Free Forest Plot Generator
A free forest plot maker for systematic reviews. Enter each study’s effect and confidence interval, optionally add your pooled result, and download a clear figure.
Your data stays in your browser. Nothing you enter is uploaded, stored or used for analytics, so export your work before closing the page.
1. Effect measure
The measure sets the scale and the line of no effect: ratio measures use a log axis with the line at 1; differences use a linear axis with the line at 0.
2. Study data
Enter values exactly as reported. Ratio estimates and limits must be greater than 0. The estimate must lie within its confidence limits. Weights are optional and are used only to size the squares; they are not calculated here.
| Study | Year | Estimate | Lower CI | Upper CI | Weight (optional) | Remove |
|---|---|---|---|---|---|---|
3. Pooled estimate (optional)
This tool does not perform a meta-analysis. If you calculated a pooled result elsewhere (for example in RevMan, R or Stata), enter it here to draw it as a diamond. It will be labelled “User-supplied pooled estimate”.
4. Plot options and export
Fix these before exporting (1)
- Add at least one study with an effect estimate and confidence interval.
Export is disabled until the 1 issue above is resolved.
What this tool does
This tool draws a forest plot from effect estimates and confidence intervals you enter or import: a square for each study, a horizontal line for its confidence interval, a vertical line of no effect and, if you supply one, a diamond for the pooled estimate.
It is a visualisation tool, not a meta-analysis engine. It does not pool studies, calculate weights, estimate heterogeneity or compute prediction intervals. Pooled results must come from your statistical software, and the tool labels them “User-supplied pooled estimate”. Marker sizes use only the weights you provide.
The scale matches the effect measure. Odds ratios, risk ratios and hazard ratios are plotted on a logarithmic axis with the line of no effect at 1. Mean differences and standardised mean differences are plotted on a linear axis with the line at 0. The tool rejects values that cannot be valid, such as a ratio of zero or an estimate outside its own confidence interval.
Who should use it
- Authors of systematic reviews who ran their meta-analysis in RevMan, R, Stata or another package and need a clean figure.
- Students and residents who want to visualise study results from a published review.
- Clinicians preparing a journal club or evidence summary.
How to use it
- Choose the effect measure: odds ratio, risk ratio, hazard ratio, mean difference or standardised mean difference.
- Enter each study’s name, year, effect estimate and lower and upper confidence limits. Add weights only if you have them from your analysis.
- If you have a pooled result from your meta-analysis software, switch on the pooled estimate and enter it and its confidence interval.
- Set the title, axis labels and optional “favours” labels, then check the plot for errors.
- Download PNG, SVG or PDF, or export the data as CSV.
Worked example (hypothetical data)
Three hypothetical trials report odds ratios of 0.72 (95% CI 0.50 to 1.04), 0.85 (0.62 to 1.17) and 1.10 (0.70 to 1.73). Each interval contains its estimate, and on the log scale each estimate sits at the middle of its interval, as expected for a Wald interval.
Suppose your software returns a pooled odds ratio of 0.84 (0.68 to 1.04). You enter it as the user-supplied pooled estimate and the tool draws a diamond spanning 0.68 to 1.04. Because that interval includes 1, the plot does not show a statistically significant pooled effect at the 5% level.
Use “Load hypothetical example” in the tool to plot these values.
Common mistakes
Plotting ratios on a linear axis
Ratios are symmetric on a log scale: an odds ratio of 0.5 and of 2 are equally far from 1. A linear axis exaggerates effects above 1.
Using the wrong line of no effect
No effect is 1 for ratio measures and 0 for differences. A line at 0 on a ratio plot would imply that the effect is a ratio of zero.
Entering a standard error as a confidence limit
This tool needs the lower and upper confidence limits. Convert from standard errors in your software first.
Leaving the direction of benefit implicit
Whether an odds ratio below 1 favours the intervention depends on the outcome. Add “favours” labels so readers do not have to guess.
Describing the figure as a meta-analysis
A forest plot of individual results is not a pooled analysis. Only describe a pooled estimate if you calculated one with a validated method.
Frequently asked questions
Is the forest plot generator free?
Yes. It is free, needs no login and has no usage limits.
Does it perform a meta-analysis?
No. It plots the numbers you supply. It does not pool studies, calculate weights, or estimate heterogeneity. Enter a pooled estimate from your own software if you want it drawn.
Which effect measures are supported?
Odds ratio, risk ratio, hazard ratio, mean difference and standardised mean difference.
Why does it reject some of my values?
Ratios must be greater than zero and the estimate must lie between its lower and upper limits. Otherwise the point cannot be placed on a log scale or is internally inconsistent.
What do the marker sizes mean?
If you supply weights, the area of each square is proportional to your weight. If you supply none, all squares are the same size. MedResearch Ninja does not calculate the weights.
Is my data stored?
No. Everything runs in your browser and nothing is uploaded or saved. Export CSV to keep a copy.
References
- Lewis S, Clarke M. Forest plots: trying to see the wood and the trees. BMJ 2001;322:1479–80. doi:10.1136/bmj.322.7300.1479
- Higgins JPT, Thomas J, Chandler J, et al., editors. Cochrane Handbook for Systematic Reviews of Interventions. Chapter 10: Analysing data and undertaking meta-analyses. Cochrane.
- Page MJ, McKenzie JE, Bossuyt PM, et al. The PRISMA 2020 statement: an updated guideline for reporting systematic reviews. BMJ 2021;372:n71. doi:10.1136/bmj.n71
More free research tools
Related: MRN Research Toolkit · Research Assistance Centre