, . TTTTTTTTTTTTI
( \CW‘ 2) LIBRARY
Michigan State
University

THE EFFECTS OF PSYCHIATRIC SYMPTOMATOLOGY ON
INVOLVEMENT AND BENEFIT FROM SELF-HELP IN A
DUALLY DIAGNOSED SAMPLE

presented by

LISA C. JORDAN

has been accepted towards fulfillment
of the requirements for

PH.D. PSYCHOLOGY-URBAN AFFAIRS

degree in

 

 

   

 

Major professor

WILLIAM s. DAVIDSON, 11
Date JUNE 10, 1996

 

MS U is an Affirmative Action/Equal Opportunity Institution 0-12771

 

PLACE IN RETURN BOX
to remove this checkout from your record.
TO AVOID FINES return on or before date due.

 

r DATE DUE

MTE DUE

DATE DUE

 

W

 

W

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

1/98 chIRCJD‘DuopGS—p“

THE EFFECTS OF PSYCHIATRIC SYMPTOMATOLOGY ON INVOLVEMENT
AND BENEFIT FROM SELF-HELP IN A DUALLY DIAGNOSED SAMPLE
By

Lisa C. Jordan

A DISSERTATION

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY
Department of Psychology
and

Urban Studies

1996

ABSTRACT
THE EFFECTS OF PSYCHIATRIC SYMPTOMATOLOGY ON INVOLVEMENT
AND BENEFIT FROM SELF-HELP IN A DUALLY DIAGNOSED SAMPLE
BY

Lisa C. Jordan

This study was designed to investigate the efficacy of Alcoholics and
Narcotics Anonymous (AA/NA) programs for persons with dual diagnoses of
substance abuse and mental illness, using correlational analyses and E08
structural equation modeling with data from a group of 474 persons with dual
diagnosis. All of the participants in this study received treatment at a public
psychiatric hospital, and many were subsequently referred to AA/NA for sobriety
support services. The current analyses were conducted with data from the hospital
admissions interview, and four follow-up interviews, concluding at 14-months
post-hospitalization.

The results indicated that from hospital discharge through 10-months
post-hospitalization, participants in this sample were attending AAINA at rates
equivalent to those reported in studies of persons with singly diagnosed substance
abuse disorders. These findings suggest that dual diagnosis is not necessarily a
deterrent to AA/NA participation. However, participants with the most severe and
chronic psychiatric problems were less likely to attend AA/NA, and were more likely
to drop out if they did attend. Results from the structural equation models did not

support the expectation of differential outcomes from AA/NA and other service use

based on psychiatric symptomatology and chronicity. Persons who attended
AAINA regularly had fewer alcohol and drug problems, and fewer psychiatric
problems at 10- and 14-months post-discharge, regardless of prior symptomatology
and chronicity. While AAINA participation was not directly related to psychological
or family/social problems, or rehospitalization, reduction in alcohol and drug
problems seemed to mediate the relationship between AAINA service use and
functioning in these other domains. Use of outpatient mental health services alone,
or in combination with AAINA was not significantly related to better outcomes.
Participants with more severe and chronic psychiatric problems were more likely to
continue using mental health services after hospitalization, although these services

were generally not related to positive outcomes.

Copyright by
Lisa C. Jordan
1996

To John Henry Jordan, Sr. the person I admire most in this world.

Thanks for you love and support. Your Baby Girl has become a doctor!

ACKNOWLEDGMENTS

Special thanks goes to my chairperson, William Davidson for all his
encouragement and support during this project. Whenever I said “no I can’t,” he
said “sure you can.” Thanks to him I am able to finish this project in a timely
manner. I also want to thank the other members of my committee: Timothy Bynum,
Robert Caldwell, Sandra Herman, Maxie Jackson, and Bertram Stoffelmayr. Each
made a significant contribution to this work. Their contributions ranged from giving
me creative ideas and input, providing me with intellectual stimulation, and helping
me with data analysis to providing me with financial support to keep me from
starving to death while I was in graduate school.

I give thanks to my family and friends for always believing in me and giving
me the inspiration to keep going through rough times. To Aunt Tina, my second
mom, thanks for always being in my comer. To Uncle Frank, thanks for introducing
me to Buddhism. Many hours of daimoku are reflected in this work. To my
Buddhist family: Gita, Rosemary, Von, Keora, Maria, and others, thanks for
chanting with me through my various crises. Special thanks to Vemita Marsh, my
sister-friend, and sometimes mentor. Thanks for all your support and guidance,

and your ever present listening ear. I couldn’t have done it without you.

vi

TABLE OF CONTENTS

LIST OF TABLES .................................................................................................. ix
LIST OF FIGURES ................................................................................................ xi
CHAPTER 1
INTRODUCTION ................................................................................................... 1
Purpose of this Study .................................................................................. 6
REVIEW OF THE LITERATURE ............ 8
Description of the Literature Search ........................................................... 11
Factors Related to ANNA Outcome .......................................................... 11
Individual Characteristics and AA Attendance ................................. 14
Outcomes when AA is Aftercare to Professional Treatment ............ 16
Levels of Involvement and Outcome ............................................... 19
Summary of Literature on AA Involvement and Outcome ................ 20
Treating Persons with Psychiatric Disorders .............................................. 22
Psychiatric Symptomatology and Interpersonal Functioning ............ 24
Psychiatric Symptomatology and Involvement in Group Therapy ..... 26
Summary ........................................................................................ 32
Limitations of Prior Studies on AA .............................................................. 33
Goals of the Present Study ........................................................................ 34
STATEMENT OF THE PROBLEM ....................................................................... 41
HYPOTHESES .................................................................................................... 43
CHAPTER 2
METHOD ................................................................................................... . .......... 46
Sample ...................................................................................................... 46
Measures .................................................................................................. 50
Individual Characteristics & Prior Experiences ........................................... 52
Aftercare Treatment .................................................................................. 54
Outcome Measures ................................................................................... 57
Procedures ................................................................................................ 58

vii

CHAPTER 3

RESULTS ............................................................................................................ 60
Data Analysis Strategy .............................................................................. 60
Missing Data Analyses .................................................................... '. ......... 62
Descriptive Statistics on Study Variable ..................................................... 66
Correlations among the Study Variables .................................................... 70
Hypotheses l & ll: Relationship of Individual Characteristics to AAINA

Attendance ..................................................................................... 72
Hypotheses III & IV: Relationship of AAINA and Other Service Use to

Functioning ..................................................................................... 75
Summary of Preliminary Analyses & Hypotheses Tests ............................. 80
The Structural Equation Models ................................................................. 81
Path Analysis Results ................................................................................ 85
Summary of Research Questions and Findings ......................................... 98

CHAPTER 4

DISCUSSION ..................................................................................................... 100
Relationship between Individual Characteristics and Service Use ............ 101
Relationship of Service Use to Outcomes ................................................ 103
Findings and Implications of the Mediated AAINA Treatment Model ........ 106
Limitations of the Current Study ............................................................... 107
Conclusions & Implications ...................................................................... 109
Future Directions ............................................................................. ‘ ........ 111

LIST OF REFERENCES .................................................................................... 112

APPENDIX ......................................................................................................... 123

viii

LIST OF TABLES

Table 1. Demographic Characteristics of Participants ............................... . .......... 48
Table 2. Distribution of DSM-lll-R Diagnoses ..................................................... 49
Table 3. Measures Used in Current Study .......................................................... 51
Table 4. Comparison among Participants with Complete vs. Missing Data .......... 64
Table 5. Descriptives on AAINA and other Service Use during Follow-up ........... 68
Table 6. Descriptives on Outcome Measures ..................................................... 69

Table 7. Correlations between Predictor Variables, Covariates, and Outcome

Variables .............................................................................................. 71
Table 8. Logistic Regression of Symptom/Chronicity and Covariates on AAINA

Ever ..................................................................................................... 74
Table 9. ANCOVA Predicting AAINA Attendance Pattern from Symptom/

Chronicity Groups and Covariates ........................................................ 76
Table 10. MANOVAs Predicting Functioning from AAINA Attendance Pattern ..... 77

Table 11. MANOVAs Predicting Functioning from Type of Services Used at
Time 3 .................................................................................................. 78

Table 12. Path Analysis Results: Integrated Model (A) with # of Days Attending
AAINA at Time 3 .................................................................................. 86

Table 13. Path Analysis Results: Integrated Model (A) with AAINA Attendance
Pattern ................................................................................................. 87

Table 14. Path Analysis Results: Integrated Model (A) with Type of Service Use
Time 3 ................................................................................................... 88

Table 15. Path Analysis Results: AAINA Treatment Model (B) with # of Days
Attending AAINA at Time 3 .................................................................... 91

Table 16. Path Analysis Results: AAINA Treatment Model (A) with AAINA
Attendance Pattern ................................................................................ 92

Table 17. Path Analysis Results: AAINA Treatment Model (A) with Type of Service
Use Time 3 ............................................................................................ 93

Figure 1.
Figure 2.
Figure 3.
Figure 4.
Figure 5.
Figure 6.

APPENDIX

Figure A-1.

Figure A-2.

Figure A-3.

Figure A-4.

Figure A-5.

Figure A-6.

Figure A-7.

LIST OF FIGURES

Model A - The Integrated Treatment Model ....................................... 36
Model B - The AAINA Treatment Model ............................................ 37
Measurement Model A1 .................................................................... 82
Measurement Model B1 .................................................................... 83
Significant Path Predicting Psychological Functioning ....................... 96
Significant Path Predicting Family/Social Functioning ....................... 97

Integrated Model: Number of Days Attending AAINA and Alcohol
Problems ....................................................................................... 123

Integrated Model: Number of Days Attending AAINA and Drug
Problems ............................................................................... p ........ 124

Integrated Model: Number of Days Attending AAINA and
Psychological Problems .................................................................. 125

Integrated Model: Number of Days Attending AAINA and
Family/Social Problems .................................................................. 126

Integrated Model: Number of Days Attending AAINA and
Rehospitalization ............................................................................ 127

Integrated Model: AAINA Attendance Pattern and Alcohol
Problems ....................................................................................... 128

Integrated Model: AAINA Attendance Pattern and Drug
Problems ....................................................................................... 129

xi

Figure A-8.

Figure A-9.

Figure A-10.

Figure A-11.
Figure A-12.

Figure A-13.

Figure A-14.

Figure A-15.

Figure A-16.

Figure A-17.

Figure A-18.

Figure A-19.

Figure A-20.

Figure A-21.

Figure A-22.

Figure A-23.

Integrated Model: AAINA Attendance Pattern and

Psychological Problems .................................................................. 130
Integrated Model: AAINA Attendance Pattern and
Family/Social Problems .................................................................. 131
Integrated Model: AAINA Attendance Pattern and
Rehospitalization ............................................................................ 132
Integrated Model: Type of Service Use and Alcohol Problems ........ 133
Integrated Model: Type of Service Use and Drug Problems ............ 134
Integrated Model: Type of Service Use and Psychological
Problems ........................................................................................ 135
Integrated Model: Type of Service Use and Family/Social
Problems ........................................................................................ 136
Integrated Model: Type of Service Use and Rehospitalization ........ 137
AAINA Treatment Model: Number of Days Attending AAINA and
Psychological Problems ................................................................. 138
AAINA Treatment Model: Number of Days Attending AAINA and
Family/Social Problems .................................................................. 139
AAINA Treatment Model: Number of Days Attending AAINA and
Rehospitalization ............................................................................ 140
AAINA Treatment Model: AAINA Attendance Pattern and
Psychological Problems .................................................................. 141
AAINA Treatment Model: AAINA Attendance Pattern and
F amily/Social Problems .................................................................. 142
AAINA Treatment Model: AAINA Attendance Pattern and
Rehospitalization ............................................................................ 143
AAINA Treatment Model: Type of Service Use and Psychological
Problems ........................................................................................ 144
AAINA Treatment Model: Type of Service Use and Family/Social
Problems ........................................................................................ 145

xii

Figure A-24. AAINA Treatment Model: Type of Service Use and
Rehospitalization ............................................................................ 146

xiii

CHAPTER 1

INTRODUCTION

Self-help or Mutual Aid for serious problems such as substance abuse has
been increasing rapidly over the past decade, both in terms of the number and
types of groups and the people being served by them. Since the 19705, Alcoholics
Anonymous (AA) membership increased by 7% annually, with the United States
and Canada accounting for the largest percent of this increase (Alcoholics
Anonymous, 1989). The number of AA chapters worldwide has increased more
than 200% over the past ten years, with current estimates of more than 80,000
groups in 150 countries (Miller & McCrady, 1993). A number of reasons have
been cited for this tremendous growth, including lack of extended family and
community supports (especially in urban areas), increasing rates of mental health
problems such as depression and addiction, and inaccessibility of professional
services for low income and uninsured persons (Borman, 1992).

For serious problems such as alcohol and drug addiction, there are often not
enough treatment options available. AA is the largest substance abuse treatment
program currently in use, and is the only treatment ever received by a majority of

those who receive any help (Bradley, 1988). When persons with substance abuse

2

problems do receive residential treatment (typically for 28 days or less),
maintenance of treatment effects is often compromised by inadequate aftercare
and/or lack of community supports. This is probably one reason why referral to
Alcoholics and Narcotics Anonymous (AA and NA) for aftercare support has
become a common practice in professional treatment programs. Self-help groups
are often the only source of continued help for persons who have begun the
process of recovery.

Another reason for the growth in AA groups is that many persons have been
helped by AAINA where professional treatment has failed. The “revolving door”
effect in many substance abuse treatment programs is testimony to the fact that
relapse is a common occurrence, even after intensive professional treatment.
Some clients keep coming back until their treatment resources are expended (e.g.,
insurance will no longer pay for treatment or the treatment facility will no longer
accept them - some programs have a limit on how many times a person can
return). When treatment resources are expended, or it becomes clear that
professional treatment is not enough, community resources such as AAINA
become more realistic alternatives. AA provides its members with an indigenous
support system, and a free therapy which is available anytime. The program goal
of AA is clear and focused: to help those who suffer from alcoholism attain
sobriety. Following the twelve steps of AA has helped many achieve this goal
(Emrick, 1987).

Although many alcoholics who attend AA do get better, more empirical

3

research is needed to determine the mechanisms of change through AAINA, and
the types of people most likely to change. To date, there have been very few
empirical investigations of the efficacy of AA. However, the burden of experimental
proof should not be shouldered by AA alone; professional treatment programs have
likewise failed to demonstrate their efficacy through empirical research. In his 1990
review of the research on substance abuse treatment programs, Peele made a
sweeping claim that no treatment program currently utilized in the US. had
sufficiently demonstrated its effectiveness via experimental research. .
Furthermore, he stated that current treatment methods (including Alcoholics
Anonymous, psychoeducation, confrontation, disulfiram, group therapy, and
individual counseling) were based on accepted historical traditions and folk beliefs
rather than proven effectiveness. Other less critical reviewers have suggested that
the provision of appropriate (i.e., to the client’s needs) treatment does substantially
improve outcomes over no treatment at all, but that no particular treatment
approach has been found to be effective for all persons with substance abuse
problems (Institute of Medicine, 1990).

In terms of mental health practice, the increasing use of AAINA as a
treatment agent exacerbates the need for research on its efficacy. Although AA
has been heralded as the most effective treatment currently available, the dearth of
scientific evidence of its efficacy poses a major concern for practitioners (cf.
Bebbington, 1976; Glaser & Ogbome, 1982 for comprehensive review and

discussion). AAINA may be a valuable resource for some people with substance

4

abuse problems, but others are unable to utilize or benefit from this resource. A
number of investigations have been conducted in attempts to identify
characteristics associated with AA involvement and outcome. Unfortunately, these
studies have been plagued by inadequate research questions and methodological
problems (see literature review for more detailed discussion), such that the efficacy
of AA for different types of alcoholics remains an elusive subject.

Some authors have argued that persons with the characteristics of the
respondents in this study - dually diagnosed African American males - are not
likely to go to AAINA, and if they do, they receive little benefit from their
involvement (cf. Bartalos, 1992; King, 1983; Swift & Beverly, 1985). However, the
issue of AAINA involvement may be complicated by other factors, in addition to
person characteristics (cf. Humphreys & Woods, 1993 for discussion of contextual
factors such as “racial fit”).

Studies on benefit from substance abuse treatment in dually diagnosed
samples are inconsistent in their findings, but generally indicate that the dually
diagnosed are a difficult to treat population (cf. Mueser, Bellack, & Blanchard for
discussion). Clinical literature indicates that persons with severe mental illness are
often perceived as difficult, non—compliant, and resistant to both psychiatric and
substance abuse treatment (Hellerstein & Meehan, 1987). The challenges posed
by clients with dual diagnosis sometimes results in their exclusion from both
systems of care.

Over the past decade, the number of studies on dual diagnosis has

5

increased dramatically (Minkoff, 1989). However, the impact of dual diagnosis on
treatment engagement and outcome is not well understood. Some authors (e.g.,
Bartalos, 1992) have postulated that dual diagnosis status would pose a particular
challenge to AA involvement. One issue that has been noted is that ANNA
members often actively promote an anti-drug policy, discouraging the use of
psychotropic medication. This policy creates a conflict for many dually diagnosed
clients. AA members are encouraged to strive for a drug-free state, use the
group, and follow the 12-steps as the primary means of dealing with personal
problems, a formidable challenge for the person with a serious mental illness.
While the official organization literature indicates that group members with
psychiatric problems should ngt be discouraged from psychotropic drug use
(Alcoholics Anonymous, 1975), the reaction of other AA members towards those
who use psychotropic medications varies and remains a very controversial topic
(cf. Buxton, Smith, & Seymour, 1987; Zweben, 1987 for more comprehensive
discussion of this issue). These, and other problems make it particularly
challenging for persons with dual diagnosis to become fully engaged with AAINA
and to receive its potential benefits.

Research on patient-treatment fit with AAINA, and data on outcome by
person characteristics would greatly enhance our ability to determine “which kinds
of individuals, under what circumstances, are likely to respond to AA, by achieving
what kinds of goals?’ (Institute of Medicine, 1990, p. 143). As more

comprehensive data is amassed to answer these questions, clinicians and

6

counselors will be better able to make appropriate referrals to these organizations,
thereby improving the chances of successful outcomes. In terms of research
strategies, randomized controlled clinical trials have been suggested as the most
optimal approach to assessing patient-treatment fit and treatment outcomes
(Institute of Medicine, 1990). Other methods include correlational studies using
predictor strategies and/or causal modelling (Miller & Hester, 1986). These
methods have been used to deal with issues of efficacy in the field of
psychotherapy (cf. Kazdin & Wilson, 1978), and are similarly appropriate to
research on substance abuse treatment outcome. The current study is an attempt
to address the need for empirical data on AAINA efficacy using correlational
methods and causal modelling.
Purpose of this Study

(1) The first purpose of this study is to provide descriptive information on
levels of AAINA participation in a dually diagnosed sample, and to determine if in
fact persons with dual diagnosis are less likely to participate in AAINA. This goal
will be accomplished by comparing rates and patterns of involvement found in this
sample with that reported in the literature on singly diagnosed substance abusing
clients.

(2) A second goal is to identify factors related to AAINA participation in a
dually diagnosed population. It is suggested that chronic and/or severe psychiatric
symptomatology will act as a deterrent to sustained participation in self-help.

Persons with high levels of psychiatric distress are expected to become only

7

marginally involved with AAINA, and to drop out quicker than their psychologically
healthier counterparts.

(3) Third, this study seeks to determine whether or not patterns of
involvement with AAINA are related to alcohol/drug use and other indices of
community functioning. The literature reviewed in the following section leads the
researcher to predict that those who do not become consistently involved with
AAINA soon after professional treatment will to not receive its benefits.

(4) Finally, this study will incorporate program theory, and compare the
mechanisms of change suggested by two different treatment models: an integrated
treatment model and an AAINA model. The goal is to determine which model is

more likely to be operating in this sample of persons with dual diagnosis.

REVIEW OF THE LITERATURE

According to Chen (1990), in order to understand why a program succeeds
or fails, the researcher must have an appropriate understanding of the underiying
program theory of the treatment being studied. Understanding program theory
offers numerous benefits including: allowing the program theory to be tested in a
way that reveals whether failure results from implementation failure or failure of the
program theory itself; and to test whether the program is being implemented under
conditions in which it is appropriate. While most prior research on AA has not
incorporated program theory into its evaluation, an attempt will be made to do so in
the present evaluation.

The program theory of AA is clearly stated in primary literature produced by
members of the fellowship, and in secondary sources. The main goal of AA is to
help those who suffer from alcoholism attain sobriety. (It should be noted that
while most evaluations of AA have focused on abstinence as the primary outcome
variable, AA makes a clear distinction between “dryness” and sobriety or serenity,
Kurtz, 1993). Those who attain sobriety are able to admit that they are powerless
in relation to alcohol, they accept help from a higher power and from other

members of the fellowship, they address constitutional and psychological

9

underpinnings of their desire to drink (e.g., self-centeredness, arrogance, etc.) and
restore their sanity, they make amends to others that they have harmed in the
process of following an alcoholic lifestyle, they make specific changes in their
physical and social environments, they stop drinking, and they carry the AA
message to others who are suffering from alcoholism. The suggested method of
attaining these sobriety-related goals is by attending AA meetings and following the
twelve steps. According to AA literature, those who thoroughly follow the AA
program w_ill succeed (Alcoholics Anonymous, 1976). Those who do not succeed
have not fully implemented the AA program.

In terms of the group structure, AA is far from a monolithic organization.
There is great diversity in types of AA groups, types of persons who participate in
AA, and in members’ perceptions of what AA does (Beckman, 1980). There is no
official “leader“ of AA, nor is there any significant input from professionals in terms
of how groups should be run or who should participate. The twelve traditions of AA
specify that the organization is to remain decentralized, so as to maintain its
autonomy from outside forces and allow individual AA groups to remain
autonomous in their structure and function, progressing toward the goal of sobriety
in a way that is most appropriate for them. AA is open to anyone who is interested
in joining; “the only requirement for membership is a desire to stop drinking"
(Alcoholics Anonymous, 1976).

