BIOEQUIVALENCE: COMPARISON OF AVERAGE VALUES WITH GENERALIZED p -VALUES

The aim of bioequivalence studies is the evaluation of bioequivalence of pharmaceutical products. The products are usually two and basic pharmacokinetic parameters such as AUC, C max and t max are used. These studies aim at investigating the “closeness” of the distributions of the pharmacokinetic parameters (responses) for the two products, which is made mainly by comparing the average values of these responses under normality or lognormality. The existing methods of comparing average bioequivalence are based on the assumption of equal variances for the pharmacokinetic parameters. In this paper, we propose the use of generalized p-values, suggested by Tsui and Weerahandi in 1989, for testing average bioequivalence in the case of unequal variances. We present exact small sample size tests of average bioequivalence for parallel study designs and demonstrate how to compute their p-values when the pharmacokinetic responses follow normal or lognormal distribution with unequal variances.