Abstract
Devices for extracting blood from donors require rigorous evaluation for the safety of donors. Complications arise from the fact that outcomes on successive donations from the same donor are not independent, adverse event rates are extremely rare, and there is substantial heterogeneity in the propensity for donors to donate over time. We develop a statistical framework for the design of a superiority trial and a non-inferiority trial, aiming to demonstrate the safety of a new donation device compared to the standard one. Historical data on the intensity and heterogeneity of donation across donors, the adverse event rate, and the serial dependence in adverse events provide information on how to plan accrual and follow-up periods to give the expected number of donors and donations. The analysis is based on a binary donation-specific outcome modeled with a Poisson approximation (i.e., log link and identity variance function) using generalized estimating equations with a working independence assumption. The historical data enables calculation of the asymptotic robust variance estimate, which is used for planning. The formulae derived are found to provide good control of the type I error rate and statistical power. We illustrate the derivations with application to a plasma donation trial aiming to investigate the safety of a new device, with the outcome being serious hypotensive adverse events.