While the explicit goal of AA is to help members attain sobriety, the implicit

goals of the organization address other domains of functioning as well. DiClemente

10
(1992) identified five domains of functioning which are targeted by AA:

(1) symptomatic/situational problems — including changes in alcohol consumption
and life consequences; (2) maladaptive cognitions - for example, pride, negative
thinking; (3) interpersonal conflicts - by making amends to others; (4)
family/systems conflicts - value of common welfare above individual’s; and (5)
intrapersonal conflicts - addressing character defects that coincide with alcoholism.
A number of studies have been published on AA (relatively few on NA);
however, a vast majority are reviews and anecdotal reports as opposed to
empirical evaluations. This literature review includes both empirical studies, clinical
reports, and other reviews, but will emphasize empirical literature on factors related
to AAINA participation, factors related to AAINA outcome, social support, and the
efficacy of group therapy for persons with severe mental illness. There have been
few clinical reports and hardly any empirical studies on NA. Thus, most of the
literature cited in this review refers to Alcoholics Anonymous. While the specific
content of AA and NA meetings may differ, AA is the historical predecessor of NA
and the groups still share many aspects in common. For instance, the groups
adhere to the same basic principles of recovery as written in the Big Book
(Alcoholics Anonymous, 1976). In fact, it is known that many AA members also
have other drug problems and that persons whose primary problem is with illicit
dmg use often attend both groups (Emener & Dickman, 1992). Until there is
sufficient empirical evidence to suggest that these groups should be considered

separately, it must be assumed that they may be simultaneously considered under

11

one broad rubric which is self-help for addiction, based on the AA model.
Description of the Literalrre Search

The literature reviewed in this section was identified through: computerized
searches on Psyclnfo (1976-1995) and MAGIC (Michigan State University
computerized library catalog). Key words used in searching these data bases
were: Alcoholics Anonymous, Narcotics Anonymous, substance abuse, drug
abuse, alcohol abuse, schizophrenia, mental illness, dual diagnosis, treatment, and
group therapy. These key words were used in various combinations and yielded
more than 200 articles and 20 books on the selected topics. References obtained
from the reference sections of the selected primary sources were also retrieved
and used in this review.

Factors Relatgl to AA/N_A Outcome

Most studies on AAINA involvement and outcome have examined personal
characteristics of participants and related those to outcome. Upon review of this
literature, one finds a mass of chaos and confusion regarding personal correlates
of treatment success. Gibbs and Flanagan (1977) conducted a meta-analytic
review of the prognostic indicators of treatment outcome and found that across 45
studies, 208 different predictors had been examined. The studies reviewed had
explored all kinds of variables from the usual (e.g., sociodemographic
characteristics and social stability, prior treatment history, baseline functioning
level, prior abstinence/sobriety, length of drinking career, MMPI and other

personality measures, etc.) to the unusual (e.g., visual-motor perception, digit

12

span, degree of fantasy ideation, frequency of contact with mother, and number of
sexual partners before marriage). Across the 45 studies, the authors were unable
to identify any stable predictors using rather modest criteria (i.e., the predictor being
tested in six or more studies, and found to predict success in at least half of these
studies). While there were a few predictors that met these criteria (e.g., good
employment status, marital stability, fewer arrests, history of AA attendance, and
higher social class), for most other predictors, there were equal numbers of studies
that found contradictory results or nonsignificant results for the same predictors.

The studies that examined outcome from AA and other types of group
treatment found socioeconomic status, age, marital status, prior involvement with
AA, history of abstinence, and psychiatric status to be related to treatment
success. While these variables were significant in some studies, in most cases,
there were other studies that found nonsignificant or contradictory results with
these same predictors. In commenting on the inconsistency of findings across
studies, Gibbs and Flanagan (1977) concluded that the task of identifying stable
(and universal?) predictors of alcoholism treatment outcome was almost .
impossible. Some of the difficulties with such an endeavor are: each study
involves different treatment agents (i.e., therapists with different orientations,
personalities, and skill levels), administering different types of treatment, and using
different outcome criteria over different follow-up time periods, with different types
of clients.

Fortunately, in the decades since Gibbs and Flanagan’s review, the range of

13

variables studied seems to have decreased somewhat and researchers are
beginning to examine some of the same predictors across studies, thereby
corroborating prior research findings. Also, researchers are tending to focus on
one type of treatment and are defining the characteristics of the treatment
population more consistently. However, in a more recent review of the literature on
AA, Emrick (1987) came to conclusions similar to those of Gibbs and Flanagan.
The findings in the more recent literature were inconsistent and contradictory.

One suggestion from these comprehensive reviews is that simplistic
questions such as, “does AA work?” are inappropriate. The researcher must
consider contextual factors of the person’s environment, their personal
characteristics and pre-existing functioning levels, and the context in which
treatment in offered (e.g., is the treatment being implemented correctly, at what
dosage, and are there any outside influences such as other concurrent treatment
which might blur the effects of AAINA involvement).

Literature on psychotherapy outcome has shown that multiple criteria are
necessary to evaluate treatment outcome and the relative value of different types
of treatment (cf. Kazdin & Wilson, 1978). Some of the most important factors in
clinical research are: the specificity of change (i.e., did the therapy ameliorate the
problem for which therapy was sought), the clinical significance or importance of
the change observed, the breadth of change (i.e., has the client experienced
improvement in other aspects of his/her everyday functioning), and the durability of

any change associated with the treatment.

14

Kazdin and Wilson (1978) have criticized prior research for emphasizing
normative group changes following treatment. They argued that group differences
in mean performance on an outcome criterion is an extremely limited method for
evaluating the effects of treatment. Empirical studies need to explicate factors
related to positive outcome in terms of the proportion and types of treated
individuals who change, as well as assessing the quality and duration of that
change. These points are particularly relevant to persons with dual diagnoses.
Dually diagnosed persons often present with a plethora of problems in various life
functioning domains. Thus, in order for treatment to be effective for persons with
dual diagnoses, it must improve their functioning in a number of affected life
domains. Prior studies of AA have tended to focus on one specific outcome:
abstinence, and have not fully examined other possible outcomes related to AA
attendance; nor have they examined the outcomes of AA participation for persons
with multiple disorders.

The remainder of this section will review AAINA outcome studies only, with a
particular emphasis on AA as aftercare to inpatient substance abuse treatment.
Individual Characteristics and AA Attendance

Persons who attend AA have been found to be different in many respects
from those who do not attend. Some studies have identified differences in
demographics, personality characteristics, problem severity, and treatment history
for attenders vs. nonattenders (Polich, Armor, & Braiker, 1980). For instance,

some studies have found that AA attenders are more likely to be older males who

15

have higher socioeconomic status, self-identify as alcoholic, are affiliated with a
church or religious group, are psychologically healthier, have adequate social
networks which they utilize, and have had prior substance abuse treatment (Elal-
Lawrence, Slade, & Dewey, 1986; O’Leary et al., 1980; Polich, Armor, 8. Braiker,
1980; Ribisl, 1995; Trice & Roman, 1970) .

While factors related to AA attendance have been examined in a number of
studies, the findings are inconsistent and more recent data generally refutes the
prior findings of significant differences (Emrick, 1989). For instance, Thurstin,
Alfano, and Nerviano (1987) failed to find any differences in demographics or
psychological characteristics of attenders vs. nonattenders. Emrick, in his 1989
literature review, indicated that sociodemographic differences were not consistently
observed (except for age - most studies find that AA attenders are likely to be
older) and that drinking-related issues (e.g., loss of control, drinking patterns, level
of physiological addiction) also do not consistently predict AA involvement. Emrick,
Tonigan, Mongtomery, and Little (1993) reviewed all the empirical literature on AA
and concluded that “systematic distinctions between AA affiliates and non affiliates
can be identified; although the literature is not currently developed enough to
provide us with a composite profile of the most likely AA affiliates” (p.53).

While motivation to engage in particular types of treatment has not been well
studied, motivation to attain sobriety has long been suggested as an integral factor
related to treatment engagement and outcome. Clinicians who treat persons with

substance abuse problems frequently lament the problem of denial in alcoholics

16

and drug addicts and often use confrontational techniques in an attempt to help
clients face the reality of their problems (i.e., “helping the alcoholic to hit bottom,”
Miller, 1985). Motivation is a key component of the stages of change model
proposed by Prochaska, DiClemente, and colleagues. These authors’ research on
addictive behaviors, including alcoholism, suggests that the client’s pretreatment
stage of change (i.e., readiness or motivation to change) is one of the most
important determinant of treatment success (Prochaska, DiClemente, & Norcross,
1992). AA program literature also highlights the importance of the alcoholic’s level
of motivation to stop drinking. The AA Big Book asserts that those who thoroughly
follow the AA program will succeed. Those who do not recover are suspected of
not giving themselves completely to the program, and are even said to be
“constitutionally incapable of being honest with themselves” (Alcoholics
Anonymous, 1976).
Ditcomes when AA is Aftercare to Professional Treatment

AA is commonly used as an adjunct to or follow-up of professional
substance abuse treatment. Treatment follow-up studies have probably
contributed the most to our fund of knowledge about the efficacy of Alcoholics
Anonymous. These studies have revealed a number of important findings.
regarding use of AA as treatment aftercare:

(1) persons who receive inpatient treatment for substance abuse problems
are likely to be referred to AA, and the majority do attend (estimates as high as

80%; Edwards et al., 1971; Knouse & Schneider, 1987; Pettinati, Sugerman,

17
DiDonato, & Maurer, 1982).

(2) there is a high drop out rate from AA, even among those who were
referred by treatment staff (estimated 25% or less will attend regulariy; Belasco,
1971; Emrick, 1987; Tomsovic, 1970). Data from the General Service Office of
Alcoholics Anonymous indicates that fewer than one half of all persons who come
to an AA meeting will remain for even three months (AA, 1989).

(3) The most positive benefits from AA are usually obtained when it follows
formal treatment (Emrick, 1987; Knouse 8 Schneider, 1987) or is used as an
adjunct to professional treatment (Emrick, Tonigan, Montgomery, 8. Little, 1992). In
addition, Knouse and Schneider (1987) found that of patients who attended AA
after inpatient treatment, those who joined AA soon after hospital discharge (i.e.,
within the first three months) were most likely to maintain sobriety. AA involvement
did not seem to make a difference if initiated after three months in the community.

While many studies have found that persons who attend AA have better
outcomes (especially abstinence rates) than those who do not (cf. EIaI-Lawrence,
Slade, & Dewey, 1986; Gregson & Taylor, 1977; Habennan, 1966; Hoffman,
Harrison, & Belille, 1983; McBride, 1991), some controlled studies have failed to
find any significant differences based on AA attendance (cf. Ditman, Crawford,
Forgy, 8 Maskowitz, 1967; Edwards et al., 1977; Walsh et al., 1991). The cited
rates of success vary across studies, depending on a number of factors including
client characteristics, patterns of AA involvement, access to other supportsystems,

and length of follow-up. Studies that follow-up clients for six months or more post-

18

hospitalization usually find that approximately 70% of respondents resumed
drinking (Bateman 8 Petersen, 1971). The factors found to be related to continued
abstinence include: age, employment status, length of prior abstinence, and prior
history of AA attendance. Factors related to relapse include: negative physical or
emotional states, social pressure, unemployment, and history of AA attendance
(ElaI-Lawrence, Slade, 8 Dewey, 1986; Ogbome 8 Bomet, 1982; Rather 8
Shennan,1989)

While most substance abuse treatment outcome studies (especially those
studies involving AA) have focused exclusively on drinking outcomes, some
authors have stressed the need to examine psychological, social, and behavioral
adjustment as well (e.g., Belasco, 1971; Bromet, Moos, Bliss, 8 Wuthmann, 1977).
Outcomes in these domains are generally less consistent than drinking outcomes,
nevertheless, they warrant mention. In a study of 429 treated alcoholics, Bromet,
Moos, Bliss, and Wuthmann (1977) found that those who attended AA after
discharge were significantly less likely to be rehospitalized. Emrick, Tonigan,
Montgomery, and Little’s (1992) meta-analysis of AA studies found positive
relationships between AA involvement and employment (I; = .12) and social/family
adjustment (5 = .13). More robust and reliable correlations were found between AA
involvement and improvement in psychological adjustment (1 = .25).

Some researchers have suggested that the relationship between AA
attendance and psychosocial functioning is moderated by attainment of abstinence.

That is, AA involvement leads to abstinence which lead to better functioning in

19

other life domains (see Laundergan, 1992 for review). Overall, these findings
suggest that the breadth of outcomes attained from AA attendance extend well
beyond simple indices of drinking or abstinence, and warrants further investigation.
Levels of Involvement and Outcome

Patterns of AA attendance have been found to predict drinking outcomes
fairly consistently. Persons who attend AA regularly report higher rates of
abstinence, fewer days drinking, and fewer episodes of intoxication than those who
do not attend regularly (Emrick, 1987; Knouse 8 Schneider, 1987; McBride, 1991;
Thurstin, Alfano, 8 Nerviano, 1987; Trice 8 Roman, 1970). McBride (1991) found
that the length of continuous AA attendance was highly correlated with months of
abstinence (r = .71) and accounted for 50% of the variance in outcome.
Conversely, infrequent or irregular attendance at AA post-treatment has been
associated with poor prognosis and outcome (McLatchie 8 Lomp, 1988).

Emrick, Tonigan, Montgomery, and Little’s (1992) meta-analysis of
quantitative studies of AA (g=107) found that frequency of AA attendance and
drinking outcomes were correlated r = .19; however, the authors noted that this
correlation might be an underestimation. (Due to inconsistency across studies, the
correlation contained a high degree of error variance.) There were too few data on
AA participation and outcome in other life domains to enter into the meta-analysis.

In an earlier review paper, Emrick (1987) suggested that of those who
become long-term (not defined), active members of AA, 40-50% will enjoy years of

sobriety, and 60-68% will improve somewhat. Similarly, Thurstin, Alfano, and

20

Nerviano (1987) found that persons who attended AA consistently throughout an
18-month follow-up period had the best outcomes of any group. Hoffman,
Harrison, and Belille (1983) found that 73% of respondents who attended AA
weekly during a 6-month post-hospital follow-up remained chemical-free, while only
33% of nonattenders abstained. One can generally conclude from the literature on
AA that people who “work the program,” and attend meetings consistently over a
longer period of time are more likely to have better outcomes, at least in terms of
their drinking (Emrick, Tonigan, Montgomery, 8 Little, 1992).
flmmaiv of Literature on AA Involvement and Outcome

The literature on AA reveals considerable evidence that AA attendance
helps at least some alcoholics (e.g., those who become regular and active
members) to attain abstinence. However, this literature has generally neglected
outcomes in other life domains. Although abstinence has been related to better
social, physical, and psychological outcomes in some reports (of. Alford, 1980; and
Akerlind, Hamquist, Elton, 8 Bjurulf, 1990 for review), neither abstinence nor level
of alcohol consumption alone are sufficient predictors of functioning in other life
domains. Researchers need to examine other areas of functioning directly instead
of using alcohol involvement as a proxy for global functioning. Additionally, there is
a need for empirical studies with dually diagnosed persons.

Many persons who seek treatment have dual problems, and it is likely that
their response to treatment will be quite different from persons with substance

abuse problems only. The literature on persons with dual diagnoses of mental

21

illness and substance abuse suggests that they experience problems across a
number of life domains, and differ in many respects from other persons with
substance abuse problems. Dually diagnosed persons are more likely to be
homeless and/or vagrant and to have been jailed (Ridgely, Goldman, 8 Talbott,
1986). Several authors have noted the high rates of criminal involvement, as
well as violent acting out and self-injurious behavior among persons with dual
diagnosis in comparison to those with singly diagnosed mental illness or
substance abuse problems (Kay, Kalathara, 8 Meinzer, 1989; Minkoff 8 Drake,
1991; Mueser, Bellack, 8 Blanchard, 1992; O’Farrell, Connors, 8 Upper, 1983).

In terms of mental health treatment, substance abuse has long been
recognized as a complicating factor in the treatment of persons with severe
mental illness (cf. Mueser, Bellack, 8 Blanchard, 1992; Ridgely, Goldman, 8
Talbott, 1986). Persons with dual problems of substance abuse and mental
illness are often more difficult to diagnose and also pose complications to
regular treatment of mental illness (e.g., noncompliance with medication
regimens). Substance abuse in persons with diagnosed schizophrenia has been
related to more severe course of illness, more severe symptomatology, more
pervasive deterioration in functioning, and increased likelihood of psychiatric
relapse (Mueser, Bellack, 8 Blanchard, 1992; Ridgely, Goldman, 8 Talbott,
1986)

Osher and Kofoed (1989) indicated that patients with dual diagnoses

were also more difficult to engage in substance abuse treatment, particularly

22

abstinence-oriented treatment. The authors suggested that dually diagnosed
patients were less able to transcend denial, a core requirement of many
substance abuse treatment programs, because of their “impaired ability to
process information due to thought disorders, depressive cognitions, or organic
brain syndromes“ (Osher 8 Kofoed, 1989, p. 1028). These authors surmised
that prolonged abstinence, the primary goal of AA, would be particularly difficult
for persons with dual diagnosis to attain.
Treating Persons with Psychiatric Disomgs

McLellan and his associates have conducted numerous studies which
indicate that high psychiatric severity (assessed by number, intensity, and duration
of symptoms) is negatively related to substance abuse treatment outcomes.
McLellan et al. (1983) found that in a VA sample of male alcoholics (n=460) and
drug addicts (n=282), scores on a psychiatric severity index were strongly and
consistently related to treatment outcomes (as measured by seven composite
scores from the Addiction Severity Index). This study revealed that patients with
low psychiatric severity experienced better treatment outcomes than those with
high or mid-range psychiatric severity, no matter what type of treatment they
received. In addition, for the high psychiatric severity group, having a greater
number of prior treatments was also associated with poorer outcomes. Those with
high psychiatric severity showed virtually no improvement in any of the treatment
conditions. Overall, psychiatric severity accounted for 48-54% of the variance in

treatment outcomes for the high psychiatric symptom group.

23

Using prior psychotherapy as a proxy for psychological health, Haberrnan
(1966) found that history of involvement in psychotherapy was negatively related to
substance abuse treatment outcome (i.e., abstinence). In Gibbs and Flanagan’s
1977 meta-analytic review, psychiatric history and symptomatology were also found
to be negative predictors of treatment outcome in a number studies. Using proxies
for psychological well-being such as: having a primary psychosis diagnosis, history
of seeking psychiatric treatment, rating oneself as “mentally ill,” scores on
personality measures such as the MMPI, and history of suicide attempts, several
studies found that better psychological health was related to better outcomes from
AA and other group therapies for alcohol abuse.

In terms of treatment engagement, some studies have found experience of
psychiatric symptoms to be negatively related to engagement in substance abuse
treatment (Liss, 1979; Panepinto, Higgins, Keane-Dawes, 8 Smith, 1970).

Persons with schizophrenia were more likely become withdrawn or drop out of
treatments which stressed collective responsibility or high levels of interpersonal
engagement with staff and other clients. If the treatment involved only minimal
engagement (e.g., ten-minute sessions with an internist that focused on symptom
update, and involved only minimal exploration of thoughts and feelings, clients with
schizophrenia were more likely to continue coming (Panepinto et al., 1970).
Persons with less severe mental illness were more likely to drop out of the. latter
type of treatment. These findings suggests some interesting relationships between

client characteristics, type of treatment, and likelihood of treatment success.

24

Eychiatric Svmptomatologymd Interpersonal Functioning

Previous studies have demonstrated that persons with chronic mental
illness or dual diagnoses are likely to suffer deficits in many areas of life
functioning, including family/social relationships (Westerrneyer 8 Neider, 1988).
Persons who report experiencing severe symptomatology are more likely to rate
their family and social relationships as poor/nonsupportive, and/or conflictual.
Participants in the current study were found to have inadequate social networks
which were unusually small and unsupportive (Ribisl, 1995). They reported that
their relationships were not stable, nor did they serve the important functions of
social support or mental health promotion. These types of relationships may
actually contribute to experienced distress instead of acting as a buffer against it.
Prior research has indicated that experience of interpersonal conflict, especially in
the family context, is often a precursor to relapse and is generally related to poorer
long-term outcomes for persons with serious mental illness (Swindle, Cronkite, 8
Moos, 1989).

While the amount of social support received is affected by contextual factors
and the availability of supportive others, it is also affected by individual skills in
accessing and maintaining supportive relationships (Heller 8 Swindle, 1983) .
Naturalistic studies suggest that persons with better mental health (e.g., social
competence and intrapersonal resources) are more likely to establish and maintain
supportive social networks (Heller, 1979). Persons with good mental health are

more likely to be sociable, assertive, comfortable with intimacy, free from

25

debilitating social anxiety, have good conversational skills, are able to talk about
their behavior and feelings, have adequate social problem-solving skills, and are
able to take the perspectives of other people (Heller 8 Swindle, 1983). All of these
qualities are essential to the initiation and maintenance of social relationships, and
social support. Persons with poor mental health are often less socially competent,
less adept at interpreting social cues or altering their interpersonal behaviors, and
more prone to inaccurate appraisals of social situations.

In addition to person characteristics, ecological and community factors also
affect the availability and quality of supportive relationships (cf. Korte, 1978 for
review). For instance, in highly urbanized areas, isolation and anonymity are
common experiences. In an environment where suspicion and hostility are more
common than helpfulness and friendliness towards others, developing support
networks outside of the family can become a formidable task. Neighborhood
attractiveness, density, and racial composition are also important factors that
influence the quality of social relationships and social support.

