weighting

From: Chantal Csajka <Chantal.Csajka@chuv.hospvd.ch> Subject: weighting Date: Mon, 20 Aug 2001 16:38:54 +0200 Dear nonmem users, I am currently modeling concentration data from routine patient care in a TDM framework. Several patients showed concentrations deviating clearly from the expectation and were therefore repeatedly sampled. I think this results in biased parameter estimates, as the few deviant individuals provided a lot of data points. Is there a way to weight the data, say by the inverse of the number of observations per patient, so that NONMEM takes into account the oversampling and limits the bias ? Thank you Chantal Csajka, PhD Division of Clinical Pharmacology Beaumont 6 CHUV 1011 Lausanne, Switzerland tel: 021/ 314.42.57 fax: 021/ 314.42.66 email: chantal.csajka@chuv.hospvd.ch

From: Nick Holford <n.holford@auckland.ac.nz> Subject: Re: weighting Date: Tue, 21 Aug 2001 09:29:03 +1200 Chantal, Chantal Csajka wrote: > > Dear nonmem users, > I am currently modeling concentration data from routine patient care in a > TDM framework. Several patients showed concentrations deviating clearly > from the expectation and were therefore repeatedly sampled. I think this > results in biased parameter estimates, as the few deviant individuals > provided a lot of data points. Your impression that there is bias from a few 'deviants' may be very reasonable but it is presumably based on prior knowledge about the drug obtained with more formal PK studies. If this is the case then you should consider the use of Bayesian prior constraints on the parameter estimates. This involves specifying the population mean and variance for each of the PK parameters. It can be done using an undocumented feature of NONMEM V version 1.1. You can find some details of this at: http://www.cognigencorp.com/nonmem/nm/99aug052000.html [search for "Winbuggy" in this page] Nick -- Nick Holford, Divn Pharmacology & Clinical Pharmacology University of Auckland, 85 Park Rd, Private Bag 92019, Auckland, New Zealand email:n.holford@auckland.ac.nz tel:+64(9)373-7599x6730 fax:373-7556 http://www.phm.auckland.ac.nz/Staff/NHolford/nholford.htm