems as less severe. Four MMPI-profile-types were respective predictors for one-year outcome differences. For patients treated in a private aversion therapy hospital, four client variables were important prognostic indicators. Clients on Medicare disability status at intake had a 33% abstinence rate at one year, while clients who were married at intake had a 62% one-year abstinence rate and those who were employed had a 65% one-year abstinence rate. Those who were both married and employed at intake had a 73% one-year abstinence rate. Treatment-outcome Sobell & Sobell, Individualized behavior therapy clients functioned ^^''3 better at one year posttreatment than did their controls who received conventional treatment. Stein etal.. Length of hospital stay beyond detoxication does ^^'^ not appear to relate to treatment outcomes. Vogel, Compton, Behavior modification treatment including actual & Weissbach, practice in controlled drinking did not result •^^'^ in significantly better gross outcomes at one year than did behavior counseling without in-treatment drinking.

PAGE 117

109 Table 14 — continued. Study Interaction Treatment-outcome Vogler, Ferstl, Kraemer, & Brengelmann, 1975 Caddy & Lovibond, 1976 Scherer & Freedberg, 1976 Sobell & Sobell, 1976 Wiens et al. , 1976 McClelland, 1977 The use of a mixed randora-and contingent-electric shock aversive counterconditioning schedule was superior to either all-random or all-contingent schedules in producing abstinence at one year posttreatment. Standard hospital treatment was inferior, by a one-year abstinence criterion, to any of the aversive counterconditioning regimens. The combination of electric shock aversion treatment and behavioral self-regulation training is superior to either method used by itself, when viewed in terras of controlled-drinking outcome at one year. Whether videotaped feedback versus verbal feedback is given in assertiveness training with alcoholics made no difference in level of assertiveness measured at one year posttreatment. Individualized behavior therapy clients functioned better at two years posttreatment than did their controls who received conventional treatment. At one year posttreatment, clients who received chemical (emetic) aversive counterconditioning had a 63% abstinence rate. Clients who received standard treatment plus "power motivation training" fared better at one year posttreatment than did their standard-treatment controls . Caddy et al . , 1978 Chaney et al. , 1978 Jackson & Smith, 1980 Individualized behavior therapy clients functioned better at three years posttreatment than did their controls who received conventional treatment. Clients receiving behavioral skill training in recognizing "drink temptation" situations and generating alternative responses to them had had shorter and less severe relapses at one year posttreatment than their standard-treatment controls. Clients for whom emetic chemical aversive counterconditioning was contraindicated and who, instead, received electric shock aversive counterconditioning had an abstinence rate at one year posttreatment

PAGE 118

no Table 14 — continued. Study Interaction Treatment-outcome favorably comparable to clients who received emetic chemical aversive counterconditioning . Costello, 1980 Maisto et al. , 1980 Miller & Taylor, 1980 Swenson et al. 1981 Valle, 1981 Clients who participated in aftercare in the first year after inpatient treatment had better treatment outcomes at one year than those who did not participate in aftercare. Controlled drinking as a goal with individualized behavior therapy has a strong positive relationship to controlled drinking as an outcome and appears to influence the number of days abstinent as well . Clients who received behavioral self-control training showed lowered alcohol consumption and improved mood states at one year posttreatment, compared with their alcohol consumption and mood states measured at entry into treatment. For DWI arrestees short-term rehabilitation efforts did no more to alter their drinking behavior patterns at one year than did minimum intervention in the form of a home study course. The higher the level of interpersonal functioning of the primary therapist, the less likely was the client to use alcohol at two years posttreatment and the fewer days of relapse was the client likely to have had at six, 12, 18, and 24 months posttreatment. Cronkite Se Moos, 1980 Clienttreatment-outcome The clients' pre-treatment occupational functioning accounts for almost all of the explained variance in occupational functioning as an outcome. Client characteristics at intake explain about one-third of the variance in drinking outcomes, while an equal amount of the variance in drinking outcomes is explained by posttreatment factors such as stressful life events and coping behaviors.

PAGE 119

Ill Table 14 — continued . Study Interaction Clienttreatment-outcome Posttreatraent stressors, however, appear to be reduced by treatment, and treatment seems to increase positive coping responses. Patterns in Table 14 Table 14 lists 40 major interactions noted in the sample studies. Two were client-treatment interactions. Nineteen were interactions of client characteristics with outcomes. Interactions of treatment with outcome were noted 18 times. One study examined the complex interactions of client, treatment, and outcome variables. Pre-treatment drinking patterns as predictors of outcomes were mentioned four times in the client-outcome relationships set, as were pre-treatment socio-economic status/stability variables. The treatment-outcome relationships set contained 14 statements of relationships between treatment methods and outcomes, two observations on the relationship of length of treatment to outcome, and one observation each on the relationship of aftercare to outcome and the relationship of treatment goal to outcome.

PAGE 120

112 theoretical position having to do with the notion of alcoholism as a way of life and the presumed critical importance of the decision by an alcoholic to do something about problem drinking. However, the emphasis in the article was not on the theoretical positions of treatment programs as variables contributing to outcomes. The other study which made direct reference to theoretical posture was Alford's 1980 report on outcomes of a program whose "treatment philosophy, procedure, and sequence was [sic] guided by the 12-step program of recovery articulated in Alcoholics Anonymous" (p. 361). Alford suggests that the success rate of the program (49% abstinent at two years) is an indication that "AA treatment can be a highly effective approach in treating alcohol dependent patients" (p. 369). However, Alford's emphasis is more on the treatment procedures used than on the relationship between theoretical position and outcome. What Are the Ways in which Client, Treatment, and Outcome Variables Have Been Measured? Client variables . Client variables were measured at intake or early in treatment 95 times by nonstandardized metods and 51 times by standardized methods. The most frequently used nonstandardized methods were structured interview forms developed by the treatment program in question. Standardized methods of measuring client variables were used most frequently in the affective-cognitive domain. The instrument used most frequently was the MMPI. Treatment variables . The categories of treatment variables under consideration in the present work were: length; locus; methods; theoretical stance; definition of alcoholism; and characteristics of treatment personnel. Most studies measured length of treatment by number of

