Content uploaded by Alison Beck
Author content
All content in this area was uploaded by Alison Beck on Feb 15, 2017
Content may be subject to copyright.
SMART RECOVERY 1
Running head: SMART RECOVERY SYSTEMATIC REVIEW
Systematic Review of SMART Recovery:
Outcomes, Process Variables and Implications for Research
Alison K. Beck, Erin Forbes and Amanda L. Baker
University of Newcastle
Peter J. Kelly and Frank P. Deane
University of Wollongong
Anthony Shakeshaft
NDARC, University of New South Wales
David Hunt
SMART Recovery Australia (Employee)
John F. Kelly
Massachusetts General Hospital, Recovery Research Institute, Harvard Medical School
Pre publication copy as accepted by Psychology of Addictive Behaviors
This article may not exactly replicate the authoritative document published in the APA journal.
It is not the copy of record.
The authoritative document can be located on the APA website by using the following link:
http://dx.doi.org/10.1037/adb0000237
© 2017 American Psychological Association
SMART RECOVERY 2
AUTHOR NOTE
Alison Beck is supported by the NHMRC Centre of Research Excellence for Mental
Health and Substance Use. The funder has no involvement in the development, conduct or
interpretation of this systematic review. We would like to thank Mary Kumvaj for her
assistance with designing the search strategy.
Alison Beck and Erin Forbes have no competing interests to declare. Amanda Baker,
Peter Kelly, Frank Deane, Anthony Shakeshaft and John Kelly are all members of the
SMART Recovery Australia Research Advisory Committee. Amanda Baker is a SMART
Recovery Australia Board Member. David Hunt was employed by SMART Recovery as the
area coordinator for South Australia, Tasmania and Victoria.
Preliminary findings from this paper were presented as part of a symposium conducted
at the 35th Conference of the Australasian Professional Society on Alcohol and other Drugs.
Perth, Australia (2015). A link to the symposium presentation is available on the SMART
Recovery Australia website: http://smartrecoveryaustralia.com.au/2015/11/
Correspondence concerning this article should be addressed to: Dr Alison Beck, Level
5, McCauley Centre, Calvary Mater Newcastle, Edith St, Waratah, 2298
Alison.Beck@newcastle.edu.au
SMART RECOVERY 2
ABSTRACT
Clinical guidelines recommend Self-Management and Recovery Training (SMART
Recovery) and 12-step models of mutual aid as important sources of long term support for
addiction recovery. Methodologically rigorous reviews of the efficacy and potential
mechanisms of change are available for the predominant 12-step approach. A similarly
rigorous exploration of SMART Recovery has yet to be undertaken. We aim to address this
gap by providing a systematic overview of the evidence for SMART Recovery in adults with
problematic alcohol, substance and/ or behavioral addiction, comprising a commentary on
outcomes assessed, process variables, feasibility, current understanding of mental health
outcomes and a critical evaluation of the methodology. We searched six electronic peer-
reviewed and four grey literature databases for English language SMART Recovery literature.
Articles were classified, assessed against standardized criteria and checked by an independent
assessor. Twelve studies (including three evaluations of effectiveness) were identified.
Alcohol related outcomes were the primary focus. Standardised assessment of non-alcohol
substance use was infrequent. Information about behavioral addiction was restricted to limited
prevalence data. Functional outcomes were rarely reported. Feasibility was largely indexed by
attendance. Economic analysis has not been undertaken. Little is known about the variables
that may influence treatment outcome, but attendance represents a potential candidate.
Assessment and reporting of mental health status was poor. Although positive effects were
found, the modest sample and diversity of methods prevents us from making conclusive
remarks about efficacy. Further research is needed to understand the clinical and public health
utility of SMART as a viable recovery support option.
Keywords: Systematic review; SMART Recovery; Mutual Aid; Self-help groups; Addiction
Protocol Registration: PROSPERO CRD42015025574
SMART RECOVERY 3
BACKGROUND
The burden of addiction is considerable, with a profound and detrimental impact on
mortality (Whiteford et al., 2010), health, relationships, employment and quality of life
(Black, Shaw, McCormick, & Allen, 2013; Laudet, 2011). Together, the harms from alcohol,
substances and behavioral addictions such as gambling have been estimated to cost over $28
billion per year (Australian Government Productivity Commission, 2010; Manning, Smith &
Mazerolle, 2013). As the course of addiction is often chronic and characterised by multiple
relapses (Sheedy & Whitter, 2009), accessible, long term support is important.
‘Mutual aid’ programmes represent one avenue for accessing such support.
‘Mutual aid’ refers to the social, emotional and informational support provided by, and to,
group members undergoing recovery from addiction (Public Health England, 2015). Twelve-
step models (e.g. Alcoholics Anonymous) are the largest and most researched source of
addiction mutual aid. Within the 12-step model, addiction is conceptualised as a medical and
spiritual disease and a key feature of the recovery process is relinquishing control to a user-
defined higher power (Donovan, Ingalsbe, Benbow & Daley, 2013). For adults with
moderate/ severe alcohol use disorder, evidence suggests that improvement following
community 12-step participation is at least equivalent to that of professional interventions
(Ferri, Amato, & Davoli, 2006; Humphreys et al., 2004; Kelly, Magill, & Stout, 2009), and in
the longer term, active participation increases the likelihood of full sustained remission and
recovery (Moos & Moos, 2006; Public Health England, 2015). However, individuals may fail
to engage with 12-step groups, for a variety of reasons including a mismatch between
personal beliefs and the 12-step philosophy (Horvath & Sokoloff, 2011). To enhance
engagement, clinical guidelines advocate for tailored addiction support that accounts for
individual needs and preferences (e.g. NICE, 2012; NICE, 2011). Choice over mutual aid
SMART RECOVERY 4
support options is therefore important and fortunately, alternatives are available (see
Humphreys et al., 2004 for a review).
One such alternative is Self-Management and Recovery Training (“SMART
Recovery”). SMART Recovery is one model recommended alongside 12-step by clinical
guidelines for both addiction (NICE, 2012; NICE, 2011) and dual diagnosis (Mills et al.,
2010). SMART Recovery is a not-for-profit organisation that provides mutual aid in group
and on-line formats (Horvath & Yeterian, 2012). SMART Recovery focuses on self-
empowerment and adopts key principles (e.g. self-efficacy) and therapeutic approaches (e.g.
motivational interviewing and cognitive behavioural therapy) shown to be effective in
promoting recovery from addiction (see Australian Psychological Society, 2010 for a recent
review of the efficacy of these approaches). Unlike 12-step approaches that offer addiction
specific support groups (e.g. alcoholics anonymous, narcotics anonymous, gamblers
anonymous), SMART Recovery offers support for a range of addictive behaviours (Horvath
& Yeterian, 2012).
OBJECTIVES AND IMPORTANCE OF THE CURRENT REVIEW
Relative to the methodologically rigorous systematic reviews of the efficacy (Ferri et
al., 2006) and potential mechanisms of change (Kelly et al., 2009) of 12-step models, to date,
reviews of SMART Recovery (e.g. Horvath & Yeterian, 2012) are narrative in nature and
tend to focus on the origins, development and principles of SMART Recovery. A systematic
approach to identifying, summarising and evaluating the quality of evidence for SMART
Recovery has yet to be undertaken. Furthermore, since Horvath’s 2012 narrative review, the
evidence base has doubled - an additional four studies have been published including the first
randomised controlled trial (RCT).
The current review is reported here following established guidelines for conducting
systematic reviews (Moher, Liberati, Tetzlaff & Altman, 2009). We advance the current
SMART RECOVERY 5
literature by using an established methodology (Higgins & Green, 2011) to provide a
comprehensive, systematic overview and critical evaluation of both published and
unpublished evidence for SMART Recovery and include recommendations for future
research. We aim to explore whether, for adults with experience of substance and/ or
behavioral addiction(s), SMART Recovery results in changes in the severity of addiction and
its consequences and whether any observed changes are influenced by process variables (e.g.
treatment engagement). To help guide our understanding of the applicability of these research
findings to ‘real world’ settings, we will also describe the feasibility of the SMART Recovery
approach, including a commentary on economic outcomes and service user satisfaction. To
better inform research and clinical care, we will also describe the treatment contexts and
clinical presentations of participants (e.g. addiction only vs. dual diagnosis). Given not only
the high prevalence, but also considerable impact of comorbid mental health conditions on
addiction recovery (Mills et al., 2010), the assessment and/ or change in mental health status
reported within the research on SMART Recovery will also be discussed.
METHODS
The current systematic review is exempt from review by a Research Ethics Committee/
Institutional Review Board as no primary data collection was undertaken from study
participants.
Criteria for Selecting Studies for this Review:
Methods were informed by Cochrane Guidelines for systematic reviews (Higgins &
Green, 2011) and are extensively detailed in the review protocol (Beck et al., 2016). The
population of interest was adults (aged > 18) attending SMART Recovery with current or past
problematic experience of at least one addictive behavior (substance and/ or behavioral).
Study participants could be residing in community, rehabilitation, treatment and/or
correctional settings. The intervention of interest (SMART Recovery), could be delivered in a
SMART RECOVERY 6
group format, of any intensity or frequency (including stand alone and/ or as an adjunct), by a
lay or professional facilitator. SMART Recovery could be compared to inactive and/ or active
conditions of any intensity, frequency and delivery method. Evaluations without a comparator
group were also eligible. Studies had to provide data for SMART Recovery participants for at
least one of the following: a) severity of addiction and its consequences, b) process variables
(e.g. treatment engagement) or c) feasibility (see Beck et al., 2016 for definitions). We
included the following designs: randomised controlled trials (cluster and parallel design);
cross-over trials; case series or case controls; one-arm trials; non-randomised trials; cross-
sectional or cohort studies and case reports. Qualitative only designs were not included.
