An alternative method for assessing average bioequivalence (ABE) using multiple crossover experiment is studied. Under a standard M-multiple sequential 2x2 crossover experiment, the test treatment is considered not equivalent to the reference if the difference of treatment effects is greater than a pre-specified limit in any two consecutive 2x2 crossover experiments (block) of the total M experiments. In this paper, power function for assessing ABE under a M 2x2 crossover this new design is studied by solving related characteristic polynomial. In addition, sample size calculation is also developed. A comparison in sample size requirement between traditional gold method and this new method is presented. It is found that when the 2 x 2 crossover block experiment is conducted 6 times with power less than 97.162%, the new method requires a smaller sample size.
Author(s): Junrong Lin, and Shein-Chung Chow