PAGE 121

113 inpatient days or number of outpatient sessions. One study (Valle, 1981) measured treatment personnel characteristics by a standardized level of interpersonal functioning rating scale. Locus, methods, theoretical stance, and definition of alcoholism were sometimes described, but not measured in any mathematical sense. Outcome variables . The most frequently used method of measuring outcome variables was by the use of either a questionnaire or a structured interview administered to the client, and often to a collateral source or sources, at the time selected for evaluation of treatment outcome. In some studies, data on subsequent treatment or posttreatment legal problems were gathered by searching the records of treatment agenci or law-enforcement agencies. The relatively few (28) instances of measuring outcome variables by standardized methods involved the administration of drinking-history or personal adjustment instruments which had previously been normed on other groups. What Research Models Have Been Used? Correlational, descriptive, experimental, and quasi-experimental research models were used. Among the 34 reports considered, nine used correlational models; six used descriptive models; ten used experimental models; nine used quasi-experimental models. What Research Models Have Yielded Significant Results? Models yielding statistically significant results . Seven out of nine (77%) reports utilizing correlational research models reported findings significant at the .05 level or better. One report out of six (17%) using descriptive research models reported two findings significant at the .05 level. All ten studies with experimental research models reported findings statistically significant at .05 or better;

PAGE 122

114 three of the ten were reports on the same subject sample at different times of evaluation. Eight out of nine (89%) quasi-experimental studies reported findings significant at .05 or better. Models yielding practically but not statistically significant results . The question of practical, or clinical, significance is a matter of judgment. In the judgment of the present author, there were a number of studies which reported findings which were significant practically but not statistically. Those studies were distributed among research models thus: correlational, three (33%); descriptive, six (100%); experimental, six (60%); quasi-experimental, six (67%). What Interactions of Client, Treatment, and Outcome Variables Have Been Observed ? Client-treatment interactions . Two major relationships were noted. Pokorny et al. (1973) found that clients with histories characterized by stability in marriage, social relationships, and employment were more likely to participate in aftercare than those with less stability in those categories. Cutter et al. (1977) reported that rebellious, nonconformist individuals are more likely to participate in AA or egalitarian therapies, while clients characterized as more compliant and conformist tend to respond well to firm medical management. Client-outcome interactions . The most frequently mentioned client-outcome interactions had to do with relationships of prior drinking histories to outcomes and relationships of socioeconomic status and/or stability to outcomes. Other relationships, each mentioned only once, had to do with occupation, prior hospitalizations for alcoholism, degree of alcoholic deterioration, level of cognitive dysfunction, MMPI type. Medicare disability status, marital status, employment status.

PAGE 123

115 ethnicity, degree of social maladjustment, predicted difficulty of treatment, and decision to seek treatment. Table 14 in Chapter IV above provides a fuller listing of client-outcome interactions. Treatment-outcome interactions . Table 14 lists 18 instances in which studies reported relationships of treatment to outcome. Fourteen of the observed interactions had to do with the effects of various treatment methods on outcomes. Of those 14 observed interactions, 11 involved behaviorally oriented treatments. While the comparisons made were often of the superiority of one behavioral treatment method over another, e.g., random versus contingent electric shock conditioning, in each case in which behavioral treatment was compared with non-behavioral treatment (e.g., milieu therapy) in the same institution, the authors found that clients receiving behavioral treatment showed better controlled-drinking outcomes and were either superior to or not significantly different from their standard-treatment comparison group in terms of abstinence rates. The general trend in the observations of interactions of treatment with outcome would suggest the superiority of behavioral over nonbehavioral treatments. The two studies which paid special attention to the effects of length of treatment on outcome found that inpatient treatment beyond three to four weeks seems to produce no better long-term outcomes than inpatient treatment of 20 to 30 days. Client-treatment-outcome interactions . One study (Cronkite & Moos, (1980) addressed itself to the relationships of client, treatment, and outcome variables, using a complex conceptual model to attempt to clarify those relationships. They found that about one-half of the variance of multidimensional treatment-outcomes could be explained by

PAGE 124

116 the complex interaction of subject and treatment variables with each other and with posttreatment experiences. Conclusions and Interpretations In general terras, the present writer concludes that, in the period 1973-1981, long-term outcomes of psychologically oriented treatment were studied by both predictive and postdictive methods, using a variety of research models and asking a number of research questions. The reported research leads one to state with confidence that subject variables are important contributors to outcomes and that inpatient treatment longer than about 30 days is no more effective than inpatient treatment of three to four weeks. One may also conclude with confidence that the studies considered herein paid only scant and peripheral attention to the relative importance of theoretical stance as a treatment variable contributing to differential outcomes. The conclusions in the preceding paragraph were virtually demanded by the data considered. The conclusions and interpretations which follow are, in the author's opinion, suggested by the data considered, with varying degrees of logical forcefulness . There were no great surprises in the various researchers' choices of variables to be considered. The client and outcome variables mentioned were generally fairly easy to fit into the 19 categories which Emrick (1973, 1974) had extracted from studies published in the period 1952-1972. The treatment variables considered were, in the main, methodological and, in some cases, in a sense, microcosmic and mechanistic; rather than considering the admittedly difficult to define and measure elements of interand intra-personal processes of treatment, most of the authors reviewed here attended to such items as reinforce-

PAGE 125

117 ment schedules or the content and methods of teaching behaviors incompatible with the maladaptive use of alcohol. It thus appears that, whatever the postures of other treatment programs not considered in the sample of research studied herein, the data from the research reports examined herein suggest that behavioral approaches predominate in the treatment of alcoholics. Such a conclusion might, however, be an artifact of sample selection. It is safer to draw the narrower conclusion that, in the reports analyzed here, behavioral treatment approaches predominate . Certain treatment methods, in combination with certain treatment goals, appear to be more effective than others. Given controlled drinking as the desired outcome, regimens which emphasize BAC discrimination training, recognition of discriminative cues for excessive drinking, and practice in generating and rehearsing alternative behaviors appear to be promising. Given abstinence as the desired outcome, aversive counterconditioning , more traditional AA-oriented milieu therapy, and behavioral skill training emphasizing self-regulatory process all appear to be useful procedures. While behavioral techniques would appear to be as effective as or more effective than other treatment methods, it should be kept in mind that the three reports of aversive counterconditioning as the main treatment (Jackson & Smith, 1978; Neubuerger et al., 1981; Weins et al., 1976) were on programs which also included other treatment interventions of a psychosocial nature, for the effects of which there was no experimental control. Moreover, most of the subjects described in those two reports were voluntary and paying for their treatment. All of the subjects were informed before admission of the unpleasant nature of the

PAGE 126

iia counterconditioning program. These facts, coupled with the observation by Neubuerger et al. that the abstinence rate at one year for persons who were on Medicaid disability status was only about one-half that of the general sample, suggest that the apparent success of the treatment may actually be a function of the high selectivity of the sample. It may be that alcoholics who, knowing what they are getting into, pay large sums of money for emetic or electric shock conditioning are so highly motivated that almost any treatment intervention would appear successful. The Azrin (1976) study leads one toward the conclusion that an unusually high rate of abstinence (98% for the experimental group versus 45% for the control group) is obtainable by way of a low cost, outpatient community reinforcement program. The project involved 20 subjects who had been treated for alcoholism in a state hospital. They were matched in pairs, with one member of each pair randomly assigned to the treatment condition and the other to a low-intervention, traditional guidance, control condition. Thus it would appear at first glance that the approach is spectacularly successful. However, it should be noted that the outcome measure of abstinence was made two years after the beginning of outpatient treatment which apparently continued up to the time of evaluation, so that the outcome figures are not properly comparable with those of projects in which the outcomes were evaluated two years after the end of treatment. Nevertheless, should the experiment be replicated with large groups of subjects across diverse sociodemographic groups and yield outcomes of the same level, it would tend to establish Azrin' s variety of operant conditioning approach as a viable, perhaps superior, alternative to other types of treatment.

