How much results differ across studies in a meta-analysis beyond chance.
Heterogeneity describes variability in effect estimates across the studies pooled in a systematic review or meta-analysis, arising from differences in populations, interventions, outcome measures, or methodological quality. It is commonly quantified with statistics such as I² or Cochran's Q, though thresholds for 'high' heterogeneity are somewhat arbitrary and should be interpreted alongside the forest plot and clinical context. High heterogeneity weakens confidence in a single pooled effect estimate and may warrant subgroup analysis or a narrative synthesis instead.
This guide was auto-drafted and is pending editorial review.