It has been suggested that persons with serious and/or chronic mental
illness are less efficient at accessing and maintaining social relationships that
provide appropriate levels of support to aid in coping. While persons with mental
illness may have access to as much social support as others, they are less likely to
give support to others or maintain reciprocal social relationships (Heller 8 Swindle,
1983). In addition, respondents in the current study were drawn from an urban

metropolitan area in the Midwest, where environmental barriers to social support

26

were likely to exist. Therefore, it was predicted that the participants in this study
would be less likely to have successful experiences in the context of a mutual self-
help group like AA or NA, where provision and receipt of social support are
necessary.
Emmetric Svmptomatologya_nd Involvement in GrogLTherapy
Clinical literature generally suggests that persons with serious mental illness
are often not appropriate for group work (cf. Kanas, 1982; Yalom, 1985). This
finding is quite relevant to investigating the appropriateness of AA for a group of
persons with problems of mental illness as well and substance abuse.
Stnrcturally and philosophically, there are some important differences

between professional psychotherapy and self-help or mutual aid groups (e.g., no
formal group “leader” or therapist, self-selection to group). Self-help groups have
traditionally been defined as being composed of members who share a common
problem, create a network of support for one another through regular social and
emotional interactions, are self-goveming, and involve egalitarian relationships
based on principles such as “helper therapy” and identification of members as
“prosumers” (i.e., producer as well as consumers of services and aid; Hedrick,
lsenberg, 8 Martini, 1992; Leiberrnan 8 Borman, 1979). Another major difference
is that self-help programs are free and open to all who wish to attend.

Nevertheless, there are many elements which AA shares in common with
other therapy groups (Knight, Wollert, Levy, Frame, 8 Padgett, 1980; Yalom,

1985). Irvin Yalom’s work has been most instrumental in identifying the curative

27

elements involved in group therapy process - that is, the mechanisms of change
that can be found in anytherapy group. Of the eleven factors identified by Yalom,
at least nine can be found in AA self-help groups: instillation of hope, universality,
imparting of information, altruism, development of socializing techniques, imitative
behavior, interpersonal learning, group cohesiveness, and catharsis. (The
remaining two factors are somewhat tangential to the emphasis of AA - corrective
recapitulation of primary family group and existential experiences.) While there is
no group leader present to ensure that these elements are in effect, there is a
standard protocol to most AA groups which includes these factors. For instance,
the use of testimonials of the effectiveness of AA serves to instill hope in other
members, especially new or struggling members. The common use of the
introduction “hi, my name is _ and I’m an alcoholic” serves to reinforce a sense
of commonality, group cohesion, and validation of the experiences of other
members. Imparting of advice and altruistic behaviors (e.g., sponsorship, 12-
stepping - saving other alcoholics by telling them about AA) are also common
experiences in AA.

One of the most important, if not the critical curative ingredient of AA is
involvement in mutual social support and fellowship with other members (Knight,
Wollert, Levy, Frame, 8 Padgett, 1980; Maxwell, 1984). As noted previously, AA
also serves as a primary support group for many members, fulfilling many of their
material, social, and psychological needs (Emrick, 1989). Those who have studied

AA groups, particularly via participatory observation, find that AA members derive

28

benefit primarily by involving themselves in fellowship with other members (e.g.,
through sponsorship, 12-stepping, etc.). Maxwell (1984), a psychologist who
conducted studies on AA process, indicated that recovery through AA “means
becoming immersed in the AA social environment. It means participating in open,
honest sharing and interaction with fellow members.” Furthermore, “the degree
and quality of involvement with other members will be a major factor in how much
will be gained from AA” (p.155). 1

Clinical literature suggests that group work is not the treatment of choice for
persons with serious and/or chronic mental illness, especially those with
schizophrenia (cf. Otteson, 1979 for review). Clinicians have noted that “among
the general characteristics of the chronically mentally ill is their difficulty in
becoming meaningfully engaged in a therapeutic endeavor” (Stone, 1991, p.13).
Often persons with severe mental illness are not able to withstand the anxieties
related to self-disclosure or confrontation, which are common elements of group
experience. Several theories have been proposed to explain this resistance to
group engagement. For example, it has been proposed that persons with mental
illness are more concerned about survival than self-realization; that the level of
developmental arrest and intrapersonal conflict inhibits the establishment of trusting
relationships and also makes it difficult for persons with mental illness to regulate
personal boundaries without feeling anxious; and finally, that psychiatric patients
experience a realistic caution against revealing themselves, since prior experience

has taught them that this is dangerous and may lead to punishment (Chacko,

29
Adams, 8 Gomez, 1985; Stone, 1991). Another hypothesis is that use of

psychotropic medications subdues both abnormal and normal behaviors, including
social responsiveness.

In terms of group process, the work of Yalom and his colleagues indicates
that when people with psychosis and other severe mental illnesses are placed in
groups where most members are not suffering from mental illness, the peculiarity of
the person with mental illness is often obvious to other group members. Oddities in
dress, mannerisms, and verbalizations (e.g., irrelevant or bizarre verbal content)
cause them to stick out like sore thumbs. Thus, they often find it difficult to fit in the
group and are quickly ostracized as deviants. (In terms of AA, the person .with
mental illness might also be taking psychotropic medication, which, if discovered
would also make them different from the other group members and also inhibit their
adherence to the group norm of abstinence from all chemicals.) Additionally,
persons with severe mental illness, especially those on psychotropic medications,
often appear to be “sealed over" (Yalom, 1985, p. 237). Their affect may be flat,
inappropriate or bizarre and they may be generally withdrawn from social
interactions. Clinicians who work specifically with psychotic clients indicate that
they often have deficits in even the most mundane social and conversational skills
(Wilson, Diamond, 8 Factor, 1990). The inability of many persons with severe
mental illness to involve themselves in an interactional group without threatening
the group norms or becoming overly anxious themselves, seriously interferes with

their ability to derive benefit from a group therapy experience. According to Yalom,

30

the psychotic person tends to interfere with the normal dynamic process of the
group and consequently, “eventually the group will extrude the deviant” (1985, p.
238). Thus, one would expect these persons to be forced out by the group or to
opt out of the group themselves.

Specific to self-help group involvement, Bartalos (1992) suggests that “in
order to be an effective member of a self-help group, one has to have adequate
perception and judgement, ability to interact meaningfully with others and a
reasonable chance to benefit from such interactions" (p. 72). Given this
proscription for successful participation, persons with severe mental illness would
be among those left out (by self-selection or a group screening process) due to
inability to participate fully.

Although much clinical evidence suggests that group therapy with seriously
mentally ill and/or dually diagnosed persons is extremely difficult, some clinicians
and researchers have been successful in developing models of group work that are
effective with these patients (e.g., Fairweather et al., 1969; Hellerstein 8 Meehan,
1987; Otteson, 1979; Wilson, Diamond, 8 Factor, 1990; Yalom, 1985). These
successful efforts generally have been developed and utilized with homogeneous
groups (e.g., hospitalized psychiatric patients) and involve special techniques
designed to address the particular needs of persons with serious mental illness.

The Fairweather Lodge studies were perhaps the first attempt by
researchers to determine whether a group of chronically mentally ill patients could

function in a community-based group. Fairweather and his colleagues found that a

31

group of chronically mentally ill patients (more than half also had substance abuse
problems), when given the chance, were able to live semi-autonomously and
ultimately take complete responsibility for their community, and their livelihood. In
terms of outcome, across a 40-month hospital follow-up period, the Lodge
members were more successful in maintaining employment and remaining in the
community (i.e., spending more days in the community than in the hospital) than a
comparison group that received traditional aftercare services. Hellerstein and
Meehan (1987) found similar results with a group of dually diagnosed patients.
Using a group model designed especially to address dual diagnosis issues, these
authors reported success in reducing the number of days participants spent in the
hospfial

Although the Lodge project was successful overall, there were differential
outcomes based on chronicity and severity of illness. Persons with more chronic
mental illness (i.e., having been in the hospital longer) spent fewer days in the
community, one of the most important measures of successful outcome, and were
less likely to be employed than those with more acute illness. Behaviorally, the less
chronic and nonpsychotic persons were more adept at handling the social
demands of Lodge living and work environments. They were more likely to chosen
by their peer to be leaders in the Lodge community (45% of nonpsychotic Lodge
members were selected as leaders vs. only 15% of psychotic members). Given
these differential outcomes, the authors concluded that "chronicity places a damper

on friendship relations, verbal contact, activity level, and freedom from emotional

32

disturbance," all of which are requirements of communal living or work settings
(Fairweather et al., 1969).

Other authors have identified successful models for group work with
hospitalized psychiatric patients (cf. Wilson, Diamond, 8 Factor, 1990), in which the
therapists use special techniques to help clients manage anxiety and learn
appropriate social skills via step-by-step skills training. In terms of AA involvement,
Minkoff (1989) found that with special preparation in advance, dually diagnosed
patients were able to participate successfully and benefit from AAINA. Staff at this
specialized program prepared clients for AAINA attendance by: providing them
with individualized education on appropriate behavior in the group, having them
participate in groups with other dually diagnosed persons, linking them up with
another person in AAINA, and selecting meetings that would be most suitable for
dually diagnosed persons. These methods eased the transition of dually diagnosed
persons into the groups, allowing them to feel more comfortable and providing them
with behavioral coaching to minimize inappropriate and/or bizarre behaviors which
might frighten other group members. Unfortunately, no specific data on patient
outcomes were reported in this study.

Summary

This review suggests that although persons with severe and chronic mental
illness are generally not expected to respond well to group therapy, under special
circumstances and with advanced training in social skills, people with chronic

mental illness or dual diagnosis may be able to benefit from mutual aid groups such

33

as AAINA. More empirical research is needed to determine the extent and

conditions under which AAINA can be utilized as a therapeutic resource by persons

with dual diagnosis. .
Limitations of Prior Studies on AA

The majority of reports on the efficacy of AA have been anecdotal and/or
inconclusive due to lack of scientific rigor. Miller and Hester (1980) identified three
very serious problems with the literature on AA: (1) lack of control groups in
experimental studies, (2) reliance on abstinence as the sole criteria of treatment
success, and (3) over reliance on self-report measures of both AA attendance and
outcomes. Another important issue is that the effectiveness of a particular
treatment will be affected by whether this is the only type of treatment received.
Concurrent involvement in other types of treatment can obscure results, and is
particularly problematic in terms of evaluating multimodal treatment programs in
which AAINA is only one component (Emrick, 1987; Thurstin, Alfano,8 Nerviano,
1987). The effects of AA are often confounded with the effects of the other
components of the treatment program.

One of the major obstacles to empirical investigation of AA is inaccessibility
of AA members. This limitation has resulted in many studies using biased samples
- either volunteers (self-selection bias) or court referred participants (who might be
quite different from other members). Additionally, most studies have been
conducted with socially stable, Caucasian males. To date, there have not been

any studies of the efficacy of AA for persons who are dually diagnosed.

34

gigala of the Present Stady

This study will address some of the weaknesses identified in prior research.
One of the major criticisms of prior research is a failure to incorporate program
theory into research evaluations. The current study will compare two treatment
models. In light of the stated program theory and goals of AA and those of the
specialized treatment received by the study participants, an attempt will be made to
test a model of integrated treatment (see Figure 1, Model A) for persons with dual
diagnoses as opposed to a model which is more consonant with AA program
theory (see Figure 2, Model B). This study will examine whether AAINA
attendance, following specialized treatment, relates directly to outcomes in other
functioning domains (e.g., psychological and family/social functioning, and
recidivism), as would be suggested by an integrated treatment model, or whether
these outcomes are mediated by abstinence or improvement in drinking/drug use,
as would be suggested by an AAINA treatment model.

The proposed model for this study (see Figure 1) was adapted from the
work of Moos and his colleagues. According to Moos, Finney, and Maude-Griffin
(1992), the individual’s personal characteristics and life context interacts with
treatment to produce differential outcomes. Specifically, participation in treatment
is influenced by the individual’s resources prior to entering the group (noted in the
model as individual characteristics/prior experiences), and the contextual factors
that concur with treatment involvement (here noted by concurrent engagement in

other forms of treatment besides AAINA). Individual characteristics and prior

35

treatment experiences are expected to correlate with involvement in aftercare
services, particularly AAINA. These characteristics will influence the level of
participation in aftercare treatment, as well as the types of services used. Service
use is expected to be related to subsequent functioning in the treated domains
(i.e., alcohol/drug use, psychological functioning, etc). The present study will
examine the extent to which pre-existing characteristics can be used to predict
subsequent service use, and whether service use has any significant impact on
outcomes. (The specific measures of each construct are depicted in the lower part

of the figure, enumerated in the square boxes).

36

 

 

gonad—833:0.“ .

2.3395 70332—9»:
.8038.— _a_oom\.£8am
nae—no.5 man

.9328 .282

v-a'oi'aiv'vn

.332 3258.5. 3338.: 2:. . < .362 A 8.83

 

 

 

 

 

 

m 25. a <Z<< £183. sea as .4
m oat. a 85 serum 8 25. .m
Ens-m once—3:2; <Z\<< .u

.26 42>)~ .2

 

 

23.8.... :8 833025 .4
«5:.— botaom .m

352.8; <z\<< com...— .N
:80 fimoioBQBSA—Ehm .—

 

 

 

383:5

 

 

 

3253; 0.82334

 

 

 

mooaotomxm no:
333280830 .32sz

 

37

 

 

3.53.3.3 .m
39:28 133.3%: .N
.532.— Eoomsma: ._

 

 

 

«Bo—no.5 uEn .N
.8035; :58: ._

 

 

     

 

moEOOwn—o

   

34:? 3933.6
3.20% 350 AH ANI «dogma; ouaouoa<

.232 .893; <z\<< 2; - m 332 .N 23:

 

 

m as? .. <22... 32.32 :5 s .4
m 08; «a com: uofltommo 093. .m
53:5 3533: <z\<< .N

51% «9:: A

 

 

   

_

 

 

30839.... DOD uni—£03m .
as: 3358 .

cone—ESE <z\<< coin .
9.95 bmomuefioxaoagm .

I—NMV

 

 

 

 

32810an St.—
\mocmt205aao 3:223:

 

38

Previous investigations with the Substance Abuse and Mental Illness
(SAMI) project data also support the inclusion of the variables selected for the
current study. The SAMI project was designed as an experimental evaluation of
the effects of a specialized treatment program for persons with dual issues of
mental illness and substance abuse. Findings from this study indicated that
persons who participated in the specialized treatment were more knowledgeable
about their substance abuse problems and were more motivated to maintain their
sobriety, as indicated by their stated plans to continue treatment and to refrain from
alcohol and drug use (Herman et al., 1996). The specialized treatment resulted in
greater motivation, which was related to subsequent AAINA involvement and
improvements in functioning (BootsMiller et al., 1996). Over an 18 month follow-up
period, persons who participated in the specialized treatment reported significantly
fewer alcohol problems, although there were no significant findings for other
functioning domains. 1

Regarding psychological functioning, Ribisl (1995) found that psychiatric
distress was the strongest cross-sectional predictor of negative outcomes,
particularly alcohol and drug abuse. Persons with greater psychiatric distress
reported significantly more substance abuse following treatment. Psychiatric
distress was a stronger predictor than sobriety support, social network factors, and
all other variables in the model.

The results of these previous investigations indicate that psychiatric

distress, participation in the specialized treatment, development of plans to

39

maintain sobriety, and AAINA attendance are important factors related to
participant outcomes. However, the nature of these relationships remains unclear.
These variables could primarily serve as main effects or as moderators for other
variables, such as pre-existing psychiatric symptomatology. The current study will
build on previous SAMI studies and examine the effects of psychiatric
symptomatology, using prior AAINA attendance, specialized treatment, and
sobriety plans as moderators. It will also examine the role of sobriety as a mediator
of functioning in other domains. Results from this study will add to our knowledge
base regarding treatment of persons with dual diagnosis.

While many prior studies have focused on homogeneous samples of middle-
class White males with alcohol problems, the current study is conducted with an
under represented sample. The participants are dually diagnosed patients
receiving treatment at a large public psychiatric hospital, most of whom are low-
income, poorly functioning, African American males. (Despite the homogeneity in
sociodemographics, the participants of this study are quite heterogeneous in terms
of their diagnoses and treatment experience.) Previous investigators have found
that ethnicity, other personal characteristics, and contextual factors influence the
patterns and types of services used, including AAINA (cf. Brown, Joe, 8
Thompson, 1985; Hanson, 1985; Humphreys 8 Woods, 1993; Moos, Finney, 8
Maude-Griffin, 1993; Thompson, 1985).

In terms of research methodology, this study will differ from prior studies in

several ways. First, a control group will be utilized - participants who received the

40

hospital treatment, but opted not to attend AA regularly as part of their aftercare.
Second, contamination of AA treatment effects by other concurrent services will be
addressed by directly testing whether use of outpatient substance abuse or mental
health services following hospitalization had any additional effect on treatment
outcomes. Third, the outcome variables in this study will include psychological and
family/social functioning and rehospitalization, in addition to drug and alcohol
involvement.

The limitations of this study are similar to those in prior investigations. This
study will be most limited by reliance on self-report data, and the lack of direct
measures of sobriety or serenity (the stated goals of AAINA). In addition, there
were no direct measures of the participants’ engagement in AAINA process (e.g.,
forming relationships with sponsors, doing 12-step work, giving testimonials, etc.)

which are thought to be factors related to successful outcome.

STATEMENT OF THE PROBLEM

There have been a number of studies examining the characteristics of
persons who engage with AA, and the pattems of involvement with AA and their
relationship with outcomes, particularIy drinking outcomes. There have been few
controlled studies of AA, few studies that have examined outcomes other than
drinking, and hardly any controlled studies of AA attendance and outcome for
persons with dual diagnoses of mental illness and substance abuse.

For this dissertation study, I propose to examine the suitability of Alcoholics
Anonymous and Narcotics Anonymous for a group of persons with dual issues of
mental illness and substance abuse. In order to assess the impact of experienced
symptomatology on participation in and benefit from AAINA, participants will be
divided into groups based on scores on the SCL-10 and history of prior psychiatric
treatment. These groups will be compared in terms of patterns of involvement with
AAINA across the three follow-up interviews (2-, 6-, and 10-months post-
discharge). Factors which might be related to AA attendance and/or outcomes -
such as history of AAINA involvement, concurrent aftercare treatment, motivation

level at discharge, etc. will be controlled.

41

42

A number of studies have found significant relationships between AA
attendance and drinking outcomes, which tends to support the explicit goal of AA -
helping alcoholics attain sobriety. Because there have been few studies that
examined outcomes in other domains, it is not clear whether AA has any direct
impact on other areas of functioning or whether improvements in other domains are
moderated by attainment of abstinence. Since abstinence/sobriety is the primary
goal of AA, a model that incorporated abstinence as a moderator of other
functioning outcomes might be more consonant with the AA program model (see
Figure 2, Model B). If sobriety is the major outcome of successful engagement in
AA, then Model B would be a more appropriate test of the effects of AA attendance
and the program theory of AA would be supported. Conversely, Model A is more
consonant with an integrated treatment approach that directly impacts a number of
functioning domains simultaneously.

The current study will attempt to address a number of issues related to AA
attendance and outcome and answer the specific research questions delineated in
the next section. While this study will not supply definitive proof of the efficacy of
AA, it will shed some light on the factors related to successful use of AA by dually-

diagnosed persons, especially African American males.

HYPOTHESES

Prior research indicates that persons who attend AAINA are psychologically
healthier, have had prior substance abuse treatment and/or AAINA involvement,
and are more likely to be motivated for sobriety. Thus, the following hypotheses
are suggested:

(1) Individual characteristics will systematically predict involvement in AAINA.
Persons who reported experiencing a high level of psychiatric symptomatology and
multiple prior psychiatric hospitalizations (high symptom/chronic group) will be less
likely to attend AAINA after hospitalization; those with lower levels of psychiatric
symptomatology and few prior psychiatric hospitalizations (low symptom/acute
group) will be most likely to attend AAINA (See Model A, path # 1, p. 35).

(2) Persons in the high symptom/chronic group who do attend AAINA will not
become regular attenders; they will be more likely to drop out than those in other
symptom/chronicity groups (Model A, path # 1).

Previous studies have indicated that persons who attend AAINA regulariy,
especially following professional substance abuse treatment, are likely to have
better outcomes than those who do not attend or attend sporadically. Several

studies have identified significant differences in alcohol use based on AA

43

44

attendance, and some have suggested that AA attendance can also lead to
improvements in psychological and family/social functioning as well. Thus, the
following hypothesis will be tested:
(3) Patterns of AAINA service use after hospitalization are expected to be
systematically related to therapeutic outcomes. Participants who begin regular
involvement with AA soon after hospital discharge (i.e., report attending AAINA
over two or more consecutive time periods, beginning at 2- or 6-months post-
discharge ) are expected to show better outcomes than those who never attend or
drop out (See Model A, path 2). Substance abuse, psychological, family/social,
and rehospitalization outcomes will be assessed at Time 3 (10-months post-
discharge) and Time 4 (14-months post—discharge). .

There have been few studies examining the effects of concurrent
professional and self-help services. In order to control for contamination of the
effects of AAINA attendance with other service use and use of other services,
these relationships will be tested directly. For persons with dual diagnosis, it is
anticipated that use of mental health services will have important implications for
outcome. Thus:
(4) Participants who report using other outpatient substance abuse services or
mental health services in addition to AA will show greater improvement on the
outcome measures than those who attend AA only (Model A, path 2).

Finally, this study will compare an Integrated or nonmediated treatment

model with a model which would be suggested according to AAINA program

45

theory. If the Integrated model is more consonant with the data, AAINA and other
service use will directly impact alcohol and drug problems, as well as psychological
and family/social problems, and rehospitalization. The AAINA Treatment model
would be supported if attainment of sobriety (or reduction in alcohol/drug problems)
led to improvements in psychological and family/social problems, and reduced the
incidence of rehospitalization. No prior research has directly compared these
models, therefore: ’

(5) Exploratory path analysis will be conducted to test the predictive strength of the
Integrated Treatment model (Model A) as opposed to the AA Treatment model
(Model B), using indices of model fit based on EQS Structural Equations Program

(Bentler, 1995).

