Back to the Top
Dear Colleagues,
I would like to know if someone uses the new (Windows) version of NPEM,
and especially if this version uses a different algorithm than the
current (DOS) version.
I am using the DOS version for a pharmacokinetic study of isepamicin and
I would like to be sure that the results will not be obsolete!
Thanks in advance.
--
Jean Debord
Laboratoire de Pharmacologie, Faculte de Medecine
2 Rue du Docteur Marcland, 87025 Limoges (France)
Back to the Top
Dear Dr. Debord:
There is no windows version of NPEM. We ARE working on one, but the
basic
IT2B and NPEM algorithms will remain the same, and the results will be just
as valid as ever.
What is perhaps new is that we have implemented the NPEM (and soon the
IT2B) algorithm on the Cray T3E at the San Diego Supercomputer Center
(SDSC). This large parallel processing machine permits population analysis
of ANY type of large ("big") model, linear or nonlinear, with effects,
saturable behavior, multiple responses,etc. - whatever model you want to
make.
You type in the differential equations, the output equations, etc.,
on the
PC. Two files are made 1) the model, and 2) the patient data files and the
instructions. One then gets into the Cray with SSH (a secure version of
Telnet), gets the 2 files from the email node for your PC, does the
analysis, and then you ftp the result files to your PC and look at them.
Thay appear just like the ones on your PC. Instead of being limited to
80000 grid points, there are now up to about 20 million grid points. This
is ready now for beta use for those who want it. There are basically 6
programs now, and we can send the software to those who would like it.
There are now:
1. Regular IT2B on the PC, as always.
2. Regular IT2B on the SDSC Cray Resource, for large data sets.
3. Big IT2B on the SDSC Cray (coming soon).
4. Regular NPEM on the PC.
5. Regular NPEM on the SDSC Cray Resopurce, for large data sets.
6. Big NPEM on the SDSC Cray Resource. This is available now.
We are also working on a new windows version of the USC*PACK programs,
which should be available (we hope) by the spring. It will use the new
"multiple model" method of dosage design, which uses nonparametric PK
models. Use of the nonparametric NPEM models permits evaluation of the
precision with which any dosage regimen is likely to achieve a desired
target goal. The regimen is then designed specifically to minimize the
expected weighted squared error with which the target goal will be
achieved. Thus the combination of nonparametric models and MM dosage design
permits maximally precise dosage regimens to be developed, for the first time.
Once again, there is no new algorithm, but there is a new resource
available for making large and nonlinear population PK/PD models using the
SDSC Resource. This new resource has been funded by the NCRR. The clinical
MM programs should be available this spring.
Very best regards,
Roger Jelliffe
PharmPK Discussion List Archive Index page
Copyright 1995-2010 David W. A. Bourne (david@boomer.org)