Content uploaded by Molly Magill
Author content
All content in this area was uploaded by Molly Magill
Content may be subject to copyright.
Addiction Research and Theory
June 2009; 17(3): 236–259
How do people recover from alcohol dependence?
A systematic review of the research on mechanisms
of behavior change in Alcoholics Anonymous
JOHN FRANCIS KELLY
1
, MOLLY MAGILL
2
,
& ROBERT LAUREN STOUT
3
1
Department of Psychiatry, MGH Center for Addiction Medicine, 60 Staniford Street, Boston,
MA 02114, USA,
2
Brown Medical School, CAAS, Providence, RI, USA, and
3
Decision Sciences Institute, Pawtucket, RI, USA
(Received 9 January 2009; accepted 18 January 2009)
Abstract
Rigorous reviews of the science on the effectiveness of Alcoholics Anonymous (AA) indicate that AA
and related 12-step treatment are at least as helpful as other intervention approaches. Exactly how AA
achieves these beneficial outcomes is less well understood, yet, greater elucidation of AA’s mechanisms
could inform our understanding of addiction recovery and the timing and content of alcohol-related
interventions. Empirical studies examining AA’s mechanisms were located from searches in Pubmed,
Medline, PsycINFO, Social Service Abstracts and from published reference lists. Thirteen studies
completed full mediational tests. A further six were included that had completed partial tests.
Mechanisms examined fell into three domains: (1) Common processes; (2) AA-specific practices; and
(3) Social and spiritual processes. Results suggest AA helps individuals recover through common
process mechanisms associated with enhancing self-efficacy, coping skills, and motivation, and by
facilitating adaptive social network changes. Little research or support was found for AA’s specific
practices or spiritual mechanisms. Conclusions are limited by between-study differences in sampling,
measurement, and assessment time-points, and by insufficient theoretical elaboration of recovery-
related change. Similar to the common finding that theoretically-distinct professional interventions
do not result in differential patient outcomes, AA’s effectiveness may not be due to its specific content
or process. Rather, its chief strength may lie in its ability to provide free, long-term, easy access and
exposure to recovery-related common therapeutic elements, the dose of which, can be adaptively self-
regulated according to perceived need.
Keywords: Self-help, mutual-help groups, addiction, alcoholism, recovery
Correspondence: John Francis Kelly, Department of Psychiatry, MGH Center for Addiction Medicine, 60 Staniford Street, Boston
MA 02114, USA. Tel.: (617) 643-1980. Fax: (617) 643-1998. E-mail: jkelly11@partners.org
ISSN 1606-6359 print/ISSN 1476-7392 online ß2009 Informa Healthcare USA, Inc.
DOI: 10.1080/16066350902770458
Introduction
For clinical professionals, addiction patients, and laypeople alike, the topic of Alcoholics
Anonymous (AA) can arouse strong emotion. Ranging from ‘‘AA is a cult’’ and ‘‘AA
doesn’t work’’ to ‘‘AA saved my life’’ and ‘‘AA is the only way to recover’’, opinions
about its value and effectiveness vary widely (Vaillant 1983; Peele 1990; Bufe 1991; Ferri
et al. 2006). Views may differ, but since its modest beginnings in the 1930s and 1940s in
Ohio and New York in the US, AA has grown into an influential international
organization (Ma¨kela 1996). Millions of individuals are current members in more than
180 countries, and more than 28 million AA ‘‘Big Books’’ (the main AA text) have sold
making it one of the highest selling non-fiction books of all time (Humphreys 2004; AA
2008; Kelly and Yeterian 2008).
In the US, AA is the most commonly sought source of help for alcohol-related problems
(Room and Greenfield 1993; Weisner et al. 1995) with roughly 55 000 groups holding
meetings at least once per week (AA 2008; Kelly and Yeterian 2008). Recent US survey data
estimate attendance at addiction mutual-help groups, such as AA, at five million persons
annually (Substance Abuse and Mental Health Services Administration (SAMSHA) 2008).
AA’s program and practices have also influenced professional treatment in the US with the
vast majority endorsing at least some tenets of 12-step philosophy (Roman and Blum 1999;
Drug Strategies 2003). Noteworthy too, is that regardless of the theoretical orientation of
formal treatment programs, referral of alcohol-dependent patients to AA is the norm in the
US (Humphreys 1997; Kelly et al. 2008b).
Rigorously conducted empirical reviews of AA-focused research indicate that AA
participation is helpful for many different types of individuals in their recovery from
alcohol dependence (Emrick et al. 1993; Tonigan et al. 1996; Kownacki and Shadish
1999; Kelly 2003; Humphreys 2004; Ferri et al. 2006; Kelly and Yeterian 2008).
Determining exactly how and why AA is helpful has been a comparatively new, yet
intriguing, line of inquiry that has gathered increasing momentum as the broader alcohol
treatment and recovery field has moved toward understanding ‘‘mechanisms of behavior
change’’ (Morgenstern and McKay 2007; Willenbring 2007). The renewed emphasis on
mechanisms of change stems from the consistent finding that, despite a burgeoning of
empirically-supported treatments for alcohol and other drug use disorders (Finney et al.
1996; Miller and Willbourne 2002; SAMHSA 2008), these theoretically disparate
interventions appear to produce very similar outcomes under rigorous testing conditions
(Project MATCH Research Group 1997; Crits-Christoph et al. 1999; Dennis et al. 2004;
Anton et al. 2006; Imel et al. 2008; UKATT Research Team 2008). Prior reviews have
covered the question of the effectiveness of AA and 12-step treatment in some detail.
The current review is therefore focused on what is known regarding how AA helps
individuals achieve sobriety.
To begin, we clarify some important distinctions between AA and 12-step-based
treatments. Subsequently, for those unfamiliar with some of the research conducted on
AA and related 12-step treatments, we present a very brief overview of the largest
prospective 12-step-focused research studies. This is followed by a description of AA’s
own purported theory of recovery-related change and a subsequent review of the
empirical literature on what is known about how AA may exert its beneficial effects.
In the final section, we discuss the limitations of existing knowledge and approaches, and
offer some theoretical elaboration on how recovery-related change might occur as
a function of AA participation.
AA mechanisms 237
AA versus ‘‘12-step treatment’’
The community fellowship, Alcoholics Anonymous, can be confused with both ‘‘Twelve-
Step Facilitation’’ (TSF) and ‘‘12-step treatment’’ (sometimes referred to as the
Minnesota Model; McElrath 1997). To clarify, AA is a non-professional, community-
based fellowship that provides help through a network of informal gatherings, convened
at rented venues, such as churches and hospitals (AA owns no property; AA 1953).
‘‘TSF’’ is the name given to a professional, manualized, intervention designed
to facilitate engagement in AA. When the phrase ‘‘12-step treatment’’ is used, it
typically refers to a residential program in which patients receive various interventions,
but also are educated in-depth about AA and the 12 steps and may formally work
through some of these steps. Patients often attend AA during treatment and are strongly
encouraged to continue post-discharge (McElrath 1997). Although professional treat-
ments may incorporate AA practices and philosophy and refer patients to AA, AA itself
is not affiliated with any professional entity or organization (AA 1953). Next, we
describe some of the findings regarding the effectiveness of AA and related 12-step
approaches.
Evidence regarding the effectiveness of AA and 12-step treatment approaches
Despite existing for more than 70 years and influencing both formal and informal
approaches to alcohol and other drug-related treatment, it is only recently that AA has been
subjected to rigorous scientific inquiry (e.g., Institute of Medicine 1990; Emrick et al. 1993;
Tonigan et al. 1996; Kownacki and Shadish 1999; Ferri et al. 2006). Empirical reviews of
the literature on AA and 12-step treatment consistently converge on the finding that AA is,
at a minimum, helpful to many as they try to recover from alcohol dependence (Emrick et al.
1993; Tonigan et al. 1996; Kownacki and Shadish 1999; Kelly 2003; Humphreys 2004;
Ferri et al. 2006; Kelly and Yeterian 2008).
Perhaps the largest and most rigorously conducted clinical trial examining 12-step
treatment and AA participation among alcohol-dependent patients was conducted in the
US (i.e., Project MATCH; Project MATCH Research Group 1993). This study found
that the TSF condition was at least as effective as the two more empirically-established
comparison treatments (cognitive-behavioral therapy [CBT], motivational enhancement
therapy [MET]) in reducing the quantity and frequency of alcohol use at post treatment,
and 1- and 3-year follow-ups (Project MATCH Research Group 1997, 1998a, 1998b).
Moreover, TSF was superior at increasing rates of continuous abstinence. Specifically,
24% of the outpatients in TSF were continuously abstinent throughout the year after
treatment, compared to 15% and 14% in CBT and MET, respectively. Abstinence rates
at 3 years continued to favor TSF, with 36% reporting abstinence, compared to 24% in
CBT and 27% in MET (Cooney et al. 2003).
Findings also suggested that TSF may be more effective with specific types of patients.
Those with less psychiatric severity (Project MATCH Research Group 1997) and those
with more severe alcohol dependence (Cooney et al. 2003) had better outcomes in TSF
than in CBT at post-treatment and 1-year follow-up, respectively. More substantial
effects were observed in relation to patients’ social networks. Outpatients with networks
supportive of drinking at intake had better 3-year outcomes in TSF than in both
comparison treatments, and this difference was related to AA group attendance during
238 J. F. Kelly et al.
the follow-up (Longabaugh et al. 1998). Also, regardless of which original treatment
patients got, individuals who attended AA had significantly better drinking outcomes
(Tonigan et al. 2003).
Compelling evidence for the cost-benefit effects of 12-step approaches to treatment
has also been shown. A large multi-site study of US Veterans Administration (VA)
intensive treatment programs compared the 1-year outcomes of patients treated in
either intensive CBT or 12-step treatment. Those treated in 12-step treatment settings
had substantially greater 12-step group participation than patients treated in CBT
programs, who received twice the number of outpatient visits and significantly more
inpatient mental health days. This difference in professional service utilization resulted
in 64% higher annual costs for CBT than for 12-step programs. Notably, the
demographic and clinical characteristics of patients across the two types of programs at
intake were comparable. The outcomes were comparable also, except that patients
treated in 12-step programs had higher rates of abstinence (46% vs. 36%; Humphreys
and Moos 2001).
A 2-year follow-up of the multi-site VA sample (Humphreys and Moos 2007) found
a substantially higher abstinence rate among patients treated in 12-step (50%) compared to
CBT (37%) programs. Again, patients from 12-step programs were more AA-involved than
CBT patients, while CBT patients relied more on outpatient and inpatient mental health
services. This resulted in 30% lower costs for those treated in 12-step programs (savings of
$2440 per patient) while achieving significantly better abstinence rates. These results were
consistent with earlier findings in a non-VA sample that showed 45% lower treatment costs
over a 3-year period for AA attendees compared to those electing outpatient care
(Humphreys and Moos 1996).