Search Methods for Identification of Studies
Figure 1 summarises the procedure used to identify studies, including databases
searched, search terms used, exclusion criteria and study classification. The full MEDLINE
search strategy is provided in Online Supplementary File 1. Abstract, title, key words and
subject headings specific to each database were searched. Subject headings were exploded.
No limits were placed on publication year. Publications had to be available in English.
Reference lists of identified publications were hand searched to identify any additional
publications. All publications were organised in reference manager Endnote. The searches
were performed in May-June 2015 and re-run in April 2016. Articles were identified and
classified according to the following steps:
Step 1: Identification and Screening
The primary author performed the searches and reviewed the titles and/or abstracts of
the identified 989 publications and used the inclusion criteria to exclude clearly ineligible
articles. If eligibility was unclear, the full text article was accessed.
Step 2: Eligibility and Classification
SMART RECOVERY 7
The full text versions of 118 publications were manually reviewed. Eighty-one publications
were excluded. The remaining 37 were classified as ‘evaluation’, ‘review’, ‘discussion’ or
‘other’ according to published definitions (see Beck et al., 2016).
Step 3: Cross Checking.
The 118 publications from step two were cross-checked by having a research assistant
(EF) blinded to the results of the initial classification, reclassify the publications. The articles
excluded in step one were not cross-checked because they were not relevant to the review.
The 12 studies independently classified by AKB and EF as ‘evaluation’ were retained for
further examination.
Data Collection and Analysis
Data extraction was performed by AKB and checked by EF. Extraction forms were
piloted on several papers and modified as needed before use. When multiple reports of the
same study were identified (Brooks & Penn, 2003; Penn & Brooks, 2000) data from each
report was extracted separately and then combined across multiple data collection forms.
Criteria for data extraction (detailed in the protocol, Beck et al., 2016) were adapted from the
Downs and Black Scale (Downs & Black, 1998) and the Cochrane Handbook for Systematic
Reviews (Higgins & Green, 2011).
Assessment of Methodological Quality and Risk of Bias
Assessment of quality and bias was undertaken independently by AKB and EF.
Downs and Black Scale
All non-randomised studies were evaluated using this 27-item checklist (which is
recommended by the Cochrane Guidelines for assessing the quality of non-randomised
trials; Higgins & Green, 2011). Consistent with previous concerns about the two items
regarding blinding of subjects and therapists (e.g. Baker, Hiles, Thornton, Hides &
Lubman, 2012) these items were not used. Scoring of item 27 (power) was unclear so the
SMART RECOVERY 8
following convention was used: 0=no power calculation reported; 1=power analysis
reported, but insufficient power achieved and 2=power analysis reported and sufficient
power achieved. Item ratings were summed for a total maximum score of 27, with higher
scores reflecting greater methodological quality. Raters achieved 80.5% consistency in their
initial independent ratings. Discrepancies were then resolved following discussion and
consensus ratings obtained for all items.
PEDro Scale
The one RCT identified was also assessed against the 11 item Physiotherapy
Evidence Database (PEDro) scale (Centre for Evidence Based Physiotherapy, 2009) - a
widely implemented and validated tool for assessing the quality of randomised trials.
Again, the two items regarding blinding were deemed inappropriate (e.g. Baker et al., 2012)
and not scored. The remaining items were assigned a yes (1 point) or no (0 points) rating
(Centre for Evidence Based Physiotherapy, 2009), generating a quality score from 0 to 8
points. Raters achieved 100% consistency in their independent ratings.
Cochrane Collaboration’s Risk of Bias Tool
Risk of bias (within and across all studies) was assessed using the Collaboration’s
Risk of Bias tool (Higgins & Green, 2011). This tool provides an overall risk of bias (‘high’,
‘low’ or ‘unclear’) based on the following methodological characteristics: sequence
generation, allocation concealment, blinding of participants and personnel, blinding of
outcome assessment, incomplete outcome data, selective outcome reporting and ‘other’
potential sources of bias. Raters achieved 89.2% consistency in their independent ratings.
Discrepancies were resolved by discussion, and consensus ratings across all items obtained.
RESULTS
Description of Studies
Twelve studies were identified (8 published in peer reviewed journals, 4 unpublished
SMART RECOVERY 9
dissertations). The studies were predominantly cross-sectional (8 of 12). The effectiveness of
SMART Recovery has been explored in one RCT (Hester, Lenberg, Campbell & Delaney,
2013), one pre-post design (described across two publications; Brooks & Penn, 2003; Penn
& Brooks, 2000) and one quasi-experimental pseudo-prospective study (Blatch, O’Sullivan,
Delaney & Rathbone, 2016). Concurrent mental illness and substance use disorder was the
focus of only one study (described across two papers, Brooks & Penn, 2003; Penn & Brooks,
2000).
The SMART Recovery intervention and comparison condition was often poorly
described. Intervention content and delivery methods were only clearly detailed for SMART
Recovery informed or adapted interventions (Blatch et al., 2016; Brooks & Penn, 2003; Penn
& Brooks, 2000; Hester et al., 2013). For community based SMART Recovery groups (and
comparison conditions), assessment and/or reporting of SMART Recovery tools, strategies,
content, delivery methods, facilitator experience and training was scarce. Thus, adherence to
SMART Recovery guidelines was unclear. Assessment and reporting of concurrent
treatment (including pharmacological and psychological) for addiction and/ or mental health
was also lacking.
Outcomes Assessed
Severity of Addiction and its Consequences
The severity of addiction and its consequences tended to be assessed in terms of
quantity, frequency and/ or duration of use. Other indices (e.g. number of hospitalisations
and recidivism) were assessed in three studies (Blatch et al., 2016; Brooks & Penn,
2003/Penn & Brooks, 2000; Milin, 2007) and quality of life in only one study (Brooks &
Penn, 2003/ Penn & Brooks, 2000). Despite high comorbidity between mental health
conditions and substance misuse, standardised assessment of mental health status occurred in
only three studies (Brooks & Penn, 2003/Penn & Brooks, 2000; Hester et al., 2013; Kelly,
SMART RECOVERY 10
Deane & Baker, 2015).
Alcohol related outcomes were the primary focus of the literature. However, only
three studies utilised standardised assessment of alcohol use (Brooks & Penn, 2003; Hester
et al., 2013; Milin, 2007). The remainder relied on subjective accounts, including self-
reported duration of ‘abstinence’, ‘sobriety’ (Atkins & Hawdon, 2007; Guarnotta, 2015;
Trumble, 2015) and ‘problems’ (Kelly et al., 2015; O’Sullivan et al., 2015). Only three
studies explicitly reported on non-alcohol substance use (Brooks & Penn, 2003; Kelly et al.,
2015; Milin, 2007). Within these, the focus was on illicit drugs, to the relative neglect of
other common forms of substance use like smoking (reported only by Kelly et al., 2015) and
misuse of prescription medication (reported only by Milin, 2007). Brooks and colleagues
were the only authors to utilise a standardised clinical interview to assess non-alcohol
substance use (Brooks & Penn, 2003). Theirs was also the only study to employ
physiological verification of alcohol and/or substance use (urine analysis; Brooks & Penn,
2003/ Penn & Brooks, 2000). The severity and impact of behavioral addictions has yet to be
assessed, but two studies did provide limited prevalence data (Kelly et al., 2015; O’Sullivan
et al., 2015).
Process Variables
Treatment engagement was the most common process variable assessed (10 of the
12), but only three studies explored its relationship to treatment outcome (Blatch et al., 2016;
Brooks & Penn, 2003/ Penn & Brooks, 2000; Hester et al., 2013). Other process variables
assessed included elements of the therapeutic process (e.g. readiness to change, group
cohesion), locus of control, spirituality/ religiosity, self-efficacy, resilience, coping and
social support, but few studies (Atkins & Hawdon, 2007; Bogdonoff, 2003; Guarnotta, 2015;
Milin, 2007) explored the relationship between these and treatment outcome.
Feasibility
SMART RECOVERY 11
Feasibility tended to be indexed by attendance, including the number of sessions
(Hester et al., 2013), duration of involvement (Brooks & Penn, 2003/ Penn & Brooks, 2000;
Kelly et al., 2015; Li et al., 2000; Milin et al., 2007; O Sullivan et al., 2015) and proportion
of participants accessing different types of mutual aid (Blatch et al.,2016). No studies
assessed economic outcomes. Two studies (Milin, 2007; O’Sullivan et al., 2015) did report
some qualitative data regarding satisfaction (Table 3).
Methodological Quality and Risk of Bias in Included Studies
The one identified RCT (Hester et al., 2013) received 6 from a possible 8 points
using the PEDro Scale and 22 from a maximum 27 points using the Downs & Black Scale.
The methodological quality of non-randomised trials varied considerably (see Table 2), with
Downs and Black ratings ranging from 8-19.
The level of risk of bias is presented separately for each study in Figure 2 and as a
combined assessment of ratings in Figure 3. Online Supplementary File 2 contains
justification for each risk assessment. Hester et al. (2013) was the only study to report both
appropriate sequence generation and allocation concealment, thereby, the only study
assessed as having a low risk of selection bias. Masking of participants and providers in
trials of psychological interventions is generally not possible, and therefore there was a high
risk of bias in this domain. However Hester et al. (2013), Brooks et al. (2000, 2003) and
Blatch et al. (2016) used objective outcome assessment and/ or collateral information, and
were therefore deemed to be at low risk of performance bias. Risk of detection bias was
assessed as low in only one paper (Hester et al., 2013) and three provided insufficient
information to make a determination (Blatch et al., 2016; Brooks & Penn, 2003/ Penn &
Brooks, 2000). Four papers adequately addressed attrition and missing data and were
deemed low risk of attrition bias (Bogdonoff, 2003; Brooks & Penn, 2003/ Penn & Brooks,
2000, Hester et al., 2013), while the remaining eight provided insufficient information. Risk
SMART RECOVERY 12
of reporting bias was deemed low in ten studies, as all planned outcomes were reported (or
an explanation provided) and post-hoc analyses were clearly specified.
