Bioequivalence and Bioavailability Forum

Dear Helmut, dear all,

one caveat: The results for the FDA highly depend on the assumption that empiric alpha is obtained if one simulates on the border of the 'implemented' acceptance ranges, i.e. 80-125% up to CVwR = 30% and for CVwR>30% at
100*exp((log(1.25)/0.25)*sqrt(log(1 + (CVwR/100)^2))) # CVwR in %
which has a discontinuity at CVwR = 30%.

This would be also my choice, but the paper

Davit et al.
"Implementation of a Reference-Scaled Average Bioequivalence Approach for Highly Variable Generic Drug Products by the US Food and Drug Administration"
AAPS Journal, Vol. 14, No. 4, December 2012

demands us to look at the widened implied limits from s0=0.25 on (CVwR ~ 25.4%), aka "FDA’s desired consumer risk model". Then the alpha inflation at the discontinuity CVwR = 30% vanishes. Its up to you to decide if this is real or some sort of hokus pokus.