Early Detection and Assessment of Epidemics by Particle Filtering

Abstract

Influenza infects from 5% to 20% of the population during an epidemic episode, which typically lasts a few weeks, and, as the conditions leading to an outbreak are not well understood, the moment of its start is difficult to foresee. The early detection of an epidemic would however make it possible to limit its impact by adopting appropriate actions—this is particularly desirable in account of the threat of a major pandemic. As a side-product of the forecasting system an estimation of the total number of infected people in each region at every time step is provided.

We first present the classical epidemiological model for contagious diseases and explain the way regionalization has been incorporated into it. Then we describe the assimilation technique that is used to process the data, which are daily reported cases of influenza-like illness. Tests on simulated data are presented to illustrate the efficiency of the process and finally real observational data from the French Sentinelles network are processed.