Re: [NMusers] nonmem question regarding condition number

From: Ayyappa Chaturvedula <ayyappach_at_gmail.com>
Date: Thu, 24 Mar 2016 07:40:05 -0500

Dear Laureen,
High condition number represents overparametrization. You must check also th=
e correlation matrix for potential correlations above 0.95 between parameter=
s. If you strongly feel on the biological basis that your drug is a 3 compar=
tment model, then you may fix some of the parameters of 3 compartment model s=
uch as Q1,Q2 based on prior data and estimate other parameters to see if you=
r diagnostics make sense. Parameters with high SEs could be your clues to th=
e problems. In my opinion, you have informative data with 12 samples per su=
bjects(hopefully at the informative time points) to have a complex model. =


Best Regards,
Ayyappa

> On Mar 24, 2016, at 5:10 AM, L.A. ten Berg - Lammers <l.a.tenberg-lammers_at_=
amc.uva.nl> wrote:
>
> Dear nmusers,
> I am currently analyzing PK profiles after a single iv adminstration. Volu=
nteers were sampled frequently: 12 samples per curve.
> For the structural model 2- and a 3-compartment models are compared. The C=
WRES vs time plot of the 2-compartment shows the typical wave, indicative fo=
r a 3-compartment model. Introduction of a third compartment eliminated the w=
ave, OFV decreased with 180 points, precision of the estimates did change no=
tably, but the condition number increased from 998 to 2.0*E16.
> Does anyone know how to interpret this large increase in condition number?=

> Best wishes,
> Laureen ten Berg
>
>
> <image001.gif>
>
>
> Laureen ten Berg-Lammers
> Hospital Pharmacist
>
> Academisch Medisch Centrum
> Universiteit van Amsterdam
>
>
> Meibergdreef 9 | 1105 AZ Amsterdam | Room: M01-224
> Tel: +31(0)20-5662726 | Pager *62726 | l.a.tenberg-lammers_at_uva.amc.nl
>
>
> AMC Disclaimer : https://www.amc.nl/disclaimer
>

Received on Thu Mar 24 2016 - 08:40:05 EDT

This archive was generated by hypermail 2.3.0 : Fri Sep 27 2019 - 16:49:29 EDT