Two Authors Reply

摘要

We thank Binder et al. (1) for their interest in and careful review of our article (2). First, we apologize that the column headings in Table 2 (2, p. 419) were mislabeled with “%,” but it should be clear from the table title that the incidence rates refer to events per 1,000 person-years of follow-up. Second, we did not use the Kaplan-Meier approach to draw our conclusions about cumulative incidence differences. Rather, we used the simple (and well-understood) plot in Figure 2 (2, p. 420) to illustrate the very small differences in dementia-specific survival between participants and nonparticipants. We used incidence density estimates (numbers of events per 1,000 person-years of follow-up) as our measures of dementia incidence, and we used these well-understood and -accepted epidemiologic estimates to make the primary comparisons of incidence between participant groups (3). We recognize that these estimators will produce higher estimates of incidence than will result from analyses that account for competing risks—for exactly the reasons outlined by Binder et al. (1).