Incorporating individual level covariates in disease mapping
Craig Anderson (University of Glasgow)
Friday 30th November, 2018 15:00-16:00 Maths 311B
Mapping disease rates over a region provides a visual illustration of underlying geographical variation of the disease and can be useful for generating new hypotheses about the causes of the disease. Disease mapping is typically carried out using a conditional autoregressive (CAR) model, which accounts for spatial correlation in the disease rates. Standard CAR models only focus on modelling data with area level covariates, but in some cases we may have access to individual level covariates which could be used to improve estimation. We propose a new algorithm which fits a CAR model in such a way as to accommodate both individual and area level covariates while adjusting for spatial correlation in the disease rates.