# 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 (

The control stream is partly listed below:
...
\$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?