|
Activity Number:
|
40
|
|
Type:
|
Invited
|
|
Date/Time:
|
Sunday, July 29, 2007 : 4:00 PM to 5:50 PM
|
|
Sponsor:
|
SSC
|
| Abstract - #308093 |
|
Title:
|
Longitudinal Data in Threshold Regression: Implementation and Relation to Cox Regression with Time-Varying Covariates
|
|
Author(s):
|
Mei-Ling Ting Lee*+ and George A. Whitmore
|
|
Companies:
|
The Ohio State University and McGill University
|
|
Address:
|
B-122 Starling-Loving Hall, Biostatistics Division, Columbus, OH, 43210,
|
|
Keywords:
|
survival analysis ; Cox regression ; longitudinal data ; wiener process ; boundary crossing ; time-varying covariates
|
|
Abstract:
|
Longitudinal survival data pose an interesting challenge. Lee and Whitmore (2007) review a new regression methodology for survival data referred to as threshold regression. The methodology is based on the concept that degradation of an item follows a stochastic process and failure occurs when the process first reaches a failure state or threshold. Breaking longitudinal records into series of single records is one strategy that has been proposed. This study looks at the formal conditions that must hold for this uncoupling procedure to be valid. The conditions are examined in terms of both theory and practical application. The uncoupling procedure modifies the time scale for the analysis and can be used in conjunction with an operational time scale. We show that the Cox proportional hazards regression model with time-varying covariates is a special semi-parametric version of the model.
|
