Rethinking and advancing meta-analyses and randomized trials of psychotherapies.

Randomized controlled trials (RCTs) and their meta-analyses remain the foundation of evidence-based psychotherapy, yet many of them are currently limited by methodological weaknesses, conceptual confusion, and interpretive challenges. This review first clarifies key distinctions-between response and effect, between minimal important change (MIC) and smallest worthwhile difference (SWD), and between absolute versus relative effect measures-showing how these distinctions contribute to valid interpretation of treatment outcomes. We argue for odds ratios (ORs) and standardized mean differences (SMDs) as the most generalizable summary indices, to be supplemented with control and experimental event rates (CER&EER) by MIC-based thresholds for clinical interpretability. Advances in evidence synthesis, particularly individual participant data (IPD) and component network meta-analyses (NMA), enable the identification of effect modifiers and the isolation of efficacious therapeutic components, although they face challenges of data availability, harmonization, and reporting standards. To strengthen the evidence base further, larger and more efficient trials are needed. Platform trials and master protocols, drawing on shared controls and adaptive features, promise greater efficiency and comparability, as exemplified by the RESiLIENT digitally centralized trial of smartphone-based CBT. Such designs, combined with adaptive interventions and AI-driven personalization, illustrate how psychotherapy research can move beyond small, fragmented RCTs towards scalable precision medicine.