Fig. 7From: How unmeasured confounding in a competing risks setting can affect treatment effect estimates in observational studiesResults from Scenario 5 Legend Back to article page