A Robust Parameterization for Unbounded Covariates Within the Cox Proportional Hazards Model

Authors

  • Richard J. Jackson Cancer Research UK Liverpool Cancer Trials Unit, University of Liverpool, Liverpool, UK
  • Trevor F. Cox Cancer Research UK Liverpool Cancer Trials Unit, University of Liverpool, Liverpool, UK

DOI:

https://doi.org/10.6000/1929-6029.2014.03.04.1

Keywords:

ESPAC 3 trial, hazard ratio, influence function, logistic function, Schoenfeld residuals.

Abstract

The Cox proportional hazards model is widely used in the analysis of medical data either for survival or time to a particular event. Factors and continuous covariates can be easily incorporated into the model and hazard ratios calculated. The model can however be distorted when extreme value observations occur within a continuous covariate and the hazard ratio can become extremely large. To overcome this, transformations of the covariate are often made, which can be simple, e.g. log, or more sophisticated such as the fitting of a fractional polynomial. This paper takes a different approach and makes a transformation based on the logistic function that has the property that the hazard ratio is bounded. The models are introduced and discussed. Model diagnostics based on Schoenfeld residuals and the influence function are established and then data from a pancreatic cancer trial are used to illustrate the model.

Author Biographies

Richard J. Jackson, Cancer Research UK Liverpool Cancer Trials Unit, University of Liverpool, Liverpool, UK

Cancer Research UK Liverpool Cancer Trials Unit

Trevor F. Cox, Cancer Research UK Liverpool Cancer Trials Unit, University of Liverpool, Liverpool, UK

Cancer Research UK Liverpool Cancer Trials Unit

References

Cox DR. Regression Models and Life-Tables. JR Stat Soc Series B 1972; 34: 187-220. DOI: https://doi.org/10.1111/j.2517-6161.1972.tb00899.x

Viviani S, Farcomeni A. Robust estimation for the Cox regression model based on trimming. Biometrical Journal 2011; 53: 956-973. http://dx.doi.org/10.1002/bimj.201100008 DOI: https://doi.org/10.1002/bimj.201100008

Royston P, Lambert PC. Flexible parametric survival analysis using Stata: Beyond the Cox model. Texas: Stata Press 2011.

Farcomeni A, Ventura L. An overview of robust methods in medical research. Stat Methods Med Res 2010; 21: 111-133. http://dx.doi.org/10.1177/0962280210385865 DOI: https://doi.org/10.1177/0962280210385865

Bednarski T. Robust estimation in Cox’s Regression models. Scand J Stat 1993; 20: 213-225.

Minder CE, Bednarski T. A robust method for proportional hazards regression. Statist Med 1996; 15: 1033-1047. http://dx.doi.org/10.1002/(SICI)1097-0258(19960530)15:10<1033::AID-SIM215>3.0.CO;2-Y DOI: https://doi.org/10.1002/(SICI)1097-0258(19960530)15:10<1033::AID-SIM215>3.0.CO;2-Y

Burton A, Altman DG, Royston P, et al. The design of simulation studies in medical statistics. Statist Med 2006; 25: 4279-4292. http://dx.doi.org/10.1002/sim.2673 DOI: https://doi.org/10.1002/sim.2673

Schoenfeld D. Partial residuals for the proportional hazards regression model. Biometrika1982; 69, 239-241. http://dx.doi.org/10.1093/biomet/69.1.239 DOI: https://doi.org/10.1093/biomet/69.1.239

Reid N, Crepeau H. Influence function for proportional hazards regression. Biometrika1985; 72: 1-9. http://dx.doi.org/10.1093/biomet/72.1.1 DOI: https://doi.org/10.1093/biomet/72.1.1

Neoptolemos JP, Stocken DD, Bassi C, et al. Adjuvant Chemotherapy With Fluorouracil Plus Folinic Acid vs Gemcitabine Following Pancreatic Cancer Resection: A Randomized Controlled Trial. JAMA 2010; 304: 1073-1081. http://dx.doi.org/10.1001/jama.2010.1275 DOI: https://doi.org/10.1001/jama.2010.1275

Downloads

Published

2014-11-06

How to Cite

Jackson, R. J., & Cox, T. F. (2014). A Robust Parameterization for Unbounded Covariates Within the Cox Proportional Hazards Model. International Journal of Statistics in Medical Research, 3(4), 331–339. https://doi.org/10.6000/1929-6029.2014.03.04.1

Issue

Section

General Articles