|
Activity Number:
|
428
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Wednesday, August 9, 2006 : 10:30 AM to 12:20 PM
|
|
Sponsor:
|
Section on Statistics in Epidemiology
|
| Abstract - #306492 |
|
Title:
|
Assessment of Capture-Recapture Methods in Estimating Populations of Oklahomans with Birth Defects
|
|
Author(s):
|
Robert C. Wild*+ and Barbara Neas
|
|
Companies:
|
The University of Oklahoma Health Sciences Center and The University of Oklahoma Health Sciences Center
|
|
Address:
|
Department of Biostatistics and Epidemiology, Oklahoma City, OK, 73109,
|
|
Keywords:
|
capture-recapture ; Indian Health Service ; multiple-records systems ; log-linear models ; bootstrap ; population size estimation
|
|
Abstract:
|
Originally developed by ecologists, capture-recapture methods are used in epidemiology to obtain accurate estimates of true population sizes from independent but overlapping samples from the same population. The Oklahoma Birth Defects Registry (ODBR) is a population-based, active surveillance registration system that includes data from various standard sources and other specialty sources such as the Indian Health Service. Recently, capture-recapture methods have been developed that estimate populations sizes from samples that have heterogeneous inclusion probabilities. These methods can be applied to Oklahomans with birth defects. This paper addresses the key statistical techniques in capture-recapture methods that are potentially applicable to the ODBR. Log-linear model selection and variance estimation methods that include MLE and empirical bootstrap techniques are also presented.
|
