This review critically evaluates methodological and conceptual limitations in RCTs and meta-analyses of psychotherapies, and proposes advances including individual participant data (IPD) meta-analyses, component network meta-analyses (NMA), platform trials, and AI-driven adaptive interventions to strengthen the evidence base.
The authors argue that odds ratios (ORs) and standardized mean differences (SMDs) are the most generalizable summary indices for psychotherapy outcomes, and should be supplemented with control and experimental event rates (CER & EER) anchored to minimal important change (MIC) thresholds for clinical interpretability; platform trials like the RESiLIENT smartphone-based CBT trial exemplify more efficient, scalable designs.
As a narrative/methodological review, it does not report primary data or pooled effect sizes; recommendations are largely conceptual and their real-world implementation has not yet been empirically validated in the psychotherapy field.
Clinicians reading psychotherapy trial results should look beyond p-values to absolute event rates (CER/EER) and MIC-based thresholds to judge whether a treatment difference is clinically meaningful. Upcoming platform trial designs and AI-personalized adaptive interventions may soon offer more reliable guidance on which therapy components work best for which patients.
Explore related topics
