Inconsistent outcome measurement in depression psychotherapy trials: A systematic historical and meta-analytic review over the past 50 years

BACKGROUND: Psychotherapy studies employ different outcome measurement instruments (OMIs) for depression, which may limit the interpretability of meta-analytic findings. This meta-epidemiological study aimed to examine trends in OMI use in randomised controlled trials (RCTs) for depression, consistency of measurement practices over time, and their influence on meta-analytic effect estimates.

METHODS: A systematic literature search was conducted in four databases up to September 01, 2024. We included RCTs comparing psychotherapy to control conditions in adults with depression using OMIs for depression, present in at least ten comparisons. Change over time in OMI usage frequencies and heterogeneity (perplexity) was analysed with Poisson and linear regression models. Confounder-adjusted meta-regression and network meta-analysis examined the dependence of treatment effect estimates on different OMIs.

RESULTS: We included 492 trials (1260 comparisons in ~51,000 patients), published between 1977 and 2024, using 17 different OMIs for depression. The proportion of trials using the Patient Health Questionnaire (PHQ-9; b = 1.04, SE = 0.30, p < 0.001) and Hamilton Rating Scale for Depression (HDRS; b = 0.29, SE = 0.12, p = 0.021), but also heterogeneity in OMI usage (b = 0.12, SE = 0.01, p < 0.001), increased significantly over time. Particularly the Centre for Epidemiological Studies Depression Scale (CES-D) and Geriatric Depression Scale (GDS) predicted increased and decreased treatment effect estimates.

CONCLUSIONS: Despite increases in PHQ-9 and HRSD administration, OMI heterogeneity in depression psychotherapy research has grown over the past 50 years. Since different OMIs, especially CES-D and GDS, affect treatment effect estimates, meta-analyses incorporating various OMIs for depression should be interpreted cautiously. Core outcome sets for depression could improve robustness of meta-analytic evidence.