Student Seminar Series - April 14, 2005
University of Minnesota
School of Statistics
College of Liberal Arts

Comparing Two Crossing Hazard Rates By Cox Proportional Hazards Modeling


Jun Sheng


Thursday, April 14, 2005
1:00 PM, 240 Amundson Hall
Minneapolis, East Bank Campus

Refreshments at 12:30 PM
300 Ford Hall


Abstract

 Motivated by a clinical trial of Zinc nasal spray for the treatment of common cold, we consider the problem of comparing two crossing hazard rates. A new method, based on modeling the crossing hazard rates, is proposed, and, is implemented under the Cox proportional hazards modeling framework. The proposed method has several benefits, as compared to some existing ones, which include: (i) it can accommodate possible covariate effect easily, and (ii) it covers a wide range of crossing patterns of the hazard rates by using the Box-Cox transformation in its model. Monte Carlo simulations show that it outperforms some of its peers. Applications to a kidney dialysis patient's data and the Zinc nasal spray clinical trial data are discussed. Related research topics proposed for further study are described.