A Parametric Cox Proportional Hazard Model with Application

Abstract

Survival analysis has become integral to clinical studies, especially in emerging diseases and terminal ailments. This study focused on improving the popular Cox PH model. The new method developed is a parametric type, incorporating the hazard rate of the exponential distribution. It was noted that though the functional form of the Cox PH model was altered, the assumptions were upheld. Additionally, the new model parameters were estimated using the same maximum partial likelihood as the Cox model. Data on the survival times of 137 patients who underwent bone marrow transplants were deployed, and the proposed parametric Cox PH model proved superior to the Cox PH model.