|
Activity Number:
|
75
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Sunday, August 6, 2006 : 4:00 PM to 5:50 PM
|
|
Sponsor:
|
General Methodology
|
| Abstract - #306858 |
|
Title:
|
Comparing the Performance of Three Asymptotic Methods in Estimating the Sample Size for a Therapeutic Equivalence Study Based on Difference of Proportions
|
|
Author(s):
|
Xiaoning Li*+ and Sara K. Vesely
|
|
Companies:
|
The University of Oklahoma Health Sciences Center and The University of Oklahoma Health Sciences Center
|
|
Address:
|
402 Wadsack Drive, Apt. A, Norman, OK, 73072,
|
|
Keywords:
|
sample size ; therapeutic equivalence study ; asymptotic methods ; difference of proportions
|
|
Abstract:
|
This study compared three asymptotic methods (Blackwelder, Dunnet and Gent, Farrington and Manning) with respect to the accuracy of sample size estimation for a therapeutic equivalence study with low dichotomous response rates (< 10%) and the impacts of misestimation of the population response rate on the empirical power (EP). EP was calculated based on the number of times the null hypothesis was rejected out of 1000 simulated sample data sets using three methods. The EP was compared with its nominal value for each method and scenario and among three methods. In most of the scenarios, the EPs were greater than their corresponding nominal powers. The effect of the misestimation on the EP decreases as delta, pis and pie increases. For pis or pie>=0.07 and delta>=3/2pis, at least one of the three tests still maintained EPs above 70% after considering all the possibilities of misestimation.
|