PAGE 127

119 While only one of the studies analyzed in the present work held treatment personnel as a major concern, two conclusions on that subject do suggest themselves. One is that neither graduate degrees, prior personal experience of alcoholism, nor lengthy, intensive training are required for treatment personnel effectively to carry out behavior modification interventions with alcoholics, provided that the procedures have been designed by highly competent personnel and carefully taught to the front-line practioners. The second conclusion regardng treatment personnel is drawn from Valle's 1981 study demonstrating a positive relationship between the level of interpersonal functioning of counselors and treatment outcomes. One concludes from that study that alcoholics whose primary therapists exhibit high levels of accurate empathy, genuineness, concreteness , and respect benefit more from treatment. From the paucity of references to the theoretical stances of treatment programs and personnel, the present writer concludes that the issue was not of primary interest to the researchers whose work has been considered herein. Whether theoretical stance is a demonstrably meaningful variable in terms of treatment outcomes remains to be seen. Nonstandardized methods for measuring subject and outcome variables were used more often than standardized methods. This is viewed as, in part, a reflection of the relative youth of alcoholism treatment evaluation as a discipline. It appears that the formative era of trial-and-error defining of salient variables and means for measuring them, which is common to the developmental stages of most new expansions of scientific knowledge, had not reached its end in the field of alcoholism studies by 1981. It is possible that the lack of widely-

PAGE 128

120 accepted standardized measurement instruments is a function of both the difficulty of defining and measuring the phenomena of alcoholism treatment and the social psychological/political ferment and fervor which surround efforts to define alcoholism and frame appropriate therapeutic responses to alcoholics. It appears that the choice of the "best" research model for studying alcoholism treatment outcomes is governed by the intent of the inquiry. Each research model used in the studies surveyed yielded some finding of interest. More will be said about choice of research model in the author's recommendations for further inquiry below. Most of what may be concluded regarding interactions of client, treatment, and outcome variables has been summarized above in the "Answers to Research Questions" segment. Suffice it to say here that the much to be desired universally applicable schema for matching treatment to client for optimal outcome did not emerge in the reports studied, individually or collectively. Implications for Theory, Research, Training, and Practice Implications for Theory Some of the findings reported appear to offer a serious challenge to the traditional concept of alcoholism as a unitary disease phenomenon characterized by chronicity, irreversibility, and an unvarying progressive course. The finding that diagnosed gamma-type alcoholics can establish and maintain patterns of controlled drinking has been and will probably continue to be hotly disputed. The present author does not wish to attempt to settle this dispute here, but mentions it as a means of highlighting the unfinished task of developing an alcoholism theory that can explain both these findings and the contention of the tradi-

PAGE 129

121 tional viewpoint, for which there is informal empirical support, that no "true alcoholic" can ever return to and maintain moderate drinking. An obvious step in developing such a theory, in the view of the present writer, would be the devising of an empirically based nosology of maladaptive alcohol use, which nosology would lend itself to differential diagnosis and differentiated treatment goals. Implications for Research The diversity and sometimes lack of definition of basic terms, e.g. "alcoholism," "controlled drinking" in the papers reviewed here implies the need for researchers to adopt an explicit and consistent vocabulary. The absence of strong support for any method of matching treatments to clients may be viewed as an indicant for careful research focused on long-term outcomes for "matched" and "mismatched" clients. Implicatins for Training Two implications for the training of alcoholism treatment personnel emerge with clarity. One is that students preparing for careers in alcoholism treatment would be well advised to become thoroughly grounded in the concepts and technology of behavior modification. Workers so prepared would have the advantages of being able to use behavioral techniques should they be employed in behaviorally oriented programs, and of being better able to comprehend the literature regardless of their employment settings. The second implication for training is that alcoholism workers appear to be more effective if they are high functioning in the interpersonal dimensions of accurate empathy, genuineness, concreteness , and respect. Training for alcoholism treatment personnel should include an emphasis on this core, regardless of what treatment methods will be used by the practioner.

PAGE 130

122 Implications for Practice While no "best" treatment for alcoholism has emerged from the studies analyzed herein, it is apparent that methods which include helping the client to recognize situations in which there is a high probability of maladaptive response and either to avoid those situations or generate and carry out competing, adaptive behaviors in them tend to eventuate in good treatment outcomes. No obvious superiority for inpatient or outpatient treatment for the entire range of drinking problems seems to have been demonstrated. Generally it seems that when the treatment goal is controlled drinking, outpatient treatment is as successful as inpatient, and more cost-effective. When the goal is abstinence, it appears that inpatient treatment of about 30 days followed by two years of outpatient aftercare counseling and AA involvement is relatively effective. Emrick's (1975) observation that all treatment approaches appear to be generally helpful to most of the alcoholics treated is upheld by the studies considered herein. This being the case, the present writer shares Emrick's inference that practioners would do well to devote considerable emergy to getting alcoholics into some treatment even although the quest for the treatment is not yet at an end, and perhaps never will be. Recommendations As was stated in Chapter I above, the first purpose of this study is to determine whether the research considered herein can provide answers to the research questions of this study. The material reviewed here provided answers to some but not all of this study's research

PAGE 131

questions, which fact leads into the second purpose, namely, to suggest ways in which the needed information might be obtained. Those suggestions follow in the form of recommendations for theory development and future research. Recommendations for Theory Development It is recommended that alcoholism treatment theoreticians, which includes just about everybody involved with alcoholism treatment and research, pay heed to the scientific tradition of inductive development of theories. It appears that traditional concepts of alcoholism, especially as embodied in the ideology of AA, have arisen from the facts of experience, informally gathered and more or less intuitively systematized into a set of propostions which are often granted the status of established theory and have been used deductively to guide thought and behavior in response to alcoholism. '. ' While the traditional concept of alcoholism is numerically dominant in the field of alcoholism, there are competing concepts, which sometimes appear also to have attained the revered status of supported theories. It seems to this author that it would be both good politics and good science if the concerned parties were to agree to a moratorium on deduction for a time and concentrate on three tasks. The first would be to systematize and evaluate the data already extant, by whatever means it has been gathered. This would mean taking seriously the experiences of the more than one million men and women who constitute the fellowship of Alcoholics Anonymous and taking seriously the data amassed by social and biological scientists using more formal, rigorous data-gathering methods.

