|
Activity Number:
|
380
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Tuesday, August 4, 2009 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
Biometrics Section
|
| Abstract - #304350 |
|
Title:
|
Two-Sided Hypothesis Testing Based on Pool Screening with Unequal Pool Sizes
|
|
Author(s):
|
Hongjiang Gao*+ and Inmaculada Aban and Charles R. Katholi
|
|
Companies:
|
The University of Alabama at Birmingham and The University of Alabama at Birmingham and The University of Alabama at Birmingham
|
|
Address:
|
1665 University Blvd, Birmingham, AL, 35294,
|
|
Keywords:
|
Pool Screening ; Sum of Positive Pools ; Exact Test ; Likelihood-based Test ; Statistical Power
|
|
Abstract:
|
Pool screening is widely used in disease control as an efficient way to estimate near zero prevalence levels. Laboratory screening test may only have the capability of handling some maximum sample. If a pool size is larger than this maximum, it needs to be subdivided into smaller pools to meet screening test requirement. We propose and compare two hypothesis testing procedures under the setting of unequal pool sizes assumed to be fixed and known: 1)exact test based on the sum of positive pools, 2)likelihood-based tests. Because the asymptotic distributions of likelihood-based tests are far from the expected chi-squared distribution when the prevalence is small, we show that simulated quantiles of likelihood-based statistics to define the new rejection region improves the performance of these tests. Exact test outperforms the other tests when the prevalence is close to zero.
|