CHAPTER 2

METHOD

This dissertation study was conducted as part of a larger project
(Substance Abuse and Mental Illness Project, SAMI) funded by the National
Institute of Mental Health, through the Michigan Department of Mental Health. The
SAMI project involved an experimental evaluation of the effects of a specialized
treatment program for persons with dual issues of mental illness and substance
abuse who received treatment at a large public psychiatric hospital. Baseline data
was collected from 474 inpatients at Northville Psychiatric Center who screened
positive for dual problems of substance abuse and mental illness. The current
study utilized-data from the baseline (hospital admission) and discharge interviews
and four of the follow-up interviews (2-, 6-, 10- and 14-months post-discharge).
§er_n_9|_e

Upon admission to Northville Regional Psychiatric Hospital, 627 patients
who screened positive for dual problems of mental illness and substance abuse
were approached to be in the SAMI study. All incoming patients were screened for
eligibility using a modified version of the Michigan Alcoholism Screening Test

(Selzer, 1971) and had to screen positive for dual problems of mental illness and

46

47

substance abuse. Patients who were severely cognitively impaired,
developmentally disabled, or who showed a predisposition for violence or
inappropriate sexual behavior were excluded from the study. An additional
requirement was that patients had to remain in the hospital for a minimum of five
days to insure that some minimum level of treatment was received. Of those
approached, 535 met the eligibility requirements and agreed to be in the study; 474
had sufficiently complete information from the hospital admissions and ward
interviews to be considered part of the baseline SAMI sample. For the purposes
of this study, all analyses were conducted With 284 participants who had 100%
complete data on pre-existing individual characteristics and prior experiences, post-
discharge AAINA attendance, and the selected outcome measures. Due to the
use of path analysis, the data sets used in the analyses had to have valid scores
for each variable used in the analyses. Missing data analyses and comparisons of
persons included in the analyses vs. those excluded are discussed in a later
section (see pp. 61-63).

Table 1 shows the demographic characteristics of the 474 patients who
comprised the baseline SAMI population. The sample consisted primarily of
African American (77%) males from the Detroit metro area. Most of the
participants were unemployed or marginally employed and presented with a

plethora of other psychosocial problems.

Table 1. Demographic Characteristics of Participants (N = 474)

 

 

Variable %/ Mean (SD)
Gender Male 74%
Ethnicity African American 77%
White 23%

Age (years) 33 (7.3)
Marital status Never Married 63%
Married 9%
Separated Divorced or Widowed 28%
Years of education 11 (2.2)

Usual employment status - Full-time 30%
Part-time 23%
Unemployed 33%
Out of workforce 14%

Monthly income $ 688 (1333.2)
Previous psychiatric hospitalizations None 33%
1 or 2 30%
3 or more 37%
Previous alcohol/drug treatment None 33%
Alcohol or Drug 38%
Both 29%
Self-labeled drug problem No problems 11%
Drug problem 26%

Addict

63%

49

Table 2. Distribution of DSMlll-R Diagnoses

 

Mental Illness

 

 

N' %
Schizophrenic Disorders 125 27.6%
Organic Mood Disorders 102 21.0%
Adjustment Disorder 73 15.1%
Mild Affective Disorders 82 16.9%
Antisocial Personality Disorder 43 8.9%
Major Depression 37 7.6%
Bipolar Depression 27 5.6%

 

Substance Dependence

 

Alcohol Dependence 220 45.4%
Cocaine Dependence 183 37.7%
Polydrug Dependence 147 30.3%
Cannabis Dependence 46 9.5%
Opioid Dependence 21 4.3%
Missing 63 13.9%

 

' Overall N = 474 _
Note. Numbers add up to more than 474 because participants could receive

multiple diagnoses.

50

In terms of diagnoses, the participants in this study were more
homogeneous that previous studies of persons with dual diagnoses. Table 2
shows the distribution of DSM-lll-R diagnoses in the sample. Diagnoses of
Schizophrenia or Schizoaffective Disorder and Organic Mood Disorders were
most common in the sample, accounting for nearly half of all mental illness
diagnoses. Alcohol and Cocaine Dependence together accounted for more than
80% of all substance dependence diagnoses, with about one-third of the
respondents receiving both diagnoses. Approximately one-third (33%) of the
respondents in this sample had both a major mental illness (e.g., Schizophrenia,
Organic Mood Disorders, Major Depression, or Bipolar Depression) and a
substance dependence diagnoses. A small percent of the sample (13. 9%,
labelled as missing under substance dependence diagnoses) were diagnosed with
mental illness, but did not meet the formal criteria for either a substance
dependence or substance abuse diagnosis.

Measures

The following is a list of instruments and measures used to test the
hypothesized relationships in Models A and B. (See Table 3 for a summarized
listing of the measures used in this study.) Reported reliabilities of the measures
were generally taken from the baseline hospital interview unless otherwise

indicated.

51

Table 3. Measures Used in Current Study

 

 

Individual Characteristics! Aftercare Treatment Outcomes
Prior Experiences ,
Symptom/Chronicity - ANNA Ever Alcohol Problems
Group - Patterns of ANNA - ASI Alcohol composite
attendance score

- Types of Services used
- # Days attending ANNA Drug Problems
at Time 3 - ASI Drug composite
score

Family/social Problems
- ASI Family/social

Covariates composite score

Prior ANNA attendance

Plans for sobriety Psychological Problems
Specialized CDU treatment - ASI Psychiatric

composite score

Rehospitalization
- # Hospital treatments

for emotional or
psychiatric problems

52

Individual Chagateristics 8 Prior Experiences

A. Psychiatric Symptom Groggs. Based on respondents' level of psychiatric
symptomatology at the time of hospital intake, and the number of times they had
been previously hospitalized for psychiatric problems, four symptom/chronicity
subgroups were created: low symptom/acute, low symptom/chronic, high
symptom/acute, and high symptom/chronic.

P_sychiatric Symptomatology. The Symptom Checklist-10 (SCL-10; Nguyen,
Atikisson 8 Stegner, 1984) was used as an index of baseline psychiatric
symptomatology. The SOL-10 is a 10-item version of the SCL-90 (Derogatis,
Lipman, 8 Covi, 1973) that is used to assess global psychopathology. The items
on the SCL-10 cluster into three factors: Depression, Somatization, and Phobic
Anxiety. Reliability and validity of the instrument have been well established and
cited elsewhere (Hoffman 8 Overall, 1978).

Psychiatric Chronicitv. Self-reported history of psychiatric hospitalization
was used as an index of chronicity of mental illness problems, the second variable
used for grouping on individual characteristics. At the baseline interview, I
participants were asked to report how many times in their life they had been
hospitalized for psychological or emotional problems.

First, participants were divided into groups based on their scores on the
SCL-10. A median split was used to establish two comparison groups: a low
symptom group and a high symptom group. The low symptom group was

composed of 203 persons who scored 2.6 or lower on the SOL-10. The high

53

symptom group was composed of 198 participants who scored between 2.61 and
5.0. Scores on the SCL-10 were normally distributed, with values ranging from 1.0
to 5.0. The mean score was 2.68; the median was 2.6.

Second, groupings on chronicity were established. Participants who
reported two or more prior psychiatric hospitalizations were categorized as a
chronic group. Those who reported fewer than two prior hospitalizations were
categorized as an acute psychiatric group.

The groupings based on the median split on the SCL-10 and history of
psychiatric hospitalization were combined to form the four groups used in the
current analyses. The four groups were: a low symptom/acute group (a = 86), a
low symptom/chronic group (a = 59), a high symptom/acute group (a = 66), and a
high symptom/chronic group (a = 73). These groupings will be used in most of the
analyses reported.

Covariates

The following is a list of other measures of individual characteristics and
prior experiences which are expected to be related to aftercare treatment and/or
treatment outcomes. While no specific hypotheses are made regarding these

variables, they are included in the models as covariates:

 

Prior AAIN_A Attendance. As part of the baseline interview, participants
were asked if they had attended AA or NA in the 30 days before they came to the
hospital. Frequency of AA attendance and NA attendance were found to be

significantly correlated (r = .42) and were therefore combined in all analyses.

54

These two questions were combined into one dummy coded variable which was
used to assess ANNA involvement prior to hospitalization. Seventy-three
participants (25.7%) had attended ANNA during the 30 days prior to .
hospitalization.

fins for Sobflehr. During the hospital discharge interview, participants
were asked to report their plans for engagement in treatment, and other efforts to
maintain sobriety and psychological health. This information was collected in a 14
item scale called Discharge Plan. There were three items on plans for sobriety
maintenance which were combined to form a separate scale, called sobriety plans,
which was used as a covariate in the analyses. Responses to these items were
coded from 1 (definitely will) to 4 (definitely not). The mean scores on this scale
was 1.38, indicating that at discharge most participants had plans to maintain their
sobriety by going to ANNA, staying sober, and staying off of street drugs. The
items on this scale showed moderate internal consistency (Chronbach alpha = .63).

Specialized Chemical Dep_e_ngencv Unit (CDU) Treatment. Receipt of
specialized treatment was measured by a two-level variable indicating whether the
respondent was assigned to the specialized treatment unit, where they participated
in ANNA and other substance abuse treatment or to one of the standard
psychiatric wards. Of the 284 participants included in this study, 192 received the
specialized treatment; 92 were assigned to a standard psychiatric ward.
Aftercg Treatment

Each of the following measures of ANNA attendance and other service use

55

were assessed at each community interview.

ANNA Attenaance. Participants were asked whether they had participated
in Alcoholics Anonymous, Narcotics Anonymous, or any other self-help groups
during the past 30 days. This item was summed across the three interview time
periods (2-, 6-, and 10-months postdischarge) and dummy coded (0 = never
attended ANNA; 1 = attended at some time points) to determine how many
participants ever attended ANNA during the follow-up. Of the 284 participants with
complete follow-up data, 107 did not report attending ANNA at any of the follow-up
interviews; 177 reported attending ANNA at one or more interview time points.

In order to assess patterns of ANNA attendance, 3 four-level variable was
created grouping participants based on ANNA attendance across the three
interview time periods. This variable was coded 0 if the participant never attended
ANNA; 1 if they reported attending at one of the interview time points; 2 if they
reported attending ANNA at two consecutive time points; and 3 if they reported
attending ANNA at all three follow-up time points. This group resulted in four
groups: nonattendees (g = 107, 37.7%), inconsistent attenders (g = 69, 24.3%),
moderately consistent attenders (a = 39, 13.7%); and regular attenders (g = 69,
24.3%).

The third measure of ANNA attendance was the number of days
participants reported attending ANNA at Time 3 (10-months post-discharge).
Participants were asked how many days during the last 30 days they had attended

AA. The same question was asked regarding NA. These two items were

56

averaged to derive a continuous measure of ANNA participation, the # days
attending ANNA at Time 3.

Type of Service Use. In addition to items about ANNA attendance,
participants were asked if they had participated in any outpatient alcohol or drug
treatment in the two weeks before the interview, such as a substance abuse group,
a dual-diagnosis group at a mental health center, individual counseling for
substance abuse problems.

The use of outpatient mental health services was assessed by asking
participants if they had been treated for psychological or emotional problems as an
outpatient or private patient since their last interview. They were also asked to
report if they had received other mental health services during the two weeks
before the interview, such as crisis assistance (via hotline, shelter, mental health
center, etc), individual therapy with a therapist, counselor, doctor, or case
manager, or if they had participated in psychotherapy/support groups where they
talked about their feelings or any emotional problems they were having.
Participants who reported receiving any of these services were coded as having
received outpatient mental health services.

In order to assess whether ANNA participation had differential effects when
combined with other services, a summary variable was created to categorize
participants by the types of services they used. Due to the low frequency of
outpatient substance abuse service use (only 8 participants reported using these

services), use of mental health and substance abuse services were combined to

57

form an index of professional service use. A four-level variable was created to
describe the types of services participants were using at Time 3. Participants who
did not use any services were coded as 0; those who attended ANNA only were
coded as 1; those who participated in ANNA and professional services were coded
as 2; and those who used professional services only were coded as 3. One
hundred people reported not using any services at Time 3, 40 attended ANNA
only, 59 attended ANNA in addition professional services, and 85 participants
reported using professional services only. These groupings were used in the
analyses testing the relationship of type of service use to outcomes (Models A and
B, paths 2, 3, 4, and 5).
Outcome Measures

The Ad_diction Severity Inga); (ASI; McLellan, Luborsky, O'Brien, 8 Woody,
1980) was the primary instrument used to measure treatment outcomes, including
alcohol/drug use, psychological functioning, and family/social functioning. The ASI
is a structured clinical and research interview that assesses the severity of seven
global domains of life functioning which are often found to be impaired in persons
with substance abuse problems. The seven domains are: Medical, Employment,
Legal, Alcohol, Drug, Family/Social, and Psychiatric. As part of the ASI interview,
participants are asked several questions related to their functioning in these areas.
For each functioning domain, objective ratings of problems are obtained by
weighted mathematical calculation of seven composite scores. Higher values on

the composite scores indicate more problems in that domain. Each of the

58

composite scores showed acceptable internal consistency (Chronbach Alphas:
medical = .86; employment =.60; legal = .70; alcohol = .81; drug = .57; family/social
= .56; and psychiatric = .82). The four composite scores used in this study were
the alcohol, drug, psychological, and family/social composites.

Rehospitalization. Respondents were asked whether they had been in a
controlled environment in the past 30 days, including alcohol or drug treatment. In
addition, the participants were asked to report how many times they had been
treated for psychological or emotional problems in a hospital since their last
interview. Due to the low frequency of inpatient substance abuse treatment, this
variable was dropped from the analyses. The number of psychiatric
hospitalizations since the last interview was used as the primary measure of
recidivism.

Procedures

As part of the procedures, incoming patients who met the criteria and
agreed to participate in the SAMI study were randomly assigned to one of two
types of hospital wards: one of the regular psychiatric wards or the specialized
Chemical Dependency Unit (CDU). (Although assignment was random, more
patients were assigned to the CDU - approximately 75% - to make use of this
specialized treatment for a population in need). Patients in the regular psychiatric
wards received individual and group psychotherapy, and could attend ANNA
groups if they were interested. On the CDU, patients received comprehensive

biopsychosocial treatment for both their mental illness and their substance abuse

59

problems. CDU participants attended 40 educational lectures which covered a
wide variety of topics related to dual diagnosis. The goal of the lectures was to
provide factual information regarding substance abuse, to change participants’
attitudes towards substance abuse, and increase motivation for sobriety. The CDU
program staff was multidisciplinary and provided treatment for physical, psychiatric,
family/social, and substance abuse problems. Specific treatment components
included: individual and group psychotherapy, family therapy, psychoeducational
lectures on Substance abuse, and ANNA groups.

Upon admission to the hospital, all participants completed a baseline
interview; at the time of discharge, a second interview was completed and follow-up
data for community tracking was provided. The five follow-up interviews were
conducted approximately every four months after hospital discharge, beginning at
2-months post-discharge. Using the information that participants provided at
discharge, and other intensive tracking measures, SAMI staff were able to locate
84 to 86% of the participants at each interview time point following hospital

discharge.

Chapter 3

Results

patflnalvsis Strat_egy

The following is a description of the statistical procedures that were used to
test the hypotheses of this study. Logistic regression was used to test the first
hypothesis, predicting post-hospital ANNA Ever. Analysis of Variance was used
to test hypotheses 2, 3, and 4 examining the relationships between
symptom/chronicity status and ANNA attendance patterns and the relationships
between the amount and types of services used and outcomes at 2- and 6-months
post-hospitalization. These analyses were conducted in SPSS for Windows
version 6.0 (Norusis, 1993) using experimental condition, prior ANNA attendance
and plans to maintain sobriety as covariates. Finally, a series of path models were
tested using the EQS Structural Equations Program (Bentler, 1995). These path
models examined the relationships between symptom/chronicity status, several
variations of ANNA and other service use, and outcomes in five domains: alcohol
problems, drug problems, psychological problems, family/social problems, and
rehospitalization. Two types of models were fit to the data: an integrated treatment

model (Model A) and an AA treatment model. Results for the two models will be

60

61

reported separately.

In order to comply with the data requirements of EQS, a few changes were
made in the data. First, all participants with missing data on any of the study
variables were excluded from the analyses. Only the 284 participants with 100%
complete data were included. (The next section discusses the missing data
analyses and comparison of those included in the analyses vs. those exclued).
Second, due to problems with non-normally distributed variables in the model (e.g.,
# days attending ANNA, psychiatric rehospitalization, and family/social problems),
all continuous measures were converted via log transformation. The variables with
highly skewed distributions tended to be negatively skewed L-shaped distributions,
with most respondents having scores of 0 (with the exception of ANNA attendance
which was bimodally distributed). Given the typically high number of near-zero
values, the following transformation was conducted: Iog,o(X =1) for each
continuous variable (Howell, 1987).

The third substantial change was the elimination of the covariates in the
path models. One variable (sobriety plans) was measured at the discharge
interview, whereas the independent variables (the symptom/chronicity groups) were
measured at baseline or hospital entry. Given the time sequence of these
measures, sobriety plans could not logically be entered into the model as a
predictor occurring before symptom/chronicity, and was therefore eliminated from
the path analyses. The other two of the three covariates were dichotomous

variables, which would have been used to predict the dichotomous design variables

62

representing the four levels of symptom/chronicity. Bentler (1995) strongly advises
against including dichotomous dependent variables in structural equation modeling
with EQS. It is noted that the path coefficients are not reliable and induce difficulty
of interpretation. Therefore, wherever it was feasible, dichotomous variable were
eliminated from the model, or used only as predictors (as in the symptom/chronicity
groups). The one exception to this was the models testing types of service use,
which was entered as three dichotomous design variables. Given the violation of
EQS model assumptions, the findings from the models using type of service use
should be interpreted with caution.
Missing Data Analyses

Participants with missing or invalid scores on any of the 23 variables used in
this study were excluded from the analyses. Most participants who were excluded
failed to complete one of the interviews from which data for this study was taken,
usually because the project interviewers were unable to locate the participant. A
less frequent cause of missing data was participant refusal or inability to answer
the survey questions.

Participants were divided into two groups based on the completeness of
their interview data. There were 284 participants who had complete data for all 22
study variables; 190 respondents had missing data for one or more variables. In
order to determine if participants with missing data differed substantially from those
with 100% complete data, a series of analyses were conducted comparing the two

groups on baseline demographic characteristics, treatment experiences, psychiatric

63

diagnoses, and symptom/chronicity grouping. Between group comparisons were
conducted using ANOVA for continuous variables and Chi-square for discrete
variables. The results of these analyses are reported in Table 4.

While there were no significant differences between the two groups on most
of the demographic, treatment, and diagnostic variables, participants with complete
data were significantly more likely to be male, to have received multiple substance
dependence diagnoses, and to have participated in the specialized Chemical
Dependency Unit treatment in the hospital. These findings are somewhat
surprising, and suggest that males with more serious substance abuse problems
were more likely to remain in the study through 14-months follow-up. One tentaive
hypothesis which might explain this finding is that males in this sample were more
likely to be jailed or rehospitalized during the follow-up period. When participants
were institutionalized, the interviewers often had an easier time of locating them

and conducting the interviews, since they were dealing with a "captive audience."

64

Table 4. Comparison among Participants with Complete vs. Missing Data

 

Level of Data Completeness

 

 

 

Missing 1 or Complete
more variables
n =190 n = 284
(40.1%) (59.9%)
Variable n % n % Statistical Test
Gender (% Male) 129 68% 220 77% x2 (1, 474) = 5.37
Ethnicity 135 71% 193 68% n.s.
(% African American)
Age (Mean) 33.8 ----- 33.1 ----- n.s.
Marital Status
Never Married 121 64% 179 63% n.s.
Married 19 10% 23 8%
Separated, etc. 50 26% 81 29%
Usual Employment Status
Employed 92 48% 158 56% n.s.
Unemployed 70 37% 84 30%
Monthly Income (Mean) $663 ----- $683 ---- n.s.
# Previous Psychiatric 3.4 ---- 4.0 ----- n.s.
Hospitalizations (Mean)
Severe Mental Illness 110 58% 148 52% n.s.
Diagnosisb
Substance Dependence 45 24% 101 36% x2 (1, 474) = 5.27
(% Polydrug Dependent)
Specialized CDU Treatment 109 57% 192 68% x2 (1, 474) = 5.15
Treatment Duration (Mean 42.8 ----- 43.7 ----- n.s.
# days spent in hospital)
Prior ANNA Attendance 47 25% 73 26% n.s.

 

65

 

Level of Data Completeness "

 

 

 

Missing 1 or Complete
more variables Data
n =190 n = 284
(40.1%) J59.9%)
Variable n % n % Statistical Test
Symptom Group
High Sx/Chronic 51 27% 73 26% n.s.
High Sx/Acute 43 23% 66 23%
Low Sx/Chronic 31 16% 59 21%
Low Sx/Acute 45 24% 86 30%

 

Note: All reported results are significant at p <.05.
' There were 22 study variables; participants with valid scores on all 22 variables were

considered to have complete data.
” Participants with Schizophrenia, Organic Mood Disorders, Major Depression, or

Bipolar Depression.

66

Descriptive Statistics on Stadv Vaflplaa

Table 5 provides descriptive information on ANNA attendance and other service
use during the follow-up period for the 284 persons included in the analyses. Almost
two-thirds of the respondents reported attending Alcoholics Anonymous or Narcotics
Anonymous at some time after hospitalization. However, the patterns of attendance
tended to be sporadic, with many respondents only attending for a brief period of time.
Of those who reported attending AA or NA at the 2-, 6-, or 10-month follow-up, only
39% were attending at each interview time-point. Those who were attending at the 10-
month time point reported attending an average of 2 days during the previous month.
Most ANNA attenders went on ten days or fewer", only 1.4% were attending ANNA
most days out of the month (i.e., 20 days or more).

In terms of types of services used, one-third of the respondents reported not
using any self-help or outpatient services at the 10-month follow-up. A small number of
people were attending AA only, while many more were receiving professional services
(typically mental health services). About one-third of the respondents were receiving
professional services only, and not attending AA or NA.

