Abstract
The presence or absence of autoimmune disease is a qualitative outcome resulting from the combined influences of genetic, environmental and stochastic factors that interact in complex ways. Because of the nature of these interactions, the effects of risk factors that contribute to the aetiology of disease cannot be adequately assessed individually. Here, I propose an approach to multiparameter modelling of autoimmune disease that incorporates Falconer's concept of liability in a generalized linear model. Specifically, a logistic model, allied to that used to assess the influence of genetic risk factors, is suggested. The limits of modelling autoimmune disease and the requirements for data collection are also briefly discussed.