Average bioequivalence for two-sequence two-period crossover design with incomplete data

Abstract

A statistical procedure for analyzing a two-sequence two-period crossover design in bioequivalence trials is discussed when some observations at the second period are missing. The maximum likelihood estimators of the parameters in the average bioequivalence model are obtained under missing at random and a modified two one-sided test is proposed. The performances of the discussed test are compared using Monte Carlo simulations.