Fig. 10From: How unmeasured confounding in a competing risks setting can affect treatment effect estimates in observational studiesResults from Scenario 8 Legend Back to article page