To the Editor Dr Chang and colleagues examined the risk of major bleeding associated with use of P-glycoprotein competitors and hepatic cytochrome (CYP) isoenzyme 3A4 (CYP3A4) inhibitors among patients taking non–vitamin K oral anticoagulants (NOACs).1 Confounding by indication, as the authors noted, is the most important challenge to address when studying drug-drug interactions. We are concerned that the authors’ approach cannot fully account for important confounders that lead to a given patient receiving the combination therapy rather than NOAC monotherapy, rendering biased results. The presence of unmeasured confounding is evident in the findings.