Modelling competing risks in cancer studies

Competing risks arise commonly in the analysis of cancer studies. Most common are the competing risks of relapse and death in remission. These two risks are the primary reason that patients fail treatment. In most medical papers the effects of covariates on the three outcomes (relapse, death in remission and treatment failure) are model by distinct proportional hazards regression models. Since the hazards of relapse and death in remission must add to that of treatment failure, we argue that this model leads to internal inconsistencies. We argue that additive models for either the hazard rates or the cumulative incidence functions are more natural and that these models properly partition the effect of a covariate on treatment failure into its component parts. We illustrate the use and interpretation of additive models for the hazard rate or for the cumulative incidence function using data from a study of the efficacy of two preparative regimes for hematopoietic stem cell transplantation.

[1]  Zhiliang Ying,et al.  Semiparametric analysis of the additive risk model , 1994 .

[2]  H. Müller,et al.  Kernel estimation of regression functions , 1979 .

[3]  Els Goetghebeur,et al.  A modified log rank test for competing risks with missing failure type , 1990 .

[4]  J. Klein,et al.  Comparison of outcome following allogeneic bone marrow transplantation with cyclophosphamide–total body irradiation versus busulphan–cyclophosphamide conditioning regimens for acute myelogenous leukaemia in first remission , 2002, British journal of haematology.

[5]  Niels Keiding,et al.  Statistical Models Based on Counting Processes , 1993 .

[6]  Zhiliang Ying,et al.  Additive Hazards Regression Models for Survival Data , 1997 .

[7]  O. Aalen.  A linear regression model for the analysis of life times. , 1989, Statistics in medicine.

[8]  M. Pepe,et al.  Summarizing data on survival, relapse, and chronic graft‐versus‐host disease after bone marrow transplantation: motivation for and description of new methods , 1993, British journal of haematology.

[9]  Susanne Rosthøj,et al.  Competing risks as a multi-state model , 2002, Statistical methods in medical research.

[10]  S. Cheng,et al.  Confidence Bands for Cumulative Incidence Curves Under the Additive Risk Model , 1999, Biometrics.

[11]  Lee-Jen Wei,et al.  Prediction of cumulative incidence function under the proportional hazards model. , 1998, Biometrics.

[12]  John P. Klein,et al.  Discretizing a Continuous Covariate in Survival Studies , 2003, Advances in Survival Analysis.

[13]  S. Zeger,et al.  Longitudinal data analysis using generalized linear models , 1986 .

[14]  M S Pepe,et al.  Kaplan-Meier, marginal or conditional probability curves in summarizing competing risks failure time data? , 1993, Statistics in medicine.

[15]  John P. Klein,et al.  Estimating the Survival Function in the Proportional Hazards Regression Model: A Study of the Small Sample Size Properties , 1996 .

[16]  J. Fine,et al.  Regression modeling of competing crude failure probabilities. , 2001, Biostatistics.

[17]  P. Grambsch,et al.  Martingale-based residuals for survival models , 1990 .

[18]  D. E. Matthews.  Efficiency considerations in the analysis of a competing‐risk problem , 1984 .

[19]  J Crowley,et al.  Estimation of failure probabilities in the presence of competing risks: new representations of old estimators. , 1999, Statistics in medicine.

[20]  J. Klein,et al.  Generalised linear models for correlated pseudo‐observations, with applications to multi‐state models , 2003 .

[21]  Ian W. McKeague,et al.  Weighted Least Squares Estimation for Aalen's Additive Risk Model , 1991 .

[22]  Robert Gray,et al.  A Proportional Hazards Model for the Subdistribution of a Competing Risk , 1999 .

[23]  O. Aalen,et al.  Further results on the non-parametric linear regression model in survival analysis. , 1993, Statistics in medicine.