Objective: Developments in working alliance theory posit that the therapist's attention to fluctuations in the alliance throughout treatment is crucial. Accordingly, researchers have begun studying the alliance as a time-varying mechanism of change rather than as a static moderator. However, most studies to date suffer from bias owing to the nonindependence of error term and predictors (endogeneity).
Method: Patients with major depressive disorder (N = 84) from a randomized trial comparing cognitive-behavioral therapy with interpersonal psychotherapy filled out the Beck Depression Inventory-II before each session. After each session, patients and therapists filled out the Working Alliance Inventory short forms. Data were analyzed using the generalized method of moments for dynamic panel data, a method commonly applied in econometrics to eliminate endogeneity bias.
Results: Improvement of the alliance predicted significant reduction of depressive symptoms by the next session (patient rating: b = -4.35, SE = 1.96, p = .026, 95% confidence interval [CI] [-8.19, -0.51]; therapist rating: b = -4.92, SE = 1.84, p = .008, 95% CI [-8.53, -1.31]). In addition, there was a significant delayed effect on the session after the next (patient rating: b = -3.25, SE = 1.20, p = .007, 95% CI [-5.61, -0.89]; therapist rating: b = -5.44, SE = 1.92, p = .005, 95% CI [-9.20, -1.68]).
Conclusion: If the quality of patient-therapist alliance is improved in a given treatment session, depressive symptoms will likely decrease by the next session. The most important limitation of this study is its relatively small sample size. (PsycINFO Database Record
Trial registration: ClinicalTrials.gov NCT01851915.
(c) 2016 APA, all rights reserved).