Cuijpers · JAMA psychiatry 2025 · systematic review and meta-analysis · n=375 trials (32,968 participants)

Cognitive Behavior Therapy for Mental Disorders in Adults: A Unified Series of Meta-Analyses.

Cited 56 times in the scientific literature.

Level 1 - systematic review of randomized trials

Systematic review and meta-analysis of randomized controlled trials

PubMed 40238104 · doi:10.1001/jamapsychiatry.2025.0482 · record verified 2026-08-30

What was done

A unified series of meta-analyses evaluated randomized clinical trials comparing cognitive behavior therapy (CBT) with inactive control conditions in adults diagnosed with 1 of 11 mental disorders via clinical interview (major depression, bipolar disorder, psychotic disorders, OCD, PTSD, bulimia nervosa, binge eating disorder, panic disorder, social anxiety disorder, generalized anxiety disorder, and specific phobia). Databases (PubMed, PsycINFO, Embase) were searched up to January 1, 2024. Data extraction, Cochrane Risk of Bias 2 assessments, and random-effects meta-analyses were standardized across all disorders. The primary outcome was disorder-specific symptom severity at posttreatment measured as standardized mean difference (Hedges g).

What was found

A total of 375 trials (423 comparisons; 32,968 patients; mean age 43.4 years, 68% female) were analyzed. Posttreatment effect sizes (g) compared to all controls were: - Below 0.50 for bipolar disorder (g = 0.31) and psychotic disorders. - Between 0.50 and 1.00 for panic disorder, social anxiety disorder, generalized anxiety disorder, bulimia nervosa, binge eating disorder, depression, and OCD. - Above 1.00 for specific phobias and PTSD (g = 1.27). Effect sizes were substantially larger in waitlist-controlled trials (g > 0.94) compared with care-as-usual trials (g = 0.22 to 1.13). Study dropout within CBT arms ranged from 8% for specific phobias to 24% for PTSD.

Why it matters

This study provides a standardized, cross-disorder benchmark demonstrating that CBT is broadly effective across major psychiatric conditions, though effects are modest for bipolar and psychotic disorders. It also quantitatively highlights that waitlist controls inflate treatment effect estimates relative to usual care comparators.

Limits

The review was restricted to adult populations, relied solely on immediate posttreatment outcomes without evaluating long-term durability, and compared CBT primarily to inactive or usual-care controls rather than alternative active psychotherapies. Specific risk of bias distributions were not reported in the abstract.

Cited by