Download PDF
Methods Inf Med 2001; 40(02): 141-147
DOI: 10.1055/s-0038-1634477
Original Article
Schattauer GmbH

Survival Analysis with Time-varying Relative Risks: A Tree-Based Approach

R. Xu
1   Department of Biostatistics, Harvard School of Public Health, and Dana-Farber Cancer Institute, Boston MA, USA
,
S. Adak
1   Department of Biostatistics, Harvard School of Public Health, and Dana-Farber Cancer Institute, Boston MA, USA
› Author Affiliations
Further Information

Publication History

Publication Date:
07 February 2018 (online)

    Permissions and Reprints

Abstract

A tree-based method for estimating time-varying effects of baseline patient characteristics on survival is introduced. A Cox-type model for censored survival data is used in which the time-varying relative risks are modelled as piecewise constants.

The tree method consists of three steps: 1. Growing the tree, in which a fast algorithm using maximized score statistics is utilized to determine the optimal change points; 2. A pruning algorithm is applied to obtain more parsimonious models; 3. Selection of a final tree, which may be either via bootstrap resampling or based on a measure of explained variation.

The piecewise constant model is more suitable for clinical interpretation of the regression parameters than the more continuously time-varying models (spline, loess, etc.) that have been proposed previously.

Keywords

Change Point - Classification and Regression Tree - Maximized Score Test - Non-proportional Hazards - Time-varying Regression Effect
 

  • References

  • 1 Cox DR. Regression models and life tables (with discussion). J Roy Statist Soc, Series B 1972; 34: 187-220.
    PubMedGoogle Scholar
  • 2 Cox DR. Partial likelihood. Biometrika 1975; 62: 269-76.
    CrossrefPubMedGoogle Scholar
  • 3 Morgan J, Sonquist J. Problems in the analysis of survey data and a proposal. J Am Statist Assoc 1963; 58: 415-34.
    CrossrefPubMedGoogle Scholar
  • 4 Breiman L, Friedman J, Olshen R, Stone C. Classification and Regression Trees. Wadsworth International Group; 1984
    Google Scholar
  • 5 Gray RJ. Flexible methods for analyzing survival data using splines, with application to breast cancer prognosis. J Am Statist Assoc 1992; 87: 942-51.
    CrossrefPubMedGoogle Scholar
  • 6 Gustafson P. Flexible Bayesian modelling for survival data. Lifetime Data Anal 1998; 4: 281-99.
    CrossrefPubMedGoogle Scholar
  • 7 Hastie T, Tibshirani R. Varying-coefficient models. J Roy Statist Soc, Series B 1993; 55: 757-96.
    PubMedGoogle Scholar
  • 8 Sargent DJ. A flexible approach to time-varying coefficients in the Cox regression setting. Lifetime Data Anal 1997; 3: 13-25.
    CrossrefPubMedGoogle Scholar
  • 9 Sleeper LA, Harrington DP. Regression splines in the Cox model with application to covariate effects in liver disease. J Am Statist Assoc 1990; 85: 941-9.
    CrossrefPubMedGoogle Scholar
  • 10 Verweij JA, Van Houwelingen HA. Time-dependent effects of fixed covariates in Cox regression. Biometrics 1995; 51: 1550-6.
    CrossrefPubMedGoogle Scholar
  • 11 Kalbfleich JD, Prentice RL. The Statistical Analysis of Failure Time Data. 1980. John Wiley and Sons.;
    Google Scholar
  • 12 Adak S, Xu R. Survival analysis with time-varying regression effects using a tree-based approach: with application to International Non-Hodgkin’s Lymphoma Prognostic Factors data. Dana-Farber Cancer Institute Department of Biostatistical Science Technical Report. 1999 No. 1026.
    PubMedGoogle Scholar
  • 13 LeBlanc M, Crowley J. Survival trees by goodness of split. J Am Statist Assoc 1993; 88: 457-67.
    CrossrefPubMedGoogle Scholar
  • 14 O’Quigley J, Flandre P. Predictive capability of proportional hazards regression. Proc Natl Acad Sci USA 1998; 91: 2310-4.
    PubMedGoogle Scholar
  • 15 Xu R. Inference for the proportional hazards model. Ph D thesis of University of California; San Diego: 1996
    Google Scholar
  • 16 Xu R, O’Quigley J. Explained variation in proportional hazards regression. In: Handbook of Statistics in Clinical Oncology. ed. Crowley. Marcel Dekker, Inc.; 2001: 397-410.
    Google Scholar
  • 17 Csörgö M, Horváth L. Limit theorems in change-point analysis. John Wiley and Sons; 1997
    Google Scholar
  • 18 Davies RB. Hypothesis testing when a nuisance parameter is present only under the alternative. Biometrika 1977; 64: 247-54.
    CrossrefPubMedGoogle Scholar
  • 19 Davies RB. Hypothesis testing when a nuisance parameter is present only under the alternative. Biometrika 1987; 74: 33-43.
    PubMedGoogle Scholar
  • 20 Lausen B, Schumacher M. Maximally selected rank statistics. Biometrics 1992; 48: 73-85.
    CrossrefPubMedGoogle Scholar
  • 21 Lausen B, Schumacher M. Evaluating the effect of optimized cutoff values in the assessment of prognostic factors. Comput Statist Data Anal 1996; 21: 307-26.
    CrossrefPubMedGoogle Scholar
  • 22 Miller R, Siegmund D. Maximally selected Chi square statistics. Biometrics 1982; 38: 1011-6.
    CrossrefPubMedGoogle Scholar
  • 23 Xu R, O’Quigley J. Estimating Average Log Relative Risk Under Nonproportional Hazards. ASA Proc Biometrics Section; 1998: 216-22.
    Google Scholar