Table 6 shows the descriptive statistics on the outcome measures. The ASI
composites were used to assess problems in the alcohol, drug, psychological, and
family/social domains, with higher scores indicating more problems. Rehospitalization
represents the number of times participants had been rehospitalized since the last
interview. This sample reported significant problems in all of the ASI functioning

domains. However, the scores on drug problems were relatively low and showed

67

little variability. In terms of rehospitalization for psychiatric problems, few participants
were rehospitalized at either time 3 or time 4. But a small number of participants
had been rehospitalized more than once since the'previous interview. Thus, there

was considerable variability in scores on rehospitalization.

68

Table 5. Descriptives on ANNA and other Service Use during Follow-up (3 =284)

 

Variable %IMean (SD)
ANNA Attendance Pattern

 

Never attended 37.7%

Attended at 1 timepoint 24.3%

Attended at 2 consecutive timepoints 13.7%
. Attended at all three timepoints 24.3%

Types of Services Used at 10-month Follow-up
None 35.2%
AA only 14.1%
AA plus professional services 20.8%

Professional services only 29.9%

Average Number of Days Attended ANNA 2.2(4.77)
during previous month at 10-month Follow-up

0 days 65.5%
1-10 days 27.8%
11-20 days 5.3%

20 days or more 1.4%

 

69

Table 6. Descriptives on Outcome Measures

 

 

Mean Standard Deviation
T3 Alcohol Problems .24 .20
T3 Drug Problems .11 .08
T3 Psychological Problems .30 .24
T3 Family/Social Problems .23 .17
T3 Rehospitalization .31 .66
T4 Alcohol Problems .24 .19
T4 Drug Problems .11 .09
T4 Psychological Problems .28 .24
T4 Family/Social Problems .21 .16

T4 Rehospitalization .33 .73

 

7O

Correlations among the Study Va_riafls_

Table 7 displays the zero-order correlations among the 23 variables used in the
analyses. There were small to moderate correlations between the independent
variables and covariates (see triangle 1 in Table 7), with sobriety plans and treatment
condition having the highest correlation ([ = .24). This suggested that the variables
were measuring relatively distinct domains. Higher correlations were found among the
ANNA service use measures (gs range from .21 to .81) due to the fact that these
variables included common items. However, all of the service use variables were
included as they were thought to measure conceptually distinct factors. None of the
service use variables were used in the same analyses. Finally, the time 3 and time 4
dependent measures were found to be significantly correlated (gs range from .25 to
.59), therefore, the measures were entered together in the analyses predicting

functioning in each outcome domain.

.225 33:03.0 as“. u w En; 25¢...ng 235%.. .i. 0 .coaucoo x ... 3.... .8) ur no: no. F .0 page So 3.95:; 60.3028 08...... .
.So.v m «a E3588 2m 0.. A m. _..o.v m «a $8583 9n 0?. A m. ”.96. 00V 0 05 «a 285:0.» 9a 3. A mm 862

 

 

71

 

 

 

 

I 00. 0v. 0r. 3. 00. 00. 0... 00. 00. F0. N0. 00. N0.- Sr .0. V0. 00.- 00. 0V- N0.- 00. 00. 009.0”. vh.0N
I 0N. 5. 0... P0. 0N. 3. F0. N0. For v0.- w0. ..0. 00. 00. N0. 3. Sr 00.- 5. N0. 00. 30...... En... E..NN
I 00. N0. .vN. 0N. 0v. 0... 0N. 00.- 50. P0. 00.- Nwr 0F.- .0... 00. 00.- NN.- 00. vo. 00. 32.". noxwa 3....N.
I N0. N. 0?. 0... N0. 0?. ..0.- NF. 3. 9.. v0.- 00.- 5.- 00. 00. 3.- N0. 00. N0.- 321020 v... .0N

I t. 00. 0e. 0... 00. 00.- 00.- V0. 00.- 00.- N0.- co. 0?. 00. 50.- «N. 00.- N0.- 32A... zo.<v._..0e

I 0?. 00. 0N. N0. 00.- 00. N0. N0.- #0.- N0. N0. 00.- 00. .Nr 0w. #0.- v0.- 000510.70?

I 0v. .N. 0... 00. .0. N0.- 00. 00.- 3.. N... 3. 00. 5.. Sr 0?. N0.- BEA. Emu. 0b.:

l- ..0. NN. 00.- v0. 3. N0.- 00.- ..0. Nw. 5. nor 5.. 00. V0.- 5. 39¢ zo>mn. 0...0F

I F0. .0.- 0w. #0. 0? 0..- 00.- N... 00. 00. 0f- 00.- 00.- F0. mooiuao 00.0..

I 0? 00.- 00.- 0%. of. Sr 00. 00. F0. 00.- 0w. 00.- 90.- 321530.23

I 00.- 00. 0v. N0. 00. ..0.- 0...- N0.- .0. 3.... N. 0... goo/42300.0?

I 0.. 0N.- 00.- 0N; 3. N0.- 00. 0? 00.- 5.- 00. 5:03...“ 0h.N—.

I _.N.- N0. 0v. 0? 3.- 00.- vor 00.- 0?. 00.- Lo.aw<<0...:

I 0v. .0. 00. 00. 00.- 00. 00.- 0... N0. cch<<0POF

I F0. 0? v0.- 00.- 3. 00.- 0N. 00. 58.371670

I N? 00. Sr 0?. 00.- 0... 3. c.o>w<<.0

I 00. N0. 00.- ..0. 00.- 00.. .o_co.=.0\xw 505$

I 00. 00.- 0...- N0.- 00. 0503.5 20.... .0

I .0.- v0. 00.- 00.. .0299.me 30.. .0

I 00.- 09.- ..0. 0520.0 26.. d

I 0... VN; wcfln. 30.500 .0

I .0; .<Z\<<._o..n..N

 

I .coaucoo fr .r

 

0N NN rN 0N 0F 0? t 0.. 0.. 3 0.. NF 3 0.. 0 0 N 0 0 v 0 N v 0.058)

 

.00N "law 3.9%; 2.850 new £23900 .mofimtm> 5.0.-poi c0223 macaw—0:00 <- 0.9m...

72

Hypotheses l 8 ll: Relationship of Individual Characteristics to ANN_A Attendance

 

It was predicted that individual characteristics, that is, psychiatric symptoms
and history of psychiatric hospitalization would be significant predictors of ANNA
attendance during the follow-up period. Persons who reported experiencing high
levels of psychiatric symptoms and multiple prior psychiatric hospitalizations were
expected to be less likely to ever attend ANNA during the 10-month follow-up
period, while those with fewer symptoms and more acute psychiatric problems
were expected to be most likely to attend ANNA. This hypothesis was tested via
logistic regression with the three symptom/chronicity design variables as the
independent variables (the low symptom/acute group was used as the constant),
treatment condition, prior ANNA attendance, and sobriety plans as the covariates,
and ANNA Ever as the outcome variable. The covariates were entered together
on step one; the symptom/chronicity groups were entered on step two. Table 8
shows the results of each step of the logistic regression predicting ANNA Ever.
Included in each step are the regression coefficients (B), the standard errors, and
the partial correlations (R) for each independent variable. The B values may be
interpreted as the log odds of the event occurring (i.e., the participant ever attends
ANNA after hospitalization), given a one-unit increase in the predictor variable.

The three covariates entered on step one were significant predictors of
ANNA Ever, classifying 62% of the respondents correctly. Of the covariates
entered on this step, prior ANNA attendance was the only significant predictor.

The log odds of a participant ever attending ANNA after hospitalization were

73

increased by .97 if they had attended ANNA prior to hospitalization. Treatment
condition (i.e., dual diagnosis treatment vs. regular psychiatric treatment) and
having plans to maintain sobriety did not significantly predict the odds of attending
ANNA. On step two, there was a slight improvement in classification of ANNA
attenders vs. nonattenders with the addition of the symptom/chronicity groups. The
percent correctly classified increased to 63%. Of the symptom/chronicity groups,
membership in the high symptom/chronic group was significantly related to ANNA
attendance. The log odds of ANNA attendance decreased by .94 for individuals in
this group. While the coefficients for the other groups were not significant,- all were
in the same direction, indicating that persons in the low symptom/chronic and high
symptom/acute groups were also less likely to attend ANNA. However, the log
odds for the constant was positive (though statistically nonsignificant), indicating
that persons in the low symptom/acute group were somewhat more likely to attend

ANNA than the other groups.

74

Table 8. Logistic Regression of Symptom/Chronicity and Covariates on ANNA

Ever (p = 283)

 

. Step1
x2 Improvement = 11.18“

6822 i

x2 Improvement = 7.79“

 

 

Classified 61.84% Classified 63.25%

B S.E. R B S.E. R
BEG—kl
Tx Condition .12 .27 .00 .12 .28 .00
Prior ANNA .97“ .32 .14 1.09“ .33 .16
Sobriety Plans -.10 .26 .00 -.04 .26 .00
Constant .35 .47
88811.2.
Low leChronic -.58 .37 -.03
High Sx/Acute -.44 .36 .00
High Sx/Chronic -.94* .35 -.12
Constant .72 .51

 

*p<.05

75

The second hypothesis was that persons in the high symptom/chronic group
would be less likely to become regular ANNA attenders, if they ever attended at all.
It was predicted that persons in the more severe/chronic groups would be more
likely to drop out of ANNA than persons in other symptom/chronicity groups. This
hypothesis was tested via ANCOVA using the 4-level symptom/chronicity variable
as the independent factor, treatment condition, prior ANNA attendance, and
sobriety plans as the covariates, and ANNA Pattern as the dependent variable.
Table 9 shows the results of this analysis. As the results for hypothesis l indicated,
persons in the high symptom/chronic groups attended ANNA less consistently
after hospitalization than members of the other symptom/chronicity groups.
Persons in the low symptom/acute group were the most consistent in their ANNA
attendance. The overall model was statistically significant and explained 8.0% of
the variance in ANNA attendance patterns.

Hymtheses III 8 IV: Relationship of ANNA amt Other Service Use to Functioning

Hypothesis three indicated that persons who became consistent ANNA
attenders soon after hospitalization would show better outcomes than those who
never attended ANNA or who attended sporadically. This hypothesis was tested
via five separate MANOVA models with ANNA attendance pattern as the
independent factor and the five sets of outcome measures (e.g., Time 3 alcohol
problems and Time 4 alcohol problems) as the dependent variables. The results of

these analyses are reported in Table 10. The MANOVA results indicate only

76

mo. v m.
.965. some 8.. names 9m 2pm. 5 82; ”9oz

 

 

 

8. L B u m emanaxm :33
255 .858
B... u m. 8565......
B. .8... u. <22... BE
8. u. 8:680 up
mm.mtm>_ oo

Eozma
S. .8...” u m 8. mm. 9.. 8.. 8:885. <23.
“mm .84 .2 u 9 .8 u s .8 u 9 .8 u s o_8.=m>

88:85 2:25 5 Si 2:955 33 28.0.5 :9: 2:25 33

 

.mmw u s «.2898 a 8:20 aaa€5§§e$ so... E88 8:88.... 4.2298628 m<>ooz< .8 28A

no. v m.
.0380 comm 6.. ENE 2... can. 5 mo:_m> No.02

 

 

77

00. 0N. 3‘. 00. N0. .V 0E...
00. NN. 00. 5‘. 0N. 0 0E...
00:050.. wocmm 050.205.”.

 

 

 

 

 

 

2... mm. mm. a. om. v as:
8. 8. cm. mm. mm. m 95
weaned 8825......
«3 mm. mm. 8. B. e 8&5
8. mm. mm. 8. on. m 9:5
mEoBBd _mo_ 29.0 mn.
me. 2. S. S. S. 48.5
o: 2. 3. NF. 3. mos:
953. 8d: 020
E. mm. 8. 8. ea. v as:
8. bad 2. m... 8. em. m 9::
weaned 6:82
New L .8 u a .8 n s .8 u 9 .B. u s 285;
95895.. 95035.. .5865... 80:92
0 __m .m 8392 90:63:00 F .m 63:32. .652
N a. 8822

 

.48 n s 58$ 88222 <22... 52.. 3885“. 8:285 252% .9 285

mo. v m.
.955 comm :2 959: 2m 999 5 mm:_m> ”Boz

 

 

78

 

 

 

 

 

 

3. mm. mm. 8. mm. v 9:5

m: 5.. mm. mm. «a. m 9:5

8:85 8%m 9.630%:

9. om. a. mm. a. v 9:5

S. cm. mm. «N. «N. m 9:5

mien. aboflzfilfl.

E. 5. mm. mu. 3.. v 9:5

8.5 5. mm. mm. 5.. m 9:5

95.205 $5 205 mn.

mad 2. N5. 8. 5. v 9:5

3. .3: 2. 9. mo. 5. m 9:5

wEmEEn. 9:0

39 mm. vm. om. 8. v 35

3. a; mm. a. 9. mm. m 9:5

mEmaQa .0282

"mm. n. a: u s an n 3 8: u s 62 u s 29%.;
2:0 wmo_>._mw mmo_>..mw 2:0 (23 0:02

.mcofimmh—oi .mcofimmBE + <z\<<

 

58 u s m 9:5 a 8m: 9.22% 5° 9:5 :5: 9.228%“. 8:069: 25292 .5 995

79

moderate support for hypothesis III. Of the ten outcome variables, patterns of
AAINA attendance were found to be significantly related to alcohol problems at
Time 3. Persons who reported attending AAINA at each of the three follow-up time
points reported significantly fewer alcohol problems at Time 3 (10-months post-
hospitalization). Time 4 alcohol problems were not significantly related to AAINA
attendance pattern, nor were variations in drug, psychological, family/social
problems or psychiatric rehospitalization significantly predicted by AAINA
attendance patterns.

Hypothesis four was that participants who report receiving outpatient
professional mental health or substance abuse services in addition to AA would
show greater improvement on the outcome measures than those who attended AA
only. This hypothesis was tested via five MANOVA models with type of service
use as the independent factor and the five sets of outcome measures as the
dependent variables. The results of these analyses are reported in Table 11.
Persons who attended AAINA Only reported significantly fewer alcohol problems at
Time 3 than persons who used Professional Services in addition to AAINA or used
Professional Services Only. Those who did not utilize any services reported the
most alcohol problems of any group. Type of service use was not significantly
related to alcohol problems at Time 4. Although the same patterns of group means
was evident, the between group differences was small and did not reach statistical
significance. Time 3 drug problems showed a similar pattern to that found for Time

3 alcohol problems. Persons who attended AAINA Only had significantly fewer

80

drug problems. Again, the results for Time 4 were in the same direction, but not
robust enough to reach statistical significance. Outcomes in the other domains --
psychological problems, family/social problems, and psychiatric rehospitalization
were not significantly related to type of service use at Time 3. However, there was
a nonsignificant trend for persons attending professional services only to report
more psychological problems and to be more likely to be rehospitalized.

Summary of Preliminary Analyses a Hypotheses Tesg

The preceding analyses indicated that respondents with high levels of
psychiatric symptomatology and multiple prior psychiatric hospitalizations were less
likely to ever attend AAINA during the 10-month follow-up time period, and that if
they did attend, they were less likely to attend consistently. Contran’ly, persons
with relatively few psychiatric symptoms and few prior hospitalizations were more
likely to attend AAINA. While these relationships were significant, they only
accounted for a small percent of the variance in AAINA attendance (approximately
4%). Thus, hypotheses l and II received moderate support.

In terms of the relationship between AAINA service use and outcomes,
persons who attended AAINA regularly after hospital discharge had fewer alcohol
and drug problems at 10-months post-hospitalization. Using AAINA Only seemed
to have the most beneficial effects on functioning. Contrary to expectations,
persons who used Professional Services in addition to AAINA did not show
additional improvement over those who used AAINA Only. In fact, the latter group

had the fewest alcohol and drug problems. AAINA service use was not found to

81

be directly related to psychological problems, family/social problems, or to
psychiatric rehospitalization. In the next phase of the analyses the indirect
relationships with these outcome measures were explored via path analysis.
The Structural Eguation Models

In addition to identifying the relationships between individual characteristics,
service use, and outcomes, a major objective of this study was to determine
whether an hypothesized Integrated Treatment model (Model A) or an AAINA
Treatment model (Model B) was more consonant with the data from this sample of
persons with dual diagnosis. In order to address these two objectives, a series of
24 different path models were tested using EQS Structural Equations Program.
These path models examined the relationships between symptom/chronicity status,
three variations of AAINA and other service use (AAINA attendance pattern over
the 10-month follow-up, the number of days attending AAINA at the 10-month
interview, and the types of services used at the 10-month interview), and outcomes
in five domains: alcohol problems, drug problems, psychological problems,
family/social problems, and rehospitalization. Two types of models were fit to the
data using the same sets of variables: an Integrated Treatment model (Model A) in
Which service use was directly related to each outcome variable (see Figure 3 for
example of a measurement model predicting psychological problems) and an AA
Treatment model in which alcohol and drug problems mediated functioning in the
other domains (see Figure 4). Results for the Integrated Treatment model will be

presented first, followed by the AA Treatment model.

82

 

m.

2330
0332.93.— v on:

 

5/m

 

23$ \m

maeuo—ofihm m 29,—.

 

 

:5 <22< m 255.

 

54 332 2868332 .m 0.35...—

 

38203 gnaw

 

 

3:23 355w

98.53. 5Q

 

 

83

            

1932.05 .. 3:.

2532;
5330—259»: n 06:.

     

«Eu—no.5
mac n 0.55:.

«89395
.2302 m 055.

   

when <Z\<< m 03;

   

_m .332 Eon—2332 .5. 9:53“

Hacioxum am:—

23§a in

92203 33

84

For each model, all variables and predicted paths were entered and tested
via maximum likelihood estimation. Since the goal of these analyses was to reach
a parsimonious model which provided the best fit with the data, the Wald test of
nonsignificant paths was implemented in each analysis. The Wald test is designed
to determine which sets of estimated parameters could be either eliminated or set
to zero without substantial loss in model fit (Bentler, 1995). Using the Wald test
results, the least significant parameters were eliminated from each model until a
parsimonious and interpretable model was attained.

In addition, the goodness of fit of each model was assessed via chi-square
test, the Bentler-Bonnett nonnonned fit index (NNFI), and the comparative fit index
(CF I). These measures of model fit are used somewhat differently in structural
equation modelling. Generally, the chi-square is a test of the null hypothesis that
the coefficients in the model are equal to zero; that is, there is no relationship
between the specified variables. A significant chi-square indicates that the null
model is the true state of affairs, and that the proposed model should be rejected.
The NNFI and CF I are commonly reported indices of incremental model fit, that is,
the degree to which the proposed model is an improvement over the null model.
While some studies utilize the NFI (normed fit index), which is very similar to the
CFI, the NNFI is reported here. In contrast to the CFI (and NFI), the NNFI is
scaled according to the degrees of freedom in the model and is generally a more
sensitive measure of model fit than the CFI or the NFI. The conventional criteria

for good model fit is NNFI and CFI indices of .90 or greater. Given a nonsignificant

85

chi-square, an NNFI and/or CFI of .90 or higher would indicate that the model was
a good fit with the data.
flth Ana_ly§es Results

Tables 12, 13, and 14 present the standardized path coefficients for the
integrated treatment models using number of days attended AAINA at time 3,
AAINA attendance pattern, and Type of service use to predict the five sets of
outcome variables: alcohol problems, drug problems, psychological problems,
family/social problems, and psychiatric rehospitalization. (See Appendix for
diagrams for these path models as well as the mediated AAINA models.) The
direct effects on each outcome variable are presented, as well as the sum of all
indirect effects via the symptom/chronicity groups. Not all indirect paths are
presented in the tables, however, the coefficients for the missing indirect paths may
be calculated from the data in the table. For instance, there was an indirect path
from Time 3 AAINA days to Time 4 alcohol problems, via Time 3 alcohol problems.
This path can be calculated by multiplying the direct effect from Time 3 AAINA days
to Time 3 alcohol problems (.15) by the direct effect from Time 3 alcohol problems
to Time 4 alcohol problems (.61). Thus, the indirect effect of Time 3 AAINA days
on Time 4 alcohol problems is .09. For each model, the fit indices are presented at

the bottom of the table.

. Ame-8395 .282 E. 2 «.8235
.05? n5 Ea: 58 2:. .u. 3 52:2. 058 2: 5 95an n5 3 33an v... 3 5.3.0505 05 E3052 «5:95:80 £3 39:. .

 

86

 

 

5. 55 moo. 358530 53 $5355 .l .3 v m .a 5:85ch E50508 5mm .
um. 2. 8. «V. 8.15
8.- 8.- 8.- No. mm. 1522
Say a $.85 :97 a .835 5:.v a .39: 53 a .853 53 a .53: u 2:9, a .2. 9 "x
IIII-HE .952
I I S. I l 2:983 :9: as
I l 5.- I 3.- ..s25 :91 as
.I l- l l- ! 5:25.96 26.. a5
Illegal-:96:
:3. .8. .3. .8. .6. 9:952: 55
No- 8. B.- 8. mo. .98 <23 3
”float 050
3. I I I No. 2:2:me :9: as
E. - I I I N:- oso<§w :9: as
5. I. -l .l 3. 552me 26.. a5
”III-.985 695:.
8.- 3. 3.- 3.- .m 5.- 98 <23. .55
.985 $95

 

ace-55.833235 mEoEoE oomemm mEoEoE anion mEmBoE 9:0 9.832.... .9582 8.35) 5:09.335

 

59:3 9:850 9:5

 

$8 I s n 9:5 :5 <22< 8292 98 a 585:2 :53 2. 38.2 85985:. ”9.33. .5932 5:: N. 995

Amp—.235 6:82 E. 2 mEoBoE
3:82 wk 89: 58 c5 .63 5an: 2:3 2: :_ 9:032: n... 5 9:0an E. .o 3:298: 9: E3892 3:29:80 53 $9:- .

