Meta-Analysis Overview

Not every systematic review includes a meta-analysis — it is only appropriate when studies are sufficiently similar (clinically and methodologically) to combine. A meta-analysis calculates a weighted average of effect sizes across studies, giving more weight to larger, more precise studies. The result is displayed in a forest plot: each study is a horizontal line (showing its confidence interval) with a square (representing the point estimate, sized proportionally to study weight). The diamond at the bottom represents the pooled effect. Heterogeneity — variation in results between studies beyond what chance would explain — is assessed using the I-squared statistic. Low I-squared (under 25%) suggests results are consistent; high I-squared (over 75%) means substantial unexplained variation and signals caution in interpreting the pooled result. When heterogeneity is high, subgroup analysis and meta-regression can explore possible explanations. Funnel plots help detect publication bias: asymmetry suggests small negative studies may be missing. A meta-analysis of poor-quality studies produces a precise but potentially misleading answer — 'garbage in, garbage out' applies.