[PDF] Cox Proportional-Hazards Regression for Survival Data in R

What is the Cox proportional-hazards regression model?

TheCox proportional-hazards regression model is the most common tool for studying thedependency of survival time on predictor variables. This appendix to Fox and Weisberg (2019)briey describes the basis for the Cox regression model, and explains how to use thesurvivalpackage inRto estimate Cox regressions.

What is stratied Cox regression model?

stratied Cox regression model is t by including a call to thestrata()function on the right-hand side of the model formula. The arguments to this function are one or more stratifying variables;if there is more than one such variable, then the strata are formed from their cross-classication.

What hazard model is used for survival data?

Models for survival data usually employ the hazard function or the log hazard. For example,assuming a constant hazard,h(t) =, implies an exponential distribution of survival times, withdensity functionp(t) =e t. Other common hazard models include which leads to theWeibull distributionof survival times.

What is the difference between a baseline hazard function and a covariate?

Each stratum is permitted to have a dierent baseline hazardfunction, while the coecients of the remaining covariates are assumed to be constant across strata.An advantage of this approach is that we do not have to assume a particular form of interactionbetween the stratifying covariates and time.