Bioequivalence and Bioavailability Forum

Dear all,

we are collecting data for a publication about the annoying ‘Group-by-Treatment interaction’ (for some thoughts see this article).
Below what we have so far in a meta-analysis of multi-group studies:

To make a long story short: As expected, significant Group-by-Treatment interactions were detected at approximately the level of the test (0.05) and below the upper 95% significance limit of the binomial test. Hence, based on our observations in well-controlled studies likely they are mere ‘statistical artifacts’, i.e., false positives. The Kol­mo­go­rov–Smirnov tests were not significant, accepting the expected standard uniform distribution of \(\small{p(G\times T)}\).
At a recent BE workshop¹ I had a conversation with a prominent member of the FDA. She told that the agency tests the Group-by-Treat­ment interaction at the 0.05 level, thus reducing the false positive rate observed at the previously used 0.1 level. But then she also said

»We often see not only p-values just below 0.05, but also some with 0.0001…«

It’s a common fallacy to regard the p-value as the probability that the null hypothesis is true (or the alternative hypothesis is false).² The outcome of a level \(\small{\alpha}\)-test is dichotomous: Hypotheses are considered true or false, not something that can be represented with a probability.

Any type of comparative BA study (BE, food-effect, DDI, dose-proportionality) is welcome.
Of course,  data will be treated strictly confidential and not published . The preferred data format is CSV (though xls(x), ODS, SAS XPT or Phoenix project files would serve as well).

Columns (any order is fine):

1. Company or individual (text)
   
2. Study code (text)
   
3. Analyte (text) if you don’t want to give this information, use not spec. X, where X is an integer 1 … number of analytes
   
4. Design (2x2x2, 3x6x3, 3x3, 4x4, 2x2x4, 2x2x3, 2x3x3)
   Simple crossover, 6-sequence 3-period Williams’ design, 3-period Latin Squares, 4-sequence 4-period Williams’ design or 4-period Latin Squares, 2-sequence 4-period full replicate, 2-sequence 3-period full replicate, partial replicate; no parallel design
   
5. Drug (integer) 1 … number of analytes
   
6. Subject (integer or text) min(n) … max(n); missings due to dropouts not a problem
   
7. Group or Site (integer) 1 … number of groups / sites
   
8. Sequence (character or integer), e.g., TR, RT or AB, BA or 1, 2 (simple crossover), e.g., TRTR, RTRT or TRT, RTR (full re­pli­cate designs), TRR, RTR, RRR (partial re­pli­cate design), ABC, BCA, CAB (Latin Squares), ABC, ACB, BAC, BCA, CAB, CBA (Williams’ design)
   Essentially any kind of coding is possible, as long as it is unambiguous.
   
9. Treatment (character) mandatory T or R (not A or B)
   
10. Period (integer) 1 … number of periods
    
11. AUC (numeric); for single dose AUC_0–t or AUC_0–72, for multiple dose AUC_0–τ.
    Missing values should be coded with NA (preferred) or Missing.
    
12. C_max (numeric)

Optional columns:

• Interval (integer) days separating groups; only if equal across groups
  
• Sex (character) f or m

If you want to participate, please send data to [email protected].
No cherry-picking, otherwise we will fall into the trap of selection bias and the outcome will be useless. Hence, if you decide to provide data, please do so irrespective of whether you ‘detected’ a significant Group-by-Treatment interaction or not.
We are primarily working on 2×2×2 crossover designs. How­ever, if you have data of replicate designs, fine as well. In Higher-Order crossover designs indicate which of the treatments is the test and the reference.
If possible, give the analyte. Once we have enough data sets, we will perform sub-group analyses.

So far we have only data of one multi-site study. If you could share some data, great.

THX in advance!

---

1. Medicines for Europe. 2^nd Bioequivalence Workshop. Session 2 – ICH M13 – Bioequivalence for IR solid oral dosage forms. Brussels. 26 April 2023.
   
2. Wasserstein RL, Lazar NA. The ASA’s Statement on p-Values: Context, Process, and Purpose. Am Stat. 2016; 70(2): 129–33. doi:10.1080/00031305.2016.1154108.