Bootstrapping in the context of Cox’s model has been discussed by Burr (1994). She considered three resampling algorithms:
1. Bootstrap resample the records for the n individuals in the study.
2. Generate a random survival time Yi for the ith individual in the study using the estimate of the survival function (Equation 13.2). Generate a random censoring time Ci for the ith individual from the Kaplan-Meier estimate of the censoring time distribution. Decide whether the ith individual is observed to die or is censored before this. Use the resulting data for all n individuals as a bootstrap sample.