Heterogeneity of Treatment Outcomes Across Therapists and Sites in a Randomized Multicentre Psychotherapy Trial

Establishing robust evidence for psychotherapeutic treatment efficacy is crucial in evidence-based medicine for mental disorders. Randomized controlled trials (RCTs) are key to minimizing biases such as selection effects and baseline imbalances between study groups. However, another challenge to robust evidence in psychotherapy research is heterogeneity in treatment outcomes due to therapists and clinical sites. While this has been frequently observed in naturalistic settings, therapist- and site-related heterogeneity in treatment outcomes has been understudied in RCTs.

The present study addresses this gap in a secondary data analysis, examining how therapists and clinical sites differ in treatment outcomes and differential/average treatment effect (i.e., outcome differences between treatment groups) within a large, multicentre RCT. We analysed data from 255 patients with chronic depression treated by 79 therapists in nine study sites, that received two different active psychotherapeutic interventions.

Therapist- and site-related variances in treatment outcomes appeared relatively small, accounting for 1.1% [0.0%, 8.1%] and 1.7% [0.0%, 9.9%] of the total variance, respectively. Notably, site-related variance in differential treatment effects appeared relatively larger at 12.5% [0.1%, 44.4%]. These variances were only partially explained by patient or therapist characteristics.

While the sample size only allowed to estimate the variance contributions of therapists and sites with high uncertainty, the relative size comparison points to the importance of considering site heterogeneity in evaluating RCTs' differential treatment effects. Further research on site characteristics' impact could enhance understanding of psychotherapeutic treatment efficacy across diverse contexts.