Purpose: To explore the heterogeneous transmission dynamics for influenza and identify the optimal serum antibody titer cutoff values for estimating its cumulative incidence.
Methods: We constructed a mathematical model describing serologically dependent disease transmission. The diagnostic performances of two serum antibody titer tests (single serum test and paired sera test) were evaluated, and cumulative disease incidence estimators were formulated. The model simulated the 2009 Japanese influenza A/H1N1 epidemic and investigated the optimal cutoff values and cumulative incidence estimates for this epidemic.
Results: Our assumed model and parameters suggested that the optimal cutoffs for A/H1N1 influenza were 1:20 for the single serum test and a 2-fold increase for the paired sera test. Using these optimal cutoff values, the paired sera test was the most reliable. The cumulative incidence estimate for the pre- and post-epidemic serological data showed that the paired serological data were also more accurately predictive.
Conclusions: From a statistical perspective, the currently used cutoff values may be too strict for diagnosing influenza and estimating its incidence. The paired sera test, which was more accurate for diagnosis and cumulative incidence estimation, is the test recommended for seroepidemiological surveillance during an epidemic.
Keywords: Diagnostic errors; Epidemiologic studies; H1N1 subtype; Influenza A virus; Models; Prevalence; Serologic tests; Theoretical.
Copyright © 2018 Elsevier Inc. All rights reserved.