Exact Properties of Some Heteroscedastic TOST Alternatives for Bioequivalence

The two one-sided tests (TOST) procedure is the common method for assessing bioequivalence between different drugs or formulations of same drug. Several solutions to Behrens-Fisher problem have been extended establish statistical equivalence under variance heterogeneity. Unlike other existing studies that rely exclusively on simulation results, analytical formulas power and Type I error rate are derived. Exact numerical investigations were conducted evaluate behavior Welch–Satterthwaite, Cochran–Cox, Banerjee–McCullough TOST methods. results show rates Cochran–Cox never exceed nominal significance level. technique actually more conservative than procedure. Welch–Satterthwaite approach maintains closest preassigned level almost equal target probability. exact levels completely nearly identical, always less method. On basis excellent control greater performance, recommended over heteroscedastic These findings provide an update previous endorse somehow inferior