Strata Academy

Meta-analysis: pooling, heterogeneity, and GRADE

I², fixed vs random effects, and when not to pool

Should studies be pooled?

Pooling assumes studies estimate a similar underlying effect. Clinically diverse populations, interventions, or outcomes may make a single summary misleading even if heterogeneity tests are non-significant.

Always read the review protocol and inclusion criteria before trusting a forest plot.

Measuring heterogeneity

Cochrane recommends reporting I² and τ² with confidence intervals where possible. I² describes the proportion of variability due to heterogeneity rather than chance – but depends on precision of included studies.

High I² suggests exploring sources via subgroup or meta-regression (pre-specified), or presenting review narratively.

• I² thresholds are guides, not rigid rules (Cochrane Handbook discusses interpretation)
• Prediction intervals show expected effect in a new study
• Check whether fixed-effect or random-effects model is justified

Small-study effects and publication bias

Funnel plot asymmetry may indicate publication bias or other small-study effects. Formal tests (e.g. Egger) are adjuncts, not proof.

AMSTAR 2 and ROBIS ask whether the review addressed risk of bias in included studies and in the review process itself.

GRADE certainty

GRADE rates certainty of evidence (high to very low) based on risk of bias, inconsistency, indirectness, imprecision, and publication bias. It applies to bodies of evidence, not single trials.

Interactive version (quizzes, walkthroughs) loads when JavaScript is enabled.