|
Activity Number:
|
472
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Wednesday, August 9, 2006 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
Biopharmaceutical Section
|
| Abstract - #306453 |
|
Title:
|
An Improved Exact Method for the Estimation and Testing of a Rate Ratio
|
|
Author(s):
|
William W. B. Wang*+ and Ivan Chan
|
|
Companies:
|
Merck Research Laboratories and Merck & Co., Inc.
|
|
Address:
|
UN-A102, Blue Bell, PA, 19422,
|
|
Keywords:
|
exact method ; Poisson distribution ; rate ratio
|
|
Abstract:
|
In controlled clinical trials evaluating treatment or prevention of diseases with low incidence, drug or vaccine effects are often characterized by the ratio of two independent Poisson rates. Many methods have been proposed for the estimation and the testing of this ratio. These include the asymptotic methods by Miettinen and Nurminen (Stat. Med. 1985) and the conditional exact methods by Guess and Thomas (Epidemiol 1990) and by Martin and Austin (Epidemiol 1996). When the study is designed as case driven, the exact methods are widely used. In this presentation, we propose an improved exact method based on a more efficient statistic. Simulation studies are performed to compare the proposed exact method with other asymptotic and exact methods in terms of size, power and coverage probability. We will illustrate these methods using a real dataset from a large-scale vaccine efficacy trial.
|
