From: Johan Rosenborg Johan.Rosenborg@astrazeneca.com 
Subject:[NMusers] Further on the handling of missing data
Date: Tue, June 29, 2004 10:28 am

Dear NONMEM Users,
I have used EVID=0 and MDV=0 to indicate when there is an observation and
EVID=2 and MDV=1 to indicate when there is not an observation in a data file
adapted for NONMEM modelling using ADVBAN8. Each subject was treated three
times and three variables were monitored in plasma and urine. My problem:
for one of the treatments the program consequently avoids to predict a value
where no observation was available. I would appreciate any suggestion on how
to solve this problem.

Kind regards,


Johan Rosenborg
AstraZeneca R&D Lund
Experimental Medicine, S-221 87 Lund, Sweden
Tel: +46 46 33 65 99 Fax: +46 46 33 71 91
E-mail: johan.rosenborg@astrazeneca.com
_______________________________________________________

From: Koup, Jeffrey Jeffrey.Koup@pfizer.com
Subject: RE:[NMusers] Further on the handling of missing data
Date: 29 june 2004 17:50 

I have occasionally had similar problems obtaining predicted values with
higher ADVANS.  
Adding a PCMT data item, giving the compartment number that you are
attempting to predict has worked for me.

Jeff Koup

_______________________________________________________

From: Johan Rosenborg Johan.Rosenborg@astrazeneca.com
Subject: RE:[NMusers] Further on the handling of missing data
Date: Thu, July 1, 2004 4:12 am  
Jeffrey,

Thank you very much, copying and pasting the CMT values under a PCMT data
item solved my problem!

Johan
_______________________________________________________