|
Activity Number:
|
292
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Tuesday, July 31, 2007 : 10:30 AM to 12:20 PM
|
|
Sponsor:
|
Section on Statistics in Epidemiology
|
| Abstract - #310210 |
|
Title:
|
Estimating the Prevalence of Disease Using Relatives of Case and Control Profanes
|
|
Author(s):
|
Kristin Javaras*+ and Nan Laird and Brian D. Ripley and James I. Hudson
|
|
Companies:
|
Harvard School of Public Health and Harvard University and Oxford University and Harvard Medical School
|
|
Address:
|
Department of Biostatistics, Boston, MA, 02115,
|
|
Keywords:
|
Disease prevalence ; Case-control family study ; Proportion
|
|
Abstract:
|
Case-control family studies are conducted to investigate familial aggregation of a disease. However, the resulting data (disease status and covariates for members of families selected via case and control probands) can also be used to estimate the prevalence of the disease. We propose estimators for overall and covariate-stratum-specific prevalence that are approximately unbiased for their population counterparts, given certain commonly made assumptions. We also propose corresponding confidence intervals that have good coverage properties, even for small prevalence's. The estimators and intervals address the over-representation of diseased individuals by excluding profanes and taking into account whether each relative was selected via a case or control probed. The estimators and intervals perform well in simulation experiments designed to mimic a real family study of major depression.
|
