Document Type


Publication Date



Logistic models are commonly used to analyze case-control data. For case-control studies, if there tends be rare disease in the control group with the risk factors, then the estimation procedure using logistic regression model for such factors becomes difficult. To overcome such situation, Chen et. al. (2003) proposed a hybrid logistic model in which they first estimate the problematic risk factor assuming that proportions having disease of such risk factor are treated equal for all permissible strata of the other risk factors, and then the residual of the risk factors are modeled by using logistic regression model. The purpose of this paper is to extend theoretically the hybrid logistic model for case-control studies for more than one rare risk factor.