SMART Recovery Participant Characteristics
A total of 7655 participants were recruited to the included evaluations (1177 SMART
Recovery, 6478 comparison conditions). Baseline demographic and clinical characteristics
of SMART Recovery participants are presented in Table 1. Mean age ranged from 34.2 to
51. The gender distribution (% males) ranged from 39% to 71%. The majority of participants
were Caucasian. Between 25% and 82% attained at least college/ graduate degree level
certification. Employment (full or part time) ranged from 30.7% to 63%. The proportion of
individuals who were single or divorced ranged from 23% to 63.9%. The dual diagnosis
population had fewer years of education (M=11.6 years education), were less likely to be
employed (full or part time; 20.4%) and more likely to be single or divorced (80%). From
the data available, mental health problems and impairment were common.
Average years of alcohol use ranged from 10 to 19.25 years. The majority of
participants reported prior treatment and/or multiple quit attempts. The two studies that used
the AUDIT at baseline both reported scores >20 (Hester et al., 2013; Milin, 2007), consistent
with hazardous alcohol use and likely dependence. Amphetamines (7.3%) and marijuana
(3.3%), were variously identified as the most common self-reported primary non-alcohol
substance of abuse. Self-reported multi-drug use was as high as 70%. In one study, 24.4% of
participants endorsed behavioral addiction (sex, pornography, food, spending) alone, or in
combination with drugs and/ or alcohol (O’Sullivan et al., 2015). In another, food (10.5%),
gambling (9.7%) and shopping (6.5%) emerged as the top three non-substance problematic
behaviours (Kelly et al., 2015).
Effects of Interventions
SMART RECOVERY 13
A summary of key findings for the four types of comparisons identified (community
SMART Recovery groups versus an online intervention, alone or in combination; SMART
Recovery informed interventions versus active and/ or control comparison conditions;
community SMART Recovery groups versus other forms of mutual aid and community
SMART Recovery groups without a comparison condition) are presented in Table 2 and
discussed in turn below.
Summary of Evidence Comparing SMART Recovery to a SMART Recovery Informed Online
Intervention (Alone or In Combination)
Hester and colleagues (2013) conducted the sole identified RCT and compared
SMART Recovery to a SMART Recovery informed web application (‘Overcoming
Addictions’; OA), alone or in combination. At three month follow-up, SMART Recovery
participants with a history of problematic alcohol use demonstrated significant improvement
in all outcome measures (percent days abstinent, standard drinks per drinking day and alcohol
related problems; Hester et al., 2013). The level of improvement did not significantly differ
between treatment conditions (Hester et al., 2013). Although mental health symptoms were
recorded at baseline (Mean BSI=19.35, SD=12.5), change across time was not assessed.
In the SMART Recovery Only condition, the number of meetings attended was
identified as a significant predictor of improvement in all three primary outcomes (Hester et
al., 2013). For the OA+SMART Recovery group the total amount of support (including
SMART Recovery/ other meetings and counsellor visits) emerged as the strongest predictor
of alcohol related change. Sixty eight participants allocated to the SMART Recovery Only
group (70%) completed 3 month follow-up assessment. 58 (85%) of these 68 had attended at
least two SMART Recovery meetings, defined by the authors as the threshold for being
considered ‘treated’ (Hester et al., 2013). Of note, the authors had to abandon their original
plan to randomise to an ‘OA Only’ condition as potential participants were unwilling to be
SMART RECOVERY 14
allocated to a treatment condition that would prevent them from attending SMART Recovery
meetings.
Summary of Evidence for Interventions Informed by SMART Recovery
Two evaluations of face-to-face interventions informed by SMART Recovery were
identified. Firstly, Brooks and colleagues (2000, 2003) used a pre-post design to compare an
intensive, outpatient/partial hospitalisation adaptation of SMART Recovery for dual diagnosis
to a similarly adapted 12-step programme for adults with serious Axis I mental illness and
concurrent substance dependence. In this dual diagnosis population, there was an overall
reduction in alcohol and substance use across time for both conditions (Brooks & Penn, 2003;
Penn & Brooks, 2000). Improvement in ASI-alcohol (but not ASI-Drug) was superior for 12-
step relative to SMART Recovery participants (Brooks & Penn, 2003). However,
interpretation is complicated since 12-step baseline ASI-Alcohol scores were also
significantly higher. Urine analysis indicated that 12-step participants were less likely than
SMART Recovery participants to use marijuana at 2-month follow-up (no other substances or
follow-up intervals reached significance; Brooks & Penn, 2003). Both groups also
demonstrated improvement across several indices of functioning (financial well-being and life
satisfaction; ASI psychiatric, employment and legal composite scores; psychiatric
hospitalisation), with between group differences on employment and number of psychiatric
hospitalisations, both in favour of SMART Recovery (Brooks & Penn, 2003/ Penn & Brooks,
2000). Observed changes in substance use, but not functional outcomes were predicted by
attendance (Brooks & Penn, 2003). Overall, (i.e. irrespective of treatment condition), greater
attendance was associated with less marijuana use, but slightly more alcohol use. This latter
finding may have been due to floor effects since participants who attended more also had less
baseline alcohol use. Between group differences emerged in the duration of attendance, with
SMART RECOVERY 15
SMART Recovery participants attending significantly fewer days and weeks of treatment
relative to 12-step participants (Brooks & Penn, 2003).
Blatch and colleagues (2016) used a quasi-experimental design to compare ‘Getting
SMART’, a SMART Recovery informed intervention for offenders (alone, or in combination
with SMART Recovery) to a propensity matched control group. For custodial offenders, all
indices of recidivism were consistently lower for ‘Getting SMART’ participants relative to
controls (Blatch et al., 2016; see Table 2). Observed reductions in reconviction (for ‘any’ and
‘violent’ crimes) were even more pronounced for participants who attended both Getting
SMART and SMART Recovery. Conversely, the improvements seen following participation
in SMART Recovery only did not significantly differ from that of controls. Completion of 10-
11 sessions (Getting SMART and/or SMART Recovery) was required to detect a significant
therapeutic effect (defined as 25% increase in days to first reconviction; Blatch et al., 2016)
and over a third of participants met this threshold (See Table 2). Neither baseline nor change
in either mental health status or alcohol/ drug use outcomes were reported.
Summary of Evidence for SMART Recovery Relative to Other forms of Mutual Aid
Five cross sectional studies compared SMART Recovery to other forms of mutual aid,
most commonly AA. Only Atkins (2007) and Milin (2007) included some index of mental
health status, with Atkins reporting on prior psychiatric hospitalisation and Milin assessing
self-reported diagnosis (see Table 2 for data on SMART Recovery participants). Atkins
(2007), Bogdonoff (2003) and Trumble (2015) all report an equivalent duration of sobriety for
SMART Recovery and AA participants. Conversely, Guarnotta (2015) found that the duration
of abstinence for AA participants was approximately double that of SMART Recovery
participants, but the statistical significance of this effect was not assessed. With the exception
of ‘years of abuse’ (which did not significantly differ) Milin (2007) described a more severe
addiction profile for AA relative to SMART Recovery participants (including greater
SMART RECOVERY 16
substance related problems, impaired functioning and poorer quality of life). However,
corrections were not made for multiple comparisons.
Milin (2007) also found that ‘readiness to change’ was greater for SMART Recovery
relative to AA participants, but contrary to expectations, it did not predict alcohol related
problems. Bogdonoff (2003) found that relative to their AA counterparts, SMART Recovery
participants’ demonstrated greater future orientation, greater approach coping skills, less
conflict and higher social support. However, contrary to prediction, none of these variables
predicted abstinence. Conversely, Guarnotta (2015) found significant, moderate, positive
correlations between abstinence and self-efficacy for both SMART Recovery and AA
participants. Atkins (2007) identified additional predictors of sobriety, including participation
and number of close friends in recovery.
When Milin (2007) asked SMART Recovery participants about what they ‘liked’
about their mutual aid group, qualitative findings revealed that both general group processes
(support, non-judgement), and key features of the SMART Recovery approach
(empowerment, tools/resources and scientific/theoretical approach) featured in the top five
themes extracted (Table 3). When SMART Recovery participants were asked about what they
disliked about prior approaches, responses again pertained to general group processes (e.g.
poor boundaries), but this time also referred to prior experience with 12-step approaches (e.g.
higher power/religion; perception of powerlessness; Table 3).
Summary of Evidence for SMART Recovery Without a Comparison Condition
Two studies without a comparison condition were identified. Kelly and colleagues
(2015) explored potential mechanisms of change in SMART Recovery by assessing the extent
to which quality of group facilitation, group cohesion and homework contributed to self-rated
use of cognitive behavioural skills. Group cohesion emerged as a significant predictor of
cognitive restructuring, while homework was identified as a significant predictor of
behavioral activation (Kelly et al., 2015). Although quality of group facilitation was not
SMART RECOVERY 17
identified as a significant predictor of either cognitive restructuring or behavioral activation, a
positive relationship was detected with group cohesion (Kelly et al., 2015). O’Sullivan and
colleagues (2015) sought to explore the recovery goals of SMART Recovery members and
facilitators; to describe the educational and training backgrounds of SMART Recovery
facilitators; and to describe SMART Recovery members and their motivations for attending.
When SMART Recovery members were asked to describe their reasons for switching from
another mutual aid approach, their responses closely mirrored those described by Milin (2007;
See Table 3). The majority of participants reported that they attended SMART Recovery
relative to other forms of mutual aid due to alignment with key features of the SMART
Recovery approach (e.g. CBT; 51.6%) or prior difficulties with 12-step approaches (e.g.
higher power/religion; perception of powerlessness; 26.8%; Table 3).
