To the Editor In the April issue of JAMA Cardiology, Dodson et al1 investigated the risk factors for traumatic intracranial bleeding among older adults with atrial fibrillation after the initiation of oral warfarin therapy. They concluded that dementia, anemia, depression, anticonvulsant use, and labile international normalized ratio control increased the risk of traumatic intracranial bleeding.1 The results also showed a high prevalence of polypharmacy and labile international normalized ratio control among the sample (87.0% and 55.9% of patients, respectively). We would like to express concern about the absence of data on patient adherence to warfarin therapy and on the presence of drug-drug interactions, which are 2 of the most common and important sources of bias in clinical studies involving drugs.