Purpose: In the past decades, the effects of problem-solving therapy (PST) for depression have been examined in several randomized controlled studies. However, until now no meta-analysis has tried to integrate the results of these studies. Methods: We conducted a systematic literature search and identified 13 randomized studies examining the effects of PST, with a total of 1133 subjects. The quality of studies varied. Results: The mean standardized effect size was 0.34 in the fixed effects model and 0.83 in the random effects model, with very high heterogeneity. Subgroup analyses indicated significantly lower effects for individual interventions in studies with subjects who met criteria for major depression, studies in which intention-to-treat analyses were conducted instead of completers-only analyses, and studies with pill placebo and care-as-usual control groups. Heterogeneity was high, and the subgroup analyses did not result in clear indications of what caused this high heterogeneity. This indicates that PST has varying effects on depression, and that it is not known to date what determines whether PST has larger of smaller effects. Conclusion: Although there is no doubt that PST can be an effective treatment for depression, more research is needed to ascertain the conditions and subjects in which these positive effects are realized. © 2006 Elsevier Masson SAS. All rights reserved.