Skip to content
Accessible Unlicensed Requires Authentication Published by De Gruyter June 22, 2011

Using Approximate Bayesian Computation to Estimate Transmission Rates of Nosocomial Pathogens

Christopher C Drovandi and Anthony N Pettitt

In this paper, we apply a simulation based approach for estimating transmission rates of nosocomial pathogens. In particular, the objective is to infer the transmission rate between colonised health-care practitioners and uncolonised patients (and vice versa) solely from routinely collected incidence data. The method, using approximate Bayesian computation, is substantially less computer intensive and easier to implement than likelihood-based approaches we refer to here. We find through replacing the likelihood with a comparison of an efficient summary statistic between observed and simulated data that little is lost in the precision of estimated transmission rates. Furthermore, we investigate the impact of incorporating uncertainty in previously fixed parameters on the precision of the estimated transmission rates.

Published Online: 2011-6-22

©2011 Walter de Gruyter GmbH & Co. KG, Berlin/Boston