Statistical and Mathematical Modeling in Biological Applications
Abstract: Serology is the gold standard approach to estimating influenza infection attack rates in human populations. A classic study design involves blood samples being collected before and after an influenza season in a cohort of individuals; a rise in antibody titers in an individual suggests infection during the season. Potential measurement errors have typically meant a 2 fold rise (1 dilution) is not considered as strong evidence for infection, with seroconersion therefore normally defined as a 4-fold or greater rise in antibody titers. This somewhat ad-hoc rule became established when those methods were first developed. Here, we revisit the basis for this assumption and explore how modern statistics for the analysis of data with measurement errors can change and improve our interpretation of serology.