On this page
A forest plot displays multiple studies' findings side by side, with each row showing one study's point estimate and confidence interval on an effect-size axis. The diamond at the bottom shows the combined result. Reading it means identifying the point estimate position, checking whether the confidence interval crosses the null value, and comparing consistency across studies.
- Each study's row shows a point estimate (square or dot) positioned on the effect-size axis and a horizontal line (confidence interval) extending through it, with square size indicating study weight.
- If a study's confidence interval crosses the null value (0 for absolute measures, 1 for ratios), that study's result is not statistically significant on its own.
- The diamond at the bottom represents the pooled effect from all studies combined, with its center as the combined estimate and its width as the confidence interval.
- When confidence intervals don't overlap or point estimates scatter widely across the null, studies disagree, which weakens confidence in the overall finding.
- High heterogeneity (measured by I-squared) signals that studies are too different to combine reliably, even if the diamond appears narrow.
Forest plots are the standard way to display results from multiple studies on the same question. They let you see at a glance whether studies agree, how precise each finding is, and what the combined evidence suggests. But the visual format only works if you know what you're looking at. This guide teaches you to read a forest plot by identifying its key elements in order, extracting the information from each study, and recognizing the patterns that signal reliable or questionable results.
What you'll be able to do
After reading this guide, you will be able to extract the effect size and confidence interval from each row, identify which individual studies reached statistical significance, read the pooled estimate from the diamond at the bottom, and spot when studies disagree enough to question the overall conclusion. You'll know how to check whether the analysis is dominated by one large study and when heterogeneity is high enough to warrant caution.
Identify the four key elements of a forest plot
A forest plot has four structural components that work together to show you what each study found and what all studies combined suggest. Learn to recognize each one, and you can read any forest plot.
The horizontal axis: what effect size is being measured
The horizontal axis shows the direction and magnitude of the effect. The scale varies depending on the type of outcome being studied. For odds ratios or relative risks, the axis shows ratios (often on a logarithmic scale). For mean differences or risk differences, it shows absolute change. The axis is labeled with the effect measure and its units. Your first step when looking at a forest plot is to read the axis label and understand what direction means benefit and what direction means harm.
The vertical axis: each row is one study
The vertical axis lists the studies included in the meta-analysis. Each horizontal line represents one study's result. Studies are usually ordered by author name or publication date, though sometimes they are ordered by effect size. The number of rows equals the number of studies included in the analysis. Read the study names on the left to understand which research is represented.
The point estimate: the dot or square showing the study's finding
The point estimate is the study's best single estimate of the effect, positioned on the horizontal axis. It is usually drawn as a square. The position of that square on the axis tells you the study's finding: how far left or right of the null value it falls. The size of the square matters. Larger squares indicate studies that contributed more weight to the pooled estimate, typically because they had larger sample sizes or narrower confidence intervals. A small square means the study had less influence on the overall result.
The confidence interval: the horizontal line showing uncertainty
The horizontal line extending left and right from the point estimate is the confidence interval. It shows the range of plausible values for the true effect based on the study's data. A narrow line means the study's result is precise. A wide line means there is more uncertainty. The width of the line reflects the study's sample size and the variability in its data: larger studies and more consistent results produce narrower intervals.
Read the point estimate and confidence interval for each study
Once you can identify the four elements, extract the information from each row. This is where you learn what each study actually found and whether that finding is statistically significant.
Locate the point estimate on the horizontal axis
Find the square (or dot) for the study you are examining. Trace straight down from it to the horizontal axis to read its value. Many forest plots also include a table column listing the exact effect size and confidence interval numerically, which is faster and more precise than reading from the axis. The position tells you the study's conclusion: if the square is to the right of the null, the study found an effect in one direction; if it is to the left, the effect goes the other way. If the square sits exactly on the null, the study found no effect.
Check whether the confidence interval crosses the line of no effect
Look at the horizontal line extending from the square. A vertical reference line marks the null value on the axis. For ratio-based measures like odds ratio or relative risk, the null is 1 [F3, F5]. For absolute measures like mean difference, the null is 0 [F3, F6]. If the confidence interval line crosses this reference line, the study's result is not statistically significant. That means the true effect could be zero (or one, for ratios), and the study did not rule it out. If the line does not cross the null, the result is statistically significant at the chosen confidence level, usually 95%. The study's data are consistent with a real effect.
