Estimating epidemic parameters: Application to H1N1 pandemic data

• A closed stochastic SIR model was fitted to data from a 2009 H1N1 influenza outbreak.
• The standard approach based on the MLE led to incorrect model trajectory dynamics due to underreported infectives at onset.
• As an alternative, the LSE procedure for simultaneous fitting of the infection rates and the model’s initial conditions was proposed.
• The study advances our understanding of parameter estimation and modeling outbreaks of infectious diseases.

This paper discusses estimation of the parameters in an SIR epidemic model from the observed longitudinal new infection count data. The potential problems of the standard MLE approaches are revealed and possible remedies suggested. The analysis is based on the epidemic data from the 2009 outbreak of H1N1 influenza on the campus of Washington State University.