Fitting a mixture distribution to complex censored survival data using generalized linear models

Mixture models may arise for a variety of reasons in survival data analysis. This paper shows how such models that involve potentially complex cross-classification by covariates may be easily fitted using a package such as GLIM. The method employs an auxiliary Poisson-binomial model in order to find the maximum-likelihood estimates of the model parameters, and has been implemented using GLIM macros.