We agree with the comments by Yoshioka and colleagues 1 about the limitations of studies using large databases, such as ours. 2 We want to note that the focus of our study was not on differences in level of testing for pulmonary embolism (PE) between patients with and without dementia, but on differences in the association of clinical factors with PE testing between these two groups of patients.Although we wanted to use reference standards for diagnosing dementia as suggested by Yoshioka and colleagues such as the Mini Mental State Examination, such data were not available. We instead used diagnosis codes, which are subject to the limitations mentioned by Yoshioka and colleagues. However, notably, such mismeasurement would normally bias our results toward the null.Second, we strongly agree that there are many clinically relevant variables to consider, some that can be calculated from our database such as a frailty index and many others that are not in our database. We agree that including such variables, either calculated from databases such as a frailty index or abstracted from databases using natural language processing, will be important for future work.