Meta-analysis of 2 × 2 cell counts. Choose risk ratio, odds ratio, or risk difference; log-scaled measures are back-transformed for the forest plot.
For binary-outcome meta-analyses, three effect measures are in use: RR (relative risk — directly interpretable, preferred for common outcomes), OR (odds ratio — symmetric in the cells, standard for case-control), RD (risk difference — interpretable in absolute terms, common in clinical trials).
Pooling happens on the log scale for RR and OR (which are positive and skewed). For RD pooling is on the raw scale. Sparse-data corrections (add 0.5 to empty cells) prevent log(0) catastrophes.
Four cell-count columns (a / b / c / d per study) + optional labels.
Pooled effect with CI, heterogeneity stats, forest + funnel plots, Egger's test.
Heterogeneity is more common in binary-outcome meta-analyses than in continuous-outcome ones; expect higher I².