The research reviewed shows abstinence and cost benefits of 12-step treatment. A large
number of studies have also examined community AA group participation in relation to
long-term outcomes following a variety of treatment approaches. These studies have
found AA attendance over follow-up to be associated with enhanced abstinence
outcomes and remission rates among different patient subgroups, including women,
youth, dually-diagnosed individuals, and patients of varying ethnic backgrounds
(e.g., Miller et al. 1997; Morgenstern et al. 1997; Ouimette et al. 1997; Kaskutas
et al. 1999, 2005; Fiorentine et al. 2000; Timko et al. 2000; Tonigan et al. 2002;
McKellar et al. 2003; Kissin et al. 2003; Moos et al. 2004, 2006; Dawson et al. 2006;
Kelly et al. 2006, 2008b).
The vast majority of existing research has been conducted in the U.S. However, although
research on AA has been rare in the United Kingdom, and clinician attitudes have been
generally less favorable toward 12-step approaches (Humphreys 1999; Laudet 2003; Day
et al. 2005; Kelly et al. 2008b), UK studies show similar AA-related recovery benefits post-
treatment (Gossop et al. 2003, 2007).
Summary
From an intervention dissemination and impact perspective (e.g., Glasgow et al. 2003), AA
has reach and effectiveness, appears to be readily adopted and implemented, and has evident
staying power. Given the consistent recovery-related and cost-effectiveness findings
associated with AA participation across time, a central question is how exactly AA helps
its participants. This is the focus of the next section.
AA mechanisms 239
How does AA help individuals recover from alcohol dependence?
Contrary to AA’s language and somewhat mystical terminology documented in its main
texts (AA 1939, 1953, 2001) one might argue that its mechanisms could be explained by
ordinary means, such as through common processes of change (e.g., by increasing and/or
maintaining motivation, self-efficacy, and coping skills). With equal validity, one might
argue that AA works through practices specific to AA (e.g., working the 12-steps) or by
providing access to a low-risk social network. It may be through some combination, all, or
none of the above. In this section, we attempt to clarify how AA works by first describing
AA’s own proposed mechanisms, and then by summarizing the empirical literature
regarding what is currently known about how AA may exert its effects.
How change occurs from AA’s own perspective
Given that AA grew out of a religio-spiritual tradition rather than an academic or scientific
one, a coherent or consistent description of AA’s purported mechanisms can be difficult to
grasp from its main texts (AA 1939, 1953). Some change mechanisms (i.e., the 12-steps)
and their effects are made clear: ‘‘The fact is just this, and nothing less: That we have had
deep and effective spiritual experiences which have revolutionized our whole attitude toward
life, toward our fellows, and toward God’s universe’’ (AA 2001, p. 25). On the other hand,
the social and fellowship aspects inherent in group meetings are more implicit in much of
AA’s writings, especially those in the Big Book (1939). This may be because the fellowship
was largely non-existent at the time this book was written and published (with less than 100
members across two states, Ohio and New York; AA 2001) and the original main text body
has not been changed, apart from forewords to each subsequent edition (AA 2001).
However, we believe it is safe to say that AA facilitates change via two broadly defined
components: the AA ‘‘program’’, exemplified in the 12 steps, and the AA ‘‘fellowship’’,
characterized by the network of formal and informal social gatherings and communications
between meetings.
The central proposed mechanism of recovery from alcohol addiction according to AA
is through a ‘‘psychic change’’ (AA 2001, p. xxvi), ‘‘spiritual experience’’, or ‘‘spiritual
awakening’’ (AA 2001, Appendix II) achieved through completion of the 12-step program
(as noted in Step 12: ‘‘Having had a spiritual awakening as the result of these steps ...’’).
A measurable definition of a spiritual ‘‘experience’’ or ‘‘awakening’’ has eluded many
(Galanter 2007b). Although AA states that this ‘‘awakening’’ can take the form of a sudden
shift in belief and perspective, it also characterizes this transformation as a gradual change of
an ‘‘educational variety’’ that leads to ‘‘... a profound alteration in [his] reaction to life’’ (AA
2001, Appendix II). This is not only associated with belief in a ‘‘higher power’’, but also
involves concrete changes in specific attitudes and behaviors. Alcohol treatment researchers
have subsequently proposed that some of these attitudinal or behavioral shifts may be
consistent with those mobilized in other behaviorally-oriented treatments (McCrady 1994).
Thus, although the language and concepts may differ on the surface, underneath the same
ultimate (common) processes may be operating. Therefore, AA identifies internal cognitive-
affective and spiritual changes arising out of engagement in AA-prescribed behaviors, and,
as will be explored in more detail shortly, varying alcohol researchers have alternately
considered these purported mechanisms as AA-specific and as common factors of behavior
change.
Although often implicit in AA’s own writings on how AA conveys its beneficial effects,
another major feature of Alcoholics Anonymous is the social or ‘‘fellowship’’ dimension.
240 J. F. Kelly et al.
This vital social context in which recovery behaviors are learned, modeled, and supported,
as noted above, is less explicitly documented as an essential curative component in AA’s
main texts, but is captured in the preamble read at the start of nearly all AA meetings
(AA 1947):
Alcoholics Anonymous is a fellowship of men and women who share their experience, strength
and hope with each other that they may solve their common problem and help others to recover
from alcoholism. The only requirement for membership is a desire to stop drinking. There are no
dues or fees for A.A. membership; we are self-supporting through our own contributions. A.A. is
not allied with any sect, denomination, politics, organization or institution; does not wish to
engage in any controversy; neither endorses nor opposes any causes. Our primary purpose is to
stay sober and help other alcoholics to achieve sobriety.
The explicit emphasis is on mutual ‘‘sharing’’, or communication, of experiences in an
attempt to help oneself and others. Sharing of addiction and recovery experiences may
impart important technical aspects of living sober, evoke a sense of empathy and belonging,
and instill hope. As alluded to in the preamble’s last sentence, a central goal of 12-step
organizations invokes the ‘‘helper principle’’ (Riessman 1965), which is quite simply
the notion that ‘‘helping you helps me’’. As AA states: ‘‘Practical experience shows that
nothing will so much insure immunity against drinking as intensive work with other
alcoholics’’ (AA 1939, 2001, p. 89). Thus, within the dimension of fellowship lie the
proposed processes of mutual sharing and helping others (Zemore and Kaskutas 2004;
2008; Pagano et al. 2007), but alternative interpretations include principles of observational
learning (Bandura, 1969) and group theory dynamics (Yalom 1995; Yalom and Leszcz
2005) as well as health-promoting social network changes (Kaskutas et al. 2002).
In summary, one can deduce that AA-related changes occur via intrapersonal, behavioral,
and social processes. A number of these are specifically tied to AA’s philosophy and
prescription for change, and treatment researchers have debated the existence of
AA-specific, versus common, processes. This latter point connects to the complexity
involved in establishing a mechanism of behavior change.
What is a mechanism of behavior change?
‘‘A mechanism of change refers to the process or series of events through which one variable
leads to or causes change in another variable’’ (Nock 2007, p. 5S). ‘‘Mechanisms’’ of
behavior change, of course, can be conceptualized at multiple levels of scale. Thus, the exact
same change in behavior can be simultaneously explained by social, psychological,
behavioral, and neurobiological processes. Thus different mechanisms thought to account
for behavior change during a defined time period might justifiably be measured at all of these
different levels, and statements such as, ‘‘AA works by exposure to abstinent role models’’,
and, ‘‘AA works by increasing self-efficacy’’ and, ‘‘AA works by increasing individuals’
density of dopamine D2 receptors’’ are all mechanisms that may be empirically supported
and occur simultaneously. Also, AA is a complex entity. Thus, no single mechanism is likely
to account for its entire effects. These mechanisms are also likely to change over time; the
mechanisms through which AA may help individuals achieve sobriety may not be the same
as those sustaining it.
Statistical mediation will satisfy some, but not all, of the criteria for identifying
a mechanism of change (Kazdin and Nock 2003). The requirements for demonstrating
the operation of a ‘‘mechanism of change’’ combines both the criteria for statistical
AA mechanisms 241
mediation (e.g., Baron and Kenny 1986; MacKinnon et al. 2002) and the criteria for
inferring causal relations (Hill 1965; Cook and Campbell 1979). Seven criteria have been
proposed: association, temporality, specificity, gradient, plausibility and coherence,
consistency, and, ideally, validation via experimental manipulation (Kazdin and Nock
2003; Nock 2007). Because research employing statistical tests of mediation (or a close
approximation) is the focus of this review, the first three of the Kazdin and Nock (2003)
criteria will be emphasized. The association is the first step (i.e., the variables must covary).
The causal inference is strengthened by temporality (i.e., the cause occurs before the effect),
and specificity (i.e., the mechanism is specific solely to the particular intervention).
Promising mechanisms of change in AA should therefore be correlated with the AA variable,
follow it in time and occur before the measured outcome, and ideally, possible confounding
variables should be eliminated as alternative explanations.
Literature search and criteria for inclusion in this review
We conducted a literature search to identify studies examining mediators of AA’s effect on
alcohol or other drug use outcomes. The search strategy included a title, abstract, and
keyword search in Pubmed, Medline, PsycInfo and Social Services Abstracts using the
following search terms: AA OR Alcoholics Anonymous OR Self-help OR 12-step AND
mediators OR mechanisms or process. References from articles identified in the preceding
step and from relevant reviews of the AA literature (e.g., Emrick et al. 1993; Kassel and
Wagner 1993; Khantzian and Mack 1994; Tonigan et al. 1996; Kelly 2003; Kelly and Myers
2007; Moos 2008) were searched to identify additional eligible studies. The studies
identified for inclusion were English language published between 1990 and 2007 (inclusive).
Of central interest were (1) primary analyses from naturalistic research on community
groups or 12-step oriented programs or (2) secondary analyses from controlled clinical trials
on TSF. Included studies conducted formal tests of statistical mediation or an
approximation.
