[Statistic modeling of geographic and epidemiologic variations].

In epidemiology, studies of the geographical variations of mortality or incidence rates for some chronic diseases have often given rise to etiological clues concerning those diseases. In this framework the variables concerned have a spatially autocorrelated structure which has to be taken into account in the statistical analysis. The statistical techniques used to study in the first place the spatial variations of mortality rates and then the joint geographical variations of mortality and exposure indices are reviewed. Emphasis is placed on the importance played by the geographical scale of the analysed data in the modelling process as well as on the interpretation problems of geographical correlation studies.