iteratively adjusted α - FDA [RSABE / ABEL]
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
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.
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.
—
Regards,
Detlew
Regards,
Detlew
Complete thread:
- EMA: adjusted α Helmut 2015-11-18 00:27 [RSABE / ABEL]
- EMA: adjusted α for scABEL (for no-nerds) d_labes 2015-11-18 08:25
- In praise of PowerTOST Helmut 2015-11-18 16:01
- iteratively adjusted α Helmut 2015-11-28 22:35
- iteratively adjusted α - FDAd_labes 2015-12-08 09:35
- EMA: adjusted α for scABEL (for no-nerds) d_labes 2015-11-18 08:25