To the Editor The recent Editorial by Drs Gill and Tonelli1 accompanying our study2 of ESRD risk following kidney donation raised several methodological questions that we believe merit response.
Gill and Tonelli suggested that “repeated inclusion of controls with long event-free survival … may have underestimated the risk of ESRD in nondonors.” We fail to see how repeated sampling could lead to bias, but it does affect variance. We conducted a bootstrap specifically designed to account for the variance of repeated sampling on rare outcomes. Using this bootstrap and other appropriate statistical tests, we reported our findings as statistically significant. Furthermore, any differences in follow-up (or duration of event-free survival) were accounted for by our use of survival analysis.
Massie AB, Muzaale AD, Segev DL. Outcomes After Kidney Donation. JAMA. 2014;312(1):94-95. doi:10.1001/jama.2014.6120