Treatment event rate versus control event rate
The L'Abbe plot maps each study as a dot with the control arm event rate on the x-axis and treatment arm event rate on the y-axis. A study where treatment and control have equal event rates (no effect) sits on the diagonal. Studies above the diagonal show treatment benefit; those below show harm. Distance from the diagonal reflects magnitude of treatment effect.
Large dots typically represent large or precise studies, small dots represent small studies. Clustering tightly around the diagonal signals consistent, small effects across studies. Scatter perpendicular to the diagonal, with some studies high and others low, suggests heterogeneous effects or measurement inconsistency.
Detecting baseline risk dependence
One advantage of L'Abbe plots is detecting whether treatment effect depends on baseline (control-arm) risk. An upward-sloping cloud (high baseline risk studies cluster upper right) suggests risk-dependent effects, where treatment provides greater absolute benefit in sicker populations. A horizontal or vertical cloud suggests consistent relative effects across baseline risks.
This visual distinction guides interpretation: a risk-dependent effect means NNT and patient selection matter (treat high-risk patients first), while a consistent relative effect means the proportional benefit applies regardless of starting risk. Spotting this pattern helps clinicians target treatment to populations where absolute gain is largest.