From: "Eltahtawy, Ahmed" <>


Date: Fri, 27 Jul 2001 14:58:24 -0400


Dear All,


I have a question for the group about Pred vs Wres plots. I pooled data for

IV, IR, SR formulations and the IV gives me the most biased PRED VS WRES

pattern (S shape trend)!! In spite of using the flexible error model

(add+prop). Does anyone have any experience or explanation for this?.


would a separate error model for each formulation be reasonable? I tempted

even to analyze the data separately? Or just remove the IV data? As I don't

have the actual sampling times.


Thanks for your time in advance.








From: Lewis B Sheiner <>

Subject: Re: PRED VS WRES

Date: Fri, 27 Jul 2001 12:32:43 -0700


WRES tends to be diagnostic for the within-individual (i.e PK) model.

If by "S shape trend" you mean that the residuals follow a trend to be

first >0 and then <0, then this suggests that you need to further

elaborate your PK structural model (e.g., a 2 compartment model instead

of a 1-compartment model).

And if this were so, it would make sense that it would be seen most readily with

the IV

data rather than with dosage forms exhibiting slower absorption.


A better plot, in my opinion, for diagnosing the adequacy of

the subject-specific mdoel is IPRED vs IWRES.



_/ _/ _/_/ _/_/_/ _/_/_/ Lewis B Sheiner, MD (

_/ _/ _/ _/_ _/_/ Professor: Lab. Med., Biophmct. Sci., Med.

_/ _/ _/ _/ _/ Box 0626, UCSF, SF, CA, 94143-0626

_/_/ _/_/ _/_/_/ _/ 415-476-1965 (v), 415-476-2796 (fax)




From: "Eltahtawy, Ahmed" <>


Date: Fri, 27 Jul 2001 15:57:49 -0400


Yes, The WRES tend to be first >0 and then <0. The fit to the IV data

alone is little better with the 2 comp Model. Unfortunately, this is not the

case with the IR and SR formulations. When all the data is pooled together

I opted for 1 cmp. Model. I don't think a mixture model would be of any

help. I will try the IPRED VS IWRES; but I still can't understand how an IV

gives the worst diagnostics.. Thanks






From: "Ferrin Harrison 301-827-3213 FAX 301-480-2825" <>

Subject: Pred vs Wres

Date: Fri, 27 Jul 2001 16:23:26 -0400 (EDT)


What models for an entry compartment would conceal a downstream compartment?


In particular, could the SR rate be sufficiently slow to conceal one of the IV

compartments, rendering one compartment comparable to infinite rate of

elimination or equilibration?


By conceal I mean render difficult or impossible to estimate for the data

available. I can't conceive of an entry compartment mathematically neutralizing

a downstream compartment.




From: "Eltahtawy, Ahmed" <>

Subject: RE: Pred vs Wres

Date: Fri, 27 Jul 2001 17:07:02 -0400


I have a dual absorption model to describe the complex absorption for the

SR data. One of these is really slow..