Regression examining whether study-level characteristics explain between-trial variation in effect size.
Meta-regression extends meta-analysis by testing whether trial-level variables (e.g., changes in body weight, blood pressure, VO2max) statistically explain variation in the pooled effect across studies. Because it operates on study-level averages rather than individual patient data, it can only support correlational claims about moderators, not causal inference for an individual patient. Findings should be treated as hypothesis-generating unless replicated with individual-patient-data analyses.
This guide was auto-drafted and is pending editorial review.