Interpreting Forest Plots in Medical Research

Key Components and Interpretation

• The American College of Cardiology suggests that forest plots are graphical tools that visually display the results of individual studies and their pooled analysis, making them essential for interpreting evidence in systematic reviews and meta-analyses 1, 2
• Study names are listed on the left side of a forest plot, identifying each individual study included in the analysis, as recommended by the American College of Cardiology 1
• Effect estimates in forest plots are represented by squares or circles, with the size proportional to the study's weight or contribution to the meta-analysis, according to the American College of Cardiology and Clinical Microbiology Reviews 1, 2
• The horizontal lines in a forest plot extend from each study's effect estimate, representing the 95% confidence interval (CI), as described by the American College of Cardiology and Clinical Microbiology Reviews 1, 2
• The vertical line in a forest plot represents the "no effect" or "line of equivalence" line, typically at 0 for mean differences or 1 for odds ratios/risk ratios, as noted by the American Academy of Sleep Medicine 3
• The diamond in a forest plot represents the pooled or summary effect estimate, with the width indicating the 95% CI of the pooled result, according to the American College of Cardiology and Clinical Microbiology Reviews 1, 2
• Tabular data in forest plots often includes sample sizes, event counts, effect sizes, and confidence intervals for each study, as recommended by the American College of Cardiology and Circulation 1, 4

Assessing Heterogeneity and Statistical Significance

• The American Academy of Sleep Medicine suggests that the direction of effect in a forest plot can be determined by the position of the effect estimate relative to the vertical line, with points to the right favoring the intervention/exposure and points to the left favoring the control/comparison 3, 5
• Statistical significance in a forest plot is indicated when a study's confidence interval does not cross the vertical "no effect" line, resulting in a p-value less than 0.05, as described by Clinical Microbiology Reviews and PLoS Medicine 2, 6
• The magnitude of effect in a forest plot is represented by the distance from the vertical line, with larger distances indicating greater effects, according to the American College of Cardiology and Clinical Microbiology Reviews 1, 2
• The I² statistic is often reported below the forest plot to assess heterogeneity, with values of 0-25% indicating low heterogeneity, 25-75% moderate heterogeneity, and >75% high heterogeneity, as noted by the American College of Cardiology and Neurosurgery 1, 7
• The p-value for heterogeneity is used to determine the significance of heterogeneity between studies, with a low p-value (<0.05) suggesting significant heterogeneity, as described by PLoS Medicine 6