PAGE 132

124 The second task would be to discover voids in the data, i.e., to find out what important questions about the phenomena of "alcoholism" were unanswered by the data at hand, and to gather the data needed to answer those questions. The third task would be inductively to derive a comprehensive theory of "alcoholism," which this writer predicts would be a theory of "alcoholisms." Recommendations for Further Research This study has been, in part, an effort toward identifying fruitful approaches to alcoholism treatment outcome research. It is apparent that each approach examined herein has been fruitful in some way. The fundamental element with regard to productivity in research is the question, "Productive of what?" The present writer set out to identify avenues of approach that tend to produce results that are clear and pragmatically, clinically significant. It now appears to him that obtaining clear and significant results is a function of asking clear and significant questions; using a research model appropriate to the questions; and analyzing, interpretng, and reporting results clearly. It is recommended, then, that future researchers pay careful attention to the selection of research questions and to the appropriate matching of questions, research method, data, and reporting. The basic criteria for good research described by Hill and Blane (1967) provide clear guidelines which future researchers would do well to follow. It is recommended that further inquiries be made into the question of alcoholics' motivation for changing and for not changing, with or without treatment. Although Hill and Blane had in their 1967 article stressed the importance of controlling for differential motivation in outcome evaluation studies, not all researchers reviewed here have done

PAGE 133

125 so. It is recommended that future researchers do control for motivation. Research reviewed herein left unanswered some questions the present writer deems important. It is recommended that further inquiry be made, by way of carefully controlled experimental studies with outcomes measured at a minimum of one year after cessation of treatment, into the following questions: (a) To what extent do the ideologies, operating assumptions, and theories of treatment personnel influence outcomes of alcoholism treatment? (b) Which types of treatment for what sorts of clients produce what kinds of outcomes? (c) What is the influence of the treatment goal preferences of treatment personnel on outcomes? (d) What is the influence of the treatment goal preferences of clients on outcomes? (e) What are objective, empirically derived guidelines for matching of client, treatment personnel, and treatment goals and procedures for optimal outcomes? Finally, it is recommended that more effort be expended in meta-analysis and synthesis of the findings of alcoholism treatment research. Specifically, it is recommended that there be published an annual, comprehensive, meta-analytic, synthetic review of progress in alcoholism treatment. Such a review should be published in two versions, one a technical version for research-oriented professionals, the other an easily-comprehended, practice-oriented, less technical summary for front-line alcoholism workers. Recommendations for Training The author offers six recommendations for the training of present and prospective alcoholism treatment personnel.

PAGE 134

126 questions, which fact leads into the second purpose, namely, to suggest ways in which the needed information might be obtained. Those suggestions follow in the form of recommendations for theory development and future research. Recommendations for Theory Development It is reconunended that alcoholism treatment theoreticians, which includes just about everybody involved with alcoholism treatment and research, pay heed to the scientific tradition of inductive development of theories. It appears that traditional concepts of alcoholism, especially as embodied in the ideology of AA, have arisen from the facts of experience, informally gathered and more or less intuitively systematized into a set of propositions which are often granted the status of established theory and have been used deductively to guide thought and behavior in response to alcoholism. While the traditional concept of alcoholism is numerically dominant in the field of alcoholism, there are competing concepts, which sometimes appear also to have attained the revered status of supported theories. It seems to this author that it would be both good politics and good science if the concerned parties were to agree to a moratorium on deduction for a time and concentrate on three tasks. The first would be to systematize and evaluate the data already extant, by whatever means it has been gathered. This would mean taking seriously the experiences of the more than one million men and women who constitute the fellowship of Alcoholics Anonymous and taking seriously the data amassed by social and biological scientists using more formal, rigorous data-gathering methods.

PAGE 135

127 The second task would be to discover voids in the data, i.e., to find out what important questions about the phenomena of "alcoholism" were unanswered by the data at hand, and to gather the data needed to answer those questions. The third task would be inductively to derive a comprehensive theory of "alcoholism," which this writer predicts would be a theory of "alcoholisms." Recommendations for Further Research This study has been, in part, an effort toward identifying fruitful approaches to alcoholism treatment outcome research. It is apparent that each approach examined herein has been fruitful in some way. The fundamental element with regard to productivity in research is the question, "Productive of what?" The present writer set out to identify avenues of approach that tend to produce results that are clear and pragmatically, clinically significant. It now appears to him that obtaining clear and significant results is a function of asking clear and significant questions; using a research model appropriate to the questions; and analyzing, interpretng, and reporting results clearly. It is recommended, then, that future researchers pay careful attention to the selection of research questions and to the appropriate matching of questions, research method, data, and reporting. The basic criteria for good research described by Hill and Blane (1967) provide clear guidelines which future researchers would do well to follow. It is recommended that further inquiries be made into the question of alcoholics' motivation for changing and for not changing, with or without treatment. Although Hill and Blane had in their 1967 article stressed the importance of controlling for differential motivation in outcome evaluation studies, not all researchers reviewed here have done

PAGE 136

128 so. It is recommended that future researchers do control for motivation. Research reviewed herein left unanswered some questions the present writer deems important. It is recommended that further inquiry be made, by way of carefully controlled experimental studies with outcomes measured at a minimum of one year after cessation of treatment, into the following questions: (a) To what extent do the ideologies, operating assumptions, and theories of treatment personnel influence outcomes of alcoholism treatment? (b) Which types of treatment for what sorts of clients produce what kinds of outcomes? (c) What is the influence of the treatment goal preferences of treatment personnel on outcomes? (d) What is the influence of the treatment goal preferences of clients on outcomes? (e) What are objective, empirically derived guidelines for matching of client, treatment personnel, and treatment goals and procedures for optimal outcomes? Finally, it is recommended that more effort be expended in meta-analysis and synthesis of the findings of alcoholism treatment research. Specifically, it is recommended that there be published an annual, comprehensive, meta-analytic, synthetic review of progress in alcoholism treatment. Such a review should be published in two versions, one a technical version for research-oriented professionals, the other an easily-comprehended, practice-oriented, less technical summary for front-line alcoholism workers. Recommendations for Training The author offers six recommendations for the training of present and prospective alcoholism treatment personnel.

