On population and individual bioequivalence

In a traditional assessment of the bioequivalence of two formulations of a drug one compares the average bioavailability from the two formulations. Anderson and Hauck argued that in some situations it is not sufficient to demonstrate average bioequivalence, and they proposed a method for the assessment of what they called individual bioequivalence, which essentially is the comparison of the individual responses to the two drug formulations within subjects. In this paper we propose a unified strategy for the assessment of bioequivalence that encompasses new approaches to the assessment of both population bioequivalence, which is the comparison of the marginal or population distributions of bioavailabilities, and individual bioequivalence, which is the comparison of the conditional or within-subject distributions of bioavailabilities. The general idea is to use a comparison of the reference formulation to itself as the basis for the comparison of the test with the reference formulation. The new approaches overcome the main weakness of the current methods for the assessment of bioequivalence by considering the variability of bioavailabilities in addition to their means. The current methods for the assessment of bioequivalence, namely the conventional assessment of average bioequivalence and the proposal by Anderson and Hauck for the assessment of individual bioequivalence, emerge as special cases. One can evaluate the new bioequivalence criteria statistically by use of bootstrap confidence intervals.