Re: Some help with simulation with uncertainty using $PRIOR TNPRI
Hi Rob,
This may help. I have used the Theophylline example that comes with the NM
installation - CONTROL5 model.
*Step 1* - You execute that model with only one minor change - addition of
MSFO at the EST step (which generates and stores the information in the
MSFI file you will need during simulation - step2)
$EST MAXEVAL=450 PRINT=5 MSFO=MSFICONTROL5
*Step 2*- The simulation control stream is very similar to the control
stream used for estimation. It has two $PROB sections and few extra
statements, I have italicized the extra statements needed in the simulation
step.
PS. Use the exact same initial estimates as you had in the estimation
control stream in Step 1 (just leave them as they were in the original
control stream). If you change the initial estimates, the simulation step
will not work.
$PROB THEOPHYLLINE POPULATION DATA
$INPUT ID DOSE=AMT TIME CP=DV WT
$DATA THEOPP
*$PRIOR TNPRI (PROBLEM 2) PLEV=.999 IVAR=1*
$SUBROUTINES ADVAN2
*$MSFI MSFICONTROL5 ONLYREAD ;; uses the msfi file generated during model
fitting*
$PK
CALLFL=1
KA=THETA(1)+ETA(1)
K=THETA(2)+ETA(2)
CL=THETA(3)*WT+ETA(3)
SC=CL/K/WT
ISIM =IREP
$ERROR
Y=F+EPS(1)
*$PROBLEM Sim part *
*$INPUT ID DOSE=AMT TIME CP=DV WT *
*$DATA THEOPP REWIND*
$THETA (.1,3,5) (.008,.08,.5) (.004,.04,.9)
$OMEGA BLOCK(3) 6 .005 .0002 .3 .006 .4
$SIGMA .4
*$SIM (12345) ONLY NSUB=5 TRUE=PRIOR*
;$EST MAXEVAL=450 PRINT=5 MSFO=MSFICONTROL5 ;; commented out in
step 2
;$COV ;; commented out in step 2
$TABLE ISIM ID DOSE WT TIME FILE=sdtab5_uncertainty
As far as I can remember, control streams from Step 1 and 2 need to be run
within same NM version. You cannot fit a model in version 7.1 and then
simulate using ver 7.2 or 7.3 to use the $TNPRI option.
Good luck!
--
Navin Goyal
Quoted reply history
On Mon, Apr 11, 2016 at 9:38 AM, <[email protected]> wrote:
> Dear all,
>
> I would like to perform a simulation (based on a finished model),
> incorporating parameter uncertainty (for all theta’s, eta’s and epsilons)
> with the $PRIOR TNPRI option in Nonmem V7.3, using the results of the
> covariance step as an input. I am still struggling how to implement this.
> Does anybody have an example control stream I could look at?
>
> Sincerely,
>
> Rob ter Heine
>
> Het Radboudumc staat geregistreerd bij de Kamer van Koophandel in het
> handelsregister onder nummer 41055629.
> The Radboud university medical center is listed in the Commercial Register
> of the Chamber of Commerce under file number 41055629.
>
--
Navin Goyal