« Back to Glossary Index

Meta-Regression Analysis


When summarizing patient or design characteristics at the individual trial level, meta-analysts risk failing to detect genuine relationships between these characteristics and the size of treatment effect. Further, the risk of obtaining a spurious explanation for variable treatment effects is high when the number of trials is small and many patient and design characteristics differ. Meta-regression techniques can be used to explore whether patient characteristics (e.g., younger or older patients) or whether design characteristics (e.g., studies of low or high quality) are related to the size of the treatment effect.

Categories: Statistics