|
Activity Number:
|
600
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Thursday, August 6, 2009 : 10:30 AM to 12:20 PM
|
|
Sponsor:
|
Section on Statistics in Epidemiology
|
| Abstract - #305015 |
|
Title:
|
A Likelihood Approach to Pooled Prevalence Estimation with Random Effects
|
|
Author(s):
|
Chengxing Lu*+ and Mikyong Shin and Adolfo Correa
|
|
Companies:
|
CDC and CDC and CDC
|
|
Address:
|
351 N Sumneytown Pike, North Wales, PA, 19454,
|
|
Keywords:
|
Pooled prevalence ; Random effect ; Region-specific effect
|
|
Abstract:
|
One of the major challenges in analyzing multi-regional epidemiological surveillance data is to provide an overall prevalence estimation and confidence interval given the heterogeneous nature of various regions. In this work, we assume the logit of the estimated prevalences for individual regions follow a normal distribution with a random effect accounting for the region-specific effect. In a real-life example estimating the overall prevalence of Down Syndrome among children and adolescents in ten regions of the U.S., we will assess the validity of the model assumptions, demonstrate the implementation of the proposed method and compare the prevalence estimations and confidence intervals with and without considering the heterogeneous regional effect.
|
