|
Activity Number:
|
175
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Tuesday, August 13, 2002 : 8:30 AM to 10:20 AM
|
|
Sponsor:
|
Section on Statistics in Epidemiology*
|
| Abstract - #300551 |
|
Title:
|
Estimating the Correlation between Left-censored Variables: Comparison between GEE Approach and MLE Approach
|
|
Author(s):
|
Jingli Song*+ and Huiman Barnhart and Robert Lyles
|
|
Affiliation(s):
|
Emory University and Rollins School of Public Health of Emory University and Rollins School of Public Health of Emory University
|
|
Address:
|
2515 NE Expressway U-5, Atlanta, Georgia, 30345, USA
|
|
Keywords:
|
Left-censored variables ; Correlation; Missing Data ; Generalized estimating equations
|
|
Abstract:
|
HIV (Human Immunodeficiency Virus) researchers are often concerned with the correlation between HIV viral load measurements and CD4+ lymphcyte counts, or the correlation between HIV viral load levels from two reservoirs or from two competing quantification assays. Due to the lower limit of detection (LOD) of such assays, HIV viral load measurements are subject to left-censoring. The maximum likelihood method is commonly used for estimating correlation of left-censored variables. In this paper, we propose a generalized estimating equations (GEE) approach to estimate the correlation coefficient between two continuous variables, where one or both of them may be left-censored. We compare the GEE approach with the MLE approach through both simulation and real data sets. We also explore the robustness to the normality assumption of the two approaches via simulation studies. We use a real data to illustrate the advantage of the GEE approach in incorporating covariates.
|
