More optimistic treatment expectations are associated with better outcomes through stronger group cohesion, but not dyadic alliance: Results from a naturalistic day clinic study in complex depression

Catherine Irniger^1,2*Johannes Vetter^2,3Steffi Weidt^2,3,4,5Erich Seifritz^2,3Martin grosse Holtforth^1,6Rainer Krähenmann^2,7

• ¹Universitat Bern, Bern, Switzerland
• ²Universitat Zurich, Zürich, Switzerland
• ³Psychiatrische Universitatsklinik Zurich, Zürich, Switzerland
• ⁴University of St Gallen, St. Gallen, Switzerland
• ⁵Psychiatry of St. Gallen, Wil, Switzerland
• ⁶Inselspital Universitatsspital Bern, Bern, Switzerland
• ⁷Psychiatrische Dienste Thurgau Psychiatrische Klinik Munsterlingen, Münsterlingen, Switzerland

More optimistic treatment expectations are typically associated with better psychotherapy outcomes, including in depression. Meta-analytic findings suggest that this relationship is mediated by the dyadic alliance. In settings including group formats, however, patients also interact and build relationships with one another. Thus, group relationship may represent an important mediator, particularly given that depression is often accompanied by interpersonal difficulties. Therefore, we aimed to investigate the potential mediating role of group relationship above and beyond the dyadic alliance, as, to the best of our knowledge, this has not yet been examined. We hypothesized that more optimistic treatment expectations would be both, directly and indirectly associated with better outcomes through a stronger dyadic alliance and better group relationship. Data were drawn from a naturalistic observational study at a local day clinic in Switzerland. Patients were assigned to one of two treatment tracks: Cognitive Behavioral Analysis System of Psychotherapy (CBASP) or Short-Term Psychodynamic Psychotherapy (STPP). Fifty adult patients with complex courses of depression were included in the analysis. Treatment expectations were assessed at baseline using the Treatment Expectation Questionnaire (TEX-Q), dyadic alliance within the first weeks using the Helping Alliance Questionnaire (HAQ), and group relationship with the subscale group cohesion of the Inpatient and Day-Clinic Experience Scale (IDES) retrospectively at treatment end. Treatment outcome was quantified as residual change scores on the revised Beck Depression Inventory (BDI-II) from baseline to treatment end. A parallel mediation analysis controlling for treatment track and duration was conducted in PROCESS using bias-corrected bootstrapping. More optimistic treatment expectations were significantly associated with higher levels of dyadic alliance and group cohesion. However, only group cohesion was associated with outcome, but not dyadic alliance. The direct association of treatment expectations with outcome was non-significant. These findings provide preliminary support for an indirect association of treatment expectations with outcome through group cohesion in depression. However, the results should be interpreted with caution due to several methodological constraints. Future research should replicate these findings in larger samples, with prospective cohesion measurements, and examine temporal dynamics. Clinically, these findings highlight the potential relevance of explicitly addressing treatment expectations and group cohesion.