EMA: TSDs [Two-Stage / GS Designs]

posted by Helmut Homepage – Vienna, Austria, 2020-06-23 18:22 (157 d 06:35 ago) – Posting: # 21568
Views: 413

Hi Mikkabel,

» Following the lectures from Helmut and the recommendations from the EMA, we decided to use the method B described by Potvin as the study will be performed in EU country.

Not a good idea (unless you want to perform a parallel design1). See this thread.

» […] the stage I will be submitted following a standard CT application including the number of subjects that will be included in the stage I.

Correct.

» But what about the stage II? Do we have to notify the CA/EC with the number of subjects to be included in the stage II according to the results obtained after stage I?

No. You write in the protocol that you plan a TSD and the sample size will be re-estimated.

» Or can we define in the study protocol a maximum number of subjects that will be allowed to participate in the study (stage I + stage II)?

Depends on the conditions of the method. F.i., Mme Potvin’s methods (of historic interest only) have no maximum sample size. If you introduce an upper limit, no problems with the type I error (cause your chances of showing BE decrease). On the other hand, you might loose power3,4 and this is not what you want for ethical & economical reasons. Simulations are a good idea.
However, I strongly recommend the method by Maurer et al. No problems any more with “we-don’t-like-simulation-based-methods” assessors. In this method you can define anything you want (futility rules, adaption not only based on the CV but also the GMR in the interim, maximum sample size). Implemented in the R-package Power2Stage.

» Furthermore, in this two-stage study, we would like to assess the biovailability of two treatment tests (two different formulations) versus a reference treatment. Could you please advice if it is acceptable to perform a two-stage design study with 3 treatments?

That’s somewhat tricky. The EMA suggests in the BE-GL a 6-sequence William’s design and evaluation as two separate Incomplete Block Designs. See the vignette of PowerTOST. Say you have T1, T2, and R you end up with one analysis T1 vs R and another one T2 vs R. In principle that should be doable with Maurer’s method, but…


  1. Fuglsang A. Sequential Bioequivalence Approaches for Parallel Designs. AAPS J. 2014;16(3):373–8.
    doi:10.1208/s12248-014-9571-1. [image] Open access.
  2. Fuglsang A. Futility Rules in Bioequivalence Trials with Sequential Designs. AAPS J. 2014; 16(1): 79–82. doi:10.1208/s12248-013-9540-0. [image] Open access.
  3. Schütz H. Two-stage designs in bioequivalence trials. Eur J Clin Pharmacol. 2015; 71(3): 271–81. doi:10.1007/s00228-015-1806-2.

Dif-tor heh smusma 🖖
Helmut Schütz
[image]

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

Complete thread:

Activity
 Admin contact
21,214 posts in 4,427 threads, 1,482 registered users;
online 9 (1 registered, 8 guests [including 6 identified bots]).
Forum time: Friday 23:58 CET (Europe/Vienna)

Biostatistician. One who has neither the intellect for mathematics
nor the commitment for medicine but likes to dabble in both.    Stephen Senn

The Bioequivalence and Bioavailability Forum is hosted by
BEBAC Ing. Helmut Schütz
HTML5