In Example 8.5, a proportional hazards model was built to the data on disease-free survival for bone marrow transplantation patients. Of primary interest in that example was the comparison of disease states, and possible factors to be adjusted for were the patients' FAB status, their waiting time to transplant, donor and recipient gender, CMV status, and age. Because patients who developed acute graft-versus-host disease may have different risk factors for disease-free survival, find the best fitting model for these factors for those patients who have experienced acute graft-versus-host disease. Compare your final model to that found in Example 8.5.

Problem 9.7 of Survival analysis techniques for censored and truncated data.