Parallel bears meeting at random in infinity [🇷 for BE/BA]

posted by ElMaestro  – Denmark, 2010-04-22 14:53 (5089 d 22:08 ago) – Posting: # 5184
Views: 55,823

Ahoy d_labes!

❝ If I got it right the code used is (f.i. AUC log-transformed):

lme(lnAUC0t ~ drug, random=~1|subj, data=TotalData, method="REML" )

❝ I am wondering where this code comes from, what this code does an why it works anyhow :ponder:.

❝ IMHO this model, one (fixed) effect for the treatment and one (random) effect for the subjects must be over-specified.


With REML, the likelihood is optimised on the covariance matrix V (=ZGZt+R). With "random=~1|subj" I think we do the most simple specification of V, as we put in sigma on the diagonal, and zeros elsewhere. Thus, subjects introduce a random effect (think sigma on the diagonal), nature of treatment not considered.

❝ We have only one value for a distinct subject treated with Test or Reference and thus we are not able to separate this uniquely into 2 effects, one part for treatment and one attributed to the subject.


In my opinion, and I could be wrong, with this specification of the model we are not trying to do exactly that through the proposed code. We are only saying that each subject is associated with a degree of variability which is described by a common sigma in V, regardless of the treatment given for any data point.

Best regards from the seven seas,
EM.

Correction: I mean sigma2 on the diagonal.

Pass or fail!
ElMaestro

Complete thread:

UA Flag
Activity
 Admin contact
22,957 posts in 4,819 threads, 1,636 registered users;
129 visitors (0 registered, 129 guests [including 15 identified bots]).
Forum time: 12:02 CET (Europe/Vienna)

With four parameters I can fit an elephant,
and with five I can make him wiggle his trunk.    John von Neumann

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