A new explanatory model of an SIR disease epidemic: A knowledge-based, probabilistic approach to epidemic analysis

Abstract

A new explanatory model for epidemic analysis is presented; it has a knowledge based component and a probabilistic computational component. The former assembles details of household characteristics, social networks and connectivity in the community–‘knowledge’–which is used to determine the structure of the computational component. The latter links individuals and households through statistically-defined opportunities for contacts and, by repeated trials, determines an average longitudinal time course (epidemic curve) of the simulated infection as it spreads through the community from inception to extinction of the epidemic. The model thus aims to describe the epidemic itself, rather than any abstraction of it. In application to a 1955–56, self-contained epidemic of an SIR disease, variola minor, the model generates 1 dominant longitudinal pattern that matches closely the epidemic curve of observed daily case rates; it is suggested that other patterns indicate different ways in which the epidemic might have evolved. The model can be used to show how differing community characteristics would affect the simulated epidemic.