Investigating Measles Dynamics in the Pre-Vaccination Era: A POMP Model Approach

Abstract

This thesis demonstrates the ability of partially observed Markov process (POMP) models to be applicable to epidemiological datasets of hundreds of cities of varying sizes. A previous investigation considered measles transmission in 20 cities and we extended this to 200. We used recently developed panel methods to assess whether parameters should be modeled separately for each city or shared between cities. Specifically, we found that the reporting rate and the rate of arrival of infected individuals from other cities were two parameters that critically varied between cities.