A meta-analysis of cognitive-behavioral therapy for alcohol or other drug use disorders: Treatment efficacy by contrast condition

Abstract

Objective: This meta-analysis examined 30 randomized controlled trials (32 study sites; 35 study arms) that tested the efficacy of cognitive-behavioral therapy (CBT) for alcohol or other drug use disorders. The study aim was to provide estimates of efficacy against three levels of experimental contrast (i.e., minimal [k = 5]; nonspecific therapy [k = 11]; specific therapy [k = 19]) for consumption frequency and quantity outcomes at early (1 to 6 months [kes = 41]) and late (8+ months [kes = 26]) follow-up time points. When pooled effect sizes were statistically heterogeneous, study-level moderators were examined.

Method: The inverse-variance weighted effect size was calculated for each study and pooled under random effects assumptions. Sensitivity analyses included tests of heterogeneity, study influence, and publication bias.

Results: CBT in contrast to minimal treatment showed a moderate and significant effect size that was consistent across outcome type and follow-up. When CBT was contrasted with a nonspecific therapy or treatment as usual, treatment effect was statistically significant for consumption frequency and quantity at early, but not late, follow-up. CBT effects in contrast to a specific therapy were consistently nonsignificant across outcomes and follow-up time points. Of 10 pooled effect sizes examined, two showed moderate heterogeneity, but multivariate analyses revealed few systematic predictors of between-study variance.

Conclusions: The current meta-analysis shows that CBT is more effective than a no treatment, minimal treatment, or nonspecific control. Consistent with findings on other evidence-based therapies, CBT did not show superior efficacy in contrast to another specific modality. (PsycINFO Database Record (c) 2019 APA, all rights reserved).

Figures

Figure 1.

Flow of primary study inclusion. Notes. K/k is defined as number of groups. CBT = cognitive behavioral therapy * E.g., dual diagnosis population; couples or self-help format; ineligible control condition.

Figure 2

Plot of assessment of publication bias. Notes. Assessment of bias in CBT effect in contrast to a minimal condition. The plot shows some asymmetry, but the rank order correlation shows a non-significant relationship between precision and effect size (τ = −.33, p > .05).

Figure 3.

Plot of assessment of publication bias. Notes. Assessment of bias in CBT effect in contrast to a non-specific therapy. The plot shows symmetry and the correlation test shows a non-significant relationship between precision and effect size (τ = .36, p > .05).

Figure 4.

Plot of assessment of publication bias. Notes. Assessment of bias in CBT effect in contrast to a specific therapy. The plot shows symmetry and the correlation test shows a non-significant relationship between precision and effect size (τ = .00, p > .05).