PAGE 137

129 1. They should be trained to high levels of interpersonal functioning on the dimensions of accurate empathy, genuineness, concreteness, and respect as discussed in various writings of Truax and Carkhuff (e.g., Carkhuff, 1969; Carkhuff & Truax, 1966). 2. They should be acquainted with the conceptual framework of behavioral therapy and with several methods of behavior modification, and it should be emphasized in the training that behavior modification can be a tool for the implementation of humanistic values and that behavioral methods are not inimical to humanistic values. 3. They should be acquainted with several conceptual models of alcoholism and with the treatment implications of each model insofar as those implications are known. 4. They should acquire skill in getting alcoholics into treatment. 5. They should learn to communicate across lines of discipline, ideology, and levels of formal training. 6. They should be encouraged to understand the practical values of research and to cooperate with research efforts. Recommendations for Practice The author proposes six recommendations for practioners of psychologically oriented alcoholism treatment. 1. Since it appears that generally alcoholics benefit somewhat from almost any kind of treatment, practioners should develop and use energetically those skills and procedures which will promote entry into treatment. 2. Especially in an era of shrinking resources, practioners should try out minimum interventions, such as brief outpatient

PAGE 138

1 130 counseling or bibliotherapy, with clients who have not undergone serious alcoholic deterioration. 3. Practioners should try to match the treatment to the client on dimensions including, but not limited to, personality, treatment-goal preference, degree of deterioration, and socioeconomic situation. 4. Practioners should offer aftercare for alcoholics who undergo inpatient or residential treatment; practioners should be aggressive and consistent in their efforts to encourage recovering alcoholics to participate in aftercare for one to two years. 5. Practioners should encourage, support, and cooperate with research efforts in their treatment settings. 6. Practioners should read research reports and seek to test and apply research findings in their practices. Summation Thirty-four studies published from 1973 through 1981, and evaluating outcomes of psychological treatment for ten or more alcoholics at 12 months or more after treatment, were reviewed. All studies included data on subject, treatment, and outcome variables. Subject variables were reported in 17 categories, with previous drinking history reported in all studies. Outcome variables were reported in 19 categories, with drinking behavior as an outcome variable considered most frequently (in 33 studies). Treatment variables were reported in six categories, with treatment method reported in all studies. No study treated the theoretical stance of the treatment program as a variable of focal importance. Subject variables were measured 95 times by nonstandardized methods and 51 times by standardized methods. Outcome variables were measured

PAGE 139

131 148 times by nonstandardized methods and 28 times by standardized methods . Correlational research models were used in nine studies, descriptive research models in six, experimental research models in ten, and quasi-experimental research models in nine. Every study in which an effort was made to compute statistical significance reported at least one finding significant at the .05 level or better. Findings deemed practically significant but not statistically significant were identified in 21 studies. Two studies reported subject-treatment interactions. Subjectoutcome interactions were reported in 13 studies. Seventeen studies reported treatment-outcome interactions. One study reported on subjecttreatment-outcome interactions. Implications for theory, research, training, and practice were described. Recommendations were made for further developments in theory, research, training, and practice.

PAGE 140

LIST OF REFERENCES Abbott, M.W., & Gregson, R.A. Cognitive dysfunction in the prediction of relapse in alcoholics. Journal of Studies on Alcohol , 1981, 42, 230-243. ~ Alford, G.S. Alcoholics anonymous: An empirical outcome study. Addictive Behaviors , 1980, 5, 359-370. Azrin, N.H. Improvements in the community-reinforcement approach to alcoholism. Behavior Research and Therapy , 1976, 14, 339-348. Baekeland, F. Evaluation of treatment methods in chronic alcoholism. In B. Kissin & H. Begleiter (Eds.), The biology of alcoholism. Vol. 5. Treatment and rehabilitation of the chronic alcoholic . New York: Plenum Press, 1977. Bowen, W.T., & Tweemlow, S.W. A follow-up study of alcoholics who failed to appear for treatment. Hospital and Community Psychiatry , 1980, 31, 349-351. Bromet, E., Moos, R.H., Bliss, F., & Wuthman C. The post-treatment functioning of alcoholic patients: Its relation to program participation. Journal of Consulting and Clinical Psychology , 1977, 45, 829-842. Caddy, G.R., & Lovibond, S.H. Self-regulation and discriminated aversive conditioning in the modification of alcoholics' drinking behavior. Behavior Therapy , 1976, 7, 223-230. Caddy, G.R., Addington, H. J. , & Perkins, D. Individualized behavior therapy for alcoholics: A third year independent double-blind follow-up. Behaviour Research and Therapy , 1978, 16^, 345-362. Camp, J.M., & Kurtz, N.R. An analysis of issues affecting the alcoholism work force. In National Institute on Alcohol Abuse and Alcoholism, Prevention, intervention, and treatment: Concerns and models (AlcohoT and Health Monograph No. 3). Rockville, Md.: the Institute, in press. Carkhuff, R.R. Helping and human relations: A primer for lay and profes sional helpers . (2 vols.). New York: Holt, Rinehart, & Winston, \9W. Carkhuff, R.R. , & Truax, C.B. Toward explaining success and failure in interpersonal learning experiences. Personnel and Guidance Journal , 1966, 44, 723-728. Chaney, E.F., O'Leary, M.R. , & Marlatt, G.A. Skill training with alcoholics. Journal of Consulting and Clinical Psychology, 1978, 46, 1092-1104. 132

