Hi,
NM-TRAN has some restrictions on its language so that it can do the
really tricky stuff behind the scenes that is needed to create the
FORTRAN that NONMEM uses. These restrictions apply tocoding conditional
statements.
First of all you cannot have nested IF-THEN-ELSE blocks with NM-TRAN if
they include random variables. The NM-TRAN error message states this if
you try this.
326 RANDOM VARIABLE IS DEFINED IN A NESTED IF STRUCTURE.
The example provided by Bernard is a nested structure containing random
effects (ETA variables). As coded by Bernard it is not legal. ENDIF
statements are always required for every IF-THEN. So the 'crash'
experienced by Bernard was caused by not following the rules.
I have re-coded it below to show the nested structure. However, this
will fail in NM-TRAN because of the use of ETA variables in the expressions.
IF (GENE.EQ.3) THEN ;GG
CL = THETA(1)* EXP(ETA(1))
ELSE
IF (GENE.EQ.2)THEN ;GT
CL = THETA(6)* EXP(ETA(3))
IF (GENE.EQ.1) THEN ;TT
CL = THETA(7)* EXP(ETA(4))
IF (GENE.EQ.4) THEN ;MISSING
CL = THETA(7)* EXP(ETA(4))
ENDIF
ENDIF
ENDIF
ENDIF
The second rule is that one line IF statements cannot include THEN so
the suggestion from Bill won't work. It needs to be written like this:
IF (GENE.EQ.3) CL = THETA(1)* EXP(ETA(1)) ;GG
IF (GENE.EQ.2) CL = THETA(6)* EXP(ETA(3)) ;GT
IF (GENE.EQ.1) CL = THETA(7)* EXP(ETA(4)) ;TT
IF (GENE.EQ.4) CL = THETA(7)* EXP(ETA(4)) ;MISSING
Best wishes,
Nick
On 25/07/2013 4:54 p.m., Denney, William S. wrote:
Hi Bernard,
Try using just IF statements instead of including the ELSE IF and END IF like
this:
IF (GENE.EQ.3) THEN CL = THETA(1)* EXP(ETA(1)) ;GG
IF (GENE.EQ.2) THEN CL = THETA(6)* EXP(ETA(3)) ;GT
IF (GENE.EQ.1) THEN CL = THETA(7)* EXP(ETA(4)) ;TT
IF (GENE.EQ.4) THEN CL = THETA(7)* EXP(ETA(4)) ;MISSING
It's also worth verifying that GENE is always 1 to 4. The crash could be
happening because you have a value other than 1 to 4. A way around this is to
have a default value and then only apply changes like:
CL = THETA(7)* EXP(ETA(4)) ; for TT or MISSING
IF (GENE.EQ.3) THEN CL = THETA(1)* EXP(ETA(1)) ;GG
IF (GENE.EQ.2) THEN CL = THETA(6)* EXP(ETA(3)) ;GT
That will ensure that it works for all cases and will generally be more robust
code.
Thanks,
Bill
-----Original Message-----
From: [email protected] [mailto:[email protected]] On
Behalf Of Bernard Ngara
Sent: Thursday, July 25, 2013 10:09 AM
To: nmusers
Subject: [NMusers] if-the-statement for 3 class categorical covariate
Dear all
I am shoked because NONMEM sents a crash message when i factor in covariate
modeling using a 4 class categorical variable genotype:
IF (GENE.EQ.3) THEN ;GG
CL = THETA(1)* EXP(ETA(1))
ELSE IF (GENE.EQ.2)THEN ;GT
CL = THETA(6)* EXP(ETA(3))
IF (GENE.EQ.1) THEN ;TT
CL = THETA(7)* EXP(ETA(4))
IF (GENE.EQ.4) THEN ;MISSING
CL = THETA(7)* EXP(ETA(4))
END IF
However if i binarize the code, nonmem runs successfully IF (GENE.EQ.3) THEN
CL = THETA(1)* EXP(ETA(1))
ELSE
CL = THETA(6)* EXP(ETA(3))
END IF
Thanks in advance
--
Bernard Ngara
Trainee Biostatistician
Africa Institute of Biomedical Science and Technology P. O Box 2294 Harare
Zimbabwe
+263 772 160 896
--
Nick Holford, Professor Clinical Pharmacology
Dept Pharmacology & Clinical Pharmacology, Bldg 503 Room 302A
University of Auckland,85 Park Rd,Private Bag 92019,Auckland,New Zealand
office:+64(9)923-6730 mobile:NZ +64(21)46 23 53 FR +33(7)85 36 84 99
email: [email protected]
http://holford.fmhs.auckland.ac.nz/
Holford NHG. Disease progression and neuroscience. Journal of Pharmacokinetics
and Pharmacodynamics. 2013;40:369-76
http://link.springer.com/article/10.1007/s10928-013-9316-2
Holford N, Heo Y-A, Anderson B. A pharmacokinetic standard for babies and
adults. J Pharm Sci. 2013:
http://onlinelibrary.wiley.com/doi/10.1002/jps.23574/abstract
Holford N. A time to event tutorial for pharmacometricians. CPT:PSP. 2013;2:
http://www.nature.com/psp/journal/v2/n5/full/psp201318a.html
Holford NHG. Clinical pharmacology = disease progression + drug action. British
Journal of Clinical Pharmacology. 2013:
http://onlinelibrary.wiley.com/doi/10.1111/bcp.12170/abstract
