Article
Nonparametric test for multiple crossing survival curves
Search Medline for
Authors
Published: | September 6, 2007 |
---|
Outline
Text
Log-rank and Wilcoxon tests are most commonly used tests for testing the overall homogeneity of survival curves, but in certain situation it appears that they have a significant loss of statistical testing power. One such case is the more than one time crossing of survival curves. The problem considered often occurs in medical research. To overcome this problem, in this article, we present and study a nonparametric test procedure based on a new weight. The proposed new weighted test has greater power to detect overall differences between more than one time crossing survival curves. Simulation studies are performed to compare the proposed method with existing methods. Furthermore, the advantage of the new test is finally exemplified in the analysis of a β-thalassaemia major data.