Properties of the estimated variance component for subject-by-formulation interaction in studies of individual bioequivalence

Characteristics of the variance component for the subject-by-formulation in teraction (sigma(D)(2)), estimated in simulated studies of individual bioeq uivalence and in three- and four-period cross-over trials reported by the F DA, were compared. sigma(D)(2) was estimated by (i) restricted maximum like lihood (REML) and (ii) the method of moments (MM). Variation of the varianc e component, estimated by both procedures (s(D)(2)) and for both the simula ted and FDA data, increased with rising intra-individual variation. Consequ ently, a constant level of s(D)(2) (such as 0.0225 suggested by the FDA) ma y not be regarded as a basis for demonstrating substantial interactions. Fe atures of the FDA and simulated parameters were similar. The results sugges ted that the FDA data were compatible with assuming sigma(D) = 0.05 or perh aps 0.00. Therefore, there is no foundation for concerns about public healt h. Both simulations and calculations demonstrated that s(D)(2) estimated by MM was unbiased and its variance was proportional to sigma(WF)(4) when sig ma(D)(2) = 0. Copyright (C) 2000 John Wiley & Sons, Ltd.