The Wisdom of Lewis Sheiner

Mark wrote: "(just went around and around with a sponsor about this - actually their stats consultant who basically just kept insisting on the theory regardless of data that we presented to the contrary)." I am glad to see you are agreeing with the data beats theory perspective. This is what I have been pushing hard in the discussion about NONMEM's estimation termination messages. Note also that I have not left it simply by saying data beats the old theory but I have proposed a new hypothesis (" In many cases NONMEM's messages are determined at random"). I have recently been re-reading and marvelling again at the wisdom contained in some commentaries of Lewis Sheiner. I encourage all nmusers to read them - especially when you need to talk to statisticians and educate them that there is life beyond the chi-square distribution and the Wald test: Sheiner LB. Clinical pharmacology and the choice between theory and empiricism. Clin Pharmacol Ther. 1989;46(6):605-15. The first page or so states the scientific method and subsequently illustrates it from a modeller's viewpoint. This is especially relevant to the use of data to challenge theoretical predictions and to continue with a new theory in cyclical fashion. Sheiner LB. The intellectual health of clinical drug evaluation. Clinical Pharmacology & Therapeutics. 1991;50(1):4-9. Makes the case that statisticians have mislead the clinicians and drug evaluation, (a broader area than drug development and regulatory hurdle jumping), has gone wrong because of that. Today, there are some areas of enlightenment in large organizations but statisticians and their ill-advised clinicians continue to growl from their dark caves. Loud noises still beat sparkles of light in this area. Rodda BE, Brooks C, Reynolds G. Statistics and clinical trials. Clin Pharmacol Ther. 1992;52(1):104-5. The senior officers of the American Statistical Society try to defend themselves. The proponents of the dull hypothesis. Sheiner LB. Statistics and clinical trials (Reply). Clin Pharmacol Ther. 1992;52(1):105-6. Lewis replies in masterful fashion and poses a thought provoking quiz. The advocate of sparkling insight. Nick Mark Sale - Next Level Solutions wrote: Leonid, I agree with your point that failure to converge/and or covariance is a message that the model is a prompt to study the model. I object to those who claim that model that fails covariance is not useful despite data to the contrary (just went around and around with a sponsor about this - actually their stats consultant who basically just kept insisting on the theory regardless of data that we presented to the contrary). But, I think that the messages are completely non-specific - they tell you something is less than ideal, but give no clue as to what. I suspect that graphics are likely to be much more consistently informative, telling you not only that something is less than ideal, but some clue what to do to fix it. As such, I'm not sure that convergence and covariance messages add anything to the process (anything that a good and thorough analyst would have known already, based on VPC, NPC, various post hoc plots etc). Mark Mark Sale MD Next Level Solutions, LLC www.NextLevelSolns.com 919-846-9185 -- Nick Holford, Dept Pharmacology & Clinical Pharmacology University of Auckland, 85 Park Rd, Private Bag 92019, Auckland, New Zealand [EMAIL PROTECTED] tel:+64(9)923-6730 fax:+64(9)373-7090 http://www.fmhs.auckland.ac.nz/sms/pharmacology/holford