Survival Analysis Using SAS®: A Practical Guide

The Proportional Hazards Model

In his 1972 paper, Cox made two significant innovations. First, he proposed a model that is standardly referred to as the proportional hazards model. That name is somewhat misleading, however, because the model can readily be generalized to allow for nonproportional hazards. Second, he proposed a new estimation method that was later named partial likelihood or, more accurately, maximum partial likelihood. The term Cox regression refers to the combination of the model and the estimation method. It didn’t take any great leap of imagination to formulate the proportional hazards model—it’s a relatively straightforward generalization of the Weibull and Gompertz models we considered in Chapter 2, “Basic Concepts of Survival ...