Power Analysis and Sample Size Determination for Crossover Trials with Application to Bioequivalence Assessment of...

To develop methods for determining a suitable sample size for bioequivalence assessment of generic topical ophthalmic drugs using crossover design with serial sampling schemes.

The power functions of the Fieller-type confidence interval and the asymptotic confidence interval in crossover designs with serial-sampling data are here derived. Simulation studies were conducted to evaluate the derived power functions.

Simulation studies show that two power functions can provide precise power estimates when normality assumptions are satisfied and yield conservative estimates of power in cases when data are log-normally distributed. The intra-correlation showed a positive correlation with the power of the bioequivalence test. When the expected ratio of the AUCs was less than or equal to 1, the power of the Fieller-type confidence interval was larger than the asymptotic confidence interval. If the expected ratio of the AUCs was larger than 1, the asymptotic confidence interval had greater power. Sample size can be calculated through numerical iteration with the derived power functions.

The Fieller-type power function and the asymptotic power function can be used to determine sample sizes of crossover trials for bioequivalence assessment of topical ophthalmic drugs.