A ‘‘formal’’ test of mediation was defined as the casual steps outlined by Baron and
Kenny (1986) and extended by MacKinnon and Dwyer (1993). To establish mediation,
the following conditions should be met: (1) the ‘‘a’’ path (e.g., AA participation) should
predict the outcome (‘‘c’’; e.g., reduced alcohol consumption/abstinence) (2) the apath
should predict the mediator (‘‘b’’; e.g., self-efficacy), (3) the mediator should predict the
outcome, and (4) the direct effect of aon cshould be reduced or eliminated when bis
controlled. The mediated effect can be identified by a reduction in coefficient magnitude
(Baron and Kenny 1986; MacKinnon and Dwyer 1993) or a test of the joint significance
of the a!band b!cpaths (Sobel 1982; MacKinnon et al. 2002). Studies formally
testing mediation were relatively few (n¼13). Therefore to allow for a broader discussion
of the literature and particularly of the range of mediators that have been considered in
relation to AA’s effects, research examining the ato bpaths and the bto cpaths was also
considered (n¼6).
The extracted studies’ mechanisms of focus generally fell into three classes: (1)
common factors (i.e., self-efficacy, commitment to abstinence, active coping efforts; see
Table I), (2) specific AA practices (i.e., AA behaviors/activities, AA beliefs/cognitions; see
Table II), and (3) more explicit constructs related to AA’s theory of change (e.g., social
network variables, spirituality variables; Table III). These are detailed next starting with
common factors.
242 J. F. Kelly et al.
Table I. Studies examining common therapeutic factors.
Study
References Type of sample (N)
Demographic
characteristics Primary outcome Mediator/s Findings
Connors et al.
(2001)
Outpatient and aftercare arms;
Project MATCH (N¼924)
Mean age 40
Male 75%
Caucasian 82%
PDA 7–12mo
alcohol use
Self-efficacy AA participation
1–6mo
!self-efficacy
6mo
!
PDA
7–12mo
a,b
Tonigan et al.
(2003)
Outpatient and aftercare arms;
Project MATCH (N¼952)
Mean age 40
Male 75%
Caucasian 82%
PDA 33–36mo
alcohol use
Self-efficacy
Spirituality
AA participation
1–6mo
!self-efficacy
6mo
!
PDA
33–36mo
a,b
Bogenschutz et al.
(2006)
Outpatient and aftercare arms;
Project MATCH (N¼1284)
Mean age 40
Male 77%
Caucasian 81%
PDA 10–15mo
alcohol use
Self-efficacy AA attendence
4–6mo
!self-efficacy
9mo
!
PDA
10–15mo
a,c
Morgenstern et al.
(1997)
Inpatient 12-step program
(N¼100)
Mean age 34
Male 58%
Caucasian 63%
PDU 1–6mo
substance use
Self-efficacy
Commitment
Coping
Primary appraisal
AA affiliation
1mo
!self-efficacy
1mo
!
PDU
1–6mo
d
commitment
1mo
coping
1mo
primary appraisal
1mo
Morgenstern et al.
(1998)
Inpatient 12-step Program
(N¼118)
Mean age 36
Male 58%
Caucasian 66%
PDU 1–6mo
substance use
Self-efficacy
Commitment
Negative
expectancy
12-step Program !self-efficacy
1mo
!
PDU
1–6mo
e
commitment
1mo
Kelly et al.
(2000)
Inpatient 12-step program
(N¼99)
Mean age 16
Male 40%
Caucasian 78%
PDA 3–6mo
substance use
Self-efficacy
Commitment
Coping
AA attendence
1–3mo
!commitment
3mo
!
PDA
3–6mo
a,f
Kelly et al.
(2002)
Inpatient 12-step program
(N¼74)
Mean age 16
Male 38%
Caucasian 70%
PDA 4–6mo
substance use
Self-efficacy
Commitment
Coping
AA affiliation
1–3mo
!commitment
3mo
!
PDA
4–6mo
a,f
Notes: Demographic data rounded to nearest whole number. PDA ¼percent days abstinent; PDU ¼percent days used; NDA ¼number days used; NDU ¼number days
used; AA ¼Alcoholics Anonymous; TSF ¼Twelve-Step Facilitation; VA ¼Veterans Administration; CB ¼Cognitive Behavioral; ASI ¼Addiction Severity Index;
mo ¼month; post ¼posttreatment; beh ¼behavior; abs ¼abstinence; dep ¼dependence symptoms.
a
Tested by Baron and Kenny (1986)/MacKinnon and Dwyer (1993) method in Structural Equation Model.
b
Intake symptomatology predicted AA participation; results not moderated by treatment group.
c
Results not moderated by alcoholism typology.
d
Tested by Baron and Kenny (1986) method in series of linear regression models.
e
Results moderated by cognitive impairment; process variables were stronger predictors of outcome for unimpaired compared to impaired individuals via linear regression
analyses.
f
Baseline severity predicted AA attendance and affiliation; impact of AA attendance on month 3 commitment mediated by AA affiliation.
AA mechanisms 243
Table II. Studies examining general AA practices and principles.
Study
References Type of sample (N)
Demographic
characteristics Primary outcome Mediator/s Findings
Morgenstern et al.
(1995)
Inpatient and outpatient
12-step programs (N¼79)
Mean age 35
Male 68%
Caucasian 68%
Abstinence 1mo
substance use
12-step cognitions
Common cognitions
1
12-step program !change in
12-step cognitions
post
a
1
12-step program !change in
general cognitions
post
a
General cognition
2
change in common cognitions
post
!
abstinence
1mo
b
Morgenstern et al.
(2002)
Inpatient and outpatient
12-step programs (N¼370)
Mean age 36
Male 62%
Caucasian 49%
PDA 9–12mo
substance use
12-step cognitions
Common cognitions
1
12-step program !change in
12-step cognitions
post
a
1
12-step program !change in
common cognitions
post
a
2
change in 12-step cognitions
post
!
PDA
9–12mo
b
Ouimette et al.
(1999)
Inpatient VA programs CB
and 12-step-oriented (N¼1873)
Mean age 43
Male 100%
Caucasian 50%
Abstinence 12mo
substance use
12-step behaviors
12-step cognitions
CB cognitions
Substance coping
General coping
1
coping X 12-step program !
substance coping
post
c
1
symptoms X 12-step program !
substance coping
post
c
1
Dep X 12-step program !
12-step behaviors
post
c
2
12-step behaviors
post
!abstinence
12mo
d
Johnson et al.
(2006)
Inpatient VA programs CB
and 12-step-oriented (N¼1873)
Mean age 43
Male 100%
Caucasian 50%
Abstinence 12mo
substance use
12-step behaviors
12-step cognitions
CB cognitions
Substance coping
General coping
12-step program !abstinence
goal
post
!abstinence
12mo
e
12-step reading
post
12-step program !abstinence
goal
12mo
!abstinence
12me
12-step reading
12mo
# self-help meetings
12mo
having a sponsor
12mo
Brown et al.
(2001)
Aftercare relapse prevention
or TSF group treatment (N¼131)
Mean age 38
Male 70%
Caucasian 93%
Composite outcome
6mo substance use
12-step behaviors
12-step cognitions
Temptation
Confidence
TSF X change in 12-step
beh/cog
post
!NDA
3–6mo
b
Crits-Christoph
et al. (2003)
Outpatient individual and
group 12-step counseling (N¼325)
Mean age 34
Male 77%
Caucasian 58%
ASI 1–6mo cocaine use 12-step behaviors
12-step cognitions
12-step counseling !12-step
beh/cog
1–6mo
!ASI
1–6mo
f
Notes: Demographic data rounded to nearest whole number. PDA ¼percent days abstinent; PDU ¼percent days used; NDA ¼number days used; NDU ¼number days
used; AA ¼Alcoholics Anonymous; TSF ¼Twelve-Step Facilitation; VA ¼Veterans Administration; CB ¼Cognitive Behavioral; ASI ¼Addiction Severity Index;
mo ¼month; post ¼posttreatment; beh ¼behavior; abs ¼abstinence; dep ¼dependence symptoms.
a
Tested by t-test.
b
Tested by binary logistic or linear regression.
c
Tested by analysis of variance.
d
Tested by bivariate correlation.
e
Tested by Baron and Kenny (1986) method in series of binary logistic regression models.
f
Tested by analysis of Baron & Kenny (1986) method in series of generalized mixed-model; final step tested by proportion decrease in Fratio for direct effect.
244 J. F. Kelly et al.
Table III. Studies examining AA-proposed intra- or interpersonal changes.
Study
References Type of sample (N)
Demographic
characteristics Primary outcome Mediator/s Findings
Humphreys et al. (1999) Inpatient VA programs
(N¼2867)
Mean age 43
Male 100%
Caucasian 45%
NDU 9–12mo
Substance use
Coping
Friendship quality
Network support
for abstinence
2-step involvement
post
!
coping
12mo
!NDU
9–12mo
a
friendship quality
12mo
network support
12mo
Kaskutas et al. (2002) Inpatient and outpatient
community programs (N¼654)
Mean age 38
Male 58%
Caucasian 67%
Problem
Severity 12mo
Alcohol use
Abstinence 36mo
Social network size
Network support
for drinking
AA involvement
0-12mo
!net sup
drink
12mo
!sev
0-12mo
a
Bond et al. (2003) Inpatient and outpatient
community programs (N¼655)
Mean age 38
Male 56%
Caucasian 60%
Alcohol use Social network size%
network drinker
Network sup abs
AA sup abs
AA involvement
0–12mo
!AA
sup abs
12mo
!abs
33–36mob
Laudet et al. (2004) Double Trouble Recovery
(DTR) group members
(N¼310)
Mean age 40
Male 72%
Caucasian 24%
PDU 12–24mo
Substance use
Social support DTR affiliation
0–12mo
!soc
support
12mo
!PDU
12–24mo
c
Owen et al. (2003) Inpatient 12-step program
(N¼112)
Mean age 38
Male 53%
Mean age 40
NDA 12mo Lifestyle changes
Response to life events
AA involvement
0–12mo
!lifestyle
ch
0–12m
!NDA
0-12mo
a
Magura et al. (2003) Double Trouble Recovery
(DTR) group members
(N¼310)
Male 72%
Caucasian 24%
Abstinence 12mo
Substance use -health
behavior 12mo
Locus of control
Sociability
Spirituality hope
DTR affiliation
0–12mo
!locus of
control
12mo
!abs
0–12mo
d
sociabilty
12mo
DTR affiliation
0–12mo
!loc
cont
12mo
health beh
0–12mo
d
sociabilty
12mo
spirituality
12mo
hope
12mo
Notes: Demographic data rounded to nearest whole number. PDA ¼percent days abstinent; PDU ¼percent days used; NDA ¼number days used; NDU ¼number days
used; AA ¼Alcoholics Anonymous; TSF ¼Twelve–Step Facilitation; VA ¼Veterans Administration; CB ¼Cognitive Behavioral; ASI ¼Addiction Severity Index;
mo ¼month; post ¼posttreatment; beh ¼behavior; abs ¼abstinence; dep ¼dependence symptoms.
a
Tested by Baron and Kenny (1986)/MacKinnon and Dwyer (1993) method in Structural Equation Model.
b
Tested by lagged panel model and Sobel (1982) test.
c
Tested by Baron and Kenny (1986) method in series of linear regression models.
d
Tested by Baron and Kenny (1986) method in series of binary logistic regression models.