 

87

 

 

. .5. car: 80— 3508508 53 mousse. I .mo.v m an ENDS-5mm «:035000 San. c

8. 2. 8. we. 8.. u to
am. 8.- B.- 8. 8. u :22
33 a .58: 53 a .882 63 a .8. 8. .8.v a .38; 28¢ a .85. u 8.9: .e: a .x
E .822
S. 5.- 8. S. I 9:285 :9: :5
I I I I I 23sz :9: as
I I 5. I I o_:2:o\xw 33 «5
I185 32...:
.8. .8. .8. .8. .6. .8532: m:
8.- 8. o :- 8. 8. 28:: 8:882 <2)?
.285 625

 

 

3.8.5...) 088430 v 9.5.

 

S. I 5. 8. 8. 9:253 :9: as

I I I I I 2836 :9: as

S. I I 3. S. "9:2:me 26.. as
”Iaoém 825:.

8.- 8. 8.- o F. o :- 288 8:855: <22<
I885 625

:oawufitmozom 2:292“. 00%.»... mEmBoE 55¢ mEofioE 9:0 mEoBoi 3:82 33§a> «59:33:.

Immfimcg 9:850 n 9:...-

 

Ava n S Eaton. 005282 (2)3 5:5 A3 .035. 98958.... ”£33m mmzas‘ 5am .9. aim-r

88

820.02: .882 v... 2 020.02:

.0:8.< n... 22.. £0: 05 :00. 50200 0200 05 5 «20.02: 2. >0 0.20.02: E. :0 88.02: 05 E8052 88.0508 50: 80:... .

 

 

 

. Po. :05 mmo. $030508 500 masofic. I .mo.v d «a uconcmmm «00.05000 50:: c

2. 8. 8. 8. 3.. a _:0
8.- 8.- 8.- 5.- 5.- .. .:22
38¢: .88: 59v: .885 53: 8:8 53: .855 5.7: .288 u 8.2,: .8 a: .0.
I9: .892
5. 5.- 8. 8. 5.- 98.85 :9: :5
5. I 5. 8. I 8.25 :9: :5
8. 5.- 8. 8. I 9:2:05 33 :5
”I880 52.2.
.8. .8. .8. .8. .8. 8.29:2: 2
5. 8.- 8. .3 F. 5. 2:0 802:: 89822:
5. 5. 8.- 8. 8. 892:0 89822: + <<
5. 5. 5.- 8. 5. 2:0 5.
I880 .890
8. 5. 5. 8. 5.- 9:285 :9: :5
5. I 8. 8. 5.- 825 :9: :5
8. I 5. 8. I 9:285 26.. :5
I825 62.2.
.3. 8. 8. .2. 3. 3.- 9.:0 892:: 89822:
3. 5. .3. 8. .8 3.- 892:0. 3:99.22: + S.
8. 8. 8.- 8.- 8.- 2:0 <<
I880 0890
8855.808: 020.02: 80an: 020.02: 85: 020.02: :20 8.0.02: .0:8_< 8.05:...) E00:0:00:.

Im0_I0m:m> 02850 n 02....

 

$8 a s n :2: .: 3: 8.2:: .o 8.; 5.; .5 .892 3252:. ”9.8:: 98:2 5:: .3 9:8

89

Overall, the hypothesized path models did not prove to be good fits with the
data. All of the Chi-squares were large and highly significant, indicating that the
null hypothesis of no relationship between the variables could not be rejected.

While the overall models were not good fits, there was some consistency
with findings from the previous analysis. For instance, persons in the high
symptom/chronic group were significantly less likely to attend AAINA regularly
(B = -.17). They were also significantly more likely to use professional services only
(B = .25), as were persons in the low symptom/chronic group (B = .19). However,
in most cases, symptom/chronicity had very negligible impact on functioning via
service use, as indicated by the low indirect effects.

The previously identified relationships between AAINA service use and
functioning were generally supported with the path analysis results. For instance,
persons who attended AAINA more often at Time 3 had fewer alcohol problems (B
= -.15, see Table 12). Some divergent findings regarding service use and
outcomes also emerged from the path analyses. For instance, regarding type of
service use (Table 14), the path from AAINA only to Time 3 drug problems was not
significant, as would be expected based on previous analysis. However, persons
who used professional services only had significantly more drug problems at Time
3 (B = .17) and Time 4 (B = .11). Also, persons who used AAINA in addition to
professional services had more psychological problems at Time 3 (B = .14). And
those who used professional services only (which were predominantly outpatient

mental health services) were more likely to be rehospitalized at Time 3 (B é .14).

90

The path models confirmed the lack of relationship between service use and Time
4 functioning. None of the direct paths from service use to Time 4 functioning
reached statistical significance. The only relationships identified with Time 4
functioning were indirect paths, via Time 3 functioning. The high correlations
between Time 3 and Time 4 functioning accounted for these relationships.

Tables 15, 16, and 17 show the path analysis results for the ANNA
Treatment models in which Time 3 alcohol and drug problems were mediators of
functioning in the other domains. None of these models were judged to be
significant fits with the data, as indicated by the large chi-squares and low values
on the fit indices. However, there were more significant paths in the AAINA
Treatment models than in the Integrated Treatment models, primarily due to the
concurrent relationships between the various indices of Time 3 functioning. For
instance, Time 3 alcohol problems were significantly related to Time 3
psychological problems. Time 3 drug problems were significantly related to Time 3
psychological, family/social problems, and psychiatric rehospitalization. However,
no significant relationships were identified for Time 4 functioning. In these. path
models, as in the Integrated Treatment models, the indirect effects of
symptom/chronicity were very small or nonexistent, as were the indirect effects of

service use.

91

Table 15. Path Analysis Results: AAINA Treatment Model (B) with Number of Days
Attended A/NA at Time 3 (fl = 284)

 

Time 3 Outcome Variables

 

 

Independent

Variables Psych Problems FamlSoc Problems Rehospitalization
Direct Effects:

T3 Alcohol Problems .15“ .08 .03

T3 Drug Problems .27* 22* .22*

Mirect Effects:
Via T3 AAINA Days -.04 -.02 -.01

Via Low leChronic --- --- ..
Via High Sx/Acute -- ... ..

 

 

 

Via High leChronic -- ... .-
Time 4 Ogtcome Variables

Dfiirect Effects:

T3 Alcohol Problems .09 .05 .05

T3 Drug Problems -.01 -.07 -.03

T3 Problems‘1 .48* .31" .36*

mgirect Effects:

Via T3 ANNA Days -.03 -.01 -.01

Via Low leChronic -- ... ..

Via High Sx/Acute -- ... . ..

Via High leChronic -- ... ..

Model Fit:

x2 (18 df), p value = 189.49, p <.001 160.39, p <.001 177.63, p <.001

NNFI = .09 -.14 -.14
CF I = .42 .27 .27

 

* Path coefficient significant at g <.05.

-- Indicated path coefficients less than .01

' These path coefficients represent the prediction of T4 problems by T3 problems in
the same domain (e.g., the path from T3 Alcohol problems to T4 Alcohol problems).

92

Table 16. Path Analysis Results: AAINA Treatment Model (B) with AAINA
Attendance Pattern (g = 284)

 

Time 3 Outcome Variables

 

 

 

Independent

Variables Psych Problems FamlSoc Problems Rehospitalization
QLreCLEffects:

T3 Alcohol Problems .15* .08 .03
T3 Drug Problems .27“ .22" .22*
lrlqirect Effects:

Via AAINA -.04 -.03 -.02
Attendance

Via Low Sx/Chronic -- -- --
Via High Sx/Acute —- -- --
Via High Sx/Chronic .01 .01 .01

Time 4 Outcome Variables
Direct Effects:

T3 Alcohol Problems .09 .05 .05 .

T3 Drug Problems -.01 -.07 -.03

T3 Problems' .48* .31" .36*

Indirect Effects:

Via AAINA -.03 -.01 -.01

Attendance

Via Low Sx/Chronic -- --- --

Via High Sx/Acute -- -- --

Via High Sx/Chronic .01 -- --

M9131;

x2 (18 df), p value = 189.27, p <.001 158.40, p <.001 175.27, p <.001

NNFI = .09 -.14 -.14
CFI = .41 .26 .27

 

* Path coefficient significant at g <.05.

-- lndicates path coefficients less than .01.

' These path coefficients represent the prediction of T4 problems by T3 problems in
the same domain (e.g., the path from T3 Alcohol problems to T4 Alcohol problems).

93

Table 17. Path Analysis Results: AAINA Treatment Model (B) with Type of Service
Use at Time 3 (Q = 284)

 

Time 3 (Artcome Variables

 

 

 

 

Independent

Variables Psych Problems FamlSoc Problems Rehospitalization

Direct Effects:

T3 Alcohol Problems .16* .08 .03

T3 Drug Problems .27" .22* .22"

Indirect Effects:

Via AAINA Only .01 .04 -.02

Via AAINA .05 .03 .02
+ Prof Services

Via Prof Services Only .07 .05 .04

Via Low Sx/Chronic .01 .02 .01

Via High Sx/Acute .01 -- .01

Via High SxIChronic .01 .02 .01

Time 4 Outcome Variables

Direct Effects:

T3 Alcohol Problems .09 .05 .05

T3 Drug Problems -.01 -.07 -.03

Time 3 Problems' .48* .31" .36*

Indirect Effects:

Via AAINA Only .02 .01 -.02

Via AAINA .03 - --
+ Prof Services - -

Via Prof Services Only .04 -- --

Via Low SxIChronic .01 -- --

Via High leAcute -- -- --

Via High Sx/Chronic .01 -- --

Model Fit:

x2 (25 df), p value = 276.94, p <.001 244.89, p <.001 262.69, p <.001

NNFI = -.14 -.34 -.34
CFI = .37 .26 .26

 

* Path coefficient significant at p <.05.

-- Indicates path coefficients less than .01.

' These path coefficients represent the prediction of T4 problems by T3 problems in
the same domain (e.g., the path from T3 Alcohol problems to T4 Alcohol problems).

94

Overall, these results indicate that the full models were very poor fits with
the data. Using the Wald test of nonsignificant paths, an attempt was made to
reach a parsimonious and well-fitted model with each set of variables. All
nonsignificant paths were eliminated from each model until an interpretable path
that met the criteria for model fit could be identified. These methods resulted in
two models that fit the data. Figure 5 shows the significant model predicting
psychological functioning, trimmed of all nonsignificant paths. (The full model with
all paths shown can be found in Appendix A.) This model showed acceptable fit
with the data - x2 (3 df) = 3.49, 9 =32; NNF I = .97; CFI = .99, indicating that it was
a plausible model. Persons who attended AAINA more often had fewer alcohol
problems at Time 3, and fewer psychological problems. Those with fewer
psychological problems at Time 3 also had fewer psychological problems at Time 4.

Symptom/Chronicity was not a significant predictor in this model. Although this
model met the criteria for acceptable fit, it only accounted for a small amount of the
variance in Time 3 alcohol problems (32 = .01) and Time 3 psychological problems
(_R_2 = .03), and a moderate amount of the variance in Time 4 psychological
problems (32 = .13).

The second model that emerged is shown in Figure 6. Persons in the high
symptom/chronic group were more likely to use professional services only, which
was related to experiencing more drug problems at Time 3 and more family/social
problems (x2 (6 df) = 8.06, 9 =23; NNFI = .94; CFI = .97). Similar to the previous

model, this trimmed model only accounted for a small percent of the variance in

95

use of Professional services (32 = .02), Time 3 drug problems (32 = .01), Time 3
family/social (i32 = .03), and Time 4 family/social problems (32 = .05). Interestingly,
use of AAINA only did not emerge as a significant predictor in the trimmed model,

as in previous analyses.

96

ha.

\

 

 

3.038."
133235 e 3:.

can.

 

 

.8030;
9.5 m 3:.

 

«6038.—
193235 n 3:. m
«MN.

 

\

so.

«803°;
3483‘ m 08C.
ma.

nausea: 333a... 3.3.3:. 5: 329:3 .m 83E

 

38:63 .3:

m :5 <2)? m 25w m 2923 SEW
A

 

 

 

 

38205 33

 

97

\

3038..
Enigma: a 2......

 

 

can.

nae—noun
30835". n as;

5a.

 

ma.

wn.

aa.

a

 

2.8395
9.5 m cam...

 

N

«80305
_2—oo—< m 08E.

w

333.5 30835.“ 3.33: a: .8055 s 2:5

am.

/

2.5 fl
«cerium idemaouoa fleeiufim gnaw
O
m 2:03am anew

masoaohm ad

 

 

 

 

 

«08.68
13.825 + <35.

m as 2.2 w

 

 

 

98

filmmarv of Research Questions and Findings

(1) Are the rates of AAINA participation among persons with dual
diagnoses comparable to those without any diagnosed mental illness?

Yes. The self-reported rates of AA/NA attendance in this sample were
Comparable to those reported in the literature on persons with substance abuse
disorders.

(2) Are persons with severe and chronic mental illness likely to attend
AAINA, and if so, how consistent will their attendance be?

Participants with severe and chronic psychiatric symptoms were
significantly less likely to ever attend AA/NA after hospital discharge. They were
also less likely to attend AAINA consistently throughout the 1 0-month follow-up
period. Participants with fewer psychiatric symptoms of acute duration reported
the most consistent AA/NA attendance.

(3) Are patterns of AAINA attendance related to alcohol/dmg use and
other indices of community functioning? .

Persons who attended AA/NA consistently over the follow-up period (e. g.,
at all 3 interview timepoints) reported the best outcomes. They had significantly
fewer alcohol problems. No significant findings were observed for AA/NA
attendance and other functioning domains, although the direction of results for
drug and psychological problems and psychiatric hospitalization were in the
same direction. The latter findings were not robust enough to reach statistical

significance.

99

(4) Do persons who receive professional services in addition to AAINA
show better outcomes than those who attend AAINA only?

No. Use of professional services was not related to better outcomes than
attendance at AA/NA only. In fact, persons who attended AA/NA only reported
the best outcomes in terms of alcohol and drug abuse. Those who did not
attend any services reported the most problems of any group. There were no
significant difference in other functioning domains based on type of service use.

(5) Are the mechanisms of change in this sample more consonant with an
integrated treatment model or an AAINA model?

Findings from the path analyses suggested that the AA/NA treatment model
was more consonant with the data. There were no direct relationships between
AAINA attendance or other service use and psychological and family/social
problems or psychiatric rehospitalization. However, persons who attended AA/NA
and reduced their alcohol use also reported significantly better psychological
problems. Similariy, there was a mediated relationship between use of

professional services, drug problems, and family/social problems.

Chapter 4

Discussion

This study sought to answer several questions related to AAINA attendance
and outcomes among dually diagnosed individuals. While there has been a
general expectation that self-help groups such as AAINA would not be appropriate
or helpful for persons with the characteristics of this study's respondents, there
was almost no empirical literature to validate this assumption. The current study is
a step towards building an empirical data base on the efficacy of self-help groups
for persons with dual diagnoses.

The first purpose of this study was to provide descriptive information on the
levels of AAINA participation in this sample, and to determine if persons with dual
diagnoses attended AAINA at different rates from persons with singly diagnosed
substance abuse disorders. Previous investigators have indicated that of persons
referred to AA following inpatient substance abuse treatment, 60-80% will attend
(Edwards et al., 1971; Knouse & Schneider, 1987). In the current study, 62% of
the respondents attended AAINA at some time during the 10-month follow-up time
period, which is comparable to the previously reported rates of attendance among

persons with single diagnoses. Pettinati, Sugerman, DiDonato, and Maurer (1982)

100

101

reported AAINA attendance rates of 74% in a dually diagnosed sample.

It is somewhat more difficult to compare gt_e_§ or patterns of attendance
across studies due to the wide divergence in definitions of “regular” vs. “irregular'
attendance. However, the finding that 24% of the respondents in this study
attended AAINA at least once every four months over the 10-month follow-up is
comparable to rates reported in the literature (Belasco, 1971; Emrick, 1987;
Pettinati, Sugerrnan, DiDonato, & Maurer, 1982; Temsovic, 1970). At 10-months
post-hospitalization, 7% of the respondents were attending AAINA two times per
week or more. These findings are rather exceptional given the fact that most prior
studies have been conducted with socially stable Caucasian males with substance
abuse problems only. These data suggest that low-functioning, dually diagnosed,
African American males are equally likely to attend AAINA.

Relationshi between Individual haracteristics and Service Use

Several studies have attempted to identify characteristics that differentiate
persons who will attend AAINA from those who will not. This study examined
several individual characteristics thought to be related to AAINA attendance:
exposure to AAINA treatment models via specialized dual diagnosis treatment,
prior AAINA attendance, plans to maintain sobriety, psychiatric symptomatology at
hospital entry, and the number of prior psychiatric hospitalizations. Of the
covariates used in this study, only prior AAINA attendance was significantly related
to AAINA attendance during the follow-up time period. Participation in the

specialized dual diagnosis treatment and motivation for sobriety (as measured by

102

stated plans to engage in treatment and other activities to maintain sobriety) were
not significant covariates in the models predicting AINA attendance.

The researcher was particularly interested in discovering whether persons
with the most severe and chronic mental illness would show different rates of
involvement with AAINA than others. Due to the heterogeneity among persons
with dual diagnosis, different characteristics and experiences might lead to different
outcomes within the same sample. Prior studies indicated: (1) that persons with
mental illness would be more difficult to engage in substance abuse treatment (of.
Liss, 1979; Osher 8 Kofoed, 1989; Panepinto, Higgins, Keane-Dawes, & Smith,
1970) and (2) psychiatric severity was a strong predictor of outcomes in other
functioning domains (cf. Gibbs & Flanagan, 1977; McLellan et al., 1983).

Findings from this study partially corroborate Osher and Kofoed’s (1989)
expectation that persons with severe psychiatric problems would be more difficult
to engage in substance abuse treatment. In the current study, participants
characterized as high symptom/chronic group were less likely to attend AAINA
during the 10-month follow-up, whereas those with fewer symptoms and more
acute problems were more likely to attend AAINA and to attend more regularly.
Persons with more chronic and/or severe psychiatric problems were more likely to
use professional services only, which were primarily outpatient mental health
services.

While symptom/chronicity was a significant determinant of AAINA

attendance, it only accounted for a modest percent of the variance (approximately

103

4.0%). In the path models predicting outcomes from AAINA attendance and
symptom/chronicity, symptom/chronicity was no longer a significant predictor. The
indirect relationships between symptom/chronicity and functioning at Time 3 were
very low or nonexistent. These findings suggest that psychiatric severity and
chronicity have some influence on who is likely to attend AAINA. However, the
data from the path models indicate that once a connection is made to the self-help
group, the severity and chronicity of mental illness are not related to outcomes
derived from AAINA participation. Regardless of individual characteristics,
persons who were able to attend AAINA more often had fewer alcohol and drug
problems, and fewer problems in other life domains. Thus, attendance at AAINA
seems to be the more important determinant of outcomes, even for persons with
severe and chronic mental illness.
Relationship of Service Use to Outcomes

Several hypotheses regarding service use and outcome were examined.
First, it was hypothesized that persons who attended AAINA regularly after
hospitalization would have better outcomes than those who did not attend or who
attended sporadically. This hypothesis was supported, but only for alcohol and
drug use at Time 3. Persons who attended AAINA had fewer alcohol and drug
problems. AAINA attendance was not directly related to better psychological or
family/social functioning, nor was it related to the chances of being rehospitalized.

This is an important finding, particularly in a dual diagnosis sample. Most of

the empirical literature on AAINA attendance and outcomes have been conducted

104

with samples of individuals with singly diagnosed substance abuse problems.
Persons with dual diagnoses experience functional deficits in multiple life domains,
most notably in their psychological functioning. Whereas previous studies of
AAINA have typically utilized alcohol and/or drug use as the only outcome.
measures, research with dually diagnosed individuals must incorporate outcomes
in other domains in order to reasonably determine the efficacy of a particular
intervention. The recovery of persons with dual diagnosis should minimally include
improvements in psychological functioning as well as decreases in alcohol and
drug use.

It is also interesting to note that AAINA attendance was only related to
concurrent alcohol and drug problems, but not to prospective problems. The lack
of relationship between AAINA attendance and latfi' alcohol or drug problems
precludes any causal inferences about the relationship between AAIl\lA service use
and sobriety. While it is tempting to infer that AAINA attendance leads to'sobriety,
it is equally plausible that persons who already committed to maintaining sobriety
gravitate towards AA to support their sobriety (Toumier, 1979). Persons who are
actively drinking may not be as motivated to attend AAINA. Previous studies with
the current data suggest that motivation (as measured by plans to maintain
sobriety) was an important factor related to AAINA attendance and outcome.
BootsMiller et al. (1996) found that persons who were more motivated for sobriety
when they were discharged from the hospital were subsequently more likely to

attend AAINA, and to attain better outcomes.

105

A second important question addressed in this study was the extent to
which the benefits of AAINA participation could be augmented with simultaneous
involvement in professional services, particulariy those where mental health issues
were primarily addressed. Previous authors had suggested that simultaneous
involvement in other forms of treatment could obscure the effects of AAINA
participation (Emrick, 1987; Thurstin, Alfano, & Nerviano, 1987). Additionally, it
was expected that professional treatment programs might be able to address the
specific mental health needs (e.g., psychotherapy, medication review) of dually
diagnosed persons that would not otherwise be met by attendance to ANNA only.

The findings regarding types of service use were contrary to what was
expected. Persons who attended AAINA only had the best outcomes. Use of
professional services was not related to improvements in any of the functioning
domains. In fact, there was some evidence suggesting that persons who used
professional services had more problems than those who attended AAINA only.
There was a nonsignificant trend for persons attending professional services to
report more psychological problems and to be more likely to be rehospitalized. A
significant path was observed between use of professional services and worse
drug problems. Again, no causal inferences can be drawn from these data as the
relationships were observed at a concurrent time point. It is quite plausible that
persons with more severe problems were more likely to use professional services,
although use of these services did not seem to impact their subsequent functioning

(professional service use at Time 3 was not related to improvements in Time 4

106

functioning).
Findings and Implications of the Mediated AAINA Treatment Model

The results from the structural equation models indicated that there was
generally more support for an AAINA treatment model than for the Integrated
model. AAINA attendance was not directly related to psychological or family/social
functioning, nor to the incidence of rehospitalization. However, there was evidence
for a mediated relationship between AAINA attendance and the other outcomes via
alcohol/drug use. As suggested by AAINA program theory, individuals who were
able to attain sobriety or reduce their alcohol and drug use had fewer problems in
the other life domains. This finding was particularly evident in the significant
pathway between the number of days of AAINA attendance, Time 3 alcohol
problems, and Time 3 psychological problems. Persons who attended AAINA had
fewer alcohol problems at Time 3; those with fewer alcohol problems also had
fewer psychological problems at Time 3.

