NONMEM Users Network Archive

Hosted by Cognigen

problems of the first stage of STS in NONMEM

From: Chenguang Wang <maurice.wang>
Date: Thu, 12 Nov 2009 14:27:31 +0100

Dear NONMEM users,
I would like to do the first stage analysis of STS in NONMEM. I followed the
tips in NONMEM Topic 35 (
http://www.cognigencorp.com/nonmem/nmo/topic035.html) with some adaptations.

The control stream is partly listed below:
...
$SUBROUTINE ADVAN11 TRANS4
$PK
   CL = THETA(1)*EXP(ETA(1))
   Q2=THETA(2)*EXP(ETA(2))
   Q3=THETA(3)*EXP(ETA(3))
   V1=THETA(4)*EXP(ETA(4))
   V2=THETA(5)*EXP(ETA(5))
   V3=THETA(6)*EXP(ETA(6)
   S1=V1
$ERROR
   IPRED = LOG(.000001)
   IF (F.GT.0) IPRED=LOG(F)
   Y=IPRED+ERR(1)*EXP(ETA(7))
$THETA
   ( 0, 0.3, ) ; TH1 CL
   ( 0, 0.3, ) ; TH2 Q2
   ( 0, 0.5, ) ; TH3 Q3
   ( 0, 0.1, ) ; TH4 V1
   ( 0, 0.1, ) ; TH5 V2
   ( 0, 0.5, ) ; TH6 V3
$OMEGA
   100 ; OMEGA1 CL
   100 ; OMEGA2 Q2
   100 ; OMEGA3 Q3
   100 ; OMEGA4 V1
   100 ; OMEGA5 V2
   100 ; OMEGA6 V3
   100 ; OMEGA7 ERROR
$SIGMA
   1 FIX ; SIGMA1
$EST NOABORT PRINT=15 MAXEVALS=0 POSTHOC METH=COND INTERACTION
...

I logtransformed the concentration, so I did the adaptation in the $ERROR
block. (In the original example, it uses "Y = F*EXP(EXP(ETA(3))*ERR(1))" for
the exponential type residual error. By logtransform it, I get
Y=LOG(F)+EXP(ETA(7))*ERR(1) which I used in my control stream. ) By the
way, it is a rich data set.

I expected to get different parameters for different individuals, but in the
result all individuals have same values. Could you tell me what is wrong?

Thanks in advance!

Chenguang

Received on Thu Nov 12 2009 - 08:27:31 EST

The NONMEM Users Network is maintained by ICON plc. Requests to subscribe to the network should be sent to: nmusers-request@iconplc.com.

Once subscribed, you may contribute to the discussion by emailing: nmusers@globomaxnm.com.