|
Activity Number:
|
323
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Tuesday, August 4, 2009 : 10:30 AM to 12:20 PM
|
|
Sponsor:
|
Biometrics Section
|
| Abstract - #303944 |
|
Title:
|
Exact and Asymptotic Weighted Logrank Tests for Interval-Censored Data: A Review and New R Package
|
|
Author(s):
|
Michael P. Fay*+ and Pamela A. Shaw
|
|
Companies:
|
National Institute of Allergy and Infectious Diseases and National Institute of Allergy and Infectious Diseases
|
|
Address:
|
6700-A Rockledge Drive, Room 5133, Bethesda, MD, 20817,
|
|
Keywords:
|
hypothesis test ; interval censoring ; permutation test ; rank test
|
|
Abstract:
|
We review several methods for generalizing the logrank and Wilcoxon-Mann-Whitney tests to interval censored data. We present a new R package for interval censored data and its validation. The calculation of the p-value for these tests may be divided into three parts: (1) estimation of the nonparametric maximum likelihood estimate (NPMLE) of the joint survival distribution, (2) using that NPMLE to create individual scores for each subject, and (3) performing inference on those scores. The new R package has been validated against existing software for parts 1 and 3. We show through a simple hypothetical example how ties in the scores can easily occur. These "true ties" can have very small differences in their calculated scores (since the scores use the iterated NPMLE estimate). We show how these ties are handled to get reproducible p-values.
|