This finding tends to support prior reports indicating that attainment of
abstinence may moderate the relationship between AA attendance and
psychological functioning (Akerlind, Hamquist, Elton, & Bjurulf, 1990; Laundergan,
1992). However, this interpretationis made with caution. Given the correlations
between the various outcome measures at the same time point (rs range from .07
to .38, with a mean correlation of .24), the significant paths between alcohbl and
drug problems and other functioning domains simply suggest that persons with

alcohol and drug problems are also likely to have problems in other domains.

107

The findings of this study also indicated that the “dosage” of AAINA was
significantly related to outcome. The more frequently person attended AAINA, the
better their outcomes were. This is consistent with prior studies (Emrick, 1987;
Knouse & Schneider, 1987; McBride, 1991; Thurstin, Alfano, & Nerviano, 1987;
Trice & Roman, 1970). In fact, the magnitude of the effect ((3 = -.16) is comparable
to that reported in Emrick, Tonigan, Montgomery, and Little’s (1992) meta-analytic
review of quantitative studies on AA. These authors reported an average
correlation of .19 between AA attendance and drinking outcomes.

Limitations of the Current Stu_d_y

This study was an improvement over previous investigations in some
respects: providing data on a sample of African American males with dual
diagnosis, incorporating AAINA program theory into the evaluation, examining other
types of service use in addition to AAINA, and assessing outcomes in domains
other than alcohol/drug use. Nevertheless, several limitations of this study warrant
discussion. First, this study, like prior reports, relied on self-report data. Previous
reports have indicated that self-report of sensitive information, such as substance
abuse, is biased by under-reporting (Mensch & Kandel, 1988; Polich, 1982; Rouse
8. Kozel, 1985). Also, some authors have expressed concern over the reliance on
self-report measures in samples with cognitive or psychological deficits (Drake et
al,1990)

In addition, some of the key variables used in this study, such as AAINA

attendance suffer from selection bias. Due to the fact that participants self-

108

selected to attend AAINA or other services after hospitalization, there is no way to
determine conclusively whether persons who attended were different in important
(and unmeasured) ways from those who did not attend, and whether these
unmeasured differences may have accounted for their outcomes. This is a
particular difficulty in conducting research on self-help participation. In order to
determine if use of particular services are differentially effective, randomized
designs with control groups are necessary. However, prior research indicates that
individuals must be personally motivated towards program engagement in order for
participation in self-help groups to have any desirable impact. Persons who are
randomly assigned or mandated to attend AAINA often do not become sufficiently
engaged, and therefore do not receive any benefits from attendance (cf. Ditman,
Crawford, Forgy, & Maskowitz, 1967; Edwards et al., 1988; Walsh et al., 1991).
The issue of selection bias is an on—going dilemma in research on self-help groups.
Perhaps the most serious weakness in this study was the insensitivity of the
measures of AAINA participation. The findings of this study are based on
noncontinuous data collected at 2-4 month intervals. In order to determine the
efficacy of AAINA, more direct and specific measures of engagement are needed.
For instance, prior studies indicate that the quality of AAINA participation may be
as important or more important than the quantity of attendance. Maxwell (1984)
indicated that in order for persons to derive benefit from AA, they must engage
themselves in the fellowship aspects of recovery. This fellowship includes forming

relationships with sponsors and other group members, giving testimonials about

109

their own recovery, and “participating in open, honest sharing, and caring
interaction with fellow members” (Maxwell, 1984, p. 155). The current study did not
assess these qualitative aspects of AAINA involvement. Additionally, engagement
in 12-step work and self-realization is an important aspect of AAINA involvement
which was not measured in this study. Without these measures, it is impossible to
determine how well participants in this study were actually “working" the AA
program.
Conclusions & Implications

Despite its limitations, this study suggests some interesting implications for
mental health policy and future research. The participants in this study with the
most severe psychiatric problems and the longest histories of psychiatric
hospitalization were less likely to engage in self-help programs following
ho'spitalization, but were more likely to continue using mental health services. Use
of mental health services seemed to have some positive relationship with alcohol
problems, but the relationship was very small and there was no positive relationship
with drug problems or problems in any of the other functioning domains. These
findings indicate a disturbing trend, suggesting that persons with chronic and
severe psychiatric problems were not receiving much support for their sobriety
once they were discharged from the hospital. Given the fact that most of the
participants in this study were not receiving outpatient substance abuse services,
nor was the use of outpatient mental health services significantly related to

improvements in alcohol and drug use, AAINA attendance seemed to be the only

110

service which was useful in helping participants continue to receive support for and
maintain their sobriety.

More research is needed to determine the impact of professional mental
health services when used in combination with AAINA. In clinical practice, persons
with substance abuse problems are often encouraged to seek individual counseling
while attending self-help groups. It is evident that persons with dual diagnosis
need to be connected with substance abuse and/or dual diagnosis networks in
order to maintain their sobriety. This study suggests that this combination of
services is more beneficial than no service use, at least for improvement in alcohol
problems. Therefore, mental health professionals may serve an important role in
helping persons with substance abuse and/or psychological problems to access
and benefit from self-help groups.

For persons with long histories of mental health problems and psychiatric
hospitalization, the mental health services network may be a more familiar and
easily assessable resource. Mental health service providers could be instrumental
in implementing linkage programs to help those who would not otherwise become
engaged with self-help and sobriety support programs. If persons with severe and
chronic mental illness are helped in the process of engagement, this study’s
findings suggest that they will receive the benefits of sobriety (or at least be able to
reduce their alcohol and drug use). Prior work by Minkoff (1989) suggests that with
special preparation prior to hospital discharge, persons with dual diagnosis were

able to successfully engage in AAINA groups. Clients receiving dual diagnosis

1 1 1
treatment. were prepared for AAINA attendance via a structured program including
social skills training, linkage with other AAINA members, and trial attendance at
selected meetings thought to be most suitable for persons with dual diagnosis.
The work of Minkoff (1989) and others is promising, however, there is a lack of
empirical data to document the efficacy of these program initiatives.
Mm Directions

The findings of this study suggest several promising avenues for future
research on AAINA attendance among persons with dual diagnosis. A few prior
studies have suggested that attainment of sobriety moderates the relationship
between AAINA attendance and psychological functioning. This study partly
corroborated prior efforts, however, the findings were only identified cross-
sectionally. No causal inferences could be made based on findings from the
current data. There is a need for prospective longitudinal studies to explicate the
relationship between psychological distress and substance abuse, and how these
are related to AAINA attendance.

While the current study indicated that persons who attended AAINA more
regularly attained better outcomes, the specific aspects of the AAINA program
related to positive outcome could not be identified. Future research should
address the following questions: Which aspects of the AAINA program are most
helpful to persons with dual diagnosis? In what ways do persons with dual

diagnosis ”work" the AAINA program? And what intrapersonal and group-related

factors are related to their successful engagement in the program?

LIST OF REFERENCES

LIST OF REFERENCES

Alcoholics Anonymous (1952). Twelve steps and twelve tra_dition§. NY:
Alcoholics Anonymous World Services.

Alcoholics Anonymous (1975). Living sober. NY: Alcoholics Anonymous
World Services.

Alcoholics Anonymous (1976). flmholiwjnonvmous: The story of how
many thousands of men and women have recovered: from alcoholism (3rd Ed.).
NY: Alcoholics Anonymous Worid Services.

Alcoholics Anonymous (1989). Apalvsis cflhe 1986 member survey.
(Available from General Service Office, 468 Park Ave., New York, NY. 10163).

Akerlind, l., Homquist, J. 0., Elton, M., 8 Bjurulf, P. (1990). Overall
functioning and criteria of progress in rehabilitation of alcohol abusers: Longitudinal
analyses of change. Alcoholism: Clinical and Experimental Research, 14(6), 856-
862.

Alford, G. S. (1980). Alcoholics Anonymous: An empirical outcome study.
A_d_<_lictive Behaviors, 5, 359-370.

Bartalos, M. K. (1992). Illness, professional caregivers, and self-helpers. In
A. H. Katz et al. (Eds) Self-help: Concepts and applications. Philadelphia: The
Charles Press, Publishers.

Bateman, N. l. & Petersen, D. M. (1971). Variables related to outcome of
treatment for hospitalized alcoholics. lntemational Journal of the Addictions, 6(2),
215-224.

Bebbington, P. E. (1976). The efficacy of Alcoholics Anonymous: The
elusiveness of hard data. British Journal of Psychiatpz, 128, 572-580.

Belasco, J. A. (1971). The criterion question revisited. British Journal of
Addiction, §_6_, 39-44.

112

113

Bentler, P. M. (1995). EQS structural eguations program manual. Encino,
CA: Multivariate Software, Inc. '

BootsMiller, B. J., Jordan, L. C., Davidson, W. S., 8 Herman, S. E.,(1996).
An evaluation of a specialized dual disorder treatment tomersons with mental
illness and substance abuse: flects of tr_eatment on service use and outcomes.
Unpublished doctoral dissertation, Michigan State University, East Lansing.

Borman, L. D. (1992). Introduction: Self-helplmutual aid groups in strategies
for health. In A. H. Katz et al. (Eds.) Self-help: Concepts and applications.
Philadelphia: The Charles Press, Publishers. ~

Boscarino, J. (1980). Factors related to "stable" and "unstable" affiliation
with Alcoholics Anonymous. lntemational Journal of the Adaictiona, 15(6), 839-848.

Boume, P. G. 8 Fox, R. (1973). Alcoholism: Progress in research and
treatment. New York: Academic Press.

Bradley, A. M. (1988). Keep coming back: A case for a valuation of
Alcoholics Anonymous. Alcohol Health and Research Worig, 12(3), 192-199.

Bromet, E.: Moos, R.; Bliss, F.; 8 Wuthmann, C. (1977). Posttreatment
functioning of alcoholic patients: Its relation to program participation. Journal of
Consulting and Cliniaal Psychology, 45(5), 829-842.

Brown, B. 8.; Joe, G. W.; 8 Thompson, P. (1985). Minority group status
and treatment retention. The lntemational Jourpal of the Ad_dictions. 2_0, 319-335.

Buxton, M. E., Smith, D. E., 8 Seymour, R. B. (1987). Spirituality and other
points of resistance to the 12-step recovery process. Journal of Psychoactive
Drugs, 19(3), 275-286.

Canter, F. M. (1966). Personality factors related to participation in treatment
by hospitalized alcoholics. Journal of Clinifil Psychology, 22(2), 114-116.

 

Chacko, R. C., Adams, G. L., 8 Gomez, E. (1985). The care of the chronic
mental patients: A historical perspective. In R. C. Chacko (Ed.), The chronic mental
patient in a community context. Washington, D. C.: American Psychiatric Press.

Chen, H. (1990). Theory-driven evaluations. Newbury Park: Sage.

Denzin, N. K. (1987). The recovering alcoholic. Newbury Park: Sage
Publications. ‘

114

Derogatis, L., Lipman, R. 8 Covi L. (1973). The SOL-90: An outpatient rating
scale. PsvchoLhamcology Bulletin, _9_, 13-29.

DiClemente, (1993). Alcoholics Anonymous and the structure of change. In
B. S. McCrady 8 W. R. Miller (Eds) Research on Alcoholics Anonymotg
Opportunities and alternatives. New answick: Rutgers Center of Alcohol Studies.

Ditman, K. 8.; Crawford, G. G.; Forgy, E. W.; Moskowitz, H.; 8 MacAndrew,
C. (1967). A controlled experiment on the use of court probations for dmnk arrests.
American Journal of Psychiatpy, 124(2), 161-163.

Drake, R. E., Osher, F. c., Noordsy, o. L., Hurlbut, s. c., Teague, G. B. &
Beaudett, M. S. (1990). Diagnosis of alcohol use disorders in schizophrenia.

Schizophrenia Bulletin, 1_6_, 5767.

Edwards, 6.; Orford, J.; Egert, S.; Guthrie, S.; Hawker, A.; Hensman, C.;
Mitcheson, M.; Oppenheimer, E.; 8 Taylor, C. (1977). Alcoholism: A controlled trial
of "treatment" and "advice." Joumal of Sttflies on Alcohol. 38(5), 1004-1031.

Elal-Lawrence, G.; Slade, P. D.; 8 Dewey, M. E. (1987). Treatment and
follow-up variables discriminating abstainers, controlled drinkers and relapsers.
Journal of Stugies on Alcohol. 48(1). 39-46.

Elal-Lawrence, G.; Slade, P. D.; 8 Dewey, M. E. (1986). Predictors of
outcome type in treated problem drinkers. Journal of Stu_dies on Alcohol. 47(1). 41-
47.

Emener, W. G. 8 Dickman, J. F. (1992). Significant demographic
characteristics of persons recovering from alcoholism and other drugs. Journal of
Rehabilitation Counseling, 23(1), 3-17. '

Emrick, C. D. (1987). Alcoholics Anonymous: Affiliation process and
effectiveness as treatment. Alcoholism: Clinical and Experimental Reseaycfl, 1_1,
416-423.

Emrick C. D. (1989). Alcoholics Anonymous: Emerging concepts. In M.
Galanter (Ed) Recent gevelopments in alcoholism. Vol. 7. New York: Plenum
Press.

Emrick, C. D., Tonigan, J. 8., Montgomery, H., 8 Little, L. (1993). Alcoholics
Anonymous: What is currently known? In B. S. McCrady 8 W. R. Miller (Eds.)
Research on Alcoholics Anonymous: Opportunities and alternatives (pp. 41-76).
New Brunswick: Rutgers Center of Alcohol Studies.

115

Fainrveather, G. W., Sanders, D. H., Maynard, H., Cressler, D. L., 8 Bleck,
D. S. (1969). Community life for the mentafl ill: An altema_tive to institutional care.
Chicago: Aldine Publishing Company.

Faulkner, W.; Sandage, D.; 8 Maguire, B. (1988). The disease concept of
alcoholism: The persistence of an outmoded scientific paradigm. Deviant Behavior,
_9_, 317-332.

Francis, D. J., Fletcher, J. M., Stuebing, K K., Davidson, K. C., 8
Thompson, NM. (1991). Analysis of change: Modeling individual growth. Journal of
Consulting and Clinical Psychology, 59(1), 27-37.

Gibbs 8 Flanagan (1977). Prognostic indicators of alcoholism treatment
outcome. lntemational Journal of the Ad_dictions. 12(8), 1097-1141.

Glaser, F. B., 8 Ogbome, A. C. (1982). Does A.A. really work? British
Journal of Acfiiction. fl, 123-129.

Gregson, R. A. M. 8 Taylor, G. M. (1977). Prediction of relapse in men
alcoholics. Journal of Strflies on Alcohol. 38(9), 1749-1760.

Haberrnan, P. W. (1966). Factors related to increased sobriety in group
therapy with alcoholics. Mmal of Clinical Psychology, 22(2), 229-235.

Hanson, B. (1985). Drug treatment effectiveness: The case of racial
minorities in America - Some research questions and proposals. The lntemational
Journal of the Addictions, _2_Q, 99-137.

Heather, N.; Rollnick, S.; 8 WInton, M. (1983). A comparison of objective
measures of alcohol dependence as predictors of relapse following treatment.
British Journal of Clinical Psychology, _2_2, 11-17.

Hedrick, H. L., lsenberg, D. H., 8 Martini, C. J. M. (1992). Self-help groups:
Empowerment through policy and partnerships. In A. H. Katz et al. (Eds.) Self-
help: Concepts and applications. Philadelphia: The Charies Press, Publishers.

Heller, K. (1979). The effects of social support: prevention and treatment
implications. In A. P. Goldstein 8 F. H. Kanfer (Eds.), Maximizing treatment gains:
Transfer enhancement in psychotherapy (pp. 87-103. NY: Academic Press.

Heller, K 8 Swindle, R. W. (1983). Social networks, perceived social
support, and coping with stress. In R. D. Felner etal. (Eds. ) Preventive psychology:
Theory research and practice. NY: Pergamon.

116

Hellerstein, D. J. 8 Meehan, B. (1987). Outpatient group therapy for
schizophrenic substance abusers. American Journal of Psychiatpy, fl, 1337-1339.

Herman, S. E., BootsMiller, B. J., Jordan, L. C., Mowbray, C. T., Brown, W.
G., Deiz, N., Bandla, H., Solomon, M., 8 Green, P. (1996). Immediate outcomes of
substance use treatment within a state psychiatric hospital. Manuscript submitted
for publication.

Hoffman, N. G. 8 Overall, P. B. (1978). Factor structure of the SCL-90 in a

psychiatric population. Journal of Consulting and Clinical Psychology, 46(6), 1187-
1 191 .

Hoffman, N. G.; Harrisbn, P. A.; 8 Belille, C. A. (1983). Alcoholics
Anonymous after treatment: Attendance and abstinence. The lntemational Journal
of the Ad_dictions. 18(3), 311-318.

Howell, D. C. (1987). %tisticamethofids for psychology (Second Edition).
Boston: Duxbury Press.

Humphreys, K. 8 Woods, M. D. (1993). Researching mutual help group
participation in a segregation society. The Journal of Applied Behavioaa! Sciences,
29(2), 181-201.

Institute of Medicine (1990). Broadening the base of treatment for alcohol
problems. Washington, DC: National Academy Press.

Kanas (1982). Alcoholism and group psychotherapy. In E. M. Pattison 8 E.
Kaufman (Eds) Encyclopedic handbook of alcoholism (pp. 1011-1021). NY:
Gardner Press.

Kay, S. R., Kalathara, M., 8 Meinzer, A. E. (1989). Diagnostic and
behavioral characteristics of psychiatric patients who abuse substances.
Hospital and Community Psychiatg, 19, 1062-1064.

Kazdin, A. E. 8 leon, T. (1978). Criteria for evaluating psychotherapy.
Archives of General Psychiatry 35 407-415.

Kenny, D. A. (1979). Correlation and causality. NY: WIIey Press.

King, L.M. (1983). Alcoholism: Studies regarding Black Americans- 1977-
1980. In T.D. Watts 8 R. Wright (Eds), Black alcoholism: Toward a
pomprehensive understanding. Springfield: Charles C. Thomas Publisher.

Knight, B., Wollert, R. W., Levy, L. H., Frame, C. L., 8 Padgett, V. P.

117

(1980). Self-help groups: The members’ perspectives. American Journal of
Community Psycholouy, a, 53-65.

Knouse, V. H. 8 Schneider, H. G. (1987). Recovering alcoholics: Personality
and aftercare factors. Psychological Reports, §_1_, 595-601.

Korte, C. (1978). Helpfulness in the urban environment. In A. Baum, J. E.
Singer, 8 S. Valins (Eds), Advances in environmentaipsycholomr. Vol 1. Tha
urban environment. Hillsdale, NJ: Lawrence Eribaum.

Kurtz, E. (1993). Research on Alcoholics Anonymous: The historical
context. In B. S. McCrady 8 W. R. Miller (Eds), Research on Alcoholics
Anonymous: Opportunities and alternatives. New Brunswick: Rutgers Center of
Alcohol Studies.

Laundergan, J. C. (1993). How could studies of Alcoholics Anonymous be
designed? Evaluation within treatment contexts. In B. S. McCrady 8 W. R. Miller
(Eds) Research on Alcoholics Anonymous: Opportunities and altematives. New
Brunswick: Rutgers Center of Alcohol Studies.

Leiberrnan, M. A. 8 Borman, L. D. (1979). Self-help groups for coping with

crisis. San Francisco: Jossey-Bass Publishers.

Leiberrnan, M. A. 8 Snowden, L. R. (1993). Problems in assessing
prevalence and membership characteristics of self-help group participants. Iaa
Journal of Applied Behavioral Scienc_e, 29(2), 166-180.

Liss, R. A. (1979). The need for homogeneity and consistency. In D. J.
Ottenberg, J.F. X. Carroll, 8 C. Bolognese (Eds) Treating mixeupsychotic drug-
addicted and alcoholic Etienta. Eagleville, PA: Eagleville Hospital and
Rehabilitation Center.

Machell, D. F. (1989). Alcoholics Anonymous: A wonderful medication with
some possible side effects. Journal of Alcohol and Drug Education, 34(3), 80-84.

 

Machell, D. F. (1992). A psychological rationale in support of the Alcoholics

Anonymous' concept of fellowship. Journal of Alcohol and Drug Education, 37(3),
1-16.

Maisto, S. A., O'Farrell, T. J., McKay, J. R., Connors, G. J., 8 Pelcovits, M.
(1989). Factors in maintaining sobriety following alcohol treatment. Alcoholism

Treatment uarterl ,6(3I4), 143-150.
Maxwell, M. A. (1984). The Alcoholics Anonymous exparience. NY:

118

McGraw-Hill Book Company.

McBride, J. L. (1991). Abstinence among members of Alcoholics
Anonymous. Alcoholism Treatment Quarterly. 8(1), 113-121.

McCrady, B. S. 8 Irvine, S. (1989). Self-help Groups. In R. K. Hester 8 W.
R. Miller (Eds), Handbook of alcoholism treatment approacheszuEffectjya
alternatives. (pp. 153-169). NY: Pergamon Press.

McLatchie, B. H 8 Lomp, K. G. E. (1988). Alcoholics Anonymous affiliation
and treatment outcome among a clinical sample of problem drinkers. American
Journal of Drug anu Alcohol Abuse, 309-324.

McLellan, A. T., Luborsky, L., Woody, G. E., O’Brien, C. P., 8 Druley, K. A.
(1983). Predicting response to alcoholic and drug abuse treatments. Archives of
GenereiLDsychiatry. fl, 620-625.

