Visualize binary outcome meta-analysis data by plotting treatment event rates against control event rates for each study, with bubble sizes proportional to sample size.
Load sample data to see how the tool works, or clear all fields to start fresh.
Drag & drop a file or
CSV, TSV, Excel (.xlsx/.xls) - max 500 rows
| Study | Events (T) | Total (T) | Events (C) | Total (C) | |
|---|---|---|---|---|---|
Input the number of events and total participants for both the treatment and control arms of each included study. You can type values directly into the table or import from a CSV or Excel file.
Adjust bubble scaling, axis ranges, and label positioning. Choose whether to display study labels, confidence regions, or additional reference annotations on the plot area.
The tool automatically computes event rates for each arm and renders an interactive D3.js scatter plot with bubbles sized proportional to total sample size. Hover over any bubble to see study details.
Studies below the diagonal line have lower event rates in the treatment group, indicating a protective effect. The farther a bubble sits from the diagonal, the larger the treatment effect for that study.
Toggle reference lines representing constant risk ratios (e.g., RR = 0.5, RR = 2.0) to quickly estimate the approximate relative risk for each study based on its position in the two-dimensional space.
Download the finished plot as PNG, SVG, or PDF for your manuscript. Copy the auto-generated methods paragraph, or export reproducible R code using the metafor labbe() function for verification.
Need this done professionally? Get a complete systematic review or meta-analysis handled end-to-end.
Get a Free QuoteAny study positioned above the diagonal (y = x) line has a higher event rate in the treatment arm than the control arm, meaning the treatment increases the outcome event. Points below the line indicate a protective treatment effect.
Larger bubbles correspond to studies with greater total sample sizes, giving you an immediate visual sense of which studies contribute most information to the pooled estimate. Small bubbles should be interpreted with caution due to imprecision.
Lines of constant risk ratio radiate from the origin. If studies cluster along one such line, the risk ratio is a stable summary measure across baseline risk levels. Divergence from these lines suggests the treatment effect varies with baseline risk.
When studies separate into visible clusters on the L'Abbe plot, this may indicate subgroup effects driven by patient characteristics, disease severity, or intervention dosage. Color-coding by subgroup makes these patterns even clearer.
The forest plot shows pooled estimates and confidence intervals on a single axis. The L'Abbe plot preserves the two-dimensional data structure, revealing baseline risk interactions and heterogeneity patterns that a forest plot cannot display.
Studies that sit far from the main cluster or on the opposite side of the equality line are immediate visual outliers. These studies warrant investigation through sensitivity analysis or leave-one-out testing to determine whether they drive heterogeneity.
The L'Abbe plot was introduced by L'Abbe, Detsky, and O'Rourke in their 1987 Annals of Internal Medicine paper as a graphical method for exploring data in clinical trial meta-analyses. For each study, the treatment arm event rate is plotted on the vertical axis and the control arm event rate on the horizontal axis, creating a two-dimensional view that preserves the original data and reveals patterns invisible in standard forest plots. Studies below the diagonal equality line (where y = x) favor treatment, and the distance from the diagonal reflects the magnitude of the treatment effect.
The primary strength of the L'Abbe plot is its ability to reveal whether heterogeneity is driven by differences in baseline risk across study populations. If studies with higher control event rates show larger treatment effects (bubbles further below the diagonal), this suggests treatment-by-baseline-risk interaction, where the intervention is more effective in higher-risk populations. This pattern connects directly to the choice of effect measure: if all studies cluster along a line through the origin (constant risk ratio), the risk ratio may be the appropriate summary measure. If studies instead cluster along a line parallel to the diagonal, this suggests a constant risk difference across populations. Compute your pooled risk ratios with our effect size calculator and visualize individual study estimates with the forest plot generator.
Understanding the relationship between binary outcomes and summary measures is central to interpreting L'Abbe plots. The relative risk (risk ratio) compares the probability of an event in the treatment group to the control group, while the odds ratio compares the odds. When control event rates vary substantially across studies, the odds ratio tends to remain more stable than the risk ratio, which can be seen directly on the L'Abbe plot by observing whether studies align better with constant RR lines (radiating from the origin) or constant OR curves. The Cochrane Handbook (Higgins et al., 2023, Chapter 10) recommends presenting L'Abbe plots alongside forest plots when investigating sources of heterogeneity in binary outcome meta-analyses.
Bubble size on the L'Abbe plot encodes study weight, typically represented by total sample size. This allows immediate visual identification of influential studies. A small study positioned far from the cluster may not substantially affect the pooled estimate, while a large study in the same position would have a major influence. Cross-reference visually identified outliers with a formal sensitivity analysis (leave-one-out) to quantify their impact on the overall pooled result.
The L'Abbe plot is particularly valuable during the exploratory phase of meta-analysis, before selecting the summary statistic or fitting subgroup models. By coloring points according to a potential effect modifier (such as intervention dose, follow-up duration, or patient severity), researchers can visually assess whether the treatment effect varies systematically. This visual hypothesis generation can then be tested formally using meta-regression or subgroup analysis within the subgroup analysis tool.
