|
Activity Number:
|
199
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Monday, August 7, 2006 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
Section on Statistics in Epidemiology
|
| Abstract - #306856 |
|
Title:
|
Analysis of Longitudinal Trinomial Outcome through a Surrogate Variable
|
|
Author(s):
|
Wenyaw Chan*+ and Yen-Peng Li and Hung-Wen Yeh
|
|
Companies:
|
The University of Texas School of Public Health and The University of Texas Health Science Center at Houston and The University of Texas School of Public Health
|
|
Address:
|
1200 Herman Pressler, Houston, TX, 77030,
|
|
Keywords:
|
continuous-time Markov chain ; EM algorithm ; hidden Markov model ; longitudinal studies ; surrogate variable
|
|
Abstract:
|
Researchers in public health studies often encounter the situation that the outcome variable can not be measured properly due to confidentiality or limitation. In this circumstance, a surrogate variable that is accessible and can gauge the character of the targeted outcome is often adopted to carry on the study. In this research, a longitudinal surrogate outcome will be used for statistical analysis in which the targeted outcome is assumed to follow a trinomial distribution. We propose to use the EM algorithm and a continuous-time hidden Markov model to analyze a longitudinal multinomial outcome that has three categories and is accessible only through a surrogate variable. The accuracy of the point estimation and the sensitivity of the computational algorithm will be examined by an empirical study. An example in public health will be demonstrated.
|