McLellan, A. T., Luborsky, L., Woody, G. E., 8 O’Brien, C. P. (1980). An
improved diagnostic evaluation instrument for substance abuse patients: The
Addiction Severity Index. The Journal of Nervous and Mental Disease, 3&8, 26-33.

Mensch, B. S. 8 Kandel, D. B. (1988). Underreporting of substance use in
a national longitudinal youth cohort. who Opinion Quarterly, Q, 100-124.

Miller, W. R. 8 Hester, R. K. (1980). Treating the problem drinker. Modern
approaches. In W. R. Miller (Ed), The addictive behaviors: Treatment of

alcoholism, drug abuse, smoking, and obesity (pp.11-141). NY: Pergamon Press.

Miller, W. R. 8 Hester, R. K. (1986). Matching problem drinkers with optimal
treatments. In W. R. Miller 8 N. Heather (Eds), Treating addictive behaviors:
Process of changa (pp.175-203). NY: Plenum Press.

Miller, W. R. 8 McCrady, B. S. (1993). The importance of research on
Alcoholics Anonymous. In B. S. McCrady 8 W. R. Miller (Eds), Research on
Alcoholics Anonymous: Opportunities and alternatives. New Brunswick: RUtgers
Center of Alcohol Studies.

Minkoff, K. (1989). An integrated treatment model for dual diagnosis of
psychosis and addiction. Hospital and Community Psychiatgy, 519, 1031-1036.

Minkoff, K. 8 Drake, R. E. (1991). Dual diagnosis of major mental illness
and substance disorder. New Directions for MentarlHaalth Services, _5_0_, 3-12.

Moos, R. H., Finney, J., 8 Maude-Griffin, P. (1993). The social climate of

119

self-help and mutual support groups: Assessing group implementatiOn, process,
and outcome. In B. S. McCrady 8 W. R. Miller (Eds), Research on Alcoholics
Anonymous: Opportunities and alternatives. New Brunswick: Rutgers Center of
Alcohol Studies.

Mueser, K. T., Bellack, A. S., 8 Blanchard, J.J. (1992). Comorbidity of
schizophrenia and substance abuse: Implications for treatment. Journal of
Consulting and Clinical Psycholggy, 65), 845-856.

Nguyen, T. D., Atikisson, C. C., 8 Stegner, B. L. (1984). Assessment of
patient satisfaction. Evaluation and Program Planning, _6_, 299-314.

Norris, J. L. (1978). Prevention of chronicity in alcoholism. Psychiatric
Annals, 8(11), 48-53.

 

Norusis, M. J. (1993). SPSS for windows. Chicago: SPSS, Inc.

O'Farrell, T.J., Connors, G. J., 8 Upper, D. (1983). Addictive behaviors
among hospitalized psychiatric patients. _Ad_dictive Befiviors. 8, 329-333.

Ogbome, A. C. (1989). Some limitations of Alcoholics Anonymous. _In Marc
Galanter (Ed) Recengevelopments in_a_lcoholism Vol.7 (pp. 55-65). NY: Plenum
Press.

Ogbome, A. C. 8 Bomet, A. (1982). Abstinence and abusive drinking
among affiliates of Alcoholics Anonymous: Are these the only alternatives?
Ad_dictive Behaviors. 1, 199-202.

Ogbome, A. C. 8 Glaser, F. B. (1981). Characteristics of affiliates of
Alcoholics Anonymous: A review of the literature. Journal of Stu_dies on Alcohol.
42(7), 661 -675.

O'Leary, M. R., Calsyn, D. A., Haddock, D. L., 8 Freeman, C. W. (1980).
Differential alcohol use patterns and personality traits among three Alcoholics
Anonymous attendance level groups: Further considerations of the affiliation

profile. Drug and Alcohol Dependence, 2 135-144.

Osher, F. C. 8 Kofoed, L. L. (1989). Treatment of patients with psychiatric
and psychoactive substance abuse disorders. Hospital and Community
Psychiatgy, up, 1025-1030.

Otteson, J. P. (1979). Curative caring: The use of buddy groups with chronic
schizophrenics. Journal of Consulting and Clinical Psychology, 47(3), 649-651.

120

Panepinto, W. C., Higgins, M. J., Keane-Dawes, W. Y., Smith, D. (1970).
Underlying psychotic diagnoses as an indicator of participation in alcoholism
therapy. Quarterly Journal of StLulies on Alcohol. 3_1, 950-956.

Peele, S. (1990). Research issues in assessing addiction treatment efficacy:
How cost effective are Alcoholics Anonymous and private treatment centers? Drug
and Alcohol Dependence, g, 179-182.

Polich, J. M. (1982). The validity of self-reports in alcoholism research.
Addictive Behaviors, 1, 123-132.

Polich, J. M.; Armor, D. J.; 8 Braiker, H. B. (1980). The course of
alcoholism: Four years after treatment. Santa Monica: Rand.

Powell, T. J. (1993). Self-help research and policy issues. The Journal of
Applied Behavioral Science. 29(2), 151-165.

Prochaska, J. O., DeCIemente, C. C., 8 Norcross, J. C. (1992). In search of
how people change. American Psychologis_t, 47(9), 1102-1114.

Rather, B. C. 8 Sherman, M. F. (1989). Relationship between alcohol
expectancies and length of abstinence among Alcoholics Anonymous members.
Addictive Behaviors, 15, 531-536.

Ribisl, K. M. (1995). A longitudinal study of the personal networks of
individuals receivinLtreatment for mental illness and substance abuse problems.
Unpublished doctoral dissertation, Michigan State University, East Lansing.

Rouse, B. A., Kozel, N. J., 8 Richards, L. G. (1985). Self-report methogs of
estimating drug use: Meeting current challenges to validity. NIDA Research
monograph 57 (DHHS Publication No. (ADM) 85-1402). Rockville, MD: U. S.
Government Printing Office.

Sanchez-Craig, M., Annis, H. M., Bomet, A. R., 8 MacDonald, K. R. (1984).
Random assignment to abstinence and controlled drinking: Evaluation of a
cognitive-behavioral program for problem drinkers. Journal of Consulting and
Clinical Psychology, gag), 390-403.

Scott, E. M. (1986). Beyond A.A. lntemational Journal of Offender Therapy
and Comparative Criminology, 30(1), 69-75.

Seixas, F. A., Washbum, S., 8 Eisen, S. V. (1988). Alcoholism, Alcoholics
Anonymous attendance, and outcome in a prison system. American Journal of

Drug and Alcohol Abuse, 14(4), 515-524.

121

Selzer, M. L. (1971). The Michigan Alcoholism Screening Test: The quest
for a new diagnostic instmment. American Journal of Psychiatpy, 1_2_7, 1653-1658.

Smart, R. 6., Mann, R. E., 8 Anglin, L. (1989). Decreases in alcohol
problems and increased Alcoholics Anonymous membership. British Journal of
Addiction, 84(5), 507-513.

Stone, W. N. (1991). Treatment of the chronically mentally ill: An opportunity
for the group therapist. lntemational Journal of Group Psychotherapy, flu), 11-22.

Swift, CF. 8 Beverly, S. (1985). The utilization of Ministers as Alcohol
Counselors and Educators: Increasing prevention and treatment resources in
the black community. In R. Wright 8 TD. Watts (Eds) Prevention of Black
alcoholism. Springfield: Charles C. Thomas Publisher.

Swindle, R. W., Cronkite, R. C., 8 Moos, R. H. (1989). Life stressors, social
resources, coping, and the 4-year course of unipolar depression. Journal of
Manna) Psychology, 98(4), 468-477.

Thompson, T. 8 Simmons-Cooper, C. (1988). Chemical dependency
treatment and Black adolescents. The Journal of Drug Issues, Ju, pp.21-31.

Thurstin, A. H., Alfano, A. M., 8 Nerviano, V. J. (1987). The efficacy of AA
attendance for aftercare of inpatient alcoholics: Some follow-up data. lntemational
Journal of the Acflictiona, 22(11), 1083-1090.

 

Tomsovic, M. (1970). A follow-up study of discharged alcoholics. Hospital
app Community Psychiauy, 21(3), 94-97.

Toumier, R. E. (1979). Alcoholics Anonymous as treatment and as ideology.
Journal of Studies on Alcohol. 40(3). 230-239.

Trice, H. M. 8 Roman, R. M. (1970). Sociopsychological predictors of
affiliation with Alcoholics Anonymous. Social Psychiatm, an), 51-59.

Vaillant, G. E. (1983). The natural histogy of alcoholism. Cambridge: Harvard
University Press.

Walsh et al. (1991). A randomized trail of treatment options for alcohol-
abusing workers. New England Journal of Medicine, 2Q. 775-782.

Westerrneyer, J. (1981). Research on treatment of drinking problems:
Importance of cultural factors. Journal of Studies on Alcohol, Suppl. 9, 44-57.

Westerrneyer, J. 8 Neider, J. (1988). Social networks and psychopathology

122

among substance abusers. American Journal of Psychiatgy, 145(10), 1265-1269.

erkey, W. W. (1986). The influence of Alcoholics Anonymous on alcoholism
treatment. Employee Assistance Quarteriy, 1(5), 1-17.

WIIson, W. H., Diamond, R. J., and Factor, R. M. (1990). The chronically
mentally ill: Group treatment for individuals with schizophrenia. Cpmmunity Mental
Health Journal, 26(4), 361-372.

Yalom, I. D. (1985). The theopy and practice of group psychotherapy. (3rd
edition). Basic Books.

Zweben, J. E. and Smith, D. E. (1987). Can the patient on medication be
sent to 12-step programs? Journal of Psychoactive Drugs, 2_1, 221-228

APPENDIX

123

338.: .282 a... <22... 33:... 33.. 83.2 ”.032 3.3.2.: .3. 83:

 

\.
32.63 .3:

 

 

m

   

nan—a8

_2—onw—< V ,Wgw /fl
:5 <22... m 3:.

 

 

33.3 3mg

 

 

 

 

 

m

 

32.—ova 33w
3.8.... m 3.

 

. 8 .S.\fl/
5.. M /..

/mm.

124

.532.— 25 ...: <22}. .58... 55% .352 ".232 3:33 .~.< 23...

fluaauim .3.me

 

ha.

2:033.“
«En v 08:.

 

   

 

 

 

 

 

 

 

 

> 8 2:23 .3:
930 <Z\<< n 25.”.
con.
2.3305 we - / . 3330\nm Bag
:5 m 3:. 2

 

 

w
/:

125

.330: 133.3%... a... $32 3322 $5 a. H3522 ”.332 3:93 .3. 23E

 

he.

\

33:8.—
wsusfiobm .. 2::

3:263 flaw

 

 

    

 

 

 

 

 

 

 

 

» 2.-
3836 .3:
55 <2)? m 3.:
on.
.8030; 3.- / 3
02—0 30
332%... n 3:. a. . :63 q

 

 

/:

126

mafia“ BoomEEm _2_- <z\<< «58.2 :5 ... .352 €32 32.3.: .3. 23;

finch—Oxnm .3me
as: flag

8
38203 33

 

\na.

«Eu—no.8
3083...: ¢ 25. ¥\4/m
:5 <2)? m 2.:
2338A .cfi\fl/
m... (We 4/:

 

     

 

 

 

 

 

 

 

033a: .0. 2:.

 

 

/cc.~

127

gouginuoaum 2: <2); mam—ES: 990 no aux—8:2 “—0.32 @3938: .n.< 83E

 

ma.

dogma—332.0%
e oars

) Nor

Such—Uim gnaw

 

 

    

 

 

3=o<>m gnaw

 

:5 <75}. m 3:.
gum.

 

 

 

 

2.863 sow

 

8.- /
«€333.23
_ m 25w 3.
{/22

 

128

m

33%... _2§< ...: E33 8:232 <75} ”.33: 33.3.: .e-< 83E

 

genuxum .3me

 

 

m

   

.332
33.2 V 2a...“

 

 

38%,» flaw

 

333%
35:38.4 <23}.

 

 

 

 

 

3853 39w
.38.... m 2...

 

no
.3. )\
.3039. / .
.w «a

o.—

129

.535 «an ...: as...“ 838.2 <22). "33: 323.: .32 23E

 

bu.

\

3:853 32g

 

 

 

 

«Eu—cam
nan— ¢ 08:.

 

 

  

 

 

 

 

 

)
a 9
N 5:: 33¢; __ mg
. 882.82 423‘
oon
2.- /
.935 . 38:63 33
«an m 3:. 8

 

 

 

/

«a.

.338.— 233303 2. 83.: 83.8.3 <23). ”.332 3:93 .a.< 2.3...

 

 

 

    

 

 

 

 

 

 

 

 

0
3
\3.
3.2.63 .32
.8032.—
193—one“: * earn. \1/
33 u.
m 53.; aim .— .mw
. 0282.33. <2):
oav
3.0382 /\ / . 33393 33
723.39»; M 26,—. «a

 

 

o4

131

.535 3835a: _a. as...” 838.2 <2)? ”.35: 33.2: «i 83E

 

\ma.
finch—Dim Ada

 

 

m,

.8038
335m . g...

 

  

 

 

8.8456 Adm—fig
noun—30:4. <E<<

 

“WA/m 5.3

 

 

 

“Haw \ 4/3.

baggage: n 2...

 

3:503 Bow

 

 

/oc.—

132

 

33:33.23 ...: 32.... 88222 <22... 33: 83:25 .2.< 23a

ma.

 

finch—Oxum Afimw

 

 

     

5:31.523
v 3:.

> no

 

 

2:33 ..ng

38303 32w

 

E03;
3. m. 352.82 56:

 

 

 

 

«3333923
m 3:.

 

133

«8038.— _2_oo_< v3 25 vacuum mo 093. ”—302 33333 ._ T< oBuE

ma.

3. /

F as
£09.30“ 1H3uu0wem
h 32.—5 a.

34002 c 06E.

 

 

fl such—oim A3:

 

    

 

 

 

 

 

 

 

 

 

 

 

 

>
«axiom 23 m.
.3. inauuououm + 5433 A sum .— .EW
2533 co.—
_2_§< m 3:.
>15 <29}. 030.39% Bo.—

 

 

 

/

¢¢.

“Eu—poi man can on: 3:35 we 2.3. ”332 3.23:: .224. 2:»;

ma.

\

2.5 A . \
323% 2550.35

 

 

 

25:83 .35

 

134

 

 

”£932.—
wEQ v 05:.

 

a

 

 

 

 

 

 

 

 

   

 

 

 

 

 

 

. 32:;
.3 :55985 + <z\<< f 23$; :9:
”Eu—no:— co.—
mao m 08:.
.. \A c
2 A .
..m— .. r

 

oo.

on.

 

«Eu—no:—
BBwo—onozmm v 08E.

135

 

>

row.

 

 

mac—98m
_aommofinommm m 2.5.

 

no.

we.

3..

:1.

_o.

 

mEoEoE _uomwfloaummm v5 3.: 333m mo 33. ”—232 33535 .m_-< EsmE

no.

\

>25
mootcom 35:80.35

 

 

 

 

f

 

«335m
323%... + <E<<

 

28:63 .3:

 

 

fl

 

/

co.—

 

zao <E<<

 

 

 

 

3:23 :3:

 

 

 

 

 

ma.

.2...

F

2:263 33

 

meozem 3031—5: 23 on: oatom Mo 2;... ”3qu 3333.: .I-< 2am;

 

 

 

 

 

136

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

no.
3.
2.5 +
auction _ucofimuuohm o_:o.:_0\xm 53m
8..
“Eu—poi .
383:5: v as: 8
>
on. 8. 323.5
9
$333.35 + <23}. A 32:35 .35
—°. /
«Bozo; cc.—
EoomEES m as;
S
2. 3:0 <z\<<
co; m . 2.822% >84
...: - r

 

g.

conun_§_%o:om EB 83 830m mo 33. “.252 33%an .m _-< 2:me

m¢.
a \

moomtom 12833.35

 

 

 

      

2.33me as:

 

 

 

 

 

 

 

 

 

 

 

 

vo.
:ouafifiamoaom E .
7 ..
3 v 05E.
mm. vo. moomiom
_., mm.» o
:32 a .5 + <Z\<< A 3=o<\xm awm:
:. /
cotanzfiamoaox mm.
m 05:. _O

 

 

 

4/ 8. :5 <2)?

3. fi .3. F

 

2:853 33

 

mm.

mane—no.5 _aofio—osuxmm can <Z\<< wcmvnou< PEG «0 Hon—E: Z ”—252 Eon—«52H <z\<< .e_-< BamE

5w. co.—

5 H

0:283 :5
\ mac—no.5 wan _.-
f/ m 25...

£8 <2)? n 05F ‘A 8. _ 33$; :5
3.

m 5032.— ?0305 3:82 .
30322:»; m 08:. A on:
*2. H m P
a. no...
oEoEExm 33
3 3.

 

 

 

 

mac—poi AW v
Rama—one»; v 08:. S

 

        

8
3
1

 

 

   

 

 

nwv.

 

 

 

 

 

 

 

 

nausea _zoomEE; E: <z.\<< 9:23; 15.? 33:52 “3.32 2.253; «.233. ..E.< 2:»:

3. co.—

 

E 28:83 :5

 

 

 

 

«8235 was
m 25...

£035 4

% ESWEEE v 2::

   

 

/——\

 

 

    

z.
. :5 5.33. m 2.: S. 2:23 .35

 

 

 

 

 

«:83on
383:2: m as: A

2538.. .23:

m 2.:

/__\

no.

mo. we;
2.8.523 Bog

no. 2.

 

 

 

 

140

conga—£32.; E; <z\<< 2:283. 35.? 325.2 ”3.32 23.53; <z\<< .m _.< 2%:

3. co.—

 

 

 

 

 

I:
.9.
‘6
.53.
.3
v s:
0 O
.E—f'fi
[—M
M
O
r—W

 

 

 

 

22:63 .3:
meozem mac _..
m 08:.

   

 

 

 

 

 

 

   

 

 

>
om.
. :5 <22: m ”a: i 2:23 :5
saga—232.3“ 2532a :28:
m 08:. A 0::
mo. r m ..F

 

 

 

a. 3.
anoEExm Bog

mo. 3.

 

1
4
1

mac—no:— Eflmo—oaommm EB 8088 unguaou< <E<< ”_ov02.~=o§a2H <z\<< d7< 23E."—

hw. ma.

 

 

tame—233m v 08:.

F 0:253 3m
2538.— AI 3:. \ 2530.5 mam

F m 25,—. ...:.-

 

 

  

 

   

x

 

 

«we.

   

58.:

23s.; .3:
352.22 <5: -

A
O.

 

 

 

 

 
 

a so. n 8m mac—no:— _oaoo;

 

 

«23°93 on: A, on:
_._,_ mm ..H. .mfijr m ....

 

 

/ /

. 3.822% Bog
3 2:

 

£032.. Eocmxzsam Ba 523 85232 <z\<< €82 Hugues <z\<< .o~-< 2:3

 

 

 

 

 

         

no. 3.
H F 25:53 Em
mac—poi A .- \ mace—no.5 wan .-
w. 3833; v ”a: 8 fi m 2:: .2

 

 

 

 

   

 

 

 

 

 

 

>
.3. 5oz; . 33$; .33.—
35222 5:: 8 -
no.
u 82 no; 80305 3:32
a A on:
:35 zaz m ”a: 8. f m ..F

 

 

 

ma. no...
3:8: 036 33

3. 2:

 

Engaging»: as 533.. 85232 <2)? H3.32 2053; <z\<< ._~-< 2%:

ma. ao.

H e 2.82 2:. En

nozszsfimogom A
v. 08:.

 

 

 

 

mEoBoE wan

\ .
fi mos—C. LL ..

not

143

     

 

 

 

 

 

 

 

 

 

  

 

 

>
..bm. 58::— . oSofixm .35
0233.32 <z\<< no -
35222335”. mac—poi .232
m 08:. A 0::
8. F m L.

 

 

/ ma. mo.-
umunza U\xm 30..—

Ma. 2:

 

£28.: 35355 Ea 3: 833m .6 25 €82 22:3: <22< «3. 25mm

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

mo.
5. 3. e
k 35 A \
...w_. moorcom 13230.35 2:8:me awi
2338.. Sr use—gem win— 7
4
4 53223:; e as; m can.
>
*hN a0 9
...mv. moatom . 8:336 :93
33:33; + <z\<< .
*2...
ac. fl
man—poi 80305 3:32 .
Homwofinohmm m 05E. m 08E. . oo _ 9
*o—. oo
/ / . ‘ 8
. ...om - ~2:0 <Z\<< omaofiuim 264
mm no. n.s.- F/

 

am.

528$ Esflbeé Ea 3: 833% mo 25 ”382 2053: <z.\<< .m~-< 05mm

 

 

 

 

 

             

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

3.
mm. ha. 6
\
F 3.5
H *3. mofltom _auofimomoi 3339.6 :33
38305 mEoEoE man— 7.
w Boomiefl v 05:. m 2:?
>
ao
_m. 323$
. 52:35 + $55. 9:23 3:
«2..
mEofioE 8030.5 .282
_soomizeé m 25 n 2:; o3
. ac. 3
mo
4/ 4/ . ‘ 8
. *ON . EGO <Z\<< omaoEUbnm 304
ha ha. *2 .- F

 

aa.

146

assign; 2.3 a: 853 L6. 25. #82 Bogs: .224 «3‘ 05mm

H e

cannuzflfimoaom 8.- mac—poi wEQ
v 25,—. m 08F

 

     

3,5

mootcom _acofimomoi 0:830me :33

 

 

 

 

 

 

 

 

  

 

 

 

 

 

  

 

 

 

 

 

>
*NN. 9
on. moetom t
.. $333.35 + <z\<< ac. 8.2336 :3:
. / ‘..-
sosafi—SEmoaom mo 8030.5 3:82 .
moEE. moEF oo_ 2
mo.
4/ 4/ { 8
. . :5 <2)? 2:253 33
mm no *2... r

 

 

 

mm.

V. LIBRRRIES

MICHIGRN STQTE U I
\NilWlHHllHlllWlNIH H M MIHIIIIIHII
31 29301 7 9

42 513