‘dlat’ for 3-period full replicate? [RSABE / ABEL]

posted by mittyri – Russia, 2023-07-07 18:04 (349 d 03:02 ago) – Posting: # 23661
Views: 2,145

Hi Helmut!

❝ I have a problem (both in Phoenix/WinNonlin, R, and – likely – in SAS as well). To calculate swR the FDA wants a linear model (in SAS-lingo proc glm) of repeated reference data with subject as a factor. However, in a 3-period full replicate they are all in one sequence and the model collapses. Now what?

you discovered another point for a lecture ;-)
So the problem is here
proc glm data=scavbe;
  class seq;
  model dlat=seq;
  ods output overallanova=dglm1;
  ods output NObs=dglm3;
  title1 'scaled average BE';
run;


I would look at the original formula

\(s_\text{wR}^2=\Large\frac{\sum\limits_{i=1}^{m}\sum_\limits{j=1}^{n_i}\left(D_{ij}-\overline{D}_{i\,\cdot}\right)^2}{2\left(n-m\right)}\)

So it is impossible to get D for `TRT` sequence. Could we just omit one sequence here?

Kind regards,
Mittyri

Complete thread:

UA Flag
Activity
 Admin contact
23,059 posts in 4,841 threads, 1,646 registered users;
31 visitors (0 registered, 31 guests [including 7 identified bots]).
Forum time: 21:06 CEST (Europe/Vienna)

Friends don’t let friends use Excel for statistics!    Jonathan D. Cryer

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