Forest Plot

Definition

A forest plot is a graphical display of effect-size estimates and their confidence intervals from multiple studies or subgroups. Each study is represented by a point estimate (usually a square) and a horizontal line showing the confidence interval. A vertical reference line at the null value (e.g., 1 for odds ratios, 0 for mean differences) helps readers judge statistical significance and consistency across studies.

Why It Matters

Forest plots are the standard visual tool in meta-analysis and systematic reviews. They allow readers to see at a glance whether studies agree or disagree, which studies drive the overall effect, and whether any study produces an outlying result. They also reveal heterogeneity — the degree to which true effects vary across studies — which is crucial for deciding whether a pooled estimate is meaningful.

Example

A meta-analysis of 12 randomised trials examines whether exercise reduces depression. The forest plot shows that 10 of 12 trials favour exercise (effect sizes to the left of the null line), while 2 small trials show no effect. The pooled estimate is a standardised mean difference of -0.66 (95% CI [-0.89, -0.43]), represented by a diamond at the bottom. The plot immediately communicates that the evidence is consistent, precise, and clinically meaningful.

Related Terms

• Meta-Analysis
• Funnel Plot
• Effect Size
• Confidence Interval
• Odds Ratio

Software Notes

• SPSS: SPSS does not produce forest plots natively for meta-analysis. Use Review Manager (RevMan), Comprehensive Meta-Analysis, or export to R.
• R: metafor::forest(rma_model) from the metafor package is the standard approach. meta::forest(metagen(...)) from the meta package offers additional customisation. ggforest() from survminer plots Cox regression hazard ratios.
• Stata: metan effect lower upper produces forest plots in Stata. metan is a user-written command installed via ssc install metan. forestplot is an alternative graphical package.