pomp-astic Inference for Epidemic Models: Simple vs. Complex

Abstract

AbstractInfectious disease surveillance data often provides only partial information about the progression of the disease in the individual while disease transmission is often modelled using complex mathematical models for large populations, where variability only enters through a stochastic observation process. In this work it is shown that a rather simplistic, but truly stochastic transmission model, is competitive with respect to model fit when compared with more detailed deterministic transmission models and even preferable because the role of each parameter and its identifiability is clearly understood in the simpler model. The inference framework for the stochastic model is provided by iterated filtering methods which are readily implemented in theR package pomp. We illustrate our findings on German rotavirus surveillance data from 2001 to 2008 and calculate a model based estimate for the basic reproduction numberR0using these data.