Bioequivalence and Bioavailability Forum

Hi Joy,

❝ Would you mind to share the guideline ?

It is not a particular good idea to rely on guidelines – especially when it comes to statistics.

❝ Because I just have Canada Guideline […]

❝ SEDifference = (2MSResidual/n)^0.5

Not the sole error in Health Canada’s guidance; the formula – as Mahmoud wrote – is only correct in the balanced case. Always use $$SE=\sqrt{\frac{MSE}{2}\left(\frac{1}{n_\textrm{TR}}+\frac{1}{n_\textrm{RT}}\right)}\tag{1}$$ HC’s example dataset is balanced only by chance (one dropout in each sequence) and imbalanced studies are not uncommon in BE. BTW, it took HC 21 (‼) years to give the correct formula for the intra-subject CV:
$$CV_\textrm{intra}=\sqrt{\exp(MSE)-1}\tag{2}$$ In HC’s first (1991) guidance – and still in the 2009 (!) draft – the wrong $$CV_\textrm{intra}=\sqrt{MSE}\tag{3}$$ was given. For AUC_T of the example data set one gets 0.275 by \(\small{(2)}\) and 0.270 by \(\small{(3)}\). So much about the “wisdom” of guidelines!

Get a statistical textbook relevant for BE (e.g., Hauschke et al.¹ or Chow & Liu²).

---

1. Hauschke D, Steinijans V, Pigeot I. Bioequivalence Studies in Drug Development. Chichester: John Wiley; 2007. doi:10.1002/9780470094778.
   
2. Chow S-C Liu J-p. Design and Analysis of Bioavailability and Bioequivalence Studies.
   Boca Raton: Chapman & Hall/CRC Press; 3^rd ed. 2009. doi:10.1201/9781420011678.