Forest plot example: six studies and a pooled risk ratio

What it shows

A forest plot of six studies and their fixed-effect meta-analysis. Studies A to F are invented for teaching and do not exist. Each study compares events in a treatment arm and a control arm, so a risk ratio below 1 favors the treatment.

Each study is a square at its risk ratio, sized by its weight, with a line for its 95% confidence interval. The pooled result is the diamond, whose width is its 95% confidence interval. Columns beside the plot give each study’s counts, risk ratio and weight.

How to read it

1. 1

   The vertical line at 1 is no effect. The axis is logarithmic, from 0.25 to 4, so equal distances mean equal ratios.

2. 2

   A study whose line crosses 1 is compatible with no effect at the 95% level: here A, B, C and E.

3. 3

   Larger squares carry more weight. Study D, at 33.4%, contributes most to the pooled estimate.

4. 4

   The diamond, a pooled risk ratio of 0.70 [0.57, 0.86], does not cross 1.

The data

• Study A: 12/100 vs 20/100, RR 0.60 [0.31, 1.16], weight 9.6%.
• Study B: 33/250 vs 44/250, RR 0.75 [0.49, 1.14], weight 24.0%.
• Study C: 9/60 vs 11/60, RR 0.82 [0.37, 1.83], weight 6.4%.
• Study D: 44/400 vs 68/400, RR 0.65 [0.45, 0.92], weight 33.4%.
• Study E: 18/120 vs 14/118, RR 1.26 [0.66, 2.42], weight 9.8%.
• Study F: 20/150 vs 36/150, RR 0.56 [0.34, 0.91], weight 16.8%.
• Pooled (fixed effect, inverse-variance weights): 136/1080 vs 193/1078, RR 0.70 [0.57, 0.86] (the Mantel-Haenszel method gives the same 0.70). Cochran’s Q = 4.65 on 5 degrees of freedom, I² = 0%; neither is drawn on the figure.

Ways to use it in teaching

• Ask which studies on their own show a clear effect, then what the pooled diamond adds.

• Have students recompute one study’s risk ratio from its counts, a quick check that they can read the columns.

• Study E points the other way, yet Q = 4.65 on 5 degrees of freedom; what would E need to look like for the studies to be clearly heterogeneous?

Sources

• Deeks, J. J., Higgins, J. P. T., Altman, D. G., McKenzie, J. E., & Veroniki, A. A. (Eds.). (2024). Chapter 10: Analysing data and undertaking meta-analyses. In J. P. T. Higgins, J. Thomas, J. Chandler, M. Cumpston, T. Li, M. J. Page, & V. A. Welch (Eds.), Cochrane handbook for systematic reviews of interventions (Version 6.5). Cochrane. https://training.cochrane.org/handbook/current/chapter-10
• Forest plot. (n.d.). In Wikipedia. Retrieved October 18, 2026, from https://en.wikipedia.org/wiki/Forest_plot