Abstract
BackgroundCommon mental disorders (CMDs) are a major cause of sickness absence and work disability. Although cognitive behavioural therapy (CBT) effectively reduces symptoms, effects on return-to-work (RTW) outcomes are limited. Work-focused CBT (W-CBT) integrates RTW processes into treatment, but the evidence base remains heterogeneous with unresolved methodological challenges. This systematic review and meta-analysis evaluated the effectiveness of W-CBT on psychiatric and work-related outcomes, examined moderators of treatment effects, assessed methodological rigor, and identified common intervention components, with particular attention to how RTW outcomes are analyzed.MethodsSystematic searches were conducted in Medline, PsycINFO, and Web of Science from inception to December 2025. Randomized and non-randomized studies evaluating CBT with an integrated work focus among working-age adults on sickness absence due to CMDs were included. Random-effects meta-analyses were performed using multilevel models with cluster-robust standard errors to address outcome dependency. Moderator analyses examined treatment content, number of sessions, and study characteristics. RTW outcomes were evaluated regarding their statistical properties, summary measures, and analytic methods used in primary studies. Risk-of-bias was assessed using Cochrane Risk of Bias toos version 2 and ROBINS-I. ResultsTwenty-four studies comprising 26 W-CBT treatment conditions (n = 3,279 participants) were included. Large within-group improvements in psychiatric symptoms were observed and maintained at follow-up, while between-group effects were statistically significant but small. At post-treatment, 56% of participants had returned to work, increasing to 87% at follow-up, with a higher likelihood of full RTW in W-CBT compared with controls. Greater CBT content, stronger work focus, and higher treatment dose were associated with larger symptom improvements. However, RTW outcomes showed substantial heterogeneity. Many studies used analytic approaches assuming normally distributed data for outcomes that were bounded, skewed, or zero-inflated, or reported hazard ratios with limited comparability, constraining valid synthesis.ConclusionsW-CBT is associated with meaningful symptom improvement and increased likelihood of RTW. However, misalignment between RTW outcome properties and analytic methods represents a major threat to validity. Future research requires improved outcome definitions, appropriate statistical modeling, and greater methodological consistency.