DISCUSSION
This review was designed to provide a comprehensive overview and critical analysis
of the current state of evidence for SMART Recovery in adults with substance and/or
behavioral addictions. We sought to comment on a) whether participation in SMART
Recovery results in changes in the severity of addiction and its consequences, b) what factors
might influence any changes observed, c) the feasibility of this approach, and d) future
research directions. Further, given the prevalence of comorbid mental health conditions and
their impact on addiction recovery (Mills et al., 2010) we also sought to comment on the
assessment, reporting and/or change in mental health status within the included studies.
A modest body of research, comprising12 studies was identified. Although
predominantly cross-sectional, three evaluations of effectiveness were identified (Blatch et al.,
2016; Brooks & Penn, 2003/ Penn & Brooks, 2000; Hester et al., 2013). Participants with
alcohol addictions were the primary focus of existing research. The relationship between
SMART RECOVERY 18
SMART Recovery and the severity and impact of behavioral addictions has yet to be
assessed. Functional outcomes were rarely reported. Feasibility was largely indexed by
attendance and economic analysis has not been undertaken. Little is known about variables
that may influence treatment outcome, although attendance (Blatch et al., 2016; Brooks &
Penn, 2003/ Penn & Brooks, 2000; Hester et al., 2013) represents a potential candidate.
Despite high rates of comorbidity between mental health and substance use disorders
within the community (Mills et al., 2010), assessment and reporting of mental health status is
limited. Only three studies (Brooks & Penn, 2003/Penn & Brooks, 2000; Hester et al., 2013;
Kelly et al., 2015) utilised standardised instruments to establish a baseline diagnosis or
severity and only one reported on change in mental health outcomes (psychiatric
hospitalisation; Brooks & Penn, 2003/ Penn & Brooks, 2000). Within the literature identified,
mental health condition(s), distressing symptoms and past suicide attempts appear common
amongst SMART Recovery participants. Given that co-morbid mental health conditions have
the potential to complicate the course and severity of addiction and vice versa (Mills et al.,
2010) improved assessment and reporting of mental health outcomes represents an important
priority for future research.
The modest sample of papers and diversity of methods prevents us from making
conclusive remarks about the efficacy of SMART Recovery, but positive effects were found
in dual diagnosis (Brooks & Penn, 2003/ Penn & Brooks, 2000) and correctional settings
(Blatch et al., 2016). Evidence from the sole identified RCT also supported the benefits of
SMART Recovery for reducing the severity and consequences of problematic alcohol use
(Hester et al., 2013). Importantly, this RCT was independently evaluated by two assessors to
be of high quality and at low risk of bias, thereby increasing our confidence in these findings.
However, an important limitation of these studies is the limited (Hester et al., 2013) or absent
(Blatch et al., 2016; Brooks & Penn, 2003/ Penn & Brooks, 2000) assessment and reporting of
SMART RECOVERY 19
concurrent treatment (pharmacological and psychological) for addiction and/or mental health.
Accordingly, the relative contribution of mutual aid and formalised treatment (alone or in
combination) to the performance of SMART Recovery remains unclear and represents an
important question for future research.
The comparative influence of SMART Recovery on addiction outcomes relative to
other forms of mutual aid and/or evidence-based treatments (alone or as an adjunct) has yet to
be systematically evaluated. This is not without methodological challenges. SMART
Recovery groups are freely accessible in the community, therefore, it may be impractical and
possibly unethical (McCrady & Miller, 1993) to randomise some participants to this resource
while prohibiting others. Indeed, the one RCT had to abandon their original research design,
as participants were unwilling to be randomised to condition where they would be unable to
continue face-to-face SMART Recovery meetings. Preference based trials, evaluation of
professionally delivered SMART Recovery groups or embedding research methods within
new community groups as they are established may help bridge this gap between
methodological rigour and real-world relevance.
The literature also suggests that the ‘sobriety’ of SMART Recovery participants is at
least equivalent to that of alternative forms of mutual aid (Atkins & Hawdon, 2007;
Bogdonoff, 2003; Trumble, 2015), with some evidence to suggest that the severity and
consequences of alcohol addiction is less for SMART Recovery relative to AA participants
(Milin, 2007). Conversely, the duration of abstinence has been identified as longer for AA
relative to SMART Recovery participants (Guarnotta, 2015). Clinical guidelines advocate
tailoring addiction support to the goals of the individual (Mills et al., 2010; NICE, 2012;
NICE, 2011), so while abstinence may be encouraged, moderated use and/or harm reduction
approaches might also be employed. Moreover, in the case of poly-substance use, some but
not all substances may be identified as an important focus of treatment. Accordingly, such
SMART RECOVERY 20
global ratings of ‘abstinence’ and ‘sobriety’ are unlikely to be adequate indicators of
clinically meaningful change.
Consistent with the broader literature (e.g. Reardon, Cukrowicz, Reeves & Joiner,
2002) attendance was identified as a significant predictor of change (e.g. Hester et al., 2013).
Further research is needed to clarify not only whether an ‘optimal’ threshold of attendance
exists, but to identify the factors involved in engaging participants and encouraging
attendance. Interestingly, despite largely comparable addiction related outcomes, current
findings suggest that the duration of attendance may be shorter for SMART Recovery relative
to 12-step participants (Brooks & Penn, 2003/ Penn & Brooks, 2000; Li, Feifer & Strohm,
2000; Milin, 2007). Although clearly in need of further investigation, this may be testimony
to the feasibility of the SMART Recovery approach. That is, SMART Recovery may
represent a more time efficient method for promoting clinically meaningful change. However,
further research on the relationship between attendance and the change process within and
across different mutual aid groups is needed before firm conclusions can be drawn.
It is important to acknowledge the methodological limitations of this review. Firstly,
this review covers a small number of heterogenous studies. Drawing comparisons between
studies was complicated by differences in outcome assessment, intervention and comparator
groups. Additionally, the studies varied in methodological quality. Only one received a high
quality rating (Hester et al., 2013) and was also the only study deemed to be at low risk of
bias. We also restricted our literature search to English language publications, so the cross-
cultural generalisability of our findings is restricted.
Authors’ Conclusions:
Implications for Practice
Given the positive effects of SMART Recovery and SMART Recovery informed
interventions, to enhance client centred, collaborative care that is tailored to the needs and
SMART RECOVERY 21
preferences of the individuals, clinicians need to be aware of the range of mutual aid support
options available, including SMART Recovery and discuss these options with their clients.
Implications for Research
To increase understanding of the role of SMART Recovery in facilitating recovery
from addiction and to consolidate our confidence in the effectiveness of this approach, future
research may benefit from improved assessment and reporting of (i) mental health status (e.g.
diagnosis, treatment history, symptoms and functioning); (ii) concurrent treatment
(pharmacological and psychological) for mental health and addiction; (iii) use and
consequences of non-alcohol substance use, including greater attention to smoking and
prescription opiate misuse; (iv) personal and social functioning (e.g. quality of life); (v)
severity and consequences of behavioral addictions and (vi) economic outcomes.
We also offer the following suggestions to improve the quality of future research.
Firstly, greater utilisation of validated data collection methods, including interviewer
administered (e.g. Time Line Follow Back), service user rated scales (e.g. AUDIT) and
biological indices (e.g. saliva) is an important priority. Secondly, there is a need for greater
attention to the relationship between ‘active ingredients’ (e.g. self-management skills),
attendance and the change process within and across different mutual aid groups. Thirdly,
where possible, future research would benefit from greater attention to the use and reporting
of random sequence generation, allocation concealment, attrition, missing data and power.
Finally, preference based trials, evaluation of professionally delivered SMART Recovery
groups and/ or embedding research methods within new community groups may help to
clarify the relative impact of SMART Recovery on addiction outcomes compared to other
forms of mutual aid and/ or evidence based treatments.
SMART RECOVERY 22
REFERENCES
Atkins, R. G., & Hawdon, J. E. (2007). Religiosity and Participation in Mutual-Aid Support
Groups for Addiction. Journal of Substance Abuse Treatment, 33(3), 321-331.
Australian Government Productivity Commission (2010). Gambling: Inquiry Report (50).
Canberra Retrieved August 5, 2015 from http://www.pc.gov.au
Australian Psychological Society (2010). Evidence-based psychological interventions in the
treatment of mental disorders: a literature review. Melbourne: APS Retrieved May 5,
2014 from https://www.psychology.org.au/practitioner/resources/interventions/
Baker, A. L., Hiles, S. A., Thornton, L. K., Hides, L., & Lubman, D. (2012). A systematic
review of psychological interventions for excessive alcohol consumption among
people with psychotic disorders. Acta Psychiatrica Scandinavica, 126, 243-255.
Beck, A. K., Baker, A., Kelly, P. J., Deane, F. P., Shakeshaft, A., Hunt, D. et al. (2016).
Protocol for a systematic review of evaluation research for adults who have
participated in the ‘SMART recovery’ mutual support programme. BMJ Open, 6(5).
Black, D. W., Shaw, M., McCormick, B., & Allen, J. (2013). Pathological gambling:
relationship to obesity, self-reported chronic medical conditions, poor lifestyle
choices, and impaired quality of life. Comprehensive Psychiatry, 54(2), 97-104.
Blatch, C., O'Sullivan, K., Delaney, J. J., & Rathbone, D. (2016). Getting SMART, SMART
Recovery© programs and reoffending. The Journal of Forensic Practice, 18(1), 3-16.
Bogdonoff, D. A. (2003). Resilience as a predictor of abstinence in the recovery from alcohol
dependence. (Unpublished Dissertation). Dissertation Abstracts International:63(9-
B), 4361.
