Subject: Re: Failing table step upon redefining ETAs as additive (or high correlations in $COV output)

From: Nick Holford n.holford@auckland.ac.nz Subject: Subject: Re: [NMusers] Failing table step upon redefining ETAs as additive (or high correlations in $COV output) Date: Fri, 22 Jul 2005 03:15:15 +1200 I suggest you think more about the purpose of your modelling and evaluate the model with a specific objective in mind rather than worrying about nuisance factors such as the shape ETA distributions and standard errors. Have you performed a visual predictive check? What do the tolerance intervals look like? Personally I think FO runs are always suspicious and I would prefer to trust the FOCE runs despite the apparently ominous termination message. Experimental investigations by two separate groups have shown NONMEM's termination messages have little relationship to the adequacy of the parameter estimates obtained. I would not place much faith in ad hoc covariate model such as AGE on volume without having used more physiological models for body size on all size related parameters. Nick-- Nick Holford, Dept Pharmacology & Clinical Pharmacology University of Auckland, 85 Park Rd, Private Bag 92019, Auckland, New Zealand email:n.holford@auckland.ac.nz tel:+64(9)373-7599x86730 fax:373-7556 http://www.health.auckland.ac.nz/pharmacology/staff/nholford/