Cost-efficient higher-order crossover designs in comparative bioavailability studies

Abstract

Objective To compare cost efficiency of five commonly used, statistically optimal or near-optimal higher-order crossover designs for comparative bioavailability studies. Methods Monte Carlo simulations were carried out to obtain empirical sample sizes of the five higher-order crossover designs using Schuirmann's Two One-Sided Tests Procedure, under a 90% power and a 5% significance level, based on the equivalence criteria (80%, 125%). The five designs were the 2-period 4-sequence (2x4), the 3-period 2-sequence (3x2), the 3-period 4-sequence (3x4), the 4-period 2-sequence (4x2), and the 4-period 4-sequence (4x4). Costs were then determined by a cost function, which takes into account recruiting and screening cost, cost associated with period, and the overhead incurred for multiple sequences. The costs of the designs were compared under different scenarios. Results No single design uniformly dominates others. The 3x2 design and the 4x4 design, especially the latter, have the best overall performance in terms of cost efficiency. The 4x4 design is often better than the 3x2 design, but the 3x2 design can outperform the 4x4 design under high sequence cost or high period cost increment. The 2x4 design had the worst performance. Conclusions The 3-period 2-sequence design and the 4-period 4-sequence design are recommended for higher-order crossover designs in comparative bioavailability studies. The 2-period 4-sequence design is least favorable in terms of cost efficiency. Clinical Pharmacology & Therapeutics (2005) 77, P90–P90; doi: 10.1016/j.clpt.2004.12.235