PAGE 141

133 Cohen, R., Appelt, H. , Olbrich, R., & Watzl, H. Alcoholic women treated by behaviorally orientated therapy: An 18-month follow-up study. Drug and Alcohol Dependence , 1979, 4, 489-498. Conley, J.J. An MMPI typology of male alcoholics: Admission, discharge, and outcome comparisons. Journal of Personality Assess ment , 1981, 45, 33-39. Costello, R.M. Alcoholism treatment and evaluation: In search of methods. International Journal of the Addictions , 1975, 10, 251-275. (a) Costello, R.M. Alcoholism treatment and evaluation: In search of methods. II. Collation of two-year follow-up studies. International Journal of the Addictions , 1975, 10, 857-867. (b) Costello, R.M. Alcoholism aftercare and outcome: Cross-lagged panel and path analyses. British Journal of Addiction , 1980, 75, 49-53. Costello, R.M., Baillargeon, J.G. , Biever, P., & Bennett, R. Secondyear alcoholism treatment outcome evaluation with a focus on Mexican-American patients. American Journal of Drug and Alcohol Abuse, 1979, 6, 97-108. Cronkite, R.C., & Moos, R.H. Determinants of the posttreatment functioning of alcoholic patients: A conceptual framework. Journal of Consulting and Clinical Psychology , 1980, 48, 305-316. Cronkite, R.C., & Moos, R.H. Evaluating alcoholism treatment programs: An integrated approach. Journal of Consulting and Clinical Psychology , 1978, 46, 1105-1119. Current literature. Journal of Studies on Alcohol , 1981, 42, 565-567. Cutter, H.S.G., Boyatzis, R.E. , & Clancy, D.D. Effectiveness of power motivation training in rehabilitating alcoholics. Journal of Studies on Alcohol , 1977, 38, 131-141. DeLuca, J.R. (Ed.) Fourth special report to the U.S. Congress on alcohol and health . Washington, D.C.: U.S. Government Printing Office, 1981. :. Emrick, CD. Psychological treatment of alcoholism: An analytic review (Doctoral dissertation, Columbia University, 1973). Dissertation Abstracts International , 1973, 34, 2926. (University Microfilms No. 73-28, 463) ~ Emrick, CD. A review of psychologically oriented treatment of alcoholism: I. The use and interrelationships of outcome criteria and drinking behavior following treatment. Quarterly Journal of Studies on Alcohol , 1974, 35, 523-549.

PAGE 142

134 Emrick, CD. A review of psychologically oriented treatment of alcoholism: II. The relative effectiveness of different treatment approaches and the effectiveness of treatment versus no treatment. Journal of Studies on Alcohol , 1975, 36, 88-108. Emrick, CD. Perspectives in clinical research: Relative effectiveness of alcohol abuse treatment. Family and Community Health , 1979, 2 (2), 71-78. Gibbs, L. & Flanagan, J. Prognostic indicators of alcoholism treatment outcome. International Journal of the Addictions , 1977, 12, 1097-1141. Glass, G.V. Primary, secondary, and meta-analysis of research. Educational Researcher , 1976, 5, (10), 3-8. Gordon, J.E. The epidemology of alcoholism. New York State Journal of Medicine , 1958, 58, 1911-1917. Gottheil, E., Thornton, CC, Skoloda, T.E., & Alterman, A.I. Followup study of alcoholics at 6, 12, and 24 months. In M. Galanter (Ed.), Currents in alcoholism (Vol. 6). New York: Grune & Stratton, 1979. Hill, M.J., & Blane, H.T. Evaluation of psychotherapy with alcoholics: A critical review. Quarterly Journal of Studies on Alcohol , 1967, 28, 76-104. Isaac, S., & Michael, W.B. Handbook in research and evaluation . San Diego: EdITS, 1971. Jackson, T.R. , St Smith, J.W. A comparison of two aversion treatment methods for alcoholism. Journal of Studies on Alcohol , 1978, 39, 187-191. Jacobson, G.R. The alcoholisms . New York: Human Sciences Press, 1976. Jellinek, E.M. The disease concept of alcoholism . New Haven: Hillhouse Press, 1960. Johnson, V.E. I'll quit tomorrow . New York: Harper & Row, 1973. Jones, S.L., & Lanyon, R.I. Relationship between adaptive skills and outcome of alcoholism treatment. Journal of Studies on Alcohol, 1981, 42, 521-525. Keeley, K.A. Introduction to evaluation and outcome. In M. Galanter (Ed.), Currents in alcoholism (Vol. 6.) New York: Grune & Stratton, 1979. Keller, M. Definition of alcoholism. Quarterly Journal of Studies on Alcohol , 1960, 21, 125-134.

PAGE 143

135 Kissin, B. , Platz, A., & Su, W.H. Social and psychological factors in the treatment of chronic alcoholics. Journal of Psychiatric Research , 1970, 8, 13-27. Kliner, D.J., Spicer, J., & Barnett, P. Treatment outcome of alcoholic physicians. Journal of Studies on Alcohol , 1980, 41, 1217-1220. Linsky, A.S. Theories of behavior and the social control of alcoholism. Social Psychiatry , 1972, 7, 47-52. Maisto, S.A., Sobell, M.B. , & Sobell, L.C. Predictors of treatment outcome for alcoholics treated by individualized behavior therapy. Addictive Behaviors , 1980, 5, 259-264. McClelland, D.C. The impact of power motivation training on alcoholics. Journal of Studies on Alcohol , 1977, 38, 142-144. McLachlan, J.F.C. Therapy strategies, personality orientation, and recovery from alcoholism. Canadian Psychiatric Association Journal , 1974, 19, 25-30. McWilliaras, J., & Brown, C.C. Treatment termination variables, MMPI scores, and frequencies of relapse in alcoholics. Journal of Studies on Alcohol , 1977, 38, 477-486. Miller, W.R. Behavioral treatment of problem drinkers: A comparative study of three controlled drinking therapies. Journal of Consulting and Clinical Psychology , 1978, 46, 74-86. Miller, W.R. , & Taylor, C.A. Relative effectiveness of bibliotherapy , individual and group self-control training in the treatment of problem drinkers. Addictive Behaviors , 1980, 5, 13-24. Moll, J.K. , & Narin, F. Characterization of the alcohol research literature. Journal of Studies on Alcohol , 1977, 38, 2165-2180. National Institute on Alcohol Abuse and Alcoholism. National directory of drug abuse and alcoholism treatment programs . (Rev. 1979) . (U.S. Department of Health, Education, and Welfare Publication No. ADM 76-321). Washington, D.C: U.S. Government Printing Office, 1979. Neubuerger, O.W., Hasha, N. , Matarazzo, J.D., Schmitz, R.E., & Pratt, H.H. Behavioral-chemical treatmet of alcoholism: An outcome replication. Journal of Studies on Alcohol , 1981, 42, 806-810. Pattison, E.M. The selection of treatment modalities for the alcoholic patient. In J.H. Mendelson & N.K. Mello (Eds.), The diagnosis and treatment of alcoholism . New York: McGraw-Hill, 1979. Pattison, E.M., Coe, R. , & Doerr, H.O. Population variation among alcoholism treatment facilities. Internatio nal Journal of Addictions, 1973, 8, 199-229.

