As pointed out by Anderson and Hauck and others, average bioequivalence doe
s not guarantee that a patient, being on a drug, may switch to a different
formulation. The consideration of individual bioequivalence was therefore s
uggested. We investigate the relationship between these concepts and show h
ow the standard test for average bioequivalence is related to the assessmen
t of individual bioequivalence. This leads to a guide of when this test bec
omes inappropriate. Copyright (C) 2000 John Wiley & Sons, Ltd.