Kudos to Jeroen, who solved this for me, needed the NOINTER option on $EST,
then you can use SAEM, which gave a reasonable answer, unlike FOCE.
Mark
Mark Sale M.D.
Vice President, Modeling and Simulation
Nuventra, Inc. ™
2525 Meridian Parkway, Suite 280
Research Triangle Park, NC 27713
Office
Hi Mark,
My first suggestion is you can start from simpler mixture model (e.g. 2
distributions) and only focus on those have AEs. 68% patients without AE is
a big disturbance to intercept. Only negative infinity of intercept in
logistic model can give a probability=0. Secondly, you can try to use
Hi Mark,
Is it indeed a logistical model or is it an ordered categorical? I
assume you refer to the latter. Not sure how you get your second
category otherwise.
Anyway, to me it reads like you are trying to have the mixture model
describe exactly what the omega is trying to describe. Perhaps
Matts,
Thanks for your insights. But, the issue isn't the post hoc values. With
the mixture model the OMEGA on the intercept is huge (680), and the entire
population is in the low intercept value group (Intercept = -11). Then to
accommodate the patients with frequent AEs, it assigns a (p
Hi Mark,
The pattern you see in the posthocs could possibly be a shrinkage
phenomenon. I.e. patients with AE most of the time will have the same ETA,
while patients with no AE will have the same ETA and there will be a third
group in between. If shrinkage is causing this, you should not expect any
Bob,
The error message I'm getting with any method other than FOCE is that you
can't use INTER, which makes sense since there is no EPS. INTER apparently is
implied with any of the NP methods.
Thanks for the offer to look at the data, but, of course, this is proprietary
data.
Mark
Mark
OK - but there's no inherent reason why LIKE could not work with NP - the
underlying theoretical NP algorithm is totally agnostic to where the
likelihoods come from or what type of observation is being used.Certainly
in Phoenix NLME this works. Not sure about USC*PACK. If you can get NON
Bob,
That certainly makes sense, but that options seems to not be available in
NONMEM, using LIKE seems to require using FOCE LAPLACE
LIKELIHOOD
This is designed mainly, but not exclusively, for use with non-
continuous observed responses ("odd-type data"). Indicates that
This sounds like a good case for a nonparametric method - if you use the one
in NONMEM, you might try
expanding Omega to counter shrinkage. The versions in USC*PACK and PHOENIX
NLME optimize over
both support point positions and probabilities, so this is not necessary with
those methods.
