|
Activity Number:
|
133
- Statistical Issues in Environmental Epidemiology and Pharmacoepidemiology
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Monday, August 9, 2021 : 1:30 PM to 3:20 PM
|
|
Sponsor:
|
Section on Statistics in Epidemiology
|
|
Abstract #319140
|
|
|
Title:
|
Hypothesis Testing and Sample Size Considerations for the Test-Negative Design
|
|
Author(s):
|
Yanan Huo* and Natalie Dean
|
|
Companies:
|
University of Florida, Department of Biostatistics and University of Florida
|
|
Keywords:
|
test-negative design;
case-control study;
vaccines;
sample size;
score test;
continuity correction
|
|
Abstract:
|
The test-negative design (TND) is a recent epidemiological study design used to evaluate vaccine effectiveness (VE). It enrolls individuals receiving diagnostic testing for a target disease as routine care. The VE is estimated as one minus the odds ratio of testing positive versus negative comparing vaccinated and unvaccinated. No guidance is available on power and sample size considerations for TND studies. TND is related to the case-control but unique in that the ratio of test-positive cases to test-negative controls is not pre-specified. We establish simulations to examine the standard Wald test, the continuity-corrected Wald test and the score test, with associated sample size calculations for case-control and TND studies. The Wald test performs poorly for high VE as the number of vaccinated test-positives can be sparse. Continuity corrections stabilize the variance but induce bias. We observe superior performance with the score test as the variance is pooled under the null hypothesis of no group differences. We recommend the score-based approach to design and analyze the TND. A TND score sample size adjusting for extra variability on the ratio of controls to cases is proposed.
|
Authors who are presenting talks have a * after their name.
