Which of the following describes common methods to evaluate heterogeneity in meta-analysis?

• A. Subgroups, stratified analyses, graphical assessment, and meta-regression
• B. Funnel plots cannot be used for heterogeneity assessment
• C. Only fixed-effect models assess heterogeneity
• D. None of the above

Answer: (A)

Heterogeneity in meta-analysis refers to variability in study results that goes beyond what would be expected by chance alone. To explore and explain that variability, several approaches are commonly used. Subgroup analyses and stratified analyses involve dividing studies into groups based on characteristics such as population, intervention type, dose, or study design, and then comparing effect estimates across these groups. If differences appear between groups, that points to sources of heterogeneity related to those traits. Graphical assessment helps you visually inspect consistency across studies; forest plots show each study’s effect and confidence interval, making it easy to see whether results cluster together or diverge, and other plots can reveal patterns suggesting heterogeneity. Meta-regression extends this by modeling the effect size as a function of study-level characteristics to test whether specific factors explain some of the between-study variability. Additionally, numerical measures like Cochran’s Q and I-squared quantify how much heterogeneity exists and how confident you can be that observed differences are not due to chance.

Funnel plots, while useful for detecting publication bias and small-study effects, are not primary tools for assessing heterogeneity, and heterogeneity can be present regardless of whether a fixed-effect or random-effects model is used. In practice, these methods collectively provide a comprehensive view of whether and why study results differ, guiding interpretation and model choice.