Bioequivalence and Bioavailability Forum

Hi John,

❝ If you want to compute the individual subject T/R Ratios from a 1 Test 2 x Reference (partial rep) study, which is the correct way IYHO?

❝

❝ 1) Test/(Average of 2x Ref treatments)

❝ 2) Average of [(Test/Ref treatment Replicate 1) and (Test/Ref Treatment Replicate 2)]

❝

❝ Yes obviously the ratios are different...

Yep. See FDA’s famous partial rep. code:
data scavbe;
  merge test ref1 ref2;
  by seq subj;
  ilat=latt-(0.5*(lat1r+lat2r));
  dlat=lat1r-lat2r;
run;

Hey, your #1 is close! Since we are in log-scale (ilat are the differences of logs):
\(T/R=e^{ilat}\)
Or if you prefer untransformed data – the ratio of T and the geometric (!) mean of Rs:
\(T/R=T/\sqrt{\bar{R}_1\times \bar{R}_2}\)