Bioequivalence and Bioavailability Forum

❝ […] advantages/disadvantages of following 6 sequences (TRR, RTR, RRT, RTT, TRT and TTR) in a RSABE design, instead of 3 sequences as per FDA recommendations.

• Additionally to the estimate of CV_wR you get the one of CV_wT (good to know esp. if your study is a pilot and you want to design a pivotal study where CV_wR ≠ CV_wT). In the partial replicate you have to assume that CV_wR = CV_wT.
  
• If s_wR <0.294 (only ABE acceptable) no convergence issues in the mixed-effects model with the FDA’s covariance-specification.

• Sample size is a multiple of six (instead of three in the partial replicate and two in the three-period full replicate).
  
• Design currently not implemented in PowerTOST’s function sampleN.RSABE().