Therapist and clinic effects in psychotherapy: a three-level model of outcome variability

Objective: The study aimed to 1) investigate the effect of treatment location on clinical outcomes for patients receiving psychological therapy (a clinic effect, akin to the concept of a therapist effect), and 2) assess the impact of explanatory individual and aggregate demographic and process variables on the clinic and therapist effects.

Method: The sample comprised 26,888 patients, seen by 462 therapists, across 30 clinics. Mean patient age was 38 years (69% female, 90% White, 92% planned ending). The dependent variable was patients’ post-therapy score on the Clinical Outcomes in Routine Evaluation – Outcome Measure. An incremental three-level multilevel model was constructed. Markov Chain Monte Carlo estimation created 95% probability intervals for the clinic and therapist effects.

Results: A three-level model with no explanatory variables detected a clinic effect of 8.2%, significantly larger than the therapist effect of 3.2%. Adding explanatory variables significantly reduced the clinic effect to 1.9% but did not significantly alter the therapist effect (3.4%). Patient-level symptom severity and employment status, and clinic-level percentage of White patients and healthcare sector explained the most clinic outcome variance and overall outcome variance.

Conclusions: Substantial variability in clinical outcomes was found between clinics providing psychological therapy. Socioeconomic mix of patients explained significant proportions of variability at the clinic level but not the therapist level. Clinical implications include the need to go beyond the therapist-patient interaction in order to deliver effective psychological therapy. Future research is also needed to identify the mechanisms by which clinic and/or area-level factors impact on clinical outcomes.