Bayesian Detection of Clusters and Discontinuities in Disease Maps: Simulations. (REVISED, June 1999)

This paper is a supplement paper to Knorr-Held and Rasser (1999),
Discussion Paper 107. "Bayesian Detection of Clusters and
Discontinuities in Disease Maps", which describes a novel approach
to disease mapping with particular emphasis on the detection of
clusters and discontinuities in disease maps. First we investigate
several features of the prior by simulations from the clustering
model both for the 544 regions of Germany and for the 88 counties
of Ohio state. Furthermore the method is applied to various
artificial datasets for Ohio with and without spatial patterns.
Sensitivity with respect to prior parameters is studied.