A proportional hazards model for informatively censored survival times

The models in Demasi et al. (1997) are adapted for informatively censored failure times to a primary event and times to followup events. The proposed model treats informative censoring as a type of risk in a competing risks setup. Inferences about the parameters in the model can be based on standard partial-likelihood theory. A simulation study shows that likelihood inference works well in practice. Use of the model is demonstrated by analyzing some data on waiting time to bone marrow transplantation (primary event) and time to relapse or death after transplantation (followup event) for leukemia patients. Residual inspection was performed to check the adequacy of the model.