Implementation of interindividual variability in residual variance

From: "Piotrovskij, Vladimir [PRDBE]" - VPIOTROV@PRDBE.jnj.com Subject: [NMusers] Implementation of interindividual variability in residual variance Date: 2/17/2004 9:37 AM Dear NONMEM users, I am trying to implement an interidividual variability in the residual variance using an additional random effect: $ERR Y = F + EXP(ETA(.))*EPS(1) It turned out the corresponding OMEGA was not iterated, and the final estimate did not differ from the initial value. Below is an example control stream and the output illustrating the problem (Note, the actual model I work with is more complicated). Thanks in advance. Best regards, Vladimir ----------------------------------------------------------------- Vladimir Piotrovsky, Ph.D. Research Fellow, Advanced PK-PD Modeling & Simulation Global Clinical Pharmacokinetics and Clinical Pharmacology (ext. 5463/151) Johnson & Johnson Pharmaceutical Research & Development

From: Leonid Gibiansky - lgibiansky@emmes.com Subject: Re: [NMusers] Implementation of interindividual variability in residual variance Date: 2/17/2004 10:21 AM Vladimir, I has to be something with your data: I ran your exact script (copy-pasted to my test script except two lines: $INPUT C ID TIME AMT WT APGR DV EVID MDV $DATA 002.CSV IGNORE=C that replaced your INPUT and DATA lines) with the test data and it is working fine Leonid

From: "Piotrovskij, Vladimir [PRDBE]" - VPIOTROV@PRDBE.jnj.com Subject: Re: [NMusers] Implementation of interindividual variability in residual variance Date: 2/18/2004 8:33 AM Leonid, Thanks for testing. I think it is not related to the data set. Most probably this is due to the compiler. Somethimes I observe strange behaviour of NONMEM in combination with Visual Fortran. Best regards, Vladimir

From: "Bhattaram, Atul" - BhattaramA@cder.fda.gov Subject: Re: [NMusers] Implementation of interindividual variability in residual variance Date: 2/18/2004 9:04 AM Hello All Could the nmusers share their experience in implementing the interindividual variability in residual variance? It would be interesting to know why it was implemented. Venkatesh Atul Bhattaram CDER, FDA.

From: Leonid Gibiansky - lgibiansky@emmes.com Subject: Re: [NMusers] Implementation of interindividual variability in residual variance Date: 2/18/2004 9:05 AM actually, I have visual fortran too. If you like, you can send me the data and I can try it here. Leonid

From: "Bachman, William (MYD)" - bachmanw@iconus.com Subject: Re: [NMusers] Implementation of interindividual variability in residual variance Date: 2/18/2004 9:08 AM Is the behavior any different when tried with another compiler?

From: "Piotrovskij, Vladimir [PRDBE]" - VPIOTROV@PRDBE.jnj.com Subject: Re: [NMusers] Implementation of interindividual variability in residual variance Date: 2/18/2004 9:27 AM Atul, This particular way of implementation of interindividual variability in residual variance was suggested by Stu Beal time ago (e.g., Intermediate-level NONMEM Workshop, Uppsala, 1999, lecture 4 "Statistical Modeling Consideration"). I did try it shortly after this workshop, and at that time it worked. Since then I got new computers, new OS. Now it does not work. Best regards, Vladimir

From: Mats Karlsson - mats.karlsson@farmbio.uu.se Subject: Re: [NMusers] Implementation of interindividual variability in residual variance Date: 2/18/2004 9:39 AM Hi Atul, The traditional assumption in NONMEM analyses has been that all subjects have the same residual error magnitude. This probably stems from early times when it was believed that residual variability was similar to assay variability (which usually is the same for all subjects!). However, assay variability usually has very little to do with the residual error we see (this conclusion can be based merely on the fact that residual variability is so much larger than assay variability). With other sources being responsible for the residual variability it makes sense not to make the rather strong assumption that it is the same for all subjects. However, if you have sparse data, it is usually not possible to distinguish different variabilities between subjects (another reason it wasn't considered in the early days of pop PK). However, whenever you have relatively rich data it is a good idea to try it. It does require that you use METH=1 INTER as the interaction allows the magnitude to vary between individuals (it is no use at all to use it with FO or FOCE without the interaction term). We introduced it in a paper back in 1998 (Assumption testing in population pharmacokinetic models: illustrated with an analysis of moxonidine data from congestive heart failure patients.J Pharmacokinet Biopharm. 1998 Apr;26(2):207-46), but there hasn't been many other publications using it. Despite this, I've often found it useful. It often makes screening for outlying data points or odd individuals easier or unnecessary as it does allow you to use all data without subjectively omit one data point or another. Precision of parameters and less (odd) influential individuals are other advantages. Hope this was helpful, Mats -- 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 Sweden phone +46 18 471 4105 fax +46 18 471 4003 mats.karlsson@farmbio.uu.se

From: Thomas_Kerbusch@sandwich.pfizer.com Subject: Re: [NMusers] Implementation of interindividual variability in residual variance Date: 2/18/2004 9:57 AM Atul, We have recently gained some experience in accounting for the individual contribution to the residual error. When we pooled data from different studies in patients, healthy volunteers and involving different formulations, it was anticipated that the residual error would not be constant across all individuals. We have published the application of this methodology in BJCP: Kerbusch T, Whlby U, Milligan PA, Karlsson MO. Br J Clin Pharmacol, 2003; 56(6): 639-652 Kerbusch T, Milligan PA, Karlsson MO. Br J Clin Pharmacol, 2004; 57(2): 170-180 With friendly regards, Thomas Kerbusch Clinical PK/PD, Modelling & Simulation Pfizer Global R&D, Sandwich, UK

From: "Piotrovskij, Vladimir [PRDBE]" - VPIOTROV@PRDBE.jnj.com Subject: Re: [NMusers] Implementation of interindividual variability in residual variance Date: 2/18/2004 10:09 AM All, It turned out the model was originally introduced by Mats and presented in 1994 at the PAGE meeting. I was not aware of that and I apologize for the confusion. Best regards, Vladimir _______________________________________________________