A methodology for analysing a repeated measures and survival outcome simultaneously.

A considerable body of literature has arisen over the past 15 years for analysing univariate repeated measures data. It is rare in applied biomedical research, however, for interest to be restricted to a single outcome measure. In previous work, Rochon considered the case of bivariate repeated measures data in which each outcome is a discrete or continuous random variable. A GEE model was prescribed to relate each set of repeated measures to important explanatory variables. The 'seemingly unrelated regression' model was then applied to combine the pair of GEE models into an overall analysis framework. In this paper, we extend this approach to the case in which one endpoint is a repeated measures outcome but the second is a survival endpoint. Estimation and hypothesis testing issues are addressed and the methodology is illustrated with an example. Copyright 2001 John Wiley & Sons, Ltd.