Interpret the overall effect (the diamond at the bottom)
Below all the individual study rows sits a diamond shape. This is the pooled estimate, the combined result from all studies. The center of the diamond is the combined effect size. The width of the diamond is the confidence interval for that pooled result [F9, F10]. A narrow diamond means the combined evidence is precise. A wide diamond means more uncertainty remains even after combining all the data. Read the diamond the same way you read an individual study: check whether it crosses the null value. If it does, the pooled result is not statistically significant. If it does not, the combined evidence supports an effect.
Recognize when studies disagree (heterogeneity)
Studies on the same question do not always reach the same conclusion. When they disagree substantially, it weakens confidence in the pooled estimate. Learn to spot disagreement visually and understand what it means for your interpretation.
Visual signs of disagreement
Look across all the study rows. Do the confidence intervals overlap, or do some extend to the left of the null while others extend to the right? Do the point estimates cluster together, or are they scattered? When confidence intervals do not overlap, or when point estimates sit on opposite sides of the null value, studies are disagreeing. Another sign is a diamond that is much wider than the individual study intervals. If individual studies are fairly precise but the diamond is very wide, that width reflects disagreement between studies rather than uncertainty within any single study.
What high heterogeneity means for your interpretation
Heterogeneity is measured by the I-squared statistic, which expresses the percentage of variability in effect sizes due to real differences between studies rather than sampling error. Cochrane classifies less than 25% as low, 25 to 50% as moderate, 50 to 75% as high, and greater than 75% as considerable heterogeneity. High heterogeneity means the studies are too different to combine meaningfully. The pooled estimate may not represent any real population. When you see high I-squared alongside substantial disagreement in the plot, treat the overall diamond with caution. The studies may differ in patient populations, interventions, or outcomes in ways that matter. Look for subgroup analyses or sensitivity analyses that break the data into more homogeneous groups.
Check the study weights and sample sizes
The pooled estimate is not a simple average of the studies. Large studies carry more weight than small ones. Verify that the meta-analysis is not distorted by one dominant study, and spot when a small study with an extreme effect is pulling the overall result. Forest plots often display study weights as percentages in a column on the right side. Compare the weights to the square sizes. If one study has a square much larger than the others and a weight much higher than the rest, that study is driving the pooled estimate. Check whether removing that study would change the conclusion substantially. If a small study has a very large effect but a tiny square and minimal weight, it is not moving the diamond much. Conversely, if a small study has an outsized effect and substantial weight, that is a red flag: the result may be driven by a single underpowered analysis rather than consistent evidence across studies.
Common pitfalls and how to spot them
Misreading a forest plot is easy when you do not know what to watch for. These three mistakes are the most common. Learn to recognize each one and you will avoid the traps that catch even experienced readers.
Assuming the overall effect is reliable when heterogeneity is high
A narrow diamond can be deceptive. If I-squared is high and the individual study results scatter widely, that narrow diamond does not mean the evidence is solid. It means the statistical model has squeezed disagreement into precision. The studies are not saying the same thing. A narrow diamond with high heterogeneity is a warning sign, not reassurance. Always check I-squared and look at the scatter of point estimates. If they do not align, the pooled estimate is not trustworthy.
Overlooking the role of study size in the pooled estimate
A large study with a moderate effect will move the diamond more than a small study with an extreme effect. If you see a tiny square with a point estimate far from the null, do not assume that study is important. Look at its weight. If it is less than a few percent, it is barely influencing the result. Conversely, if one or two large studies make up half the weight, the pooled estimate is really just those studies combined. The meta-analysis adds little value if the evidence base is dominated by a handful of large trials. Check the weight distribution. Ideally, no single study should account for more than a quarter of the total weight.
Do this automatically
Interpreting a forest plot manually requires extracting effect sizes and confidence intervals from each study row, identifying statistical significance thresholds, and synthesizing results across multiple studies, a process that demands careful attention to the horizontal and vertical axes, point estimates, and confidence interval lines. Software can automate the extraction of these numerical values and visual elements from forest plots, reducing the manual effort required to compile and analyze meta-analysis data. Healthcare research runs this kind of analysis end to end and cites every figure back to its original source, across as many subjects as you need, with Mimir attaching a citation to every figure. It will not predict a trial's outcome or advise on anyone's treatment. Those decisions stay with clinicians and regulators, and deliberately so. Every plan includes it, including the free one.
Start your first report and run it on something you already know well.
This article is general information about healthcare and life-sciences topics. It is not medical advice, diagnosis, or treatment guidance, and it is not a substitute for a qualified professional. Any treatment decision should be discussed with a clinician.
