Bioequivalence and Bioavailability Forum

Hi ElMaestro,

❝ this looks to me like an ugly work-around rather than a solution.

Right. Solution was my term. Linda wrote:

Set up the Repeated subtab of the Variance Structure tab as:
Repeated Specification: <blank> (see below)
Variance Blocking Variables (Subject): Subject
Group: Formulation
Type: Variance Components
However, in the current version of Phoenix, if the Repeated Specification is blank, Phoenix 1.3 will ignore the rest of this tab, so a workaround must be used. […] After completion of these steps, execution of Bioequivalence will produce the same values as SAS.

❝ […] I'd like to understand why that type of coding fixes it.

So do I.

❝ I believe WNL will fit this model via REML, …

REML is the only language WinNonlin speaks.

❝ … and on basis of the syntax I imagine this means a covariance matrix with two squared sigmas on the diagonal and zero elsewhere.

You are the Matrazenkaiser. Reasonable assumption.

❝ But to get through to the 'correct' CI there must be a step for derivation of the denominator DF which happens to coincide with the DF you get via the Welch correction when just doing ordinary unpaired t-test with unequal variances. Perhaps that's similar to the SATT option in SAS (which I also don't use). Could be checked?

Sure. Satterthwaite is WinNonlin’s standard setup. I’m not in the office so can’t check what will happen with the other option ‘Residual’.

❝ So, what is really going on within WNL when that weird coding principle is applied - Why does it give the seemingly right result? Anyone knows?

Maybe. Too many cooks spoil the broth. Pharsight’s statisticians ≠ coders.