AA mechanisms 245
A review of research on mechanisms of behavior change in AA
Common factors
A number of studies have examined cognitive (e.g., self-efficacy, motivation for abstinence)
and behavioral (e.g., active coping) changes thought to be important to recovery regardless of
the treatment received, but that have been more explicitly implicated in treatments other than
12-step approaches (e.g., CBT). Using data from Project MATCH, Connors et al. (2001)
looked at 6-month self-efficacy as a mediator of the relationship between AA participation
(including attendance, step-work, spirituality) during the first 6 months of treatment and
abstinence at 7–12 month follow-up. Self-efficacy accounted for a significant reduction in the
direct effect of AA participation on later drinking, and this mediational effect held in both the
outpatient and aftercare samples and across treatment conditions (i.e., TSF, CBT, MET;
Connors et al., 2001). The effect of self-efficacy was maintained at 3-year follow-up, and
showed that AA step-work in particular was related to perceived confidence to avoid drinking
in social situations and when experiencing negative affect (Tonigan 2003). Thus, one way that
AA appears to work is by boosting confidence in participants’ perceived ability to handle
common relapse-related situations or circumstances. In both of these studies, greater alcohol
use severity was associated with more AA participation. The benefit of AA for these patients
appeared to be at least partially explained by changes in beliefs in their capacity to abstain from
alcohol.
Research with community samples has examined common therapeutic factors as
mechanisms of behavior change. In an inpatient sample, Morgenstern et al. (1997) found
that the positive effects of an aggregated measure of AA affiliation (e.g., AA attendance,
talking with sponsor, group service, step-work) on 6-month substance use were mediated by
a set of common factors measured at 1 month (i.e., self-efficacy, commitment to abstinence,
active coping efforts, primary appraisal [i.e., recognition of sustained or anticipated
consequences of use]). A subsequent study found that these factors were stronger predictors
of later abstinence among 12-step treatment program patients without cognitive impairment
when compared to impaired individuals (Morgenstern and Bates 1999). In an adolescent
sample, Kelly et al. (2000) examined self-efficacy, motivation for abstinence, and coping as
mechanisms of the effect of AA attendance in the first 3 months following inpatient treatment
on subsequent 4 to 6-month substance use outcomes. Here, only motivation for abstinence
mediated the effect of AA, and similar to the finding among adults receiving TSF (Connors
et al. 2001; Tonigan 2003), more severely substance-involved patients attended AA more
frequently. In a follow-up study, active AA involvement (e.g., talking with sponsor, group
service, step-work) was shown to be a more important predictor than attendance alone, and
the central mediating role of motivation for abstinence remained consistent (Kelly et al.
2002). Thus, similar to adults, baseline alcohol use severity in adolescents is a predictor of AA
attendance and affiliation, and both age groups show increases in common therapeutic factors
of self-efficacy, motivation, and active coping. However, in younger patients, only motivation
for abstinence mediated the effect of 12-step involvement on outcome, suggesting AA-derived
therapeutic benefits may differ developmentally. Findings from these studies are strengthened
by formal meditational tests and temporality maintained across causal pathways (Nock 2007).
Specific AA practices
A series of studies has examined and compared AA-specific cognitive and behavioral
changes to other potential non-specific change mechanisms. Morgenstern et al. (1995)
246 J. F. Kelly et al.
tested three groups of cognitive change factors post-treatment: AA-specific (e.g., acceptance
of powerlessness, belief in a higher power, commitment to AA, disease attribution);
abstinence specific (i.e., commitment to abstinence, intent to avoid high-risk situations), and
general psychotherapy (i.e., reduction in self-criticism), in relation to 1-month follow-up
abstinence. The study found that inpatient 12-step treatment resulted in both AA-specific
and general psychotherapy changes, but it was commitment to abstinence and intention to
avoid high-risk situations that predicted abstinence at follow-up. However, among patients
who relapsed, AA-related commitment and belief in a higher power were associated with
lower relapse severity. A follow-up study demonstrated that while patients made both AA-
specific and common cognitive changes, AA cognitions at discharge predicted better
outcomes at 6 and 12 months, and findings held with baseline use severity and
psychopathology controlled (Morgenstern et al. 2002). These studies did not conduct
formal tests of statistical mediation, but showed that 12-step oriented treatment resulted in
AA-related cognitive changes and such changes predicted reduced severity of relapse in the
short- as well as longer-term.
Research with VA samples has also considered AA-specific proximal changes, and has
provided more stringent tests via comparison of 12-step to cognitive-behavioral (CB)
treatment. In a naturalistic study across five 12-step and five CB inpatient programs,
Ouimette et al. (1999) examined a series of retrospective treatment matching effects among
male veterans. This study evinced some surprising findings. Specifically, compared to CB
patients, patients treated in 12-step programs made greater changes in an important
proximal outcome thought to be specific to CB treatment (i.e., substance-specific coping),
and this effect was moderated by lower coping and higher psychological severity at intake.
Patients in 12-step programs with higher alcohol use severity made greater changes in
12-step behaviors (e.g., meeting attendance, reading AA literature, step-work), and these
behaviors were related to better 12-month outcomes (Ouimette et al. 1999). A subsequent
study with this sample conducted tests of mediation and ‘‘dismantled’’ these AA-related
cognitions and behaviors. Only abstinence as a goal and reading AA literature at post-
treatment were mediators of the effect of program type on 12-month substance use. At
follow-up (with mediators measured at the same time-point), in addition to these variables,
having a sponsor and attending a greater number of AA meetings, were additionally
predictive of better outcomes (Johnson et al. 2006). It is noteworthy that patients treated in
both types of programs did not differ on CB-related proximal changes, which speaks to the
accumulating body of evidence identifying the importance of not only 12-step cognitive
shifts and engagement in prescribed behaviors, but also of common factors of change.
Two final studies tested composite measures of change in 12-step cognition and behavior.
A comparison of TSF to CB-based outpatient group therapy found that participants may
make similar changes in variables thought to be treatment-specific. Here, TSF patients had
greater within treatment change in the composite 12-step measure than CB patients and CB
patients similarly changed more in confidence/self-efficacy. However, these differences were
not maintained at 6-month follow-up. The interaction of TSF group and change in 12-step
proximal outcome, however, accounted for a significant portion of the variance in 6-month
addiction-severity (Brown et al. 2001). Therefore, differences in treatment-specific proximal
outcomes were not maintained at follow-up, but 12-step within-treatment changes were
associated with better outcome among TSF patients.
In a study that included outpatient 12-step individual and group counseling with
individuals with cocaine use disorders, Crits-Christoph et al. (2003) found that a measure of
12-step cognition and behavior was a partial mediator of 6-month drug use. The significant
relationship between the 12-step-oriented treatment condition and the composite 12-step
AA mechanisms 247
mediator path was accounted for primarily by the comparison with the supportive-
expressive, and not the cognitive therapy treatment conditions that were the other
treatments in the trial. However, the authors further noted more fine-grained temporal
analyses that did not show changes in the mediator preceding changes in drug use.
This series of studies provides a complex picture of putative mechanisms of AA and
professionally-delivered 12-step treatments. While comparison to other psychosocial
treatments, formal meditational tests, and attention to temporality are present in some
analyses, it is difficult to know the extent to which these changes are due exclusively to AA.
Thus, determining the mechanisms’ ‘‘specificity’’ (Kazdin and Nock 2003) is not possible.
Individuals appear to make predicted AA-specific cognitive and behavioral shifts, but it is
unknown whether these same changes would be accounted for equally well by other
‘‘common process’’ measures.
Social and spiritual changes
The literature reviewed thus far has examined common therapeutic factors as well as
measures of AA-specific factors. Some explicit beliefs (i.e., abstinence as a goal; Johnson
et al. 2006) and behaviors (i.e., step-work, AA reading; Tonigan 2003; Johnson et al. 2006)
have been identified and it appears that common processes (e.g., self-efficacy, commitment
to abstinence, active coping; Morgenstern et al., 1997; Kelly et al. 2000; Connors et al.
2001; Kelly 2001; Tonigan 2003) may also play a role in recovery through AA. Relatively
less work has been done in relation to constructs specifically highlighted in the AA literature.
The concept of fellowship in AA has been studied, but has most often been reframed as
social networks. Spirituality has also been of interest, but mediation studies are rare. First,
we examine the research on social factors followed by emerging work on spirituality.
Social change mechanisms. Using a large male VA sample, Humphreys et al. (1999b) found
12-month cognitive and behavioral coping, friendship quality, and network support for
abstinence to partially mediate the relationship between 12-step involvement post inpatient
treatment and 12-month substance use outcomes. Similarly, in a mixed sample of inpatients
and outpatients, Kaskutas et al. (2002) found that social network influences were partial
mediators of 12-month abstinence (see Table III). Specifically, a larger social network and
greater network support were shown to explain the relationship between greater AA
involvement and better substance use outcomes. In tests of individual predictors, reducing
pro-drinking influences and increasing AA-related support for abstinence were shown to be
important (Kastukas et al. 2002). These two studies examined variables during the same
time-period, which limits casual conclusions, but research has supported the role of social
networks, particularly AA-related abstinent role models, in ongoing recovery. In fact,
a subsequent analysis showed that among variables related to social network size (i.e., the
percent of drinkers in the network and network support for abstinence), only AA-specific
network support mediated alcohol abstinence 3 years following treatment (Bond et al.
2003). Although mutual support was not specifically measured, social support in the first
12 months of mutual-help involvement has also been shown to mediate the relationship
between affiliation and substance use 2 years later among individuals with co-occurring
substance and mental health disorders (Laudet et al. 2004).
There are a number of potential dimensions of ‘‘fellowship’’ as a mechanism of behavior
change, but with respect to social network support for abstinence, the evidence derived from
mediation research is quite compelling. The task of further research is to continue to parse
248 J. F. Kelly et al.
out the construct of fellowship, to maintain temporality in pathways, and to provide evidence
of specificity through mediation analyses that include comparison to other psychosocial
treatments or psychosocial processes. Critical too will be to place these social mechanisms
within a multi-level theoretical framework that describes how social changes influence
change mechanisms at other levels (e.g., individual-psychological and neurobiological) and
vice-versa, in what is most likely a reciprocal process that changes dynamically over time.
