An Exponential Model Relating Censored Survival Data and Concomitant Information for Prostatic Cancer Patients2

Abstract

An exponential survival model was developed and used to estimate the relative importance of presenting signs and symptoms in prediction of survival of patients with advanced (Stages III and IV) carcinoma of the prostate gland. Based on this information, one can predict survival for patients with various combinations of presenting signs and symptoms. Survival predictions based on the exponential model agreed remarkably well with observed survival rates.