This tool generates reproducible R code using the metafor package (Viechtbauer, 2010) and its labbe() function, which recreates the plot from your data with confidence regions. The generated code includes proper escalc() calls for computing log risk ratios and log odds ratios, allowing seamless integration with your broader analysis pipeline. Translate your risk ratios into clinically meaningful numbers with our number needed to treat calculator, and assess potential publication bias using the funnel plot and publication bias tool.
When reporting L'Abbe plots in publications, include the equality line (y = x) as the primary reference, annotate any constant risk ratio or odds ratio lines that you overlay, and specify whether bubble size represents total sample size or inverse-variance weight. Following L'Abbe et al. (1987) and the Cochrane Handbook recommendations ensures that readers can correctly interpret the visual relationship between treatment and control event rates across the included studies.
A L'Abbe plot (L'Abbe, Detsky, and O'Rourke, 1987) is a scatter plot that shows the event rate in the treatment group (y-axis) against the event rate in the control group (x-axis) for each study in a binary outcome meta-analysis. Each study appears as a bubble, with bubble size proportional to total sample size. You should use a L'Abbe plot whenever your meta-analysis involves dichotomous outcomes (such as mortality, response rates, or adverse events) and you want a visual overview of treatment effects and between-study heterogeneity before computing pooled effect estimates.
The diagonal line running from the bottom-left corner (0, 0) to the top-right corner (1, 1) represents no treatment effect, where the treatment event rate equals the control event rate (risk ratio = 1). Studies that fall below this line have a lower event rate in the treatment group than in the control group, meaning the treatment reduces the event. Studies above the line show a higher treatment event rate, meaning the treatment increases the event. The farther a study sits from the diagonal, the larger the treatment effect for that study.
If all studies show a similar treatment effect, their bubbles will cluster along a line parallel to or below the diagonal (for a protective treatment). When studies scatter widely across the plot, some above and some below the diagonal, or at very different distances from it, this visual spread indicates substantial heterogeneity in treatment effects across studies. The L'Abbe plot is especially good at revealing whether heterogeneity is related to baseline risk, because you can see if the treatment effect changes at different control event rates. For example, a treatment may reduce events more in studies with high baseline risk (high control rate) than in those with low baseline risk.
The optional reference lines on the L'Abbe plot show where studies would fall if they had a specific constant risk ratio (RR). For instance, the RR = 0.5 line represents studies where the treatment halves the event rate compared with the control group. These lines radiate from the origin (0, 0) because RR is defined as the treatment event rate divided by the control event rate. Lines below the diagonal (RR < 1) represent treatments that reduce events, while lines above the diagonal (RR > 1) represent treatments that increase events. These reference lines help you quickly gauge the approximate risk ratio for each study based on its position in the plot.
A forest plot displays individual study effect estimates (such as risk ratios or odds ratios) with confidence intervals along a single axis, making it ideal for showing the pooled estimate and study-level precision. A L'Abbe plot, by contrast, displays the raw event rates from both groups on a two-dimensional scatter plot, which preserves the original data and reveals patterns that a forest plot cannot show. Specifically, the L'Abbe plot reveals the relationship between baseline risk and treatment effect, the absolute event rates in each group, and the overall distribution of studies across the outcome spectrum. These two plots complement each other: the forest plot is the standard for reporting pooled results, while the L'Abbe plot provides deeper insight into heterogeneity and treatment-by-baseline-risk interaction.
The traditional L'Abbe plot is designed specifically for binary (dichotomous) outcomes, where each study reports events and totals in two groups. It plots event rates (proportions) on both axes, so it requires count data. For continuous outcomes like mean differences, the equivalent visualization would be a scatter plot of mean outcomes in the treatment group versus the control group, but this is not conventionally called a L'Abbe plot. If your meta-analysis uses continuous outcomes, consider using a forest plot, funnel plot, or Galbraith plot instead. Our effect size calculator and forest plot generator support both binary and continuous outcome types.
Compute pooled effect sizes for binary and continuous outcomes with our effect size calculator. Visualize individual study estimates and pooled results with our forest plot generator for meta-analysis. Detect potential publication bias with contour-enhanced funnel plots using our funnel plot and publication bias tool. Convert risk ratios into clinically meaningful numbers with our number needed to treat calculator.
Reviewed by
Dr. Sarah Mitchell holds a PhD in Biostatistics from Johns Hopkins Bloomberg School of Public Health and has over 15 years of experience in systematic review methodology and meta-analysis. She has authored or co-authored 40+ peer-reviewed publications in journals including the Journal of Clinical Epidemiology, BMC Medical Research Methodology, and Research Synthesis Methods. A former Cochrane Review Group statistician and current editorial board member of Systematic Reviews, Dr. Mitchell has supervised 200+ evidence synthesis projects across clinical medicine, public health, and social sciences. She reviews all Research Gold tools to ensure statistical accuracy and compliance with Cochrane Handbook and PRISMA 2020 standards.
Whether you have data that needs writing up, a thesis deadline approaching, or a full study to run from scratch, we handle it. Average turnaround: 2-4 weeks.