ABSTRACT
When explanatory variables are incorporated in a model for survival data, the values taken by such variables are those recorded at the time origin of the study. For example, consider the study to compare two treatments for prostatic cancer rst described in Example 1.4 of Chapter 1. Here, the age of a patient, serum haemoglobin level, size of the tumour, value of the Gleason index, and of course the treatment group, were all recorded at the time when a patient was entered into the study. The impact of these variables on the hazard of death is then evaluated.
