A measure of explained risk is developed for application with the proportional hazards model. The statistic, which is called the estimated explained relative risk, has a simple analytical form and is unaffected by censoring that is independent of survival time conditional on the covariates. An asymptotic confidence interval for the limiting value of the estimated explained relative risk is derived, and the role of individual factors in the computation of its estimate is established. Simulations are performed to compare the results of the estimated explained relative risk to other known explained risk measures with censored data. Prostate cancer data are used to demonstrate an analysis incorporating the proposed approach.