Scaling of Cmin [Regulatives / Guidelines]

posted by VStus – Poland, 2017-02-03 14:04 (2638 d 02:53 ago) – Posting: # 17005
Views: 10,790

(edited by VStus on 2017-02-03 15:50)

Dear Helmut,

Have you solved this puzzle?

I must confess that Liu's article is too hard for me. I understand that the idea of scaling is based on assessment of within-subject variability after the administration of only Reference formulation.

But article is targeting repeated crossover studies (RRTT|TTRR) for assessment of population/individual bioequivalence. Due to this, a more complicated statistical model was proposed and justified. But we are not interested in population bioequivalence now.

Would it be correct to analyze only a reference subset of data to estimate the variability of reference PK parameter only? Let's assume I'm interested only in possibility of widening of acceptance limits for Ctau_ss, do I need test formulation data at all?

We will not have different formulations in resulting subset, just factors of period + sequence + subject + repetition.

I have reduced Liu's dataset to have:
subject - 20 levels (1-20)
sequence - 2 levels (RT and TR)
period - 2 levels (1 or 2)
repetition - 2 levels (1 or 2)
AUC for Reference formulation.

Which gives me:
anova(lm(log(AUC) ~ prd + seq + rep + subj:seq, data=data, na.action=na.exclude))
## Residuals 19 0.60602 0.031896
100*sqrt(exp(0.031896)-1)
## [1] 18.00281


Residual MS is higher than reported by Lui for Reference formulation ('MS(R) = 0.03094'), thus it seems that I'm overestimating the variability of the Reference in this simplified model.

Thank you very much in advance!

Regards, VStus

Sorry for confusion: this is not a replicate design

Complete thread:

UA Flag
Activity
 Admin contact
22,993 posts in 4,828 threads, 1,651 registered users;
110 visitors (0 registered, 110 guests [including 6 identified bots]).
Forum time: 17:57 CEST (Europe/Vienna)

Never never never never use Excel.
Not even for calculation of arithmetic means.    Martin Wolfsegger

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