Brooks, A. J., & Penn, P. E. (2003). Comparing treatments for dual diagnosis: twelve-step and
self-management and recovery training. American Journal of Drug & Alcohol Abuse,
29(2), 359-383.
SMART RECOVERY 23
Centre for Evidence Based Physiotherapy, C. f. E.-B. (2009). PEDro Scale: Centre for
Evidence-Based Physiotherapy. Retrieved May 5, 2015 from http://www.pedro.org.au.
Donovan, D. M., Ingalsbe, M. H., Benbow, J., & Daley, D. C. (2013). 12-step interventions
and mutual support programs for substance use disorders: an overview. Social Work in
Public Health, 28(3-4), 313-332.
Downs, S. H., & Black, N. (1998). The feasibility of creating a checklist for the assessment of
the methodological quality both of randomised and non-randomised studies of health
care interventions. Journal of Epidemiology & Community Health, 52(6), 377-384.
Ferri, M., Amato, L., & Davoli, M. (2006). Alcoholics Anonymous and other 12-step
programmes for alcohol dependence. Cochrane Database of Systematic Reviews(3).
Guarnotta, E. (2015). A Comparison of Abstinence and Perceived Self-efficacy for
Individuals Attending SMART Recovery and Alcoholics Anonymous. (Unpublished
Dissertation). ALLIANT INTERNATIONAL UNIVERSITY.
Hester, R. K., Lenberg, K. L., Campbell, W., & Delaney, H. D. (2013). Overcoming
Addictions, a Web-based application, and SMART Recovery, an online and in-person
mutual help group for problem drinkers, part 1: three-month outcomes of a
randomized controlled trial. Journal of medical Internet research, 15(7), e134.
Higgins, J. P. T., & Green, S. (2011). Cochrane Handbook for Systematic Reviews of
Interventions (V. 5.1.0.) Retrieved May 5, 2015 from http://handbook.cochrane.org/
Horvath, A. T., & Sokoloff, J. (2011). Individuals seeking non-12-step recovery. In G. W.
Lawson & A. W. Lawson (Eds.), Alcoholism & Substance Abuse in Diverse
Populations (2nd ed., pp. 75-90). Austin: PRO-ED.
Horvath, A., & Yeterian, J. (2012). Smart recovery: Self-empowering, science-based
addiction recovery support. Journal of Groups in Addiction & Recovery, 7(2-4), 102-
117.
SMART RECOVERY 24
Humphreys, K., Wing, S., McCarty, D., Chappel, J., Gallant, L., Haberle, B., . . . Kivlahan, D.
(2004). Self-help organizations for alcohol and drug problems: Toward evidence-
based practice and policy. Journal of substance abuse treatment, 26(3), 151-158.
Kelly, J. F., Magill, M., & Stout, R. L. (2009). How do people recover from alcohol
dependence? A systematic review of the research on mechanisms of behavior change
in Alcoholics Anonymous. Addiction Research & Theory, 17(3), 236-259.
Kelly, P. J., Deane, F. P., & Baker, A. L. (2015). Group cohesion and between session
homework activities predict self-reported cognitive–behavioral skill use amongst
participants of SMART Recovery groups. Journal of Substance Abuse Treatment, 51,
53-58.
Laudet, A. B. (2011). The Case for Considering Quality of Life in Addiction Research and
Clinical Practice. Addiction Science & Clinical Practice, 6(1), 44-55.
Li, E. C., 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.
Manning, M., Smith, C., & Mazerolle, P. (2013). The societal costs of alcohol misuse in
Australia Trends and Issues in Crime and Criminal Justice. Canberra: Australian
Institute of Criminology. Retrieved August 5, 2015 from http://www.aic.gov.au
Mc Crady B.S., Miller, W.R., eds. (1993). Research on Alcoholics Anonymous: Opportunities
and Alternatives. New Brunswick, NJ: Rutgers Center of Alcohol Studies
Milin, M. (2008). Which types of consequences of alcohol abuse are related to motivation to
change drinking behavior? (Unpublished Dissertation). Dissertation Abstracts
International: Section B: The Sciences and Engineering, 68(9-B), 6322.
Mills, K. L., Deady, M., Proudfoot, H., Sannibale, C., Teeson, M., Mattick, R., & Burns, L.
(2010). Guidelines on the management of co-occurring alcohol and other drug and
SMART RECOVERY 25
mental health conditions in alcohol and other drug treatment settings Sydney:
NDARC, Retrieved August 18, 2015 from https://ndarc.med.unsw.edu.au
Moos, R. H., & Moos, B. S. (2006). Participation in Treatment and Alcoholics Anonymous: A
16-Year Follow-Up of Initially Untreated Individuals. Journal of Clinical Psychology,
62(6), 735-750. DOI: 10.1002/jclp.20259
Moher, D., Liberati, A., Tetzlaff, J., & Altman, D. G. (2009). Preferred reporting items for
systematic reviews and meta-analyses: the PRISMA statement. PLoS Medicine, 6(7),
e1000097.
National Institute for Health Excellence. (2011). Alcohol-use disorders: diagnosis, assessment
and management of harmful drinking and alcohol dependence. London: National
Institute for Health and Clinical Excellence Retrieved August 18, 2015 from
https://www.nice.org.uk/guidance/cg115.
National Institute for Health Excellence (2012). NICE Quality standard for drug use
disorders. London: National Institute for Health and Clinical Excellence Retrieved
August 18, 2015 from https://www.nice.org.uk/guidance/qs23
O’Sullivan, D., Blum, J. B., Watts, J., & Bates, J. K. (2015). SMART Recovery: Continuing
Care Considerings for Rehabilitation Counselors. Rehabilitation Counseling Bulletin,
58(4), 203-216.
Penn, P. E., & Brooks, A. J. (2000). Five years, twelve steps, and REBT in the treatment of
dual diagnosis. Journal of Rational-Emotive & Cognitive-Behavior Therapy, 18(4),
197-208.
Public Health England (2010, January). Improving mutual aid engagement: A professional
development resource. London: Public Health England. Retrieved August 18 2015
from http://www.nta.nhs.uk
SMART RECOVERY 26
Reardon, M. L., Cukrowicz, K. C., Reeves, M. D., & Joiner, T. E. (2002). Duration and
Regularity of Therapy Attendance as Predictors of Treatment Outcome in an Adult
Outpatient Population. Psychotherapy Research, 12(3), 273-285.
Sheedy, C. K., & Whitter, M. (2009). Guiding principles and elements of recovery-oriented
systems of care: What do we know from the research? HHS Publication No. (SMA)
09-4439. 2009. Rockville, MD: Center for Substance Abuse Treatment, Retrieved
August 17 2015 from http://store.samhsa.gov
Trumble, R. G. (2015). Comparison of drinking-related locus of control in Alcoholics
Anonymous members and smart recovery participants. (Unpublished Dissertation).
THE UNIVERSITY OF THE ROCKIES.
Whiteford, H. A., Degenhardt, L., Rehm, J., Baxter, A. J., Ferrari, A. J., Erskine, H. E., . . .
Vos, T. (2010). Global burden of disease attributable to mental and substance use
disorders: findings from the Global Burden of Disease Study. The Lancet, 382(9904),
1575-1586.
SMART RECOVERY 27
FIGURE CAPTIONS
Figure 1. Flow chart of study selection process
Figure 2. Risk of bias summary: Review authors’ judgements about each risk of bias item for
each included study
Figure 3. Risk of bias graph: Review authors’ judgements about each risk of bias item
presented as percentages across all included studies
Identification
EligibilityScreening
110recordsreturnedfromelectronicscientificdatabases:Medline(4);
PubMed(9);Embase(7);CINAHLComplete(4);PsychInfo(83);CENTRAL
(3).Withsearchterms
“SMARTRecovery”OR“SelfManagementAndRecoveryTraining”
AND
alcohol*ORalcoholrelateddisorderORalcoholabuseORalcohol
dependenceORsubstanceusedisorderORsubstanceabuseORsubstance
dependen*ORgamblingORAddictivebehav*rORaddict*
AND
addictionseverityORabstinen*ORharmreductionORrecurrenceOR
relapseOR
alcoholdrinkingORalcoholconsumptionORsubstanceus*OR
“dollarslost”ORexpenditureOR“hoursspent”OR“timespent”
OR
patientcomplianceORadherenceORparticipationORattendanceOR
engagement
879recordsreturnedfrom
non‐scientificelectronic
databases:GoogleScholar
(855);Virginia
CommonwealthUniversity
(19);ProjectCork(5);
Prevention,Informationand
EvidenceLibrary(0),search
terms:“SMARTRecovery”
17Duplicatesremoved
5recordsfrom
othersources
(referencelists/
author
correspondence)
972Manualsearchof
titles/abstract
859excludedbecause
NotaboutSMARTRecovery
NotpublishedinEnglish
Notajournalarticle,report,book
chapterornewsletter(e.g.book
review)
118fulltextarticles
assessedforeligibility
81excludedbecause
NotaboutSMARTRecovery(62)
QualitativeOnlyDesign(4)
Evaluation‐Don’tfocusonadults
withproblematicaddictivebehaviour
(2)
Evaluation–Donotreportoutcome
measuresofinterestforSMART
participants(3)
Notajournalarticle,report,book
chapterornewsletter(e.g.
conferenceabstract,bookreview)(4)
Conferenceproceedingslater
published(6)
37publicationsclassified
Discussion
15
Evaluation
12
Reviews
4
12EvaluationStudies
SelectedforInclusion
Other
6
Classification
CriticalAnalysis
Random sequence
generation (selection bias)
Allocation concealment
(selection bias)
Blinding of participants and
researchers (performance
bias) - All Outcomes
Blinding of outcome
assessment (detection
bias) - All outcomes
Incomplete outcome data
(attrition bias)
Selective reporting
(reporting bias)
Other bias
Hester et al (2013)
Brooks & Penn (2003)
Penn & Brooks (2000)
Blatch et al. (2016)
Li et al. (2000)
Atkins & Hawdon (2007)
O’Sullivan et al. (2015)
Kelly et al. (2015)
Guarnotta (2014)
Milin (2007)
Trumble (2015)
Bogdonoff (2002)
Risk of bias: Low Unclear High
? ?
