Fig. 9From: How unmeasured confounding in a competing risks setting can affect treatment effect estimates in observational studiesResults from Scenario 7 Legend Back to article page