|
Activity Number:
|
199
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Monday, July 30, 2007 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
Biopharmaceutical Section
|
| Abstract - #310327 |
|
Title:
|
Nonproportional Hazards in Time-to-Event Analysis
|
|
Author(s):
|
Yanqiong Zhang*+ and Amarjot Kaur
|
|
Companies:
|
Merck & Co., Inc. and Merck & Co., Inc.
|
|
Address:
|
81 Shelley Circle, East Windsor, 08520,
|
|
Keywords:
|
quantitative nonproportionality ; qualitative nonproportionality ; asymptotic relative efficiency
|
|
Abstract:
|
Cox proportional hazard (PH) model (1972) is the most commonly used method for time-to-event analysis and is referred to as a semiparametric model. Many a time in practice, the proportional hazard assumption does not always hold due to potentially two reasons. One reason is due to crossover of two hazard functions, where the hazard ratio could be >1 during one time period and <1 in the other period, also referred to as qualitative nonproportionality. The other reason of nonproportionality could arise where the hazard ratio doesn't keep constant over the time periods but no crossover of the hazard ratios occurs, also referred to as quantitative nonproportionality. An exploratory investigation of the power loss as well as asymptotic relative efficiency for the score test will be carried out in the presence of the quantitative nonproportionality.
|
