Seventeen RCTs (n reported as 2,712 in table and 2,302 in the text) including 46 treatment comparisons were included in the meta-analysis. The majority of included studies were considered to be Type 1 quality.
The results of individual treatments were presented in the paper. Other studies were reported without data in addition to the studies included in the meta-analysis. The authors suggested that there was well-established evidence in favour of MDFT, MST, FFT, CBT-G, and CBT-I, and that BSFT, BFT were probably efficacious. Promising support was identified for the Minnesota 12-step approach, TFT and SOFT.
Pooled results showed that there was a statistically significant reduction in drug use arising from the minimal treatment control group, effect size 0.19 (p<0.05). Pooled results of the three active treatments (CBT-I; CBT-G, and family therapy) compared with the minimal treatment control showed that mean effect size for the active interventions was significantly higher at 0.45 (p<0.01). This was specifically in terms of marijuana use. There was statistically significant heterogeneity (p<0.005). This was largely among studies of CBT-G and could not be explained by demographic variables alone.
Sensitivity analysis failed to identify any significant differences between the active treatment groups. A discussion of potential treatment mediators and moderators was presented in the paper.