Spiritual change mechanisms. The 12-step philosophy is explicitly spiritual in nature, and
the stated mechanism of action for recovery, as noted previously, is a ‘‘spiritual awakening’’
attained by working the 12-step program. A majority of AA meetings also include prayers
adopted from religious organizations (AA 1984). Yet very few studies have formally tested
spiritual mediators of AA or professionally delivered 12-step treatment. Research has shown
that atheists or agnostics who attend AA do not benefit any less than more religious patients,
but they are less likely to seek recovery help through AA (Winzelberg and Humphreys 1999;
Tonigan et al. 2002; Kelly et al. 2006). Another study of mutual-help group participants
from five different organizations, including 12-step, found that degree of spiritual/religious
beliefs was unrelated to staying sober (Atkins and Hawdon 2007). A study that considered
both self-efficacy and the extent of endorsed ‘‘spiritual awakening’’ as mediators among
Project MATCH patients found that greater AA participation in the first 6 months was
associated with greater 6-month spirituality, but spirituality did not predict subsequent
outcome 3 years following treatment (Tonigan 2003). Among community mutual-help
participants with co-occurring substance use and mental health disorders attending an AA-
based dual-focused recovery group (i.e., Double Trouble in Recovery), 12-month locus of
control and sociability were mediators of 12-month abstinence, while spirituality and hope
mediated other health behaviors (Magura et al. 2003). Other research examining specific
links in the causal chain suggests a complex picture. For example, AA members were found
to have more spiritual beliefs and external locus of control than SMART Recovery (Ellis and
Velton, 1992) members (Li et al. 2000), but internal locus of control has been associated
with long-term sobriety among AA members (Murray et al. 2003). Long-term AA sobriety
has also been related to a shift from recovery-specific, to more general community, helping,
as well as endorsement of religiosity or ‘‘theism’’ (Kaskutas et al. 2003).
The evidence to date regarding spirituality is as yet too limited to support or refute
a central role for spirituality in recovery through AA; further research is needed (Galanter
2007a). From the standard scientific perspective, defining and measuring spirituality has
presented a consistent challenge. It may be also that spiritual processes interact with other
mechanisms and spiritual changes in particular may be more important at specific stages of
recovery, such as maintaining, rather than attaining, sobriety, or be related to indices other
than abstinence, such as enhancing quality of life (Spaulding and Metz 1997).
Discussion
AA has been shown to be beneficial for many different types of individuals seeking help for
alcohol problems and professionally-delivered 12-step treatment shows at least comparable
efficacy and appears more cost-effective than other treatments to which it has been
compared. AA’s pragmatic community approach was never designed to facilitate empirical
validation and many AA constructs have eluded explicit operationalization. Consequently,
what we know is colored by the research lens that observes it. However, a limited, but
AA mechanisms 249
growing body of research provides support for three broad areas through which AA may
exert its beneficial effects: (1) common factors; (2) specific AA practices; and, (3) social and
spiritual factors related to AA’s theory of change.
Common factor mechanisms
Research examining common processes tells a beginning story about the effects of AA as at
least partially transmitted through these constructs. This line of research has its theoretical
roots in social cognitive learning theory (Bandura 1986) and the more specifically related
cognitive-behavioral relapse prevention (RP) theory explicated by Marlatt and Gordon
(1985). RP highlights the importance of coping and self-efficacy as critical mediators in
increasing or decreasing the probability of relapse. Studies measuring these constructs have
primarily conducted proper mediational tests and maintained temporality of the constructs
in assessment and analysis. Findings suggest that self-efficacy, motivation for abstinence and
commitment to recovery, and behavioral coping, are mechanisms through which AA exerts
its beneficial effects. The effects of AA participation among Project MATCH treatment
participants were partially explained by self-efficacy to resist drinking for up to 3 years
following treatment (Connors et al. 2001; Tonigan 2003). Also of note is this finding’s
salience with more severely involved alcohol patients. Changes not only in self-efficacy, but
also in commitment for abstinence and active coping have mediated the effects of AA
affiliation on 6-month substance use (Morgenstern et al. 1997). Although similar constructs
were examined with adolescent inpatients, only abstinence commitment was found to
mediate subsequent outcome (Kelly et al. 2000, 2002). Mechanisms may thus differ
developmentally, but for both adolescents and adults, greater addiction severity predicts
more AA involvement.
AA practice mechanisms
Regarding the specificity of AA’s effects (Nock 2007), research on AA practices has moved
forward by conducting comparative tests of the treatment to mediator (a!b) path. These
efforts to identify ‘‘active AA ingredients’’ have, however, met with mixed results. Most
often, 12-step treatment will predict changes in AA-specific cognitions and behaviors, but
such changes will not necessarily lead to changes in alcohol or other drug use. Variables
predictive of outcome often reflect a return to proposed common factors such as
commitment to abstinence (Morgenstern et al. 1995; Johnson et al. 2006) or intention to
avoid high-risk situations (Morgenstern et al. 1995). AA-specific cognitive commitment and
belief in a higher power were associated with reduced severity of relapse (Morgenstern et al.
1995), but these were not mediational tests. One specific behavior, reading AA literature,
was found to mediate 12-month substance use (Johnson et al. 2006). However, as
mentioned previously, this immediately evokes the subsequent question as to what exactly it
is about reading AA literature that leads to better outcomes. For example, it is possible that
reading AA literature leads to increased coping, self-efficacy, and motivation for abstinence.
In sum, research on AA-specific practices is less conclusive than that on common
therapeutic factors. This is partially due to less methodological consistency regarding
mediational tests and difficulties in construct definition and measurement.
250 J. F. Kelly et al.
Social and spiritual change mechanisms
The majority of extant research on other constructs of relevance to AA’s theory of change
has addressed social and spiritual themes. It may be due to the comparative difficulty
inherent in explicating these concepts, but the evidence regarding the importance of social
support and social network changes in recovery through AA is, to date, more compelling
than that on the role of spirituality. Specifically, these measures have fallen into two classes:
general social support and changes in health-promoting networks. General friendship quality
(Humphreys et al. 1999b), network support for abstinence (Humphreys et al. 1999a), and
reduced pro-drinking influences (Kaskutas et al. 2002) have shown partial mediating effects
on later reductions in drinking, and AA-specific network support has demonstrated
mediation effects 3 years later (Bond et al. 2003). Moreover, general social support plays
a similarly vital role among individuals with co-occurring substance and mental health
disorders (Laudet et al. 2004). These aspects fit within the broader social-cognitive learning
theory framework (Bandura 1986; Moos 2008).
With respect to spirituality, a large number of studies have examined a number of themes
including religiosity, meaning seeking, hope, and internal versus external locus of control.
However, in the two mediation studies that we were able to locate, spirituality did not
mediate 3-year outcomes (Tonigan 2003), but was important to other health behaviors
among an AA-based, but dual-diagnosis focused, mutual-help group (Magura et al. 2003).
Spiritual change or awakening through acceptance of one’s condition and ultimately through
service to others is a core process in the AA literature. Research in this regard, is an
important direction for future mechanisms investigations (Pagano et al. 2004; Zemore and
Kaskutas 2004, 2008; Galanter 2007a).
Limitations of existing mechanisms research
The available research in this area contains several limitations. The vast majority of studies
are naturalistic and thus possess inherent problems of self-selection (Rosenbaum 1995).
They also mostly involve treatment samples in the early phases of stabilization and recovery.
Little is known about non-treatment AA samples or the processes involved in maintaining
recovery-related changes over the long-term, which may well differ. Also, some treatment
programs may strongly emphasize AA participation while others do not, and this is rarely
assessed, reported, or included in analyses. There may also be large differences on baseline
demographic and clinical characteristics across study samples, and on the constructs
measured as well as the measures used. Available measures frequently lack adequate
psychometric validation and may have insufficient content validity in terms of capturing the
construct of interest. Also, not all studies included here conducted mediational analyses that
took temporality into account, thus weakening firm conclusions about cause and effect
relations. All of these additional variables can create inconsistencies across findings making
conclusions difficult, especially when the number of studies is small. However, this is an
important, emerging field, and appropriate methodology is becoming better known (Stout
2007). We anticipate that studies in this area will increase in quantity, consistency, and
quality.
Finally, due to the bias on measuring purely alcohol/drug outcomes in most studies of
recovery from substance dependence, this review contained studies that focused
exclusively on how AA affects changes in alcohol/drug use behavior. Noteworthy,
however, is the fact that almost half of all AA attendees have been sober for five or more
years (AA 2008). This suggests that there may be other positive quality of life factors that
AA mechanisms 251
keep many attending AA well beyond the achievement of full sustained remission from
alcohol dependence. The influence of AA participation on other indicators of functioning
such as physical and psychological well-being may be worthy areas for future mechanisms
investigation.
Toward a more comprehensive theory of AA-related change
Perhaps existing research reflects an approach of feasibility rather than comprehensiveness
or theoretical cohesiveness. It may be out of convenience that we have looked to common
psychological constructs. These measures are tangible, have been assessed with some
success, and possess established psychometric properties (e.g., self-efficacy). Yet, we have
ignored a complex reality and phenomenology of individuals who speak of ‘‘hitting bottom’’,
‘‘guilt’’, ‘‘shame’’, ‘‘loneliness’’, ‘‘fear’’, ‘‘anger’’, ‘‘self-loathing’’, ‘‘hope’’, ‘‘freedom’’, and
‘‘gratitude’’. Researchers have struggled with defining and measuring constructs such as
these and even defining recovery itself (Betty Ford Institute Consensus Panel 2007; Laudet
2007; White 2007). As Albert Einstein once remarked: ‘‘Not everything that can be counted
counts, and not everything that counts can be counted’’. These constructs would appear to
be common aspects of what is described under the phenomenological rubric of the
addiction-recovery experience. However, these experiences are rarely explicated in theories
of recovery-related change.
The general model of therapeutic change proposed by Howard et al. (1993) may provide
a useful transtheoretical developmental framework for examining AA-related change. It
emphasizes a sequential recovery process beginning with ‘‘remoralization’’ (the enhance-
ment of subjective well-being), followed by ‘‘remediation’’ (symptomatic relief), and
‘‘rehabilitation’’ (the unlearning of pervasive, maladaptive patterns of functioning and the
learning of more adaptive approaches). This broader framework provides a structure in
which AA practices may operate producing different stage-related benefits. Table IV shows
this developmental framework, along with some potential parallel AA-related dimensions
and elements. Also described are some potential intermediate outcomes, or mechanisms,
associated with the framework and AA-related dimensions.
