Bioequivalence and Bioavailability Forum

Hi Helmut,

❝ For Example 2 I got:

Where is example 2??

❝ When we specified all effects fixed (EMA?), Phoenix/WinNonlin (tested 6.2.51, 6.3beta, 6.3) spits out:

❝ ERROR 11050: Parsing error: The containing term must be in the model.

This should absolutely work, I think. Either it is the syntax that was wrong or PHX/WNL is flawed?

❝ OK, since in PHX/WNL the BE-module sits on top of the LME-engine I set up the model directly in LME. Could reproduce the mixed model:

(and blah)

❝ … giving back-transformed 101.4544 (88.4452, 116.3770), identical to BE-wizard’s results.

❝ I get the same values in the all-fixed model – but could not request LSMeans any more; otherwise I got the same error like in the BE-wizard. Though we don’t need these values in the BE assessment I would be interested how other software deals with it (SAS, R?).

You don't need a mixed model when you have only one sigma in play (this isn't a replicated study I presume?). In the simple non-replicated case fitting a mixed model with subject as random should give the same result as the normal linear model based on all fixed. If there are subtle differences at some decimal then this sounds most likely just like either a rounding phenomenon, or a convergence issue where somehow the optimiser in this case is capable of finding smaller sigma when the mixed model is applied.
The normal linear model can be solved exactly with e.g. R at least to the extent we can rely on its numerical stability for inversion of X^tX; I would take a look at sigma that way and see what the truth might look like.