model validation

Dear all, First of all, as new comer to the list, I would like to say Hi to everyone. I am planning to validate an PK model with a new dataset, my question is that, what is the basic steps of using NONMEM to accomplish it? Many thanks in advance.

Dear Ethan, External model evaluation consists of a comparison between the validation dataset (your new dataset) and the predictions obtained by applying the model built on the learning dataset. An usual way to do that is to perform Visual Predictive Checks (VPC), which graphically compares the observations with their predictive distribution according to the model. Others technics, as NPDE could be applyied too for external model evaluation. An add-on package for the open source statistical package R, designed to computed NPDE, is available at www.npde.biostat.fr http://www.npde.biostat.fr/ . Best regards. Karl Refs: N.H. Holford. The visual Predictive Check-Superiority to standard diagnostic (Rorschach) plots. PAGE 14 (2005) Abstr 738 [www.page-meeting.org/?abstract=738]. M. O. Karlsson, N. H. Holford. A tutorial on visual predictive checks. PAGE 17 (2008) Abstr 1434 [www.page-meeting.org/?abstract=1434]. K. Brendel, E. Comets, C. Laffont, C. Laveille, F. Mentré. Metrics for external model evaluation with an application to the population pharmacokinetics of gliclazide. Pharm. Res. 23:2036-2049 (2006). E. Comets, K. Brendel, F. Mentré. Computing normalised prediction distribution errors to evaluate nonlinear mixed-effect models: the npde add-on package for R. Comput. Methods Programs Biomed. 90:154-166 (2008). ________________________________

Ethan, There are many different methods and tools to perform model validation. In addition to what Karl has already mentioned, I will also recommend the following presentation by Dr. Marc Gastonguay as a good overview. http://www.ecpag.org/presentations/2004/workshops/2_gastonguay_Session_I.pdf For tools, you can also try WFN ( http://wfn.sourceforge.net/), PsN( http://psn.sourceforge.net/index.php) or Xpose4 ( http://xpose.sourceforge.net/). Good luck.

Thanks Jun and Karl for the help! -- I almost gave up already after "long" wait After going through both suggestions, I think I will do VPC first as graphical comparison, then use NPDE for more formal evaluation of the whole profile prediction.