Survival Analysis Using SAS

CHAPTER 5

Estimating Cox Regression Models with PROC PHREG

Introduction

The Proportional Hazards Model

Partial Likelihood

Tied Data

Time-Dependent Covariates

Cox Models with Nonproportional Hazards

Interactions with Time as Time-Dependent Covariates

Nonproportionality via Stratification

Left Truncation and Late Entry into the Risk Set

Estimating Survivor Functions

Testing Linear Hypotheses with CONTRAST or TEST Statements

Customized Hazard Ratios

Bayesian Estimation and Testing

Conclusion

INTRODUCTION

PROC PHREG (pronounced P-H-REG, not FREG) implements the regression method first proposed in 1972 by the British statistician Sir David Cox in his famous paper “Regression Models and Life Tables” (Journal of the Royal Statistical Society, Series ...