MIXTURE-MODELS FOR THE JOINT DISTRIBUTION OF REPEATED-MEASURES AND EVENT TIMES

Many long-term clinical trials collect both a vector of repeated measu rements and an event time on each subject; often, the two outcomes are dependent. One example is the use of surrogate markers to predict dis ease onset or survival. Another is longitudinal trials which have outc ome-related dropout, We describe a mixture model for the joint distrib ution which accommodates incomplete repeated measures and right-censor ed event times, and provide methods for full maximum likelihood estima tion, The methods are illustrated through analysis of data from a clin ical trial for a new schizophrenia therapy; in the trial, dropout time is closely related to outcome, and the dropout process differs betwee n treatments. The parameter estimates from the model are used to make a treatment comparison after adjusting for the effects of dropout. An added benefit of the analysis is that it permits using the repeated me asures to increase efficiency of estimates of the event time distribut ion.