Effectiveness and acceptability of different psychotherapies for obsessive-compulsive disorder: network meta-analysis

Various psychotherapies are available for individuals with obsessive-compulsive disorder (OCD). However, the comparative effectiveness and acceptability of these therapies remain unclear.

To examine the comparative effectiveness and acceptability of different psychotherapies for OCD.

A living database of psychological interventions for OCD was utilised, and randomised controlled trials comparing psychotherapies with each other/control groups were included. Pairwise and network meta-analyses were conducted using a random-effects model. Comparative standardised mean differences (SMDs) were pooled for effectiveness in reducing OCD symptom severity post-treatment. Relative risks were calculated for acceptability. Sensitivity analyses were conducted by repeating the main analyses while controlling for specific variables to test the robustness of the findings.

A total of 68 controlled trials (76 comparisons, 4019 patients) were included, involving 7 psychotherapeutic approaches. All psychotherapies were significantly more effective than both waitlist (SMD -1.40 to -0.96) and pill placebo (SMD -1.44 to -1.00). Except for mindfulness-based therapy, all approaches were more effective than both care-as-usual (SMD -0.98 to -0.67) and psychological placebo (SMD -0.95 to -0.63). Sensitivity analyses excluding outliers, studies with comorbidities, comparisons with waitlist controls and comparisons supported by only a single study, as well as analyses restricted to adults with OCD, yielded results consistent with the main analyses. When restricted to studies rated as low risk of bias, all therapies except mindfulness-based therapy and the inference-based approach remained significantly more effective than waitlist. No significant differences were found among psychotherapies regarding effectiveness and acceptability.

Psychotherapies are similarly effective and acceptable for treating OCD. However, these findings should be interpreted with caution due to limited statistical power, substantial heterogeneity and a high risk of bias across many included studies. More methodologically rigorous research is needed to validate and strengthen the current evidence base.