Forest PlotConfidence Interval (CI)Error Bars
About Forest Plot
Few visualizations carry as much weight in shaping clinical guidelines and health policy as the forest plot. When researchers need to distill findings from dozens — sometimes hundreds — of independent studies into a single, interpretable graphic, the forest plot is the gold standard. Each horizontal line tells the story of one study's effect estimate and uncertainty, while the diamond at the bottom delivers the verdict: the pooled result that can change how doctors treat patients, how drugs get approved, and how public health strategies are designed.
The elegance of a forest plot lies in its transparency. Unlike a simple summary statistic, it lays bare the heterogeneity across studies, letting readers see at a glance whether results converge on a clear answer or scatter across a range of possibilities. The vertical line of no effect serves as a visual anchor — studies whose confidence intervals cross it suggest no statistically significant finding, while those landing firmly on one side point toward a meaningful effect. Subgroup analyses, sensitivity checks, and cumulative meta-analyses all find natural expression in this format, making the forest plot an indispensable tool in the evidence synthesis toolkit.
Need to create a forest plot for your next systematic review or meta-analysis? Plottie's AI editor at ai.plottie.art can help you build publication-ready forest plots in minutes — just describe your data and let the AI handle the layout, scaling, and formatting. Browse our curated collection below for inspiration from real-world examples published in leading journals.
When to Use Forest Plot
- Presenting pooled effect sizes and confidence intervals in systematic reviews and meta-analyses
- Comparing treatment effects across multiple clinical trials or observational studies in medical research
- Conducting subgroup analyses to explore whether treatment effects vary by patient population, dosage, or study design
- Performing sensitivity analyses to assess how excluding specific studies influences the overall pooled estimate
- Summarizing diagnostic test accuracy measures such as odds ratios, risk ratios, or hazard ratios across multiple datasets
Key Features
- Individual study effect estimates displayed as squares, with square size proportional to study weight in the meta-analysis
- Horizontal lines extending from each square representing 95% confidence intervals for each study's estimate
- A vertical reference line (line of no effect) at the null value, typically at 0 for mean differences or 1 for ratio measures
- A diamond at the bottom representing the pooled summary effect, with its width indicating the confidence interval of the combined estimate
- Study labels and numeric data (effect sizes, CIs, weights) aligned alongside the graphical elements for easy cross-referencing
- Optional heterogeneity statistics (I², Q statistic, tau²) and subgroup headers to organize studies into meaningful categories
Frequently Asked Questions
What does the diamond at the bottom of a forest plot represent?
The diamond represents the pooled (overall) effect estimate from the meta-analysis. Its center indicates the point estimate of the combined effect, while the width of the diamond spans the confidence interval. If the diamond does not cross the line of no effect, the pooled result is considered statistically significant at the chosen confidence level.
How do I interpret heterogeneity in a forest plot?
Heterogeneity refers to the variability in effect estimates across included studies. Visually, if the confidence intervals of individual studies overlap substantially and cluster around the pooled estimate, heterogeneity is low. The I² statistic quantifies this — values above 50% suggest moderate to high heterogeneity, meaning the studies may not be measuring the same underlying effect. When heterogeneity is high, a random-effects model is typically preferred over a fixed-effect model.
What is the difference between a fixed-effect and random-effects forest plot?
A fixed-effect model assumes all studies estimate the same true effect, and differences are due to sampling error alone. A random-effects model assumes the true effect varies across studies and accounts for both within-study and between-study variance. In practice, the random-effects model tends to produce wider confidence intervals for the pooled estimate and gives relatively more weight to smaller studies. The choice between models should be guided by the expected clinical and methodological diversity of the included studies.
Can forest plots be used for outcomes other than treatment effects?
Absolutely. While forest plots are most commonly associated with meta-analyses of randomized controlled trials, they are equally useful for summarizing diagnostic accuracy studies, prevalence estimates, correlations, and even survey results across populations. Any research question where multiple independent estimates of the same quantity need to be synthesized and compared can benefit from the clarity and transparency of a forest plot.