Meta-analysis of the alliance-outcome relation in couple and family therapy

This article reviews meta-analytic evidence for the alliance-outcome relation in couple and family therapy (CFT), with implications for clinical practice. We begin by describing the unique features of CFT alliances and their measurement, followed by two case descriptions. We explain that due to the systemic context of CFT, each patient's personal alliance with the therapist affects and is affected by other family members' levels of collaboration. Because family members often seek help when they are in conflict with one another, "split" alliances are common, as are problematic within-system alliances, defined as the degree to which family members agree on the nature of their problems and value participating in therapy together to achieve shared goals. In our meta-analysis of 48 studies with 40 independent samples, we used a three-level random effects model (Ns = 2,568 families, 1,545 couples, and 491 effect sizes) and found r = .297. In another analysis with seven independent samples and 31 effect sizes, the split alliance-outcome association was also significant, r = .316, indicating that the more split or unbalanced the alliance, the poorer the outcome. Moderator analyses showed that alliance-outcome associations did not differ for couple versus family therapy, but correlations were significantly stronger in samples with younger problem children, older adults, proportionally more male youth and adults, and when the families voluntarily sought help (as compared with involuntary or mandated families). The article concludes with research-informed strategies for building and sustaining strong systemic alliances in CFT. (PsycINFO Database Record (c) 2018 APA, all rights reserved).