PAGE 144

Pattison, E.M., Coe, R., & Rhodes, R.A. Evaluation of alcoholism treatment: Comparison of three facilities. Archives of General * Psychiatry , 1969, 20, 478-488. Pattison, E.M., Sobell, M.B., & Sobell, L.C. Emerging concepts of alcohol dependence . New York: Springer Publishing Company, 1977. Pokorny, A.D., Miller, B.A. , Kanas, T. , & Valles, J. Effectiveness of extended aftercare in the treatment of alcoholism. Quarterly Journal of Studies on Alcohol , 1973, 34, 435-443. Poley, W. , Lea, G. , & Vibe, G. Alcoholism: A treatment manual . New York: Gardner Press, 1979. Polich, J.M., Armor, D. J. , & Braiker, H.B. The course of alcoholism . New York: John Wiley and Sons, 1981. Scherer, S.E., & Freedberg, E.J. Effects of videotape feedback on development of assertiveness skills in alcoholics: A follow-up study. Psychological Reports , 1976, 39, 983-992. Schuckitt, M.A. Drug and alcohol abuse: A clinical guide to diagnosis and treatment . New York: Plenum Medical Book Company, 1979. Smart, R.G. Do some alcoholics do better in some types of treatment than others? Drug and Alcohol Dependence , 1978, 3, 65-75. Sobell, M.B., & Sobell, L.C. Alcoholics treated by individualized behavior therapy: One year treatment outcome. Behavior Research and Therapy , 1973, 11, 599-618. Sobell, M.B., & Sobell, L.C. Second year treatment outcome of alcoholics treated by individualized behavior therapy: Results. Behavior Research and Therapy , 1976, U, 195-215. Stein, L.I., Newton, J.R. , & Bowman, R.S. Duration of hospitalization for alcoholism. Archives of General Psychiatry , 1975, 32, 247-252. Swenson, P.R. , StruckmanJohnson, D.L., Ellingstad, V.S., Clay, T.R. , & Nichols, J.L. Results of a longitudinal evaluation of court-mandated DWI treatment programs in Phoenix, Arizona. Journal of Studies on Alcohol , 1981, 42, 642-653. Tomsovic, M. "Binge" and continuous drinkers: Characteristics and treatment follow-up. Quarterly Journal of Studies on Alcohol. 1974, 35, 558-564. — Trice, H.M., Roman, P.M., & Belasco, J. A. Selection for treatment: a predictive evaluation of an alcoholic treatment regimen. International Journal of Addictions , 1969, 4, 303-307. Valle, S.K. Interpersonal functioning of alcoholism counselors and treatment outcome. Journ al of Studies on Alcohol. 1981 42 783-790. ~ [ ' ' — '

PAGE 145

137 Vogler, R.E., Compton, J.V., & Weissbach, T.A. Integrated behavior change techniques for alcoholics. Journal of Consulting and Clinical Psychology , 1975, 43, 233-243. Vogler, R.E., Ferstl, R., Kraemer, S., & Brengelmann, J.C. Electrical aversion conditioning of alcoholics: One year follow-up. Journal of Behaviour Therapy and Experimental Psychiatry , 1975, 6, 171-173. Weins, A.N. , Montague, J.R., Manaugh, T.S., & English, C.J. Pharmacological aversive counterconditioning to alcohol in a private hospital: One-year followup. Journal of Studies on Alcohol , 1976, 37, 1320-1324. Willems, P. J. A., Letemendia, F.J.J. , & Arroyave, F. A two-year followup study comparing short with long stay inpatient treatment of alcoholics. British Journal of Psychiatry , 1973, 122, 637-648. Wolf I., Chafetz, M.E., Blane, H.T. , & Hill, M.J. Social factors in the diagnosis of alcoholism: Attitudes of physicians. In Chafetz, M.E., Blane, H.T. , & Hill, M.J. (Eds.), Frontiers of alcoholism . New York: Science House, 1970. Zimberg, S., Wallace, J., & Blume, S.B. (Eds.). Practical approaches to alcoholism psychotherapy . New York: Plenum Press, 1978.

PAGE 146

APPENDIX A DATAGATHERING PROTOCOL 1. Author(s), date: , 2. Treatment type: 3. Treatment duration: 4. Outcome criteria: 5. Time of evaluation: 6. Sample size and character: 7. N evaluated: 8. Drinking outcome: class ; % ; uncommoness 9. Treatment locus: 10. Treatment methods: 11. Treatment personnel: 12. Definition of alcoholism if given: 13. Theoretical stance: 14. Research model: 15. Statistically significant major findings: 16. Practically but not statistically significant major findings 17. Interactions: ^ , . ' 18. Research questions: 19. Subject and outcome variables: Subject-baseline Outcome Variable How measured How measured Drinking amount or frequency Affective cognitive 138

PAGE 147

Work situation Home situation: relationships Inpatient treatment Physical Legal problems Mixed Social situation A. A. Finances Global adjustment Miscellaneous Other drinking behavior Home situation: residence Outpatient treatment Leisure time Religious life Treatment: locus unclear

PAGE 148

APPENDIX B INTERJUDGE AGREEMENT After the data-gathering protocol had been completed on all sample articles, the second reader was given 45 minutes of instruction on the use of the protocol. The author selected five articles for their variety in length, content, research model, and manner of reporting. The second reader read the selected articles and completed the protocol on each. In one instance the second reader consulted with the author on the computation of an outcome percentage; otherwise there was no further discussion of the use of the protocol until after the second reader had read and completed the protocols on the five articles. Each protocol form allowed for 59 possible reader responses, so that the five completed protocols allowed a total of 295 opportunities for disagreement. Twenty-eight instances of disagreement were counted when the two sets of protocols were compared, leaving 267 instances of agreement. The interjudge agreement rate was computed to be 90.5%. 140

