Variability in treatment effects in an English national dataset of psychological therapies: the relationships between severity, treatment duration, and therapy type

Background: Comparisons between bona fide psychological therapies generally report no effects or small effects favoring cognitive behavioral therapy (CBT), suggesting that differences between therapy modalities are of little importance.

Methods: An observational, cohort study of patients (N = 11,116), drawn from the English Talking Therapies program. Patients presented with depression and were treated by CBT or person-centered experiential therapy (PCET), 67 % were female and the age range was 15–94. Multilevel modeling was used to identify variables associated with outcomes and CBT and PCET outcomes were compared dependent on the number of sessions patients attended and the severity of their depression at intake.

Results: Although overall a small effect size of 0.14 (0.10, 0.18) favored CBT, we found differing effect sizes depending on patient severity and the number of sessions they had attended at outcome. For non-clinical and moderately depressed patients no significant differences between therapies were found. For moderately-severe and severe patients there was a crossover in effectiveness with PCET more effective up to 5 or 6 sessions and CBT more effective thereafter. However, small effects in favour of PCET were unreliable. The largest effects were found for CBT for patients who were moderately-severe and had >10 sessions (d = 0.30, 0.12, 0.48) or were severe and had >12 sessions (d = 0.26 (0.02, 0.49).

Limitations: No data was available on therapy allocation decisions or therapists and the results may not be generalizable to non-NHS Talking Therapies services.

Conclusions: Small effects can mask important contextual differences between therapies and their study can inform allocation decisions.