AA attendance and abstinence for dually diagnosed patients: a meta-analytic review

Background and aims: There is consensus that best clinical practice for dual diagnosis (DD) is integrated mental health and substance use treatment augmented with Alcoholics Anonymous (AA) attendance. This is the first quantitative review of the direction and magnitude of the association between AA attendance and alcohol abstinence for DD patients.

Method: A systematic literature search (1993-2017) identified 22 studies yielding 24 effect sizes that met our inclusion criteria (8075 patients). Inverse-variance weighting of correlation coefficients (r) was used to aggregate sample-level findings and study aims were addressed using random- and mixed-effect models. Sensitivity and publication bias analyses were conducted to assess the likelihood of bias in the overall estimate of AA-related benefit.

Results: AA exposure and abstinence for DD patients were associated significantly and positively [r_w = 0.249; 95% confidence interval (CI) = 0.203-0.293; tau = 0.097). There was also significant heterogeneity in the distribution of effect sizes and high between-sample variance (I² = 74.6, P < 0.001). Subgroup analyses indicated that the magnitude of AA-related benefit did not differ between 6- (k = 7) and 12- (k = 12) month follow-up (Q = 0.068, P = 0.794), type of treatment received (in-patient k = 9; intensive out-patient, out-patient, community k = 15; Q = 2.057, P = 0.152), and whether a majority of patients in a sample had (k = 11) or did not have (k = 13) major depression (Q = 0.563, P = 0.453). Sensitivity analyses indicated that the overall meta-analytical estimate of AA benefit was not impacted adversely or substantively by pooling randomized controlled trial (RCT) and observational samples (Q = 0.763, P = 0.382), pooling count, binary and ordinal-based AA (Q = 0.023, P = 0.879) and outcome data (Q = 1.906, P = 0.167) and reversing direction of correlations extracted from studies (Q = 0.006, P = 0.937). No support was found for publication bias.

Conclusions: Clinical referral of dual diagnosis patients to Alcoholics Anonymous is common and, in many cases, dual diagnosis patients who attend Alcoholics Anonymous will report higher rates of alcohol abstinence relative to dual diagnosis patients who do not attend Alcoholics Anonymous.