ABSTRACT
This chapter focuses on relative risk estimation and modeling of risk both in terms of correlation between diseases and in terms of comparison. In general, once multiple diseases are admitted into an analysis there is a need to consider relations between the diseases. The basic likelihood derived for the multitype situation can be seen as a special case of an ordinal logistic formulation where a probability of a disease type is to be modeled. There are few published examples of Bayesian analysis of multi-type spatial disease realizations. A multivariate analysis of the residential locations of death certificates for respiratory disease and air-way cancers was proposed by Andrew B. Lawson and F. Williams. A classic approach to modeling cross-correlation between spatial fields is to adopt a simple model for the relation between selected fields. Within geostatistics, the linear model of coregionalisation. The chapter considers the both case event and count data.
