Activity Number:
|
133
|
Type:
|
Contributed
|
Date/Time:
|
Monday, August 1, 2016 : 8:30 AM to 10:20 AM
|
Sponsor:
|
Biopharmaceutical Section
|
Abstract #319335
|
|
Title:
|
Conditional Power for Testing Equivalence of Two Proportions
|
Author(s):
|
Zejiang Yang* and Kalyan Ghosh
|
Companies:
|
INC Research and INC Research
|
Keywords:
|
Conditional Power ;
Equivalence Testing ;
Proportion
|
Abstract:
|
Calculation of conditional power is of major interest when interim analyses are undertaken in order to facilitate decisions and /or modifications to an ongoing clinical study. The conditional power calculation for equivalence studies based on the ratio of two proportions is a challenge since the sample ratio is not normally distributed. In this talk, we will present an 'exact' method to calculate the conditional power by summing probabilities for all possible outcomes after the interim look. We will also show how simpler calculation via the first-order and second-order Taylor expansion approximations can be applied to estimate the conditional mean and variance to enable the conditional power calculation via asymptotic normality of the log sample ratio. A simulation shows that these approximations provide results consistent with the 'exact' calculation. The same methods can be applied to the equivalence studies where the odds ratio is of interest.
|
Authors who are presenting talks have a * after their name.