To the Editor We read with interest the article by Zissimopoulos et al 1 that found that Alzheimer disease (AD) risk was 10% lower among Medicare beneficiaries exposed to higher levels of statins compared with lower levels of statins. While the authors have designed the study to address certain biases, we are concerned that they overlooked other potential biases in ascertaining AD after the submission of Medicare claims for statins that may have contributed to the lower observed incidence among people using high levels of statins.Ascertainment biases arise from the particular complexities of Medicare claims for medical conditions, such as AD, that are only available for periods when the beneficiary is covered by fee-for-service plans and not participating in a health management organization. Zissimopolous et al 1 limit their cohort to those covered by fee-for-service for 2 or more years. It is, however, possible that after the 2006 to 2008 period used to define statin use, some beneficiaries could have switched from fee-for-service to health management organizations, which would prevent the ability to ascertain their claims for AD. This could bias results, if, for example, those with higher statin use more often switched to health management organizations, possibly because they were healthier. While we do not know if this occurred, it would be informative to at least partially address this potential bias by adjusting for months covered by fee-forservice among cases and controls during the entire follow-up period.Additionally, we suggest adjusting for the number of visits to doctors (only knowable during fee-for-service coverage). 2 Because statin use may be related to the subsequent frequency of visits to doctors and an AD diagnosis cannot be made without a doctor's visit, controlling for medical surveillance intensity could affect the observed associations between statins and AD. Finally, in evaluating ascertainment bias, it may be useful to examine the relationship between AD and cholesterol-lowering medications other than statins that affect circulating cholesterol levels through mechanisms of action that differ from statins. Thus, if the relationship between statins and AD reflects the particular mechanism of action of statins, one would not expect to replicate the association with medications reliant on other biological processes.