[Epidemiological analysis of rank category exposure effects using cumulative logistic models]

Logistic regression is the fundamental procedure to analyze the dependence of the frequency of a binary disease indicator from exposure and other factors. Disease indicators with ranked levels contain a larger amount of information and permit a more precise prediction. For that, the cumulative logistic model was developed. Its application, demonstrated with an epidemiologic study for the dependence of arthritic changes in the knee-joint from intensity of kneeling stress and age, is very recommendable for the evaluation of well-designed epidemiologic studies concerning the prevalence of chronic diseases.