ABSTRACT
When the Cox regression model is used in the analysis of survival data, there is no need to assume a particular form of probability distribution for the survival times. As a result, the hazard function is not restricted to a specic functional form, and the model has exibility and widespread applicability. On the other hand, if the assumption of a particular probability distribution for the data is valid, inferences based on such an assumption will be more precise. In particular, estimates of quantities such as relative hazards and median survival times will tend to have smaller standard errors than they would in the absence of a distributional assumption. Models in which a specic probability distribution is assumed for the survival times are known as parametric models, and parametric versions of the proportional hazards model, described in Chapter 3, are the subject of this chapter.
