A Parametric Test for Correlated Time-to-Event Data with a Failure Rate Change

• Activity Number: 75
• Type: Contributed
• Date/Time: Sunday, August 1, 2010 : 4:00 PM to 5:50 PM
• Sponsor: ENAR
• Author(s): Gang Han and Ji-Hyun Lee*+
• Companies: Moffitt Cancer Center & Research Institute and Moffitt Cancer Center & Research Institute
• Keywords: log-rank test ; Gamma distribution ; uniformly most powerful unbiased test ; multiple comparison adjustments ; HPV incidence and clearance

Abstract

The log-rank test is widely used as a standard approach to compare survivorship. It assumes that the survival times are mutually independent. However, in some applications, the independence assumption may be invalid with multiple events per subject being possible. This study proposes a parametric test to compare the survivorship based on the exponential distribution. Specifically, using the memory-less property of the exponential distribution, we derive Gamma distributions of two total-time-on-tests (TTOTs), with the number of events as the shape parameter and the failure rate as the scale parameter. A uniformly most powerful unbiased (UMPU) test is used to compare the two failure rates. Multiple comparison adjustments are used if more than two groups are being compared. This method will be applied to the HPV in men (HIM) study to test the effect of several factors.