Fig. 3From: How unmeasured confounding in a competing risks setting can affect treatment effect estimates in observational studiesResults from Scenario 1. Legend Back to article page