Methods for Vaccine Efficacy in a Fixed Duration Design with Censoring

• Activity Number: 29 - Statistical Issues Specific to Therapeutic Areas, Power and Sample Size Calculations, and Trial Monitoring
• Type: Contributed
• Date/Time: Sunday, August 8, 2021 : 1:30 PM to 3:20 PM
• Sponsor: Biopharmaceutical Section
• Author(s): Ying Zhang* and G. Frank Liu and Radha A Railkar
• Companies: Merck & Co., Inc. and Merck Sharp & Dohme Corp. and Biostatistics and Research Decision Sciences, Merck & Co., Inc.
• Keywords: vaccine efficacy; Poisson regression; censoring; controlled-based imputation; survival regression

Abstract

To evaluate the vaccine efficacy (VE) of a new vaccine, a fixed duration trial may be considered to show the vaccine reduces the incidence of the disease compared to the control within a given duration of follow-up, e.g., influenza. VE is defined as the relative risk (RR) reduction 1-(Rv/Rp), with Rv and Rp as the incidence rates for the disease in the vaccine and control groups, respectively. A conditional exact method from Chan and Bohidar (1998) is often used to estimate VE, and Poisson regression can also be used to estimate RR. In the first part, we compared the conditional exact method with Poisson regression, modified Poisson regression by Zou (2004), and Cox regression through simulations. In the second part, we considered sensitivity analyses to evaluate the impact of censoring due to early dropout in a vaccine study. We evaluated several controlled-based imputation (CBI) methods based on survival models through simulations, including 1) analytic approach; 2) multiple imputation (MI); and 3) bootstrapping + MI. The simulation study compared these methods under different assumptions on sample size, VE, dropout rate, incidence rate, and data generation.