AA itself, at least in its core texts, may have ignored explicating perhaps its most potent
influence on individuals’ recovery – that of social group dynamics in the AA meeting, the
broader fellowship, and the expression of support that can be healing to many. Explicit in
its meeting preamble, the ‘‘...fellowship of men and women who share their experience,
strength and hope with each other ...’’ may be the most critical element of AA’s
effectiveness. The potency of the non-specific group dynamic elements associated with
the 12-step and 12-tradition organizational template (AA 1953) may be supported by the
fact that in addressing a diverse set of problems and disorders and problems of living (e.g.,
drug addiction, sexual compulsivity, gambling, mental illness, shopping, eating, debt)
popular support groups have emerged that use the same template (e.g., Narcotics
Anonymous, Sex Addicts Anonymous, Schizophrenics Anonymous, Depression
Anonymous, Gamblers Anonymous, Overeaters Anonymous). However, research on
these other organizations remains scarce. AA-influenced adaptive social context changes
are supported by available evidence. The ‘‘down-stream’’ mechanisms of these social
changes are likely to be stimulus control (i.e., absence of alcohol and related cues),
recovering role models, access to low-risk social activities, and strong social reinforcement
for abstinence (Moos 2008).
252 J. F. Kelly et al.
One prime function of AA is that the reported testimony and personal stories of ‘‘...what
happened, what we use to be like, and what we are like now’’ (AA 2001, p. 58) expose
meeting attendees to stories similar to their own that can revive fading or suppressed aversive
memories regarding past painful alcohol-related negative experiences. These ongoing
reminders in conjunction with positive (sober) experience and observable examples of
recovery may serve to continually re-motivate individuals tipping the decisional balance
toward ongoing recovery and AA participation. This common aspect of AA meetings may be
a more specific active ingredient of AA worthy of investigation.
Conclusions
Why have spiritually-oriented organizations, like AA, become so popular among alcohol and
other drug addicted individuals and not as popular among those suffering from other kinds
of mental disorders? One reason perhaps is that the disinhibiting effects from heavy alcohol
and drug use frequently generate deviations from one’s own moral code or set of values.
Repeated over and over this ‘‘Jekyll and Hyde’’ scenario can lead to a sense of profound
moral failing, self-blame and self-loathing. As structural and functional brain alterations
Table IV. Theoretical elaboration of AA’s active ingredients and mechanisms.
Therapeutic developmental stage
AA elements
Remoralization
(abstinence)
Remediation
(sobriety)
Rehabilitation
(recovery)
AA program-, or
fellowship-related
activity
Meeting attendance;
observational learn-
ing: (hearing and
seeing how others
recovered); positive
attention and strong
social reinforcement
for abstinence and
participation.
Sharing at meetings;
having and using AA
sponsor; AA connec-
tions between meet-
ings by phone or in
person; self-assess-
ment and collabora-
tive appraisal of
personal strengths
and weaknesses.
Sharing of personal
story: addiction and
recovery experiences
(chairing meetings);
continued self-
assessment, increase
spiritual practice;
helping and sponsor-
ing others; AA
service work.
A Relevant AA steps 1–3 4–9 10–12
Effect (intermediate
outcome or
mechanism)
Instillation of hope;
decreased sense of
aloneness; increase
self-esteem; realiza-
tion that personal
experience can be of
value to others.
Decreased shame, guilt,
and fear; increased
sense of belonging;
human connected-
ness (spiritual
experience); consis-
tency of behavior
with personal values
(cognitive
consonance).
Increased self-esteem;
mastery; increased
identity as recovering
person; community
and social
re-engagement.
Potential explanatory
theoretical principle
Instillation of hope;
interpersonal learn-
ing; social-cognitive
learning.
Universality; dimin-
ished self-discrepan-
cies between ‘‘actual-
ideal-ought’’ sense of
self.
Altruism (Group ther-
apy theory); helper
principle; spirituality.
AA mechanisms 253
caused by continued alcohol use exacerbate impairment over the regulation of drinking
behavior negative feelings may intensify and be deepened and reinforced further by the
reproach of affected onlookers and significant others. For many, the sense of ‘‘salvation’’,
historically and implicitly embedded within AA philosophy and in many other recovery
mutual-help societies throughout US history (White 1998), may feel cleansing and self-
soothing, providing a framework for self-forgiveness. This may account for the rather odd
AA mix of spirituality and, ‘‘alcoholism as a medical disease’’ (see the Doctors’ Opinion,
xxv; AA 1939, 2001), which may have been initially a purely pragmatic hybrid constructed to
alleviate common feelings of guilt, shame, and self-loathing that serve as barriers to salutary
change. The attenuation of these feelings and self-perceptions as a function of AA exposure
may be more specific to the spiritual/disease framework of AA and worthy of future
mechanisms research.
Despite AA’s clearly ‘‘spiritual’’ roots, language, and emphasis, this central aspect of AA
has received very limited research attention. The few studies that have examined spirituality,
have not found it to be a clear mechanism. This important dimension warrants more
research. It may be that, for the most part, underneath there is little unique or miraculous
about the specific content or format of AA; the same common therapeutic elements inherent
in most, if not all, formal and informal group and individual interventions may be similarly
operating in AA (Moos 2007, 2008). These elements appear to help individuals stay
motivated and gain the ability and confidence to cope with the demands of recovery within
a supportive context. Arguably, in the same way that we observe similar outcomes from
studies of different professional treatments, if alcohol-dependent individuals were to attend
other recovery-focused mutual-help groups (e.g., SMART Recovery, Secular Organization
for Sobriety) at the same intensity, it is conceivable that they may experience benefits similar
to those derived from AA. However, these other organizations have seldom been studied
and, as yet, cannot approach AA’s widespread availability (Kelly and Yeterian 2008). The
main benefit of AA in aiding addiction recovery may lie in its accessibility and its long-term,
‘‘extensive’’, focus (Humphreys and Tucker 2002). The fact that individuals can gain
exposure to these therapeutic elements ‘‘on demand’’ at a self-regulated dosing intensity for
as long as desired, makes AA a highly adaptive recovery management tool (White 2008).
The fact that it is free and widely available in almost every community on every day of the
week, notably during high-risk relapse periods when professional services are often not
available (i.e., on Holidays/nights/weekends), also makes AA a highly cost-effective public
health resource in helping alleviate the massive burden of disease attributable to alcohol
problems.
Declaration of interest: The authors report no conflicts of interest. The authors alone
are responsible for the content and writing of this article.
References
Alcoholics Anonymous. 1939. Alcoholics Anonymous: The story of how thousands of men and women have
recovered from alcoholism. 3rd ed. New York: Alcoholics Anonymous World Services.
Alcoholics Anonymous. 1953. Twelve steps and twelve traditions. New York: Alcoholics Anonymous World
Services.
Alcoholics Anonymous. 1947. Alcoholics Anonymous preamble. New York: Alcoholics Anonymous Grapevine,
Inc. June.
254 J. F. Kelly et al.
Alcoholics Anonymous. 1984. Pass it on: The story of Bill Wilson and how the AA message reached the world.
New York: Alcoholics Anonymous World Services.
Alcoholics Anonymous. 2001. Alcoholics Anonymous: The story of how thousands of men and women have
recovered from alcoholism. 4th ed. New York: Alcoholics Anonymous World Services.
Alcoholics Anonymous. 2005. 2004 Membership survey: A snapshot of A.A. membership. New York: Alcoholics
Anonymous World Services.
Alcoholics Anonymous. 2008. 2007 Membership survey: A snapshot of A.A. membership. New York: Alcoholics
Anonymous World Services.
Anton RF, O’Malley SS, Ciraulo DA, Cisler RA, Couper D, Donovan DM, Gastfriend DR, Hosking JD, Johnson
BA, LoCastro JS, et al. 2006. Combined pharmacotherapies and behavioral interventions for alcohol
dependence: The COMBINE Study: A Randomized controlled trial. Journal of the American Medical
Association 295(17):2003–2017.
Atkins RG, Hawdon JE. 2007. Religiosity and participation in mutual-aid support groups for addiction. Journal of
Substance Abuse Treatment 33(3):321–331.
Bandura A. 1969. Principles of behavior modification. New York: Holt, Rinehart, & Winston, Inc.
Bandura A. 1986. Social foundations of thought and action: A social cognitive theory. Englewood Cliffs,
NJ: Prentice-Hall.
Baron RM, Kenny DA. 1986. The moderator-mediator variable distinction in social psychological research:
Conceptual, strategic, and statistical considerations. Journal of Personality and Social Psychology
51(6):1173–1182.
The Betty Ford Institute Consensus Panel. 2007. What is recovery? A working definition from the Betty Ford
Institute. Journal of Substance Abuse Treatment 33(3):221–228.
Bogenschutz MP, Tonigan JS, Miller WR. 2006. Examining self-efficacy as a mediator of AA’s effect
on abstinence: Moderating effects of alcoholism typology. Journal of Studies on Alcohol and Drugs
67(4):562–567.
Bond J, Kaskutas LA, Weisner C. 2003. The persistent influence of social networks and alcoholics anonymous on
abstinence. Journal of Studies on Alcohol 64(4):579–588.
Brown BS, O’Grady KE, Farrell EV, Flechner IS, Nurco DN. 2001. Factors associated with frequency of 12-step
attendance by drug abuse clients. American Journal of Drug and Alcohol Abuse 27(1):147–160.
Bufe C. 1991. Alcoholics Anonymous: Cult or cure? Tucson, AZ, US: See Sharp Press.
Connors GJ, Tonigan JS. 2001. A longitudinal model of intake symptomatology, AA participation and outcome:
Retrospective study of the project MATCH outpatient and aftercare samples. Journal of Studies on Alcohol
62(6):817–825.
Cook TD, Campbell DT. 1979. Quasi-experimentation: Design & analysis issues for field settings. Chicago: Rand
McNally & Co.
Cooney N, Babor T, DiClemente C, Del Boca F. 2003. Clinical and scientific implications of Project MATCH.
Treatment matching in alcoholism. New York: Cambridge University Press. pp 222–237.
Crits-Christoph P, Gibbons MC, Barber J, Gallop R, Beck AT, Mercer D, Tu X, Thase ME, Weiss RD, Frank A.
2003. Mediators of outcome of psychosocial treatments for cocaine dependence. Journal of Consulting and
Clinical Psychology 71(5):918–925.
Crits-Christoph P, Siqueland L, Blaine J, Frank A, Luborsky L, Onken LS, Muenz LR, Thase ME, Weiss RD,
Gastfriend DR, et al. 1999. Psychosocial treatments for cocaine dependence: National Institute on Drug Abuse
Collaborative Cocaine Treatment Study. Archives of General Psychiatry 56(6):493–502.
Dawson DA, Grant B, Stinson FS, Chou PS. 2006. Estimating the effect of help-seeking on achieving recovery
from alcohol dependence. Addiction 101(6):1360–0443.
