A Framework for Measuring Population Immunity Against Influenza Using Individual Antibody Titers

Measuring population immunity is crucial for epidemic control and prevention of infectious diseases. While correlates of protection have been identified at the individual level for some pathogens, methods to translate these into population-level immunity metrics remain underdeveloped. We developed and validated a framework to construct population immunity estimators derived from individual serological measurements. Using influenza as a model system, we analyzed 36,150 serum samples across 4 studies covering 19 epidemics in 2009-2020, establishing four hemagglutination-inhibiting (HAI) antibody titer-based estimators: geometric mean titer, proportion of non-naive individuals, proportion of population immune, and relative reduction in reproductive number. We found that subtype-specific relative changes of these estimators from previous seasons predicted predominant subtypes in upcoming seasons with up to 80% sensitivity and 100% specificity. In a longitudinal cohort spanning eight influenza seasons with serum collection during epidemics, we found significant negative correlations between each estimator and subsequent cumulative incidence for H1N1 viruses, but weaker correlations for H3N2 viruses. These relationships remained consistent regardless of the specific protection threshold assigned to different antibody titer levels, provided individual-level protection was at least moderate. Simulation studies revealed that lower effective reproductive numbers and higher antibody waning rates diminished the correlation strength. Our framework provides a systematic approach for evaluating population immunity estimators that could inform preparedness efforts for seasonal epidemics and potentially extend to other infectious diseases.