Sixteen RCTs (756 participants) were included in the review. Sample sizes ranged from 18 to 108.
Compared with control groups, CBT was significantly more effective at reducing OCD symptoms post-treatment (g 1.39, 95% CI 1.04 to 1.74; 16 RCTs) and at follow-up (g 0.43, 95% CI 0.12 to 0.74; three RCTs). There was significant heterogeneity for this analysis.
CBT was also more effective than control groups at reducing depression symptoms post-treatment (g 0.51, 95% CI 0.21 to 0.82; nine RCTs). Insufficient data was reported to enable analysis at follow-up.
Sub-group analyses reported that studies using waiting list controls had larger effect sizes than studies using placebo controls. Effect sizes were smaller for RCTs including adults than for RCTs including children. No significant differences in effect sizes were reported for a range of other moderator variables. Full details of sub-group analyses were reported in the review.
There was no significant evidence of publication bias.