Day E, Gaston RL, Furlong E, Murali V, Copello A. 2005. United Kingdom substance misuse treatment workers’
attitudes toward 12-step self-help groups. Journal of Substance Abuse Treatment 29(4):321–327.
Dennis M, Godley SH, Diamond G, Tims FM, Babor T, Donaldson J, Liddle H, Titus JC, Kaminer Y, Webb C,
et al. 2004. The Cannabis Youth Treatment (CYT) Study: Main findings from two randomized trials. Journal of
Substance Abuse Treatment 27(3):197–213.
Dirac PM. 1999. The principles of Quantum Mechanics. International Series of Monographs on Physics. Oxford,
England: Oxford University Press.
Drug Strategies. 2003. Treating teens: A guide to adolescent programs. Washington, DC: Author.
Ellis A, Velton E. 1992. When AA doesn’t work for you: Rational steps to quitting alcohol. New York: Barricade
Books.
Emrick CD, Tonigan JS, Montgomery H, Little L. 1993. Alcoholics Anonymous: What is currently known?.
In: McCrady BS, Miller WR, editors. Research on Alcoholics Anonymous: Opportunities and alternatives.
Piscataway, NJ: Rutgers Center of Alcohol Studies. pp 41–76.
AA mechanisms 255
Ferri M, Amato L, Davoli M. 2006. Alcoholics Anonymous and other 12-step programmes for alcohol
dependence. Cochrane Database of Systematic Reviews 2006 Jul 19 (3), Art. No.: CD005032. DOI: 10.1002/
14651858.CD005032.pub2.
Finney JW, Hahn AC, Moos RH. 1996. The effectiveness of inpatient and outpatient treatment for alcohol abuse:
The need to focus on mediators and moderators of setting effects. Addiction 91(12):1773–1796.
Fiorentine R, Hillhouse MP. 2000. Exploring the additive effects of drug misuse treatment and twelve-step
involvement: Does twelve-step ideology matter?. Substance Use and Misuse 35(3):367–397.
Galanter M. 2007a. Spirituality and recovery in 12-step programs: An empirical model. Journal of Substance Abuse
Treatment 33(3):257–264.
Galanter M. 2007b. The concept of spirituality in relation to addiction recovery and general psychiatry.
In: Galanter M, editor. Recent developments in alcoholism: Research on Alcoholics Anonymous and spirituality
in addiction recovery. New York: Springer. pp 125–135.
Glasgow RE, Lichtenstein E, Marcus AC. 2003. Why don’t we see more translation of health promotion research
to practice? Rethinking the efficacy to effectiveness transition. American Journal of Public Health
93(8):1261–1267.
Gossop M, Harris J, Best D, Man L, Manning V, Marshall J, Strang J. 2003. Is attendance at Alcoholics
Anonymous meetings after inpatient treatment related to improved outcomes? A 6-month follow-up study.
Alcohol and Alcoholism 38(5):421–426.
Gossop M, Stewart D, Marsden J. 2007. Attendance at Narcotics Anonymous and Alcoholics Anonymous
meetings, frequency of attendance and substance use outcomes after residential treatment for drug dependence:
A 5-year follow-up study. Addiction 103:119–125.
Hill AB. 1965. The environment and disease: Association or causation? Proceedings of the Royal Society of
Medicine 58:295–300.
Howard K, Lueger R, Maling M, Martinovich Z. 1993. A phase model of psychotherapy outcome: Causal
mediation of change. Journal of Consulting and Clinical Psychology 61(4):678–685.
Humphreys K. 1997. Clinicians’ referral and matching of substance abuse patients to self-help groups after
treatment. Psychiatric Services 48(11):1445–1449.
Humphreys K. 1999. Professional interventions that facilitate 12-step self-help group involvement. Alcohol
Research and Health 23(2):93–98.
Humphreys K. 2004. Circles of recovery: Self-help organizations for addictions. Cambridge, UK: Cambridge
University Press.
Humphreys K, Huebsch PD, Finney JW, Moos RH. 1999a. A comparative evaluation of substance abuse
treatment: V. Substance abuse treatment can enhance the effectiveness of self-help groups. Alcoholism:
Clinical and Experimental Research 23(3):558–563.
Humphreys K, Mankowski E, Moos RH, Finney JW. 1999b. Do enhanced friendship networks and active
coping mediate the effect of self-help groups on substance abuse?. Annals of Behavioral Medicine 21(1):54–60.
Humphreys K, Moos R. 1996. Reduced substance abuse-related health care costs among voluntary participants in
Alcoholics Anonymous. Psychiatric Services 47:709–713.
Humphreys K, Moos RH. 2001. Can encouraging substance abuse patients to participate in self-help groups reduce
demand for health care? A quasi-experimental study. Alcoholism: Clinical and Experimental Research
25(5):711–716.
Humphreys K, Moos RH. 2007. Encouraging post-treatment self-help group involvement to reduce demand for
continuing care services: Two-year clinical and utilization outcomes. Alcoholism: Clinical and Experimental
Research 31(1):64–68.
Humphreys K, Tucker JA. 2002. Toward more responsive and effective intervention systems for alcohol-related
problems. Addiction 97(2):126–132.
Imel Z, Wampold BE, Miller SD, Fleming PR. 2008. Distinctions without a difference: Direct comparisons of
psychotherapies for alcohol use disorders. Psychology of Addictive Behaviors 22(4):533–543.
Institute of Medicine. 1990. Broadening the base of treatment for alcohol problems. Washington, DC: National
Academy Press.
Johnson JE, Finney JW, Moos RH. 2006. End-of-treatment outcomes in cognitive-behavioral treatment and
12-step substance use treatment programs: Do they differ and do they predict 1-year outcomes?. Journal of
Substance Abuse Treatment 31(1):41–50.
Kaskutas L, Ammon L, Delucchi K, Room R, Bond J, Weisner C. 2005. Alcoholics Anonymous careers: Patterns
of AA involvement five years after treatment entry. Alcoholism: Clinical and Experimental Research
29(11):1983–1990.
Kaskutas LA, Bond J, Humphreys K. 2002. Social networks as mediators of the effect of Alcoholics Anonymous.
Addiction 97(7):891–900.
256 J. F. Kelly et al.
Kaskutas L, Turk N, Bond J, Weisner C. 2003. The role of religion, spirituality and Alcoholics Anonymous in
sustained sobriety. Alcoholism Treatment Quarterly 21(1):1–16.
Kaskutas LA, Weisner C, Lee M, Humphreys K. 1999. Alcoholics Anonymous affiliation at treatment intake
among Whites and African Americans. Journal of Studies on Alcohol 60:810–816.
Kassel JD, Wagner EF. 1993. Processes of change in Alcoholics Anonymous: A review of possible mechanisms.
Psychotherapy: Theory, Research, Practice, Training. Special Issue: Psychotherapy for the Addictions
30(2):222–234.
Kazdin AE, Nock MK. 2003. Delineating mechanisms of change in child and adolescent therapy:
Methodological issues and research recommendations. Journal of Child Psychology and Psychiatry
44(8):1116–1129.
Kelly JF. 2001. Do adolescents affiliate with 12-step groups? A multivariate process model of effects. Dissertation
Abstracts International: Section-B: The Sciences and Engineering 62(3-B):1583.
Kelly JF. 2003. Mutual-help for substance use disorders: History, effectiveness, knowledge gaps & research
opportunities. Clinical Psychology Review 23(5):639–663.
Kelly JF, McCrady BS. 2009. Twelve-step facilitation in non-specialty settings. In: Galanter M, editor. Research on
alcoholism: Alcoholics Anonymous and spiritual aspects of recovery. New York: Springer. pp 1–24.
Kelly JF, Myers MG. 2007. Adolescents’ participation in Alcoholics Anonymous and Narcotics Anonymous:
Review, implications, and future directions. Journal of Psychoactive Drugs 39(3):259–269.
Kelly JF, Myers MG, Brown SA. 2000. A multivariate process model of adolescent 12-step attendance and
substance use outcome following inpatient treatment. Psychology of Addictive Behaviors 14(4):376–389.
Kelly JF, Myers MG, Brown SA. 2002. Do adolescents affiliate with 12-step groups? A multivariate process model
of effects. Journal of Studies on Alcohol 63(3):293–304.
Kelly JF, Myers MG, Rodolico J. 2008. What do adolescents exposed to Alcoholics Anonymous think about
12-step groups?. Journal of Substance Abuse 29(2):53–62.
Kelly JF, Stout R, Zywiak W, Schneider R. 2006. A 3-year study of addiction mutual-help group participation
following intensive outpatient treatment. Alcoholism: Clinical and Experimental Research 30(8):1381–1392.
Kelly JF, Yeterian JD. 2008a. Mutual-help groups. In: O’Donohue W, Cunningham JR, editors. Evidence-based
adjunctive treatments. New York: Elsevier. pp 61–105.
Kelly JF, Yeterian JD, Myers MG. 2008b. Treatment staff referrals, participation expectations, and perceived
benefits and barriers to adolescent involvement in 12-step groups. Alcoholism Treatment Quarterly
26(4):427–449.
Kissin W, McLeod C, McKay J. 2003. The longitudinal relationship between self-help group attendance and
course of recovery. Evaluation and Program Planning 26(3):311–323.
Khantzian EJ, Mack JE. 1994. How AA works and why it’s important for clinicians to understand. Journal of
Substance Abuse Treatment 11(2):77–92.
Kownacki RJ, Shadish WR. 1999. Does Alcoholics Anonymous work? The results from a meta-analysis of
controlled experiments. Substance Use & Misuse 34(13):1897–1916.
Laudet AB. 2003. Attitudes and beliefs about 12-step groups amount addiction treatment clients and clinicians:
Toward identifying obstacles to participation. Substance Use and Misuse 38(14):2017–2047.
Laudet AB. 2007. What does recovery mean to you? Lessons from the recovery experience for research and
practice. Journal of Substance Abuse Treatment 33(3):243–56.
Laudet AB, Cleland CM, Magura S, Vogel HS, Knight EL. 2004. Social support mediates the effects of dual-
focus mutual aid groups on abstinence from substance use. American Journal of Community Psychology
34(3–4):175–185.
Li EC, Feifer C, Strohm M. 2000. A pilot study: Locus of control and spiritual beliefs in Alcoholics Anonymous
and SMART Recovery members. Addictive Behaviors 25(4):633–640.
Longabaugh R, Wirtz PW, Zweben A, Stout RL. 1998. Network support for drinking, Alcoholics Anonymous and
long-term matching effects. Addiction 93(9):1313–1333.
MacKinnon DP, Dwyer JH. 1993. Estimating mediated effects in prevention studies. Evaluation Review
17:144–158.
