Simulation-based methods acceptable for the FDA [Regulatives / Guidelines]

Hi ElMaestro & Khagesh,

» » Potvin method is acceptable or not for FDA,

It is.1,2 All authors of the FDA…

» But the answer to your question is yes, Potvin is acceptable.

Yep.3

» Do a controlled corr. with FDA to make sure you get everything right.

Not sure whether this is required cause you have to submit the protocol to the OGD anyhow.

» If you decide to cap sample size […] the method has new properties in terms of power and type I error and you are expected to be able to present data for that if the agency asks.

Correct when it comes to power, wrong concerning the Type I Error. If you use the same $$\small{\alpha_\textrm{adj}}$$ than in a published method, any futility criterion reduces the chance to proceed to the second stage and hence, the empiric TIE.
On the other hand, that means one may find a suitable $$\small{\alpha_\textrm{adj}}$$ with less adjustment acc. to the given conditions in own simulations. Since one has to simulate the empiric TIE in a reasonable narrow grid4 of n1-CV-combinations, this will be time-consuming and is rarely worth the effort (limited increase in power).

» In general, your power drops a lot if you use such caps. You need to be absolutely aware of it, otherwise you'll be initiating a study that may have a very low chance of success even if power for stage 2 is set at 80% or higher.

Correct again. Last year we submitted a protocol to the FDA, where we provided extensive simulations in an appendix. Was not a problem.

When one has to deal with a 2×2×2 crossover I suggest to opt for Maurer’s method, which is not only exact (control of the Type I Error analytically proven; no simulations required) but also extremely flexible when it comes to futility criteria and even sample size re-estimation based on the T/R-ratio observed in the first stage (i.e., fully adaptive).5 Though exact, simulations are useful to find suitable conditions (futility rules, weights of the stages, selection between standard combination and maximum combination test), which give a reasonably high chance to show BE already in the first stage whilst maintaining the desired overall power.
The method is implemented in the  package Power2Stage.6

1. Davit B, Braddy AC, Conner DP, Yu LX. International Guidelines for Bioequivalence of Systemically Available Orally Administered Generic Drug Products: A Survey of Similarities and Differences. AAPS J. 2013; 15(4): 974–90. doi:10.1208/s12248-013-9499-x. Free Full text.
2. Lee J, Feng K, Xu M,Gong X, Sun W, Kim J, Zhang Z, Wang M, Fang L, Zhao L. Applications of Adaptive Designs in Generic Drug Development. Clin Pharm Ther. 2020; 110(1): 32–5. doi:10.1002/cpt.2050.
3. Tsang YC, Brandt A (moderators). Session III: Scaling Procedure and Adaptive Design(s) in BE Assess­ment of Highly Variable Drugs. EUFEPS/AAPS 2nd International Conference of the Global Bioequivalence Harmonization Initiative. Rockville, MD: 14–16 September 2016.
4. Last year I had to deal with a deficiency letter of the Czech agency SÚKL. Potvin ‘Method B’ passed with flying colors in the first stage. The agency questioned the reliability of the simulations ‘because the grid in the publication’s simulations was too sparse’. Bizarre. Of course, a step size of 2 for n1 and 2% for the CV showed no inflated TIE as well (maximum 0.048856 for n1 12 and CV 24%). Hurray, 1.116 billion simulated studies!
5. Maurer W, Jones B, Chen Y. Controlling the type 1 error rate in two-stage sequential designs when testing for average bioequivalence. Stat Med. 2018; 37(10): 1587–1607. doi:10.1002/sim.7614.
6. Labes D, Lang B, Schütz H. Power2Stage: Power and Sample-Size Distribution of 2-Stage Bioequivalence Studies. R package version 0.5-4. 2021-11-20. CRAN.

Dif-tor heh smusma 🖖
Helmut Schütz

The quality of responses received is directly proportional to the quality of the question asked. 🚮
Science Quotes