ABSTRACT
Covariates are commonly included in analyses of longitudinal clinical trials. In Section 8.1, it was noted that it is usually preferable, or at least it should be considered, to allow a full interaction of covariates with time. Fitting covariate-by-time interactions avoids imposing restrictions that the dependence of response on the covariates is the same at all assessment times. Therefore, throughout this chapter, examples will be used that include not only the covariate but also the covariate interaction with time.
