Bioequivalence and Bioavailability Forum

Hi SMA,

THX for posting!

An interesting statement in 3.16.2 Statistical Programming and Data Analysis:

1. The sponsor should ensure that appropriate and documented quality control of statistical programming and data analysis is implemented (e.g., for sample size calculations, …).

This would be a tough cookie in connection with 4.5 Validation of Computerised Systems.

Good luck! For parallel groups, 2×2×2 cross-overs, and RSABE/ABEL sample size tables were published. Regrettably some contain typos or the number of simulations for RSABE/ABEL was too small for a stable result (see this article). You are left out in the rain if you want anything not covered (say, power 85% and/or a T/R-ratio of 0.93). AFAIK, nothing is published for Higher-Order crossovers, Balaam’s design, ABE in replicate designs, the FDA’s ref­er­ence-scaling for NTIDs, and Non-Inferiority / Non-Superiority. IMHO, the only solution would be to get yet another software and perform cross-validation.

Of course, requiring software validation in general is not new. Don’t trust the vendor. Validation always lies in the hand of the user.^1–3

---

1. ICH. Statistical Principles for Clinical Trials. E9. 5 February 1998. Section 5.8.
   
2. FDA. Statistical Software Clarifying Statement. May 6, 2015. Download.
   
3. WHO. Guidance for organizations performing in vivo bioequivalence studies. Geneva. May 2016. Technical Report Series No. 996, Annex 9. Section 4.4. Online.