? ? ? ? ?
?
?
?
?
?
?
? ?
? ?
?
?
0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%
Otherbias
Selectivereporting(reportingbias)
Incompleteoutcomedata(attritionbias)
Blindingofoutcomeassessment(detectionbias)‐Alloutcomes
Blindingofparticipantsandresearchers(performancebias)‐Alloutcomes
Allocationconcealment(selectionbias)
Randomsequencegeneration(selectionbias)
Low Unclear High
SMART RECOVERY 1
Table 1.
Demographic and Clinical Characteristics of SMART Recovery Participants
N Age
M (SD)
Gender
(%
Male)
Ethnicity
(%
Caucasian)
Education/
Degree Employment Marital
Status Mental Health Addiction
Alcohol Substance Behavioral
Established Community SMART Groups versus SMART Informed Online Intervention (Alone or In Combination)
Hester et al
(2013) 86 43.4
(10.6)
39% 88.4% M=15.93
(SD=2.5)
years
Not reported Not
reported
Brief Symptom
Inventory:
M=19.35(SD=12.5)
Hazardous Alcohol use
(see baseline measures
reported in Table 2)
Substance
Dependence Excluded
Not
assessed
SMART Informed Interventions versus Active and/ or Control Comparison Conditions
Blatch et al
(2016) 2882aNot
reported
68% 27% ATSI
Not
reported
Not reported Not
reported
Not assessed Not assessed Not assessed Not
assessed
Brooks & Penn
(2003) 58 34.2(8.4) 67.2% 72.4% M=11.6
(SD=2.4)
years
20.4% (Full
Time or Part
Time in the
past 3 yrs)
80% single
or divorced
44.8% mood disorder
20.7% thought
disorder
15.5% personality
disorder
Mean number of times
of psychiatric
hospitalisation=8.2
(SD=10.5
ASI: M=0.2825
Years of use:
M=10.5(SD=9.6)
Longest duration of
abstinence= 6.5monthsb
Mean number of times of
alcohol
treatment=3.4(SD=6)
Substance Dependence:
8.8% polysubstance
14% cocaine
10.5% amphetamines
8.8% Marijuana
Mean Years of use:
Polysubstance: 9.1(SD=8.4)
Marijuana: 8.6(SD=8.6)
Cocaine: 4.4(SD=5.7)
Mean number of times of
substance use
treatment=3.2(SD=5.5)
Not
assessed
Established Community Groups: SMART Recovery versus Other Forms of Mutual Aid
Atkins &
Hawdon (2007) 321a 47
(no SD)
58.1%
b
90.3% Not
reported
Mean income
$55000
43.4%
single or
divorced
21.4% reported prior
psychiatric
hospitalisation
Years of use
>10 (78.8%)
Past hospitalisation (35%)
Past outpatient treatment
(48.7%)
>70% Polysubstance Use Not
assessed
Bogdonoff (2002;
Unpublished
Dissertation)
53 36
(no S.D.)
24.5% 32.4% 5.9%
bachelors/
graduate
12.9% (Full
Time or Part
Time)
23% Single,
divorced,
separated
Not reported Past treatment
(60.4%)
N
o independent assessment
of substance use
Not
assessed
SMART RECOVERY 2
Guarnotta (2014;
Unpublished
Dissertation)
58 42.1
(13.4)
45.1% 87.7% 51.7%
college/
graduate
Not reported 63.9%
Single or
divorced
Not reported Mean Number of Days
Abstinent=322.4
(SD=323.79)
N
o independent assessment
of substance use
Not
assessed
Li et al (2000) 33 45.79
(11.8)
67% Not reported 82%
college/
graduate
Not reported Not
reported
Not reported Not assessed Not assessed Not
assessed
Milin (2007;
Unpublished
Dissertation)
60 44
(no SD)
56.7kell
%
97% 30.2%
college
64.6% (Full
Time or Part
Time)
52.6%
Single or
divorced
Self-Reported
Diagnosis:
Major depression
(40%); Severe
anxiety/ panic
(26.7%); ADHD
(8.3%); Bipolar
(11.7%); SZ (0%);
OCD (5%); Other
(1.7%)
History of problematic
alcohol use (see Table 2)
Mean number of months
abstinent=7.13
(SD=10.93)
Past Treatment:
Inpatient (25%)
Residential (15%)
Individual (46.7%)
Outpatient (25%)
Drug of Choice:
Cocaine (1.7%)
Prescription Meds (1.7%)
Current Abuse:
Marijuana (3.3%) and
Prescription Medication
(6.7%) only
Ever Abuse:
Marijuana (56.7%)
Cocaine (46.7%)
Heroin (8.3%)
Methamphetamine (25%)
Hallucinogens (33.3%)
Prescription Meds (31.7%)
Not
assessed
Trumble (2015;
Unpublished
Dissertation)
70 51.62
(11.74)
64% 95% 41%
bachelors
48%: Income=
$35,000 to
$100,000
Not
reported
Not reported Mean number of days
abstinent=1417.6
(SD=1985.28)
N
o independent assessment
of substance use
Not
assessed
SMART Recovery Without a Comparison Condition
Kelly et al (2015) 124 40.65
(11.38)
56.5% 6.5% ATSI Not
reported
30.7% Full
Time or Part
Time
Not
reported
Self-Reported
Diagnosis
46.7% depression;
29% anxiety; 5.6%
bipolar; 4.8 % PTSD;
3.2% SZ or psychotic
disorder; 6.5% other
29% reported a prior
suicide attempt
Mean K10:
21.74(SD=4.91)
85.6% used alcohol within
the preceding 12 months
Mean years of problems
(alcohol or
substance)=18.11
(SD=10.97)
Primary substance of abuse:
Amphetamine (7.3%)
Heroin (5.6%)
Tobacco (4.8%)
Marijuana (3.2%)
Use < 12 months:
Tobacco (63%)
Marijuana (44%)
Heroin (32.3%)
Amphetamines (27.4%)
Analgesics (22.6%);
Food
(10.5%)
Gambling
(9.7%)
Shopping
(6.5%)
Pornography
(4.8%)
Sex (3.2%)
SMART RECOVERY 3
47% in high or very
high range (>22)
66.4% reported prior
treatment
46.8% reported
current medication
management
Cocaine (12.1%);
Ecstacy (11.3%)
O’Sullivan et al.
(2015) 81 48 (13.1) 66.7% 90.1% white 66.6%
graduate/
bachelors
63% employed 34.6%
single
51% endorsed
‘psychiatric disability’
Mean years of abuse:
15.62(SD=11.5)
Sobriety attempt: 1st
(26.6%); 2-5 (50.6%); 6-10
(11.4%)
Mean years of individual
counselling
5.14(SD=7.39)
Drugs (14.8%)
+ Alcohol (9.9%)
+ Behavioural (2.4%)
+ Both (7.4%)
7.3%
+ Alcohol
(7.3%)
+ Drugs
(2.4%)
+ Both
(7.4%)
Note. ADHD = Attention Deficit Hyperactivity Disorder; ATSI = Aboriginal or Torres Strait Islander; M = Mean; PTSD = Post traumatic Stress Disorder; SD = Standard Deviation; SZ =
Schizophrenia. aData only available across all treatment conditions (SMART Only not available); bFull sample gender distribution (41.2% male) skewed by Women For Sobriety (women only
group) – revised gender distribution removing WFS participants reported
SMART RECOVERY 4
Table 2
Summary of Methodology and Key Findings from Evaluations of SMART Recovery
Study
Quality Rating
Risk of Bias
Aim
Target Population
Country
Design
Treatment/
Comparison Groups
Key Findings
Severity of Addiction and its Consequences Process Variables Feasibility
Established Community SMART Groups versus SMART Informed Online Intervention (Alone or In Combination)
Hester et al
(2013).
Downs & Black:
22 (max=27)
PeDRO:
6 (max=8)
Overall Risk of
Bias:
Low
To evaluate the effectiveness of a
web application informed by
SMART Recovery “Overcoming
Addictions” (OA) and SMART
Recovery in a sample of problem
drinkers new to SMART
Recovery
Participants had <4 weeks
attendance at SMART and
hazardous alcohol use as indexed
by >8 on AUDIT and alcohol
consumption outside
recommended guidelines
USA
RCT
OA (n=19)
vs..
SMART (n=86)
vs.
OA+SMART (n=83)
Significant improvement in all conditions at 3-
month follow-up (no between group
differences):
Percent days abstinent (PDA): Significant
Main Effect (44% vs 72%)
F1,149=160.93, p<.001
Group differences F<1.0
Standard drinks per drinking day (DDD):
Significant Main Effect (8.0% vs 4.6%)
F1,149=61.73, p<.001
Group differences F<1.0
Inventory of Drug and Alcohol Use
Consequences (InDUC):
Significant Main Effect (40.8% vs 19.5%)
F1,149=122.28, p<.001
Group differences F<1.0
Change in mental health status not assessed
Number of days of face-to-face
meetings, online meetings and/ or
‘any support’ were identified as
significant predictors of change in
alcohol use (the relationships that
emerged varied according to
treatment condition and outcome
measure).
OA: Number of days of online
SMART meetings identified as a
significant predictor of PDA
(p=0.25).
