Bayesian disease mapping
Duncan Lee (University of Glasgow)
Friday 8th October, 2010 15:00-16:00 326
Mapping the spatial distribution of disease risk over a set of contiguous small-areas is a common problem in spatial epidemiology, and the primary aim is to determine if there are areas that exhibit excess levels of risk. The spatial risk surface is typically estimated using a Bayesian hierarchical model, where the spatial correlation in the data is represented by a set of random effects with a conditional autoregressive (CAR) prior distribution. This talk proposes an extension to this class of models, and compares it via simulation to a number of the models that are commonly used. All the models are then applied to a study mapping cancer incidence in Greater Glasgow, Scotland, between 2001 and 2005.