Activity Number:
|
21
|
Type:
|
Contributed
|
Date/Time:
|
Sunday, August 11, 2002 : 2:00 PM to 3:50 PM
|
Sponsor:
|
Biometrics Section*
|
Abstract - #301095 |
Title:
|
The Longitudinal Measurement Error Models with Applications to Cardiovascular Risk Factors
|
Author(s):
|
Shumei Guo*+ and Wei Wu and Cameron Chumlea
|
Affiliation(s):
|
Wright State University School of Medicine and Wright State University School of Medicine and Wright State University School of Medicine
|
Address:
|
3171 Research Blvd., Kettering, Ohio, 45420, USA
|
Keywords:
|
longitudinal model ; measurement error ; cardiovascular risk factors ; attenuation
|
Abstract:
|
Random effects models with a time-varying independent variable were used to demonstrate how the error in the independent variable attenuates the longitudinal relationships of the dependent and independent variables. Longitudinal data for low-density lipoprotein cholesterol (LDL) and high-density lipoprotein (HDL) and %BF from 22 to 40 years of age for 86 men and 100 women in the Fels Longitudinal Study were used. LDL and HDL are the dependent variables and %BF is the independent variable. The reliability for %BF is 0.75 for men and 0.79 for women. A unit increase in %BF increases LDL by 1.28 mg/dl in men and 0.77 mg/dl in women. After error adjustment for %BF, the regression estimate increases to 1.71 mg/dl for men and 0.97 mg/dl for women for LDL. A unit increase in %BF, reduces HDL from 0.47 mg/dl to 0.62 mg/dl for men and from 0.30 mg/dl to 0.38 mg/dl for women. The correction for error improves the longitudinal relationships between dependent and independent variables. Supported by HD 27063, NIH.
|