SMART: Number of days of face to
face meetings identified as a
significant predictor of all three
outcome measures (PDA: r=.358,
p=.003; DDD: r=-.250, p=.039;
InDUC: r=-.244, p=.045) and
change in these from baseline
(PDA: r=.274, p=.024; DDD:
r=.478, p<.001; and InDUC:
r=.403, p=.001)
OA+SMART: Number of days of
‘any support’ identified as a
significant predictor of PDA
(r=.306, p=.012) and improvement
in InDUC (r=.305, p=.012)
OA: Logged onto
the program on
average 7.2 times
(SD=6.4)
SMART: 71%
attended at least two
online meetings.
Days of face to face
meetings (M=3.31);
days of online
meetings (M=5.90);
days of ‘any
support’ (M=14.85)
OA+SR: 85%
attended at least two
face-to-face
meetings
Days of face to face
meetings:
(M=1.82); days of
online meetings
(M=4.42); days of
‘any support’
(M=12.8)
SMART RECOVERY 5
SMART Informed Interventions versus Active and/ or Control Comparison Conditions
Blatch et al
(2016)
Downs & Black:
16 (max=27)
Overall Risk of
Bias:
High
To determine reconviction
outcomes for male and female
offenders who participated in
Getting SMART (a 12 session
adaptation of SMART for
custodial offenders) and/ or
SMART Recovery relative to a
propensity score-matched control
group who did not participate in
either programme
Participants were offenders who
served custodial sentences in New
South Wales between 2007 and
2011
AUS
Quasi-experimental,
pseudo-prospective
study design
Getting SMART
(n=2343)
vs
Getting
SMART+SMART
(n=306)
Vs
SMART (n=233)
Vs
Control (n=2882)
Relative to controls:
Time to “any” reconviction:
Getting SMART: ~8% reduction (HR=0.918*;
CI=0.848-0.995).
Neither reduction in SMART (~13%) nor
Getting SMART+SMART (~8%) reached
significance
Time to “violent” reconviction:
Getting SMART:~13% longer (HR=0.867*;
CI=0.763-0.985)
Neither change in SMART (~16% longer) nor
Getting SMART+SMART (~25% longer)
reached significance
Reconviction rate (any):
Getting SMART: ~19% reduction
(HR=0.808**; CI=0.747-0.875)
Getting SMART + SMART: ~22% reduction
(HR=0.784*; CI=0.647-0.950)
SMART did not reach significance
Reconviction rate (violent):
Getting SMART:~30% lower (HR=0.704**;
CI=0.621-0.799)
Getting SMART+SMART: ~42% lower
(HR=0.578**; CI=0.407-0.821)
SMART did not reach significance
Mental Health Status not reported
Attendance at 10-11 sessions (of
either programme, alone or in
combination) was associated with a
significant therapeutic effect (25%
reduction in reconviction rate;
HR=0.764*; CI=0.612-0.953).
Brief exposure (1-6 sessions)
mirrored the control group
81% attended
Getting Smart Only
8% attended
SMART Only
11% attended
Getting
SMART+SMART
37% attended >12
sessions
19% attended
between 9 and 11
sessions
Brooks & Penn
(2003)
Downs & Black:
14 (max=27)
To compare the effectiveness of a
SMART and 12-step Informed
intervention for dual diagnosis in
an intensive outpatient/ partial
hospitalisation setting
Participants had severe Mental
Illness (Schizophrenia, Bipolar
USA
Multivariate multiple
baseline comparison
(Alternate allocation)
SMART Informed
(n=58)
vs.
Improvement over time for both groups on:
Addiction Severity Index (Alcohol, Substance,
Employment, Legal and Psychiatric
Subscales);
Urine Analysis: marijuana and ‘other’
(cocaine, heroin, amphetamines and
barbiturate use);
Attendance was identified as a
significant predictor of marijuana
use (better attendance, less likely to
use at 2 months; odds ratio=.05)
Greater attendance increased the
slope indicating that alcohol use
(ASI) decreased less with greater
SMART
participants
attended
significantly fewer
days than 12-step:
M=81(SD=18.3) vs.
M=94(SD=21.6)
SMART RECOVERY 6
Overall Risk of
Bias:
High
Penn & Brooks
(2000)
Downs & Black:
8 (max=27)
Overall Risk of
Bias:
High
Disorder, Schizoaffective
Disorder, Major Depression) and
Concurrent Substance Use
disorder (as indexed by
Diagnostic Interview and
Collateral Information)
12-step Informed
(n=54)
Lehman Quality of Life (Financial and Life
Satisfaction Subscales);
Number of Psychiatric Hospitalisations
Significant between group differences only
emerged on
Addiction Severity Index: Alcohol (in favour
of 12-step; Intercept coefficient -0.0076;
SE=.0033, t=-2.28, p<.05)
2 month Urine Analysis: Marijuana (in favour
of 12-step; odds ratio = 0.05)
Addiction Severity Index: Employment (in
favour of SMART; coefficient = -0.0076,
SE=.0033, t = -2.28*)
Psychiatric Hospitalisation (in favour of
SMART): Significant time x treatment
interaction F(2,78.6)=4.239, p<.024; M = 0 vs
5.52 (SD=13.7)
attendance (Time linear slope:
Attendance Coefficient=.0001,
SE=.0001, t=2.51*)
t(48)=2.26; p<.028*
SMART
participants
attended
significantly fewer
weeks than 12-step
M=26(SD=3.2) vs.
28(SD=4.7)
t(48)2.46; p<.018*
Established Community Groups: SMART Recovery versus Other Forms of Mutual Aid
Atkins &
Hawdon (2007)
Downs & Black:
16 (max=27)
Overall Risk of
Bias:
High
To explore the relationship
between participants’ personal
religious/spiritual beliefs, the
religious/spiritual beliefs of their
mutual aid group and level of
participation
In participants who identified as
being “In Recovery” and were
able to identify a “primary
recovery group”
USA
Cross sectional
(National Survey)
SMART (n=321)
vs
12-step (n=161)
vs
Women for Sobriety
(WFS; n=236)
vs
Secular Organisations
for Sobriety (SOS;
n=104)
Mutual aid group was not predictive of
‘number of days clean and sober’ (Wald
X2=1.11, p=.267)
Significant predictors of sobriety
identified were
Age: (β=.272); Coefficient=.054***
SE=.009
Number of close friends in
recovery: (β=.240); Coefficient =
.418***SE=.093
Participation (as indexed by a study
specific instrument): (β=.177);
Coefficient = .045***
SE=.013
Religiosity and belief in a higher
power did not emerge as significant
predictors of sobriety
Mutual Aid Group
= significant
predictor of
participation
SMART: (β=-.193)
Coefficient=-3.02
SE=1.09
WFS: (β=-.211)
Coefficient = -3.63*
SE=1.04
SOS: (β=-.191)
Coefficient = -4.73*
SE=1.71
Authors concluded
that relative to 12-
step all other groups
SMART RECOVERY 7
were less likely to
participate
Bogdonoff
(2002;
Unpublished
Dissertation)
Downs & Black:
17 (max=27)
Overall Risk of
Bias:
High
To explore the predictive
relationship between recovery and
resilience (including self-efficacy,
coping skills and internal locus of
control- constructs embedded
within resilience) and to compare
these characteristics in SMART
Recovery and 12-step groups to
see if either type of group was
more effective in supporting
abstinence during the early period
of the first 90 days of recovery
Participants had a history of
alcohol and substance abuse and/
or dependence and were in the
“early stage of recovery” (< 30
self-reported days abstinent)
USA
Cross sectional
Prospective
Quasi-experimental
SMART (+/- short or
long term residential
rehabilitation; n=53)
Vs
12-step (+/- short or
long term residential
rehabilitation; n=86)
At 90-day follow-up ‘Sobriety’ did not
significantly differ between 12-step (39.5%)
and SMART (39.6%) groups (as indexed by
dichotomous self report assessment (yes/no) of
abstinence over the preceding 90 days)
Neither mutual aid group, nor any
of the following factors were
identified to be significant
predictors of abstinence:
Resilience and optimism (Adult
resiliency belief system), self-
efficacy regarding drinking ( Drug
Taking Confidence Questionnaire),
locus of control (drinking related
internal-external locus of control),
coping ( oping response inventory)
and social support resources
(Family/ Social Composite score on
the addiction severity index)
Of potential relevance to the
differing underlying philosophies
(i.e. the role of relinquishing to a
higher power in 12-step) SMART
Recovery participants demonstrated
greater ‘approach’ coping skills
(including logical analysis, seeking
guidance and problem solving)
F(1,132)=7.11, p=.009
Did not report on
feasibility
Guarnotta
(2015;
Unpublished
Dissertation)
Downs & Black:
17 (max=27)
Overall Risk of
Bias:
To explore the relationship
between abstinence and self-
efficacy for individuals in
SMART Recovery and how this
compared to the relationship
between self-efficacy and
abstinence for individuals
attending AA
Participants demonstrated a
history of problematic alcohol use
USA
Cross sectional
Quasi-experimental
SMART (n=58)
vs
AA (n=64)
Self-reported duration of abstinence for AA
(M=677.2, SD=1576.4; Range=30 to 4589)
was approximately double that of SMART
(M=322.4, SD=323.79; Range=30 to 1012),
but significance not reported
Self efficacy (as indexed by the
General Self Efficacy Scale) did not
significantly differ between AA
(M=30.58, SD=6.3) and SMART
(M=30.28, SD=5.9), p=.79.
