adjusted alpha for RSABE [RSABE / ABEL]

posted by d_labes  – Berlin, Germany, 2017-03-29 18:23 (2556 d 02:50 ago) – Posting: # 17208
Views: 3,182

Dear Yura

❝ So, is it right: for ABEL I should adjust Type I Error (alpha)

To protect against an alpha-inflation which otherwise may occure, it's exactly what we recommend in our paper1.

❝ or in other words to decrease to 0.025?

That's not what we recommend in our paper. We recommend to adjust alpha based on the outcome of the study. For details see our paper, especially the examples.

❝ What I can do if I’ll use RSABE

Here I can't help since nobody has thoroughly investigated alpha inflation of the FDA recommended method for scaled ABE. This has to do with the controversies regarding where to locate the region of the type I error. See the paper of Davit et al.2

❝ (because here alpha is absent in formula)?

That's not correct.
Alpha is contained in the request to calculate 95% upper confidence limit of the linearized RSABE criterion. If you adjust alpha you have to calculate 1-alphaadj upper confidence limit. Looking at the SAS code change all alpha=0.1 to 2*alphaadj you will use.
But as already said: Nobody knows or can calculate the value of alphaadj to use.


  1. Labes D, Schütz H. Inflation of Type I Error in the Evaluation of Scaled Average Bioequivalence, and a Method for its Control. Pharm Res. 2016;33(11):2805–14. doi:10.1007/s11095-016-2006-1. full-text view-only
  2. 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
.

Regards,

Detlew

Complete thread:

UA Flag
Activity
 Admin contact
22,957 posts in 4,819 threads, 1,636 registered users;
89 visitors (0 registered, 89 guests [including 6 identified bots]).
Forum time: 20:14 CET (Europe/Vienna)

Nothing shows a lack of mathematical education more
than an overly precise calculation.    Carl Friedrich Gauß

The Bioequivalence and Bioavailability Forum is hosted by
BEBAC Ing. Helmut Schütz
HTML5