Forest Plot Maker

Build publication-ready systematic review and meta-analysis forest plots in seconds. Computes Fixed-Effect (IV) and Random-Effects (DerSimonian-Laird) pooled estimates, weights, and heterogeneity statistics automatically.

Data Input

6 rows parsed

Chart Configuration

Live Vector Preview

SVG · PNG · PDF
Meta-Analysis: Early Anticoagulation vs. Standard Care Meta-Analysis: Early Anticoagulation vs. Standard Care Study Odds Ratio (95% CI) Effect [95% CI] Weight Chen et al. (2021) 0.68 [0.49, 0.94] 15.6% Patel & Gomez (2022) 0.81 [0.58, 1.13] 14.9% Müller et al. (2023) 0.61 [0.45, 0.83] 17.6% Okafor et al. (2023) 0.75 [0.52, 1.08] 12.4% Sato et al. (2024) 0.64 [0.47, 0.87] 17.4% Lindqvist et al. (2025) 0.72 [0.55, 0.95] 22.1% Pooled (Random, DL) 0.69 [0.61, 0.79] 100.0% 0.5 1 ← Favors Treatment Favors Control → Heterogeneity: I² = 0.0%, τ² = 0.0000, Q = 2.03 (df = 5, p = 0.8448) | Test for overall effect: Z = -5.56, p < 0.0001

Live Statistical Readout

Pooled Effect0.69 [0.61, 0.79]
Heterogeneity (I²)0.0%
Tau² (τ²)0.0000
Cochran Q2.03 (p=0.8448)
Overall p-value< 0.0001

What is a forest plot in meta-analysis?

A forest plot is the standard visual summary of a systematic review and meta-analysis. Each horizontal row represents an individual study: a square marks the point estimate (such as an Odds Ratio, Risk Ratio, Hazard Ratio, or Mean Difference) sized proportionally to its statistical weight, and the horizontal whisker shows its 95% confidence interval.

At the bottom, a diamond displays the pooled overall effect and its 95% confidence interval across all studies. If the diamond does not cross the vertical line of no effect (1.0 on ratio scales, 0.0 on difference scales), the pooled effect is statistically significant at α = 0.05. Always report heterogeneity (I², τ², and Cochran's Q) alongside the pooled diamond.

How to build this chart

  1. Enter studies as four columns: study name, effect size, lower 95% CI, and upper 95% CI.
  2. Select Log scale (OR/RR/HR) or Linear scale (MD/SMD) and choose Random-Effects (DerSimonian-Laird) or Fixed-Effect pooling.
  3. Verify live I², τ², Q, Z, and overall p-value readouts, then export vector SVG or print PDF for your manuscript.

Frequently asked questions

When should I use Random-Effects (DerSimonian-Laird) vs. Fixed-Effect?

Use Random-Effects when clinical or methodological diversity exists across trials (especially when I² > 25–50%). Fixed-Effect assumes all studies estimate one identical true effect.

Why does Log scale require positive numbers?

Ratio measures (Odds Ratio, Risk Ratio, Hazard Ratio) are symmetric on the natural-log scale with null value 1.0, so values must be strictly greater than zero. For Mean Differences that include zero or negative values, switch to Linear scale.

How are study weights computed?

Standard errors are derived from each study's 95% CI width: SE = (upper − lower) / (2 × 1.96) on the analysis scale. Fixed weights equal 1/SE² and random weights equal 1/(SE² + τ²).

Can I customize the 'Favors treatment / Favors control' axis labels?

Yes. Edit the Left axis note and Right axis note fields directly in the configuration panel before downloading your SVG or PDF.

Explore other chart makers