Self Efficacy was identified as a
significant predictor of abstinence
for AA (r=.345, p<.01) and
SMART (r=.378, p<.01)
participants, explaining
Did not report on
feasibility
SMART RECOVERY 8
High
(as indexed by Michigan Alcohol
Screening Test), were abstinent
from alcohol or illicit substances
for a maximum of 30 days and
reported a “strong commitment”
to attend mutual aid
approximately 10.4% of the
variance in abstinence time (R2
=.104)
Li et al (2000)
Downs & Black:
14 (max=27)
Overall Risk of
Bias:
High
To investigate whether AA’s
Higher Power concept encourages
externally dependent
b
ehaviour by
testing whether AA and SMART
members are equal on measures of
locus of control
Participants were “In Recovery”
(at least 8 weeks mutual aid
attendance)
USA
Cross sectional
(Survey)
SMART (n=33)
vs
AA (n=48)
This study only explored process measures Significantly higher percentage of
AA (96%) than SMART (48%)
participants reported belief in a
higher power (X2=24.42, df=1,
p<.0001)
AA participants demonstrated a
more external locus of control
(DRIE: M=5, SD=3.23, Range=0-
13) than SMART participants
(M=2.09, SD=2.66, Range=0-13),
p=.00003
The relationship between process
variables and outcome measures
was not explored
Self-reported
duration of
involvement
(months) was
significantly longer
for AA (M=66.48,
SD=76.24;
Range=3-252)
relative to SMART
(M=18.76,
SD=15.54;
Range=2-48)
t=3.58, p=.0006
Milin (2007;
Unpublished
Dissertation)
Downs & Black:
17 (max=27)
Overall Risk of
Bias:
High
To examine the relationship
between consequences of alcohol
abuse and motivation to change
drinking behaviour and to explore
similarities and differences
between members of AA and
SMART
Participants were currently
attending mutual aid
USA
Cross sectional
Between Subjects
Correlational
SMART (n=60)
vs
AA (n=56)
Outcomes for SMART were consistently
superior relative to AA including
Less hazardous use of alcohol as indexed by
the AUDIT (M=22.5, SD=6.64 vs M=26.57,
SD=7.31; p=.002);
Lower severity and less functional impact of
addiction as indexed by The Addiction
Severity Index (ASI)
ASI-Alcohol (M=.48, SD=.28
vs M=.64, SD=.19; p=.001)
ASI-Drug (M=.05, SD=.09
vs M=.27, SD=.23, p<.001)
ASI-Psychiatric (M=.29, SD=.21
vs M=.42, SD=.25; p=.003)
ASI-Employment (M=.39, SD=.24
Levels of pre-contemplation were
significantly greater in AA ( M=-
5.59, SD=3.95) relative to SMART
( M=-7.42, SD=3.42), p<.01
Similarly, levels of contemplation
were significantly higher in
SMART ( M=5.77, SD=2.68)
relative to AA ( M=3.98, SD=4.48),
p<.05
However, readiness to change was
not identified as a significant
predictor of alcohol related
problems (All models failed to
reach significance)
Duration of
involvement was
significantly longer
for AA (4.95,
SD=1.63) relative to
SMART (3.55,
SD=1.8)
participants, where
1=< 30days,
2=30days to 3
months; 3=3-6
months; 4=6
months to 1 year;
5=1 to 2 years; 6=2
years or more
SMART RECOVERY 9
vs M=.57, SD=.33, p=.001)
ASI-Family/Social Problems (M=.29, SD=.23
vs M=.49, SD=.25 p=.001)
ASI-Legal (M=.06, SD=.21
vs. M=.24, SD=.37 p=.002)
Fewer alcohol related problems:
DrINC-R-Total (M=55.27, SD=23.43 vs
M=88.48, SD=25.34 p<.001
DrINC-Lifetime M=31.88, SD=7.62 vs
M=37.11, SD=7.39 p<.001
Trumble (2015;
Unpublished
Dissertation)
Downs & Black:
19 (max=27)
Overall Risk of
Bias:
High
The purpose of the study was to
replicate the results from Li et al’s
2000 study, (that AA will be more
externally controlled and SMART
more internally oriented) and to
explore the relationship to faith in
a higher power
Participants were currently
attending mutual aid
USA
Cross sectional
Quasi-experimental
SMART (n=70)
vs
AA (n=36)
The duration of sobriety did not significantly
differ between groups (p=.09)
AA: M=2506.17days (6.87 years) vs SMART:
M=1417.60 days (3.88 years)
Both groups demonstrated internal
locus of control (as indexed by low
scores on the Drinking Related
Internal-External Locus of Control
Scale) SMART was more internal (
M=2.08 SD=2.47) relative to AA
(M=4.67, SD=4.01), p=.001
The relationship between locus of
control treatment outcome was not
assessed
Did not report on
feasibility
SMART Recovery Without a Comparison Condition
Kelly et al
(2015)
Downs & Black:
13 (max=27)
Overall Risk of
Bias:
High
To provide a description of
participants, including potential
clinical complexities; to examine
how frequently participants used
cognitive and behavioural skills
outside of meetings; to examine
the variables that may predict
participants’ self-reported use of
cognitive and behavioral skills
Participants were currently
attending SMART Meetings
AUS
Cross sectional
(Survey)
SMART Alone
(n=124)
Duration of Alcohol/ substance use problems
was 18.11 years (SD=10.97)
Group cohesion = significant
predictor (17% of variance) of
cognitive restructuring (β=0.23).
F(3,113)=8.42, p<.001
Homework = significant predictor
(21% of variance) of behavioral
activation (β=0.26).
F(3,113)=10.99, p<.001
Significant positive correlation
between quality of facilitation and
group cohesion (r=.38)
Relationship to treatment outcome
was not assessed
The majority of
participants
attended weekly
(72.8%)
Duration of
attendance:
M=8.78 months,
SD=14.11;
Range=1 week-96
months
SMART RECOVERY 10
Note. AA = Alcoholics anonymous ASI = Addiction Severity Index; AUDIT = Alcohol Use Disorders Identification Test; CI = Confidence Interval; DDD = Standard drinks per drinking day;
DRIE = Drinking Related Internal-External Locus of Control Scale; HR = Hazard Ratio; InDUC = Inventory of Drug and Alcohol Use Consequences; M = Mean; OA = Overcoming addictions;
PDA = Percent days abstinent; SD = Standard Deviation
*p <.05;
** p < .01
*** p < .001
OSullivan et al.
(2015)
Downs & Black:
10 (max=27)
Overall Risk of
Bias:
High
To describe members of the
SMART Recovery Community,
their motivations for membership;
describe SMART facilitators and
their educational and training
backgrounds; rank order of
members’ and facilitators’
recovery goals
Participants had >3 months
attendance at SMART Meetings
USA
Cross sectional
(two sample
exploratory
descriptive survey;
n=81)
Duration of problematic addiction was
M=15.62, SD=11.5; Range=3 months to 40
years
The relationship between process
variables and treatment outcome
was not explored
Frequency of
attendance: M=4.69
meetings per month
(SD = 2.64)
Duration of
attendance:
M=1.58, SD=1.81
Range: 3 months to
10 years
On a nine point
scale (higher scores
= greater
confidence) mean
confidence in
SMARTs ability to
meet recovery goals
8.16 (SD=1.24)
SMART RECOVERY 11
Table 3.
Summary of Qualitative Findings
Study Treatment Condition
Milin (2007) What do you like about your current primary self-help groupa?
12-Step (n=56) SMART Recovery (n=60)
Supportive environment (e.g. helping others, people trying to do the right thing; n=29)
Fellowship (n=12)
12-steps give a sense of direction/ purpose (e.g. plan of action, structure; n=10)
People have common problem (e.g. shared experiences, relate with other alcoholics,
sober people; n=7)
Availability of groups (e.g. always there, somewhere to go and not drink; n=4)
Internal locus of control (e.g. self-directed, self-empowered; n=22)
Supportive environment (e.g. giving and getting help; positive
reinforcement; n=20)
Many tools/ resources for relapse prevention (n=17)
Scientific nature, theoretical (e.g. CBT, REBT; n=14)
Non-judgemental (e.g. absence of guilt, slip is not catastrophic; n=7)
What did you dislike about self-help groups you attended in the pasta
Prior 12-Step:
Negative attributes of group members (e.g. complaining/ whining; cussing/ vulgarity;
closed off; dishonesty; n=8)
Disparity among different types of 12-step groups (e.g. NA too rigid about alcohol;
lack of sponsorship in MA; could not identify with CA/ NA; too much mixing of NA
and AA; n=6)
Repetitious (e.g. no new information, retell same stories; n=5)
Lack of seriousness (e.g. some not serious about sobriety; n=5)
Frequent relapses (n=3)
Prior SMART
Lack of sponsorship (n not reported)
Prior 12-Step
Higher power, religious (n=21)
Powerlessness (n=19)
Dogmatic, authoritative, rigid (e.g. have to do it one way, problem for life,
moderation not an option; n=18)
Labeling (e.g. ‘alcoholic’, ‘disease’; n=11)
People with poor boundaries (intimidating/ domineering people, ask me for
money, unwanted advances from men, disrespectful, not trustworthy/
dishonesty; n=10)
Prior SMART
I don’t like counselling/ advice (n=1)
Abstinence is not required (n=1)
Same stories repeated (n=1)
Easy to be facetious on-line (n=1)
O’Sullivan
et al. (2015) Reasons for switching from another mutual aid approach to SMART Recoveryb
---
Alignment with SMART philosophy, principles and format (e.g. CBT 51.6%)
Difficulties with surrendering to religious affiliations such as a higher power/
adoption of a powerlessness identity (26.6%)
Still attending both types of mutual aid (18.8%)
Outlier responses (3%)
Note. CA = Cocaine Anonymous; CBT = Cognitive Behavioral Therapy; MA = Marijuana Anonymous; NA = Narcotics Anonymous; REBT = Rational Emotive Behavioral Therapy; atop
five themes from the 13 identified reported here; b79% had switched from another approach (primarily 12-step), qualitative findings are derived from thematic analysis
