RE: medians of the simulation in VPC are systematically below the medians of the observations when the model fit looks fine.

From: Kenneth Kowalski Date: February 06, 2012 technical Source: mail-archive.com
Hi Shirley, What this suggests is that the distributions of the empirical Bayes' predictions of the ETAs (which are used to generate the IPREDs) have some departures from the normality assumptions (including zero mean with constant variance omega^2) when you use the model in simulations. I suggest that you look at box plots of the ETAs for each parameter stratified by key features of the design such as dose and generate histograms of the ETAs to verify that the distributions are symmetrically distributed about zero. If you observe departures from the normality assumptions that are being used in your simulations (e.g., trends in ETAs versus dose, right-skewed distribution, etc.) this may provide some insight as to where the deficiencies lie in your model. While I agree that ETA predictions are an imperfect diagnostic since they can suffer from shrinkage, as long as you continue to use simulations and a VPC approach to verify the adequacy of your model, a cautious look at these ETA plots to provide some guidance as to how to modify your model to improve the VPC may be helpful. Best regards, Ken
Quoted reply history
From: [email protected] [mailto:[email protected]] On Behalf Of Hou, Xiaoli Shirley Sent: Friday, February 03, 2012 5:48 PM To: nmusers Subject: [NMusers] medians of the simulation in VPC are systematically below the medians of the observations when the model fit looks fine. Dear All, I have built a non-compartment model, in which the medians of IPREDs and Observations are very close. But in VPC plot, the medians of simulations are systematically below the medians of observations Any idea why it happens? Thanks. Shirley Notice: This e-mail message, together with any attachments, contains information of Merck & Co., Inc. (One Merck Drive, Whitehouse Station, New Jersey, USA 08889), and/or its affiliates Direct contact information for affiliates is available at http://www.merck.com/contact/contacts.html) that may be confidential, proprietary copyrighted and/or legally privileged. It is intended solely for the use of the individual or entity named on this message. If you are not the intended recipient, and have received this message in error, please notify us immediately by reply e-mail and then delete it from your system.
Feb 03, 2012 Xiaoli Shirley Hou medians of the simulation in VPC are systematically below the medians of the observations when the model fit looks fine.
Feb 04, 2012 Nick Holford Re: medians of the simulation in VPC are systematically below the medians of the observations when the model fit looks fine.
Feb 06, 2012 Kenneth Kowalski RE: medians of the simulation in VPC are systematically below the medians of the observations when the model fit looks fine.