PAGE 149

APPENDIX C REFERENCE LIST FOR STUDIES ANALYZED Abbott, M.W. , & Gregson, R.A. Cognitive dysfunction in the prediction of relapse in alcoholics. Journal of Studies on Alcohol , 1981, 42, 230-243. Alford, G.S. Alcoholics anonymous: An empirical outcome study. Addictive Behaviors , 1980, 5, 359-370. Azrin, N.H. Improvements in the comjuunity-reinforcement approach to alcoholism. Behavior Research and Therapy , 1976, 14, 339-348. Bowen, W.T., & Tweemlow, S.W. A follow-up study of alcoholics who failed to appear for treatment. Hospital and Community Psychiatry, 1980, 31, 349-351. Caddy, G.R. , & Lovibond, S.H. Self-regulation and discriminated aversive conditioning in the modification of alcoholics' drinking . behavior. Behavior Therapy , 1976, 7, 223-230. Caddy, G.R., Addington, H.J. , & Perkins, D. Individualized behavior therapy for alcoholics: A third year independent double-blind follow-up. Behaviour Research and Therapy , 1978, 1_6, 345-362, Chaney, E.F. , O'Leary, M.R. , & Marlatt, G.A. Skill training with alcoholics. Journal of Consu lting and Clinical Psychology, 1978. 46, 1092-1104. ^ ^ Cohen, R. , Appelt, H. , Olbrich, R. , & Watzl, H. Alcoholic women treated by behaviorally orientated therapy: An 18-month follow-up study. Drug and Alcohol Dependence , 1979, 4, 489-498. Conley, J.J. An MMPI typology of male alcoholics: Admission, discharge, and outcome comparisons. Journal of Personality .Assessment, 1981, 45, 33-39. "~~ Costello, R.M. Alcoholism aftercare and outcome: Cross-lagged panel and path analyses. British Journal of Addiction , 1980, 75, 49-53. Costello, R.M., Baillargeon, J.G., Biever, P., & Bennett, R. Secondyear alcoholism treatment outcome evaluation with a focus on Mexican-American patients. American Journal of Drug and Alcohol Abuse , 1979, 6, 97-108. Cronkite, R.C., & Moos, R.H. Determinants of the posttreatment functioning of alcoholic patients: A conceptual framework. Journal of Consulting and Clinical Psychology , 1980, 48, 305-316. Cutter, H.S.G., Boyatzis, R.E., & Clancy, D.D. Effectiveness of power motivation training in rehabilitating alcoholics. Journal of Studies on Alcohol . 1977, 38, 131-141. 141

PAGE 150

142 Jackson, T.R., & Smith, J.W. A comparison of two aversion treatment methods for alcoholism. Journal of Studies on Alcohol , 1978, 39, 187-191. ~ Jones, S.L., & Lanyon, R.I. Relationship between adaptive skills and outcome of alcoholism treatment. Journal of Studies on Alcohol , 1981, 42, 521-525. Kliner, D. J. , Spicer, J., & Barnett, P. Treatment outcome of alcoholic physicians. Journal of Studies on Alcohol , 1980, 41, 1217-1220. Maisto, S.A., Sobell, M.B., & Sobell, L.C. Predictors of treatment outcome for alcoholics treated by individualized behavior therapy. Addictive Behaviors , 1980, 5, 259-264. McClelland, D.C. The impact of power motivation training on alcoholics. Journal of Studies on Alcohol , 1977, 38, 142-144. McWilliams, J., & Brown, C.C. Treatment termination variables, MMPI scores, and frequencies of relapse in alcoholics. Journal of Studies on Alcohol , 1977, 38, 477-486. Miller, W.R. Behavioral treatment of problem drinkers: A comparative study of three controlled drinking therapies. Journal of Consulting and Clinical Psychology , 1978, 46, 74-86. Miller, W.R., & Taylor, C.A. Relative effectiveness of bibliotherapy, individual and group self-control training in the treatment of problem drinkers. Addictive Behaviors , 1980, 5, 13-24. Neubuerger, O.W., Hasha, N. , Matarazzo, J.D. , Schmitz, R.E., & Pratt, H.H. Behavioral-chemical treatment of alcoholism: An outcome replication. Journal of Studies on Alcohol , 1981, 42, 806-810. Pokorny, A.D. , Miller, B.A. , Kanas, T. , & Valles, J. Effectiveness of extended aftercare in the treatment of alcoholism. Quarterly Journal of Studies on Alcohol . 1973, 34, 435-443. Scherer, S.E., Se Freedberg, E.J. Effects of videotape feedback on development of assertiveness skills in alcoholics: A follow-up study. Psychological Reports . 1976, 39, 983-992. Sobell, M.B., & Sobell, L.C. Alcoholics treated by individualized behavior therapy: One year treatment outcome. Behavior Research and Therapy , 1973, U, 599-618. " Sobell, M.B., & Sobell, L.C. Second year treatment outcome of alcoholics treated by individualized behavior therapy: Results. Behavior Research and Therapy , 1976, 14, 195-215. Stein, L.I., Newton, J.R. , & Bowman, R.S. Duration of hospitalization for alcoholism. Archives of General Psychiatry , 1975, 32, 247-252.

PAGE 151

143 Swenson, P.R., StruckmanJohnson, D.L., Ellingstad, V.S., Clay, T.R., & Nichols, J.L. Results of a longitudinal evaluation of court-mandated DWI treatment programs in Phoenix, Arizona. Journal of Studies on Alcohol , 1981, 42, 642-653. Tomsovic, M. "Binge" and continuous drinkers: Characteristics and treatment follow-up. Quarterly Journal of Studies on Alcohol, 1974, 35, 558-564. Valle, S.K. Interpersonal functioning of alcoholism counselors and treatment outcome. Journal of Studies on A lcohol, 1981, 42, 783-790. — Vogler, R.E., Compton, J.V., & Weissbach, T.A. Integrated behavior change techniques for alcoholics. Journal of Consulting and and Clinical Psychology , 1975, 43, 233-243. Vogler, R.E., Ferstl, R. , Kraemer, S., & Brengelmann, J.C. Electrical aversion conditioning of alcoholics: One year follow-up. Journal of Behaviour Therapy and Experimental Psychiatry , 1975, 6, 171-173. Wiens, A.N., Montague, J.R. , Manaugh, T.S., & English, C.J. Pharmacological aversive counterconditioning to alcohol in a private hospital: One-year followup. Journal of St udies on Alcohol, 1976, 37, 1320-1324. — Willems, P. J. A., Letemendia, F.J.J. , & Arroyave, F. A two-year followup study comparing short with long stay inpatient treatment of alcoholics. British Journal of Psychiatry , 1973, 122, 637-648.

PAGE 152

BIOGRAPHICAL SKETCH James D. Poage was born November 7, 1935, in Bartlesville, Oklahoma. He attended public schools in Oklahoma and graduated from Oklahoma Baptist University in 1958. He earned graduate degrees in theology from Southeastern Baptist Theological Seminary and worked variously as a pastor, psychotherapist, and drug abuse program administrator before entering the University of Florida in 1973. At this writing Mr. Poage is director of .Anabasis, Inc., an alcoholism treatment center in Sarasota, Florida. He is married and the father of three sons. 144

PAGE 153

I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. Professor of Counselor Education I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. Paul W. Fitzgerald Professor of Counselor Education This dissertation was submitted to the Graduate Faculty of the Department of Counselor Education in the College of Education and to the Graduate Council, and was accepted as partial fulfillment of the requirements for the degree of Doctor of Philosophy. May, 1982 F. G. Stehli Dean for Graduate Studies and Research