Objective This study systematically reviewed and meta-analyzed the differential attrition and utilization of occupational mental health interventions, specifically examining delivery methods (internet-based versus in-person).

Methods The research, with papers spanning 2010–2024, involved filtering criteria and comprehensive searches across PubMed, Scopus, and Web of Science Core (PROSPERO registration n. CRD42022322394). Of 28 683 titles, 84 records were included in the systematic review, with 75 in meta-analyses. Risk of bias was assessed through the revised Cochrane risk of bias tool for randomized control trials and funnel plots. Differential attrition across studies was meta-analysed through a random-effects model with limited maximum-likelihood estimation for the degree of heterogeneity.

Results Findings reveal higher mean differential attrition in the intervention group, indicating a potential challenge in maintaining participant engagement. The attrition rates were not significantly influenced by the mode of intervention delivery (internet versus in-person). Compensation for participation and year of publication could potentially influence differential attrition from baseline to follow-up measurements.

Conclusions These results suggest a need for cautious consideration of attrition in occupational mental health intervention study designs and emphasize the importance of adapting statistical analyses to mitigate potential bias arising from differential attrition.

Key terms differential attrition; drop-out; employee; intervention; mental health; meta-analysis; methodology; psychology; randomized controlled trial; systematic review

• Supplementary material