Meta-analysis of couple therapy in non-randomized clinical trial studies: Individual and couple level outcomes.

Abstract

Couple therapy has outperformed control conditions in randomized clinical trials (RCTs). However, there have been some questions whether couple treatment in naturalistic settings is as effective as those with more rigorous controls. The current meta-analysis examined 48 studies of couple therapy in non-randomized clinical trials. The pre-post effect size was Hedge's g = 0.522 for relational outcomes and Hedge's g = 0.587 for individual outcomes. However, there was significant heterogeneity in the results. Several moderators explained some of the variance in these estimates. For relationship outcomes, studies who had older couples and longer length of relationship had better outcomes. Studies with a higher percentage of racial/ethnic minority (REM) couples and studies in Veteran Affairs Medical Centers (VAMC) had lower relational outcomes. For individual outcomes, studies that had more sessions, older couples, and VAMC had better outcomes. Studies with a higher percentage of REM couples also had worse individual outcomes. Trainee status was not consistently related to relational or individual outcomes. Implications for research and practice are provided.