ABSTRACT
To set up the framework for survival data with right censoring, two random variables need to be defined
Tsurv : the survival time, Tcens : the censoring time.
The censoring time Tcens is often denoted by C. Different censoring mechanisms can be distinguished. A common one for clinical data is administrative censoring, where the censoring time is determined by the termination of the study. For most purposes it suffices to consider the censoring to be random. The crucial condition for statistical analysis is that survival time Tsurv and censoring time Tcens are independent. In the presence of explanatory
can be weakened to independence of Tsurv and Tcens conditional on the explanatory variables.