MacKinnon DP, Lockwood CM, Hoffman JM. 2002. A comparison of methods to test mediation and other
intervening variable effects. Psychological Methods 7(1):83–104.
Magura S, Knight E, Vogel HS, Mahmood D, Laudet AB, Rosenblum A, et al. 2003. Mediators of effectiveness
in dual-focus self-help groups. American Journal of Drug and Alcohol Abuse 29(2):301–322.
Ma¨ kela K. 1996. Alcoholics Anonymous as a mutual-help movement: A study in eight societies. Madison, WI:
University of Wisconsin Press.
Marlatt GA, Gordon JR. (Eds.). 1985. Relapse prevention: Maintenance strategies in the treatment of addictive
behaviors. New York: Guilford Press.
AA mechanisms 257
McCrady BS. 1994. Alcoholics Anonymous and behavior therapy: Can habits be treated as diseases? Can diseases
be treated as habits?. Journal of Consulting and Clinical Psychology 62(6):1159–1166.
McKellar JD, Stewart E, Humphrey KN. 2003. Alcoholics Anonymous involvement and positive alcohol-related
outcomes: Cause, consequence, or just a correlation? A prospective 2-year study of 2,319 alcohol-dependent
men. Journal of Consulting and Clinical Psychology 71(2):302–308.
McElrath D. 1997. The Minnesota model. Journal of Psychoactive Drugs 29(2):141–144.
Miller NS, Ninonuevo FG, Klamen DL, Hoffmann NG. 1997. Integration of treatment and post-treatment
variables in predicting results of abstinence-based outpatient treatment after one year. Journal of Psychoactive
Drugs 29(3):239–248.
Miller WR, Willbourne P. 2002. Mesa Grande: A methodological analysis of clinical trials of treatment for alcohol
use disorders. Addiction 97(3):265–277.
Moos R. 2003. Social contexts: Transcending their power and their fragility. American Journal of Community
Psychology 31(1-2):1–13.
Moos R. 2007. Theory-based active ingredients of effective treatments for substance use disorders. Drug and
Alcohol Dependence 88:109–121.
Moos RH. 2008. Active ingredients of substance use-focused self-help groups. Addiction 103(3):387–396.
Moos RH, Moos BS. 2004. Long-term influence of duration and frequency of participation in Alcoholics
Anonymous on individuals with alcohol use disorders. Journal of Consulting and Clinical Psychology
72(l):81–90.
Moos RH, Moos BS. 2006. Participation in treatment and Alcoholics Anonymous: A 16-year follow-up of initially
untreated individuals. Journal of Clinical Psychology 62(6):735–750.
Morgenstern J, Bates ME. 1999. Effects of executive treatment. Journal of Studies on Alcohol 60:846–855.
Morgenstern J, Bux D, Labouvie E, Blanchard KA, Morgan TJ. 2002. Examining mechanisms of action in 12-step
treatment: The role of 12-step cognitions. Journal of Studies on Alcohol 63(6):665–672.
Morgenstern J, Frey RM, McCrady BS, Labouvie E, Neighbors CJ. 1995. Examining mediators of change in
traditional chemical dependency treatment. Journal of Studies on Alcohol 57(1):53–64.
Morgenstern J, Kahler CW, Epstein E. 1998. Do treatment process factors mediate the relationship between Type
A–Type B and outcome in 12-step oriented substance abuse treatment?. Addiction 93:1765–1776.
Morgenstern J, Labouvie E, McCrady BS, Kahler CW, Frey RM. 1997. Affiliation with Alcoholics Anonymous
after treatment: A study of its therapeutic effects and mechanisms of action. Journal of Consulting and Clinical
Psychology 65(5):768–777.
Morgenstern J, McKay J. 2007. Rethinking the paradigms that inform behavioral treatment research for substance
use disorders. Addiction 102(9):1377–1389.
Murray T, Malcarne V, Goggin K. 2003. Alcohol-related god/higher power control beliefs, locus of control, and
recovery within the Alcoholics Anonymous paradigm. Alcoholism Treatment Quarterly 21(3):23–39.
Nock M. 2007. Conceptual and design essentials for evaluating mechanisms of change. Alcoholism: Clinical and
Experimental Research 31(3):4S–12s.
Ouimette PC, Finney JW, Gima K, Moos RH. 1999. A comparative evaluation of substance abuse treatment III.
Examining mechanisms underlying patient-treatment matching hypotheses for 12-step and cognitive-behavioral
treatments for substance abuse. Alcoholism: Clinical and Experimental Research 23(3):545–551.
Ouimette PC, Finney JW, Moos RH. 1997. Twelve-step and cognitive-behavioral treatment for substance abuse:
A comparison of treatment effectiveness. Journal of Consulting and Clinical Psychology 65(2):230–240.
Owen PL, Slaymaker V, Tonigan JS, McGrady BS, Epstein EE, Kaskuatas LA, Humphreys K, Miller WR. 2003.
Participation in AA: Intended and unintended change mechanisms. Alcoholism: Clinical and Experimental
Research 27:524–532.
Pagano ME, Friend KB, Tonigan JS, Stout R. 2004. Sponsoring others in Alcoholics Anonymous and avoiding a
drink in the first year following treatment: Findings from Project MATCH. Journal of Studies on Alcohol
65(6):766–773.
Pagano ME, Phillips KA, Stout RL, Menard W, Piliavin JA. 2007. Impact of helping behaviors on the course of
substance-use disorders in individuals with body dysmorphic disorder. Journal of Studies on Alcohol and Drugs
68(2):291–295.
Peele S. 1990. Research issues in assessing addiction treatment efficacy: How cost effective are Alcoholics
Anonymous and private treatment centers?. Drug Alcohol Depend 25(2):179–182.
Project MATCH Research Group. 1993. Project MATCH (Matching Alcoholism Treatment to Client
Heterogeneity): Rationale and methods for a multisite clinical trial matching patients to alcoholism treatment.
Alcoholism: Clinical Experimental Research 17(6):1130–1145.
Project MATCH Research Group. 1997. Matching alcoholism treatments to client heterogeneity: Project MATCH
posttreatment drinking outcomes. Journal of Studies on Alcohol 58(1):7–29.
258 J. F. Kelly et al.
Project MATCH Research Group. 1998a. Matching alcoholism treatments to client heterogeneity:
Treatment main effects and matching effects on drinking during treatment. Journal of Studies on Alcohol
59:631–639.
Project MATCH Research Group. 1998b. Matching alcoholism treatments to client heterogeneity:
Project MATCH three-year drinking outcomes. Alcoholism: Clinical and Experimental Research
22(6):1300–1311.
Riessman F. 1965. The ‘helper therapy’ principle. Social Work 10:27–32.
Roman PM, Blum TC. 1999. National Treatment Center Study (Summary 3). Athens, GA: University of Georgia.
Room R, Greenfield T. 1993. Alcoholics Anonymous, other 12-step movements and psychotherapy in the US
population, 1990. Addiction 88(4):555–562.
Rosenbaum PR. 1995. Observational studies. New York: Springer-Verlag.
Sobel ME. 1982. Asymptotic confidence intervals for indirect effects in structural equation models. In: Leinhardt S,
editor. Sociological Methodology. Washington DC: American Sociological Association. pp 290–312.
Spaulding AD, Metz GJ. 1997. Spirituality and quality of life in Alcoholics Anonymous. Alcoholism Treatment
Quarterly 15(1):1–14.
Stout RL. 2007. Advancing the analysis of treatment process. Addiction 102:1539–1545.
Substance Abuse & Mental Health Services Administration. 2008. National registry of substance abuse programs
and practices. Washington, D.C: US Department of Health and Human Services.
Timko C, Moos RH, Finney J, Lesar MD. 2000. Long-term outcomes of alcohol use disorders: Comparing
untreated individuals with those in Alcoholics Anonymous and formal treatment. Journal of Studies on Alcohol
6(4):529–540.
Tonigan JS. 2003. Spirituality and Alcoholics Anonymous practices three and ten years after project MATCH.
Alcoholism: Clinical and Experimental Research 26(5, Supplement):660A.
Tonigan JS, Connors GJ, Miller WR. 2003. Participation and involvement in Alcoholics Anonymous. In: Babor T,
DelBoca F, editors. Treatment matching in alcoholism. New York: Cambridge University Press. pp 184–204.
Tonigan JS, Miller WR, Schermer C. 2002. Atheists, agnostics, and Alcoholics Anonymous. Journal of Studies on
Alcohol 63(5):534–541.
Tonigan JS, Toscova R, Miller WR. 1996. Meta-analysis of the literature on Alcoholics Anonymous: Sample and
study characteristics moderate findings. Journal of Studies on Alcohol 57(1):65–72.
UKATT Research Team.. 2008. UK alcohol treatment trial: Client-treatment matching effects. Addiction
103(2):228–238.
Vaillant GE. 1983. The natural history of alcoholism: Causes, patterns, and paths to recovery. Cambridge, MA:
Harvard University Press.
Weisner C, Greenfield T, Room R. 1995. Trends in the treatment of alcohol problems in the US general
population, 1979 through 1990. American Journal of Public Health 85(1):55–60.
White WL. 1998. Slaying the dragon: The history of addiction treatment and recovery in America. Bloomington:
Chestnut Health Systems.
White WL. 2007, Addiction recovery: Its definition and conceptual boundaries Journal of Substance Abuse
Treatment 33(3):229–241.
White WL. 2008. Recovery management and recovery-oriented systems of care: Scientific rationale and promising
practices. Northeast Addiction Technology Transfer Center, the Great Lakes Addiction Technology Transfer
Center, and the Philadelphia Department of Behavioral Health/Mental Retardation Services.
Willenbring ML. 2007. A broader view of change in drinking behavior. Alcoholism: Clinical and Experimental
Research 31(Suppl 3):84S–86S.
Winzelberg A, Humphreys K. 1999. Should patients’ religiosity influence clinicians’ referral to 12-step self-help
groups? Evidence from a study of 3,018 male substance abuse patients. Journal of Consulting and Clinical
Psychology 67(5):790–794.
Yalom ID. 1995. The theory and practice of group psychotherapy. 4th ed. New York: Basic Books.
Yalom ID, Leszcz M. 2005. The theory and practice of group psychotherapy. 5th ed. New York: Basic Books.
Zemore SE, Kaskutas LA. 2004. Helping, spirituality and Alcoholics Anonymous in recovery. Journal of Studies on
Alcohol 65(3):383–391.
Zemore SE, Kaskutas LA. 2008. Peer helping and 12-step involvement in day hospital and residential programs.
Substance Use and Misuse 43(12):1882–1903.
AA mechanisms 259
