Sample-Size Determination for Equivalence Trial of Continuous Responses

• Activity Number: 179
• Type: Contributed
• Date/Time: Monday, August 5, 2013 : 10:30 AM to 12:20 PM
• Sponsor: Biopharmaceutical Section
• Author(s): Yu-Wei Chang*+ and Xiaoyu Dong and Yi Tsong
• Companies: Temple University and FDA and FDA
• Keywords: Equivalence Test ; Sample Size ; Three-arm Clinical Trial

Abstract

In a three-arm clinical trial, an equivalence assessment may be carried out by demonstrating superiority of test and reference drugs to placebo, as well as having the equivalent mean effects of test and reference drugs. One may also assess the equivalence by comparing the ratio of mean difference between the test drug and the placebo to mean difference between the reference drug and the placebo under the assumption that the mean difference between reference and placebo is positive. This equivalence problem can be performed with the hypothesis testing approach and the sample size can be determined by solving the power function based on either exact t-test or approximate normal test. The sample size will become a function of the given type I error rate, power, the expected ratio value in the alternative space, the variance, the upper and lower equivalence margins, and the unknown but estimable mean difference between the reference drug and the placebo. With historical reference-placebo trial, we can define the mean range for the difference while calculating the sample size. In this paper, We will show that the sample sizes for exact t and approximate normal tests are very similar.