RE: Calculating shrinkage when some etas are zero
Resent with fewer trailing messages. Sorry for the spam if both reached NMusers
Hi Marco,
I suspect what is happening in PsN is that IWRES has not been defined in your
model. If you ask for shrinkage then PsN will compute ETA shrinkage (removing
the ETA=0 values) as well as “EPS” shrinkage, which is computed using IWRES.
The IWRES is computed as:
IWRES = (DV – IPRED)/Residual_variance
Where the Residual_variance is the total expected residual variance of each
data point. For an additive+ proportional error model on the normal scale the
IWRES would be defined as:
IPRED = F
W = SQRT((IPRED*THETA(prop))**2+THETA(add)**2)
; log transformed data
; IPRD = F
;DEL = 0.0001
;IF(IPRD.LE.DEL) IPRD=DEL
;IPRED = LOG(IPRD)
;W = SQRT(THETA(prop)**2+THETA(add)**2/IPRD**2)
IRES = DV - IPRED
IF(W.EQ.0) W = 1
IWRES = IRES/W
Y = IPRED + W*EPS(1)
$SIGMA 1 FIX
Of course, if you are just interested in calculating the shrinkage of an ETA
without the zero values then you can print ETA values out in a table file and
compute via:
Shrinkage_eta = 1 – sd(ETA)/Omega
Removing the ETA=0 before computation of the sd(ETA).
Best regards
Andy
Andrew Hooker, Ph.D.
Associate Professor of Pharmacometrics
Dept. of Pharmaceutical Biosciences
Uppsala University
Box 591, 751 24, Uppsala, Sweden
Phone: +46 18 471 4355
www.farmbio.uu.se/research.php?avd=5
Quoted reply history
From: [email protected] [mailto:[email protected]] On
Behalf Of [email protected]
Sent: December 07, 2010 11:16
To: Gastonguay, Marc
Cc: Eleveld, DJ; Ribbing, Jakob; [email protected];
[email protected]; Pyry Välitalo
Subject: Re: [NMusers] Calculating shrinkage when some etas are zero
Hi to all,
Do you know if there is a quick method to exclude subjects with ETA=0 from the
calculation of ETA shrinkage using NONMEM 7?
I also tried to use the option –shrinkage of PsN, but I get the following
error:
AN ERROR WAS FOUND IN THE CONTROL STATEMENTS.
187 $TABLE RECORD REQUESTS AN UNKNOWN ITEM.
at /usr/lib/perl5/site_perl/5.8.8/PsN_3_2_4/nonmem.pm line 40
Kind Regards
Marco
------------------------------------------------------------------------------
Marco Campioni, PhD
Modelling & Simulation Senior Scientist
Exploratory Medicine
Merck Serono S.A. - Geneva