Dealing With Non-proportional Hazards in Survival Analysis

Abstract

For the last few decades,exceptional concentration has been given to the field of survival analysis for its techniques applied in different areas of research.One of the main assumptions regarding a number of survival analysis techniques is that of proportional hazards.If the hazards rates are found to cross,non parametric analyses such as Cox proportional hazards regression,Kaplan Meier and Log- rank test will either be rendered scarce or lose power.One way to suffiently analyze survival data with crossing hazard rate is to obtain inference using the Renyi test statistic and Stratified Cox Regression Model,interaction model with time dependent covariate.