Fig. 3From: How to perform prespecified subgroup analyses when using propensity score methods in the case of imbalanced subgroupsBias in the estimation of the treatment effect under PS matching without replacement using the “across subsets” or the “within subsets” strategy, according to the treatment prevalence (A), the relative risk to be treated in subset 1 (B) and the treatment effect in subset 1 (C) (Scenario 1)Back to article page