|
Activity Number:
|
395
|
|
Type:
|
Invited
|
|
Date/Time:
|
Wednesday, August 6, 2003 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
SSC
|
| Abstract - #300708 |
|
Title:
|
Semiparametric Maximum Likelihood Estimation in Case-Control Studies of Gene-Environment Interaction
|
|
Author(s):
|
Nilanjan Chatterjee*+ and Raymond J. Carroll
|
|
Companies:
|
National Institute of Health and Texas A&M University
|
|
Address:
|
Room 8038, Bethesda, MD, 20892-0001,
|
|
Keywords:
|
profile likelihood ; case-only design ; semiparametric efficiency ; conditional independence
|
|
Abstract:
|
It is now believed that the risks of many complex diseases are determined by the joint effect of genetic susceptibility and environmental exposures. We consider the problem of maximum likelihood estimation in case-control studies of gene-environment when genetic and environmental exposures can be assumed to be independent in the underlying population. Traditional logistic regression analysis, which is well-known to be fully efficient for analysis of case-control data when the exposure distribution is unspecified, may not be efficient in this setting. By exploiting certain "missing data structure" of the retrospective likelihood, we develop an algorithm for obtaining the semiparametric maximum likelihood estimator (SPMLE) of logistic-regression parameters that leaves the distribution of the environmental exposures to be unspecified. We develop the asymptotic theory for the SPMLE. We use simulation study to investigate small sample properties of the estimator. An application of the method is illustrated using data from a case-control study designed to investigate the interplay of BRCA1/2 mutations and oral contraceptive use in the etiology of ovarian cancer.
|
