Which method is used to explain heterogeneity by incorporating study-level covariates?

• A. Meta-regression
• B. Funnel plot
• C. Sensitivity analysis
• D. Random-effects model

Answer: (A)

Heterogeneity across studies can reflect differences in study characteristics. Meta-regression explicitly tests whether that variation in effect sizes can be explained by study-level covariates (like publication year, population characteristics, or study design features). By modeling the effect size as the outcome and regressing it on these covariates, you can see if changes in the covariates are associated with shifts in the observed effect, and whether including them reduces unexplained between-study variation. This is different from a random-effects model, which accounts for heterogeneity but does not use covariates to explain it. It’s also distinct from a funnel plot, which assesses publication bias, and from sensitivity analyses, which test robustness by altering data or methods rather than explaining heterogeneity with covariates. Keep in mind that meta-regression uses study-level data and can be limited by ecological bias and the number of studies available.