Governing Equations

Two-Period Two-Sequence Crossover (2×2)

TOST Procedure (two one-sided t-tests):
  H01: μT - μR ≤ -θ  →  reject if t1 > t(α,df)
  H02: μT - μR ≥ +θ  →  reject if t2 < -t(α,df)

90% Confidence Interval:
  (x̄T - x̄R) ± t(0.05,df) · SE
  where SE = sqrt(2·MSE/n)
  MSE = intra-subject variance

Power:
  n = 2 · (Zα + Zβ)² · CV² / (ln(θ) - |ln(μT/μR)|)²

BE accepted if 90% CI within 0.80–1.25