Objective: The aims of this study were to examine labor supply effects of interventions for major depressive disorder and to compare these effects with a summary measure of clinical effectiveness.
Methods: Research articles published in English-language journals from 1980 through May 2004 were searched by using five research databases. Only randomized trials that included a placebo group or a usual care group were eligible for the study, regardless of the specific type of intervention. Valid trials were those that enrolled adult patients with major depressive disorder and assessed changes in labor output by using a measure of time worked or labor market participation. From a total of 706 trials uncovered from the database searches, only four met all inclusion criteria. Trial outcomes were transformed into standardized effect sizes on the basis of Cohen's d. Hierarchical linear models were used to separately pool work outcomes and clinical outcomes.
Results: An improvement of .34 standard deviation was found in the size of the clinical effect of interventions compared with placebo or usual care among 1,261 unique patients with depression. An improvement of .12 standard deviation was found in the size of the effect on labor supply among 1,848 unique patients.
Conclusions: Although the interventions studied were associated with reduced symptoms of depression and increased labor output, the labor benefits were small according to standard benchmarks used in interpreting the substantive significance of values of Cohen's d. The difference in effects may have been due to different underlying efficacies, brief durations of follow-up, or extrinsic factors that affect labor supply.