The Within-Client Alliance-Outcome Relationship: A Response Surface Analysis

The authors examined how stability/change in working alliance predicted subsequent symptoms, and how stability/change in symptoms predicted subsequent alliance in a sample of 188 adult clients with 44 doctoral student therapists over the course of 893 eight-session time periods of individual psychodynamic psychotherapy. Clients completed the Working Alliance Inventory-Short Revised (WAI-SR; Hatcher & Gillaspy, 2006) after every session and the Outcome Questionnaire-45 (OQ; Lambert et al., 1996) before intake and every eighth session. The authors used multilevel polynomial regression and response surface analyses to examine within-client effects. The authors found that change in the alliance across an eight-session period did not have an immediate temporal effect on symptoms, but when alliance was stable and stronger compared to other periods, subsequent symptoms were lower. Similarly, change in symptoms across an eight-session period did not have an immediate temporal effect on alliance, but when symptoms were stable and lower compared to other periods, subsequent alliance was stronger. These results suggest that sustained improvements in the alliance contribute to subsequent symptom improvements, and vice versa. The authors conclude that it is important to work to improve and maintain improvements in the working alliance and symptoms. Limitations and future directions are discussed.