A Risk Population Estimation Method for Interregional Comparison

An analysis on the influenza data reported by the surveillance system

Abstract

This study proposes a risk population estimation method based on the reported counts of patients of influenza in each area. A risk population estimation method is developed by regarding the sum of the weekly observations in a season following the binomial distribution, which enables estimations of population parameters and probability parameters through maximum likelihood method. The results show that the estimated infection probability is consistently correlated with population density, and the estimated population parameters in areas with high populations are obviously small in comparison with the proportion of population.