ABSTRACT
A case could be an incident or prevalent case or could be a death certificate address. We assume that the cases are geo-coded down to a point. Hence they form a point process in space. This is called a realization of the disease process, in that we assume that all cases within the study area are recorded. This is a common form of data available from government agencies. Sub-samples of the spatial domain, where incomplete realizations are taken are not considered at this point. The first component makes allowance for the underlying population in study region, and the second component is usually specified with the modelled components. The analysis reduces to that of a logistic likelihood, and this has the advantage that the "at risk" population nuisance function does not have to be estimated. This model is ideally suited to situations where it is natural to have a control and case realization, where conditioning on the spatial pattern is reasonable.
