Abstract:
|
One critical step in bioequivalence studies is to establish the equivalence margin, which defines a range of values for which two methods or therapies are close enough to be considered equivalent. In practice, it is often necessary to initially establish the equivalence margin statistically and then further update it based on scientific justifications and regulatory considerations. In the framework of testing the equivalence between means of Normal Distributions, we propose an approach to the determination of the equivalence margin from an empirical Bayesian perspective and also propose a corresponding equivalence test. We derive our method in a one-parameter model where mean is the only unknown parameter and also in a more realistic two-parameter model where both mean and variance are considered unknown. In both of these models, we first derive a theoretical equivalence margin with certain probability properties, and then provide a method to estimate the theoretical margin. Simulation studies were performed to evaluate the proposed method and a case study will be presented to illustrate the method.
|
ASA Meetings Department
732 North Washington Street, Alexandria, VA 22314
(703) 684-1221 • meetings@amstat.org
Copyright © American Statistical Association.