From: "michel.tod" <>
Subject: NONPAR
Date: Tue, 27 Mar 2001 11:53:53 +0200

Dear NM users,

In the very good paper entitled "Use of a PK/PD model to design an optimal dose input profile" by Kyungsoo Park et al. (JPB 1998, 26, 471-492), there is a rather enigmatic statement (p 480): " ... a discrete nonparametric estimate of the density of eta was used... The option NONPAR in NONMEM provides estimates of the probability mass for each eta(j) ..."

It seems that the option NONPAR is not documented in the users guide (at least I did not found it). Could anyone explain how to use this option and/or provide an example of a control file?

Michel Tod.
Dpt of Pharmacotoxicology
Avicenne Hospital
Bobigny, France


From: Mats Karlsson <>
Subject: Re: NONPAR
Date: Tue, 27 Mar 2001 12:13:53 +0200

Dear Michael,

The option NONPAR is not available in the released version of NONMEM.. I've used it (JPB 26: 207-246, 1998) and found it useful. I recently asked Stuart Beal, whether it would be included in the next release of NONMEM, but he said he was unsure about that. Essentially, what it does is to use Alain Mallet's method implemented in NPML for the second step, but use the distribution of posthoc parameter estimates as "points of support" for which the nonparametric density is estimated (this I think is the best I can do in describing it).

Best regards,
Mats Karlsson, PhD
Professor of Pharmacometrics
Div. of Pharmacokinetics and Drug Therapy
Dept. of Pharmaceutical Biosciences
Faculty of Pharmacy
Uppsala University
Box 591
SE-751 24 Uppsala
phone +46 18 471 4105
fax +46 18 471 4003