Figure 1.

A simulation study to illustrate the concept of confounding by indication and effect measure modification.

Notes: Presented data originate from a simulation of binary outcomes in dependence of nine covariates acting as (continuous) confounders (X), (binary) effect modifiers (Z), or both in 1500 virtual patients. The simulation framework thus addresses a typical situation of observational data from health insurance claims. The panels below show the results for each simulation scenario (A–D), where event numbers (predicted in adjusted analyses) are given within circles referring to both treatment options A (red) and B (blue) (odds ratios were derived from logistic regression models, the size of the circles is proportional to the absolute numbers of patients, whereas no statistical comparison between the groups is shown).