Improving precision of vaccine efficacy evaluation using immune correlate data in time-to-event models.

Abstract

Understanding potential differences in vaccine-induced protection between demographic subgroups is key for vaccine development. Vaccine efficacy evaluation across these subgroups in phase 2b or 3 clinical trials presents challenges due to lack of precision: such trials are typically designed to demonstrate overall efficacy rather than to differentiate its value between subgroups. This study proposes a method for estimating vaccine efficacy using immunogenicity (instead of vaccination status) as a predictor in time-to-event models. The method is applied to two datasets from immunogenicity sub-studies of vaccine phase 3 clinical trials for zoster and dengue vaccines. Results show that using immunogenicity-based estimation of efficacy in subgroups using time-to-event models is more precise than the standard estimation. Incorporating immune correlate data in time-to-event models improves precision in estimating efficacy (i.e., yields narrower confidence intervals), which can assist vaccine developers and public health authorities in making informed decisions.