Mats,
When I referred to a change of 50 being needed to detect something of
practical importance I was not saying that was of clinical relevance.
That cannot be judged from the OFV alone. But small OFV changes are
rarely if ever indicators of something that is clinically relevant.
I expect y
Leonid,
I did not say NONMEM stops at random. Whether or not the stopping point
is associated with convergence or a successful covariance step appears
to be at random. The parameter values at the stopping point will
typically be negligibly different. Thus the stopping point is not at
random.
Nick,
Concerning "random stops at arbitrary point with arbitrary error" I was
referring to your statement: "NONMEM VI will fail to converge or not
complete the covariance step more or less at random"
For OFV, you did not tell the entire story. If you would look only on
OF, you would go for th
Nick,
I too would use OFV as the most important goodness-of-fit diagnostic when
comparing models, especially when deeming something to be redundant. If
adding a component doesn't reduce OFV, I see no reason to include it (I
think we're agreeing on something!). However, you write
" Small (5-10) c
Leonid,
I do not experience "random stops at arbitrary point with arbitrary
error" so I don't understand what your problem is.
The objective function is the primary metric of goodness of fit. I agree
it is possible to get drops in objective function that are associated
with unreasonable para
Nick,
I think it is dangerous to rely heavily on the objective function (let
alone on ONLY objective function) in the model development process. I am
very surprised that you use it as the main diagnostic. If you think that
nonmem randomly stops at arbitrary point with arbitrary error, how can
Mats, Leonid,
Thanks for your definitions. I think I prefer that provided by Mats but
he doesn't say what his test for goodness-of-fit might be.
Leonid already assumes that convergence/covariance are diagnostic so it
doesnt help at all with an independent definition of
overparameterization.
Joachim,
Thanks for your proposal to make our lives as PK modellers even more
challenging :-)
I have indeed thought of the idea of just trying to get the raw
measurement data e.g. absorbance, and including a 'standard curve' model
in addition to the PK model. But there are practical problems
Mark et al.
yes I am thinking of instrument output being fed directly into the NONMEM data
file. That output could be radioactivity count, absorption intensity, some MS
signal... any thing numeric.
Joachim
--
AstraZenec
Joachim et al. Moreover, assays are typically done in duplicate or triplicate, with only the mean reported. We had many discussions at GSK with the chemists about getting "all" the data (BQL, all replicates) promising everything imaginable (first, the data would never end up in a regulatory docu
Dear all,
I am thinking back to early days of pharmacodynamics. The clinical people would
report to us "early" modellers response data in the form of percent change from
baseline. Very soon we asked for the raw data rejecting their "model" of
generating response data as being subjective and bia
Hi Nick,
I am not sure how you build the models but I am using convergence,
relative standard errors, correlation matrix of parameter estimates
(reported by the covariance step), and correlation of random effects
quite extensively when I decide whether I need extra compartments, extra
random
Hi, my approach has been to use both LTBS and un-transformed data and then
see which one characterizes the data better. Then change initial estimates
and see how the model predicts.
My previous experience was when using untransformed with INTER the model was
not able to always converge specially wh
Sense and nonsense.
Analysts are constrained certainly but good analytical science is good
analytical science. During the method development stage several parameters
are examined.
One parameter for example is recovery of analyte from a set of spiked
individual (matrices). Another exercise
Nick,
Pls see below.
Mats Karlsson, PhD
Professor of Pharmacometrics
Dept of Pharmaceutical Biosciences
Uppsala University
Box 591
751 24 Uppsala Sweden
phone: +46 18 4714105
fax: +46 18 471 4003
-Original Message-
From: owner-nmus...@globomaxnm.com [mailto:owner-nmus...@globomaxnm.com]
Steve,
And we need to design our studies to what is reported, which means taking
into account that for some observations, the only information you will get
is that they are below a limit.
Best regards,
Mats
Mats Karlsson, PhD
Professor of Pharmacometrics
Dept of Pharmaceutical Biosciences
Uppsal
Nick,
So the advice regarding handling of error models actually concerns a study
that is to be done in your lab where negative concentrations are to be
reported. Not the million of studies with LOQ limits. Could have been useful
to know.
Best regards,
Mats
Mats Karlsson, PhD
Professor of Pharmac
Mats
I agree with Nick. Negative "observed" concns do occur for assays, even in my
limited time working with HPLC I have seen them, however due to LOD/LOQ they
are never really looked for and certainly never reported...
Steve
> -Original Message-
> From: owner-nmus...@globomaxnm.com [
18 matches
Mail list logo