Comparison of Methods for Two Crossing Survival Curves Público

Chen, Yuqing (Spring 2018)

Permanent URL: https://etd.library.emory.edu/concern/etds/6108vb275?locale=pt-BR
Published

Abstract

In clinical research, it is not uncommon to see that the proportional hazard assumption is violated, particularly when survival curves cross. The log-rank test which is widely used in comparing survival curves is known to have less power under such circumstance. In this thesis, we introduce three methods to test the difference between two survival curves which are (1) Gehan’s weighted log-rank test, (2) Renyi type test, and (3) Lin and Xu’s test. We give a brief background of the statistical theory underpinning these methods. Then, we conduct a simulation study to compare the three tests with log-rank test under different situations and censor rates. We also apply all three methods to a real data example from a kidney dialysis trial. In our comparison of the three methods, we found that Gehan’s test does not perform well in the setting situations. The weight function of the weighted log-rank test must be specified prior to the analysis, and an inappropriate weight function may result in a misleading conclusion. Renyi test is suitable to use when two survival curves separate largely and cross, because its power stays above 90% even under a high censor rate. Lin and Xu’s test is appropriate when two survival curves do not separate largely, and its power is influenced by the censor rate.

Table of Contents

Table of Contents 1. Introduction........................................................................................................ 1 2. Weighted Log-rank Test ...................................................................................... 3 3. Renyi Type Test ................................................................................................... 6 4. Lin & Xu’s Test .................................................................................................... 8 5. Simulation Study ............................................................................................... 11 5.1 Estimation of Type I Error ........................................................................................... 11 5.2 Estimation of Power ................................................................................................... 11 5.2.1 Situation 1 .................................................................................................................... 12 5.2.2 Situation 2 .................................................................................................................... 13 5.2.3 Situation 3 .................................................................................................................... 14 6. Real Data Example from a Kidney Dialysis Trial .................................................. 16 7. Discussion and Conclusion ................................................................................ 18 8. Reference ......................................................................................................... 20 9. Appendix .......................................................................................................... 21

About this Master's Thesis

Rights statement
  • Permission granted by the author to include this thesis or dissertation in this repository. All rights reserved by the author. Please contact the author for information regarding the reproduction and use of this thesis or dissertation.
School
Department
Degree
Submission
Language
  • English
Research Field
Palavra-chave
Committee Chair / Thesis Advisor
Committee Members
Partnering Agencies
Última modificação

Primary PDF

Supplemental Files