|
Activity Number:
|
548
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Thursday, August 2, 2007 : 10:30 AM to 12:20 PM
|
|
Sponsor:
|
Section on Statistics in Epidemiology
|
| Abstract - #309502 |
|
Title:
|
Using Surrogate Outcomes for Improving Power To Detect Gene-Environment Interactions
|
|
Author(s):
|
Tamanna Howlader*+ and Michal Abrahamowicz and Yogendra P. Chaubey
|
|
Companies:
|
Concordia University and McGill University and Concordia University
|
|
Address:
|
Dept. of Mathematics and Statistics, 1400 de Maisonneuve Blvd West, Montreal, QC, H3G 1M8, Canada
|
|
Keywords:
|
Gene-environment interaction ; Surrogate outcome ; Power
|
|
Abstract:
|
This study explores the use of quantitative surrogates of a clinical binary outcome to improve low power to detect gene-environment (G × E) interaction which is of major concern in genetic-epidemiologic research. We consider hypothetical models of the relationship between the binary and quantitative surrogate outcomes, and their relationships to genetic susceptibility, exposure, and other risk factors. Simulations are used to estimate power of the test for G × E interaction in linear and logistic regression models. Sensitivity analyses are performed to assess the impact on power of important parameters, such as strength of the underlying (G × E) interaction effect, and measurement errors in the outcomes. It is found that under certain conditions, higher power can be achieved by replacing the binary outcome by a quantitative surrogate outcome.
|
