Spatial mixture relative risk models applied to disease mapping.

Abstract

An important issue within health services research is the correct allocation of resources within health authority regions and the capability of public health professionals to make such allocation appropriately. This allocation is often based on a mapping of relevant disease incidence and the assessment of the geographical distribution of relative risk of disease in small areas within the health authority administrative domain. Existing methods for the statistical analysis of small area risk are mostly based on smoothing methods. However, these methods often smooth over large discontinuities in the risk surface which might be important to maintain for the purposes of resource allocation. In this paper we propose a method that involves the use of spatial mixtures of components that can provide a balance between smoothness and the maintenance of discontinuity. The method is applied to a sudden infant death incidence data set.