COVARIATE ADJUSTED MIXTURE-MODELS AND DISEASE MAPPING WITH THE PROGRAM DISMAPWIN

The analysis and recognition of disease clustering in space and its re presentation on a map is an important problem in epidemiology. An appr oach using mixture models to identify spatial heterogeneity in disease risk and map construction within an empirical Bayes framework is desc ribed. Once heterogeneity is detected, the question arises as how expl anatory variables could be included in the model. A mixed Poisson regr ession approach to include covariates is presented. The methods are il lustrated using data for tuberculosis from Berlin in 1991.