ABSTRACT
This chapter reviews some of the many possible models that may be used to represent the effect on failure time of explanatory variables. The explanatory variables may be classified also in other ways, in particular as for each individual constant or time-dependent. Explanatory variables measuring intrinsic properties of individuals include, in a medical context, such demographic variables as sex, age on entry, and variables describing medical history before admission to the study. Considerable care is needed in introducing as explanatory variables time-dependent variables that may be influenced by the treatment variables under investigation. The central property of the accelerated life model can be re-expressed in various ways that can be used as a basis for testing the adequacy of the model. For constant explanatory variables the question naturally arises as to when the proportional hazards model is also an accelerated life model.
