Therapist effects in psychotherapy: A random-effects modeling of the National Institute of Mental Health Treatment of Depression Collaborative Research Program data

Abstract

Data for completer and intent-to-treat samples from the two psychotherapy conditions of the National Institute of Mental Health Treatment of Depression Collaborative Research Program were analyzed to estimate the proportion of variability in outcomes resulting from therapists. Therapists, who were nested within treatments, were considered a random factor in multilevel analyses. These analyses, which modeled therapist variability in several different ways, indicated that about 8% of the variance in outcomes was attributable to therapists, whereas 0% was due to the particular treatment delivered. When therapist effects were appropriately modeled, previously detected differences in efficacy between the two psychotherapy conditions for more severely depressed patients disappeared, as predicted by methodological considerations.