Abstract
BACKGROUND: A previously published article in this journal reported the service effects from 103 services within the UK Improving Access to Psychological Therapies (IAPT) initiative and the comparative effectiveness of CBT and Counselling provision. All patients received High-intensity CBT or High-intensity Counselling, but some also received Low-intensity CBT before being stepped-up to High intensity treatments. The report did not distinguish between patients who received low-intensity CBT before being stepped-up. This article clarifies the basis for collapsing low- and high-intensity interventions by analysing the four treatment conditions separately. METHOD: Data from 33,243 patients included in the second round of the National Audit of Psychological Therapies (NAPT) were re-analysed as four separate conditions: High-intensity CBT only (n = 5975); High-intensity Counselling only (n = 3003); Low-intensity CBT plus High-intensity CBT (n = 17,620); and Low-intensity CBT plus High-intensity Counselling (n = 6645). Analyses considered levels of pre-post therapy effect sizes (ESs), reliable improvement (RI) and reliable and clinically significant improvement (RCSI). Multilevel modelling was used to model predictors of outcome, namely patient pre-post change on PHQ-9 scores at last therapy session. RESULTS: Significant differences obtained on various outcome indices but were so small they carried no clinical significance. Including the four treatment groups in a multilevel model comprising patient intake severity, patient ethnicity and number of sessions attended showed no significant differences between the four treatment groups. Comparisons between the two high-intensity interventions only (N = 8978) indicated Counselling showed more improvement than CBT by 0.3 of a point on PHQ-9 for the mean number of sessions attended. However, this result was moderated by the number of sessions and for 12 or more sessions, the advantage went to CBT. CONCLUSIONS: This re-analysis showed no evidence of clinically meaningful differences between the four treatment conditions using standard indices of patient outcomes. However, a differential advantage to high-intensity Counselling for fewer than average sessions attended and high-intensity CBT for more than average sessions attended has important service implications. The finding of equivalent outcomes between high-intensity CBT and Counselling for more severe patients also has important policy implications. Empirically-informed procedures (e.g., predictive modelling) for assigning patients to interventions need to be considered to improve patient outcomes.