Nineteen studies were included in the review: 5 comparative studies (n=262) and 14 single group-studies (n=300).
One study was considered to have a low risk of bias, 5 studies were considered to have a low-to-moderate risk of bias, 10 studies were considered to have a moderate risk of bias, and 3 studies were considered to have a moderate-to-high risk of bias.
In the controlled studies, effect sizes for the severity of OCD ranged from 0.79 to 2.73. The estimates were not pooled because of substantial statistical heterogeneity (I-squared 89.3%). In the single-group studies, the mean effect sizes ranged from 0.78 to 4.38.
Low risk of bias was associated with lower adjusted effect sizes; one study (20%) in the low or low-to-moderate bias categories reported an adjusted effect size above 1, while 11 studies (92%) in the moderate or moderate-to-high bias categories reported effect sizes around 1 and seven (58%) reports effects above 2.
Three controlled studies assessed CBT alone against medication. No between-group difference was found for treatment groups for OCD severity (WMD -3.87, 95% CI: -8.15, 0.41), based on 2 studies
(CBT versus clomipramine and CBT versus sertraline). No between-group differences were found for the proportion of participants continuing to have OCD at post-treatment follow-up (RR 0.75, 95% CI: 0.54, 1.05), based on 2 studies.
A result in favour of the combined treatment compared with medication alone was found (WMD -4.55, 95% CI: -7.40, -1.70), based on 2 controlled studies. One study found that CBT combined with medication was significantly less likely to continue to have OCD at post-treatment follow-up than medication alone (RR 0.59, 95% CI: 0.38, 0.92) and placebo (RR 0.48, 95% CI: 0.32, 0.72), but no significant difference was found in comparison with CBT alone (RR 0.76, 95% CI: 0.47, 1.26).