Figure 1.

L'Abbé plot showing the relationships between Rc (x axis) and Rt (y axis) presented in the condition that the 'effect model' is linear. The identity line (Rt = Rc) corresponds with no treatment effect. Below this line, Rt is lower than Rc and the treatment is beneficial. Above this line the treatment is deleterious. A regression line may be estimated assuming the relative risk (RR) as a constant 'a' (Rt = a*Rc). In case of a full linear model (Rt = a*Rc + b), the treatment is beneficial for levels of Rc greater than the threshold, and deleterious for levels of Rc lower than the threshold. X axis: risk of event under control conditions (Rc), Y axis: risk of event under treatment (Rt).