Thank you! I followed your advice by ignoring bql values at first and I managed
to get my model to work with satisfying results.
I did chose not to exclude any bql DV in my final dataset, including those
pre-first dose samples, since removing them only slightly changes my parameter
estimates while increasing the objective function, what's the usual approach
when dealing with predose observations?
And finally while my model seems to be a good fit, during its execution nonmem
displayed several error messages in the output window that weren't reported in
the res file such as:
OCCURS DURING SEARCH FOR ETA AT INITIAL VALUE, ETA=0
PK SUBROUTINE: ERROR IN COMPUTATION
ATTEMPT TO COMPUTE BASE*POWER WITH BASE < 0
PRED EXIT CODE= 1
OCCURS DURING SEARCH FOR ETA AT INITIAL VALUE, ETA=0
ERROR In TRANS4 ROUTINE, CL (could also be V2 or Q) IS NEGATIVE
All these errors involved the first subject ID=1 exclusively
Should these messages be taken into consideration or can they be overlooked
considering they weren't transcribed in the report file and I'm satisfied with
my model?
Thank you again
regards
-----Original Message-----
From: Leonid Gibiansky <lgibian...@quantpharm.com>
Sent: Tuesday, 30 May 2023 04:17
To: Hiba Sliem <hiba.sl...@pharmalex.com>
Cc: nmusers <nmusers@globomaxnm.com>
Subject: Re: [NMusers] Problem with estimating sigma when using M3 method
The code is incorrect, $ERROR should be
$ERROR
LOQ=0.1
IPRED = F
W1 = THETA(8)*IPRED
W2 = THETA(9)
W = SQRT(W1**2+W2**2)
IRES = DV - IPRED
IWRES = IRES/W
DUM = (LOQ -IPRED)/W
CUMD = PHI(DUM)
IF(BLQ.EQ.0) THEN
F_FLAG=0
Y= IPRED +W1*ERR(1) + W2*ERR(2)
ELSE
F_FLAG=1
Y=CUMD
MDVRES = 1
ENDIF
My be it makes sense to show the entire $PK and $EST blocks: easier to check.
Which ADVAN do you use? If ADVAN6, switch to ADVAN13
Use MATRIX=S on $COV step
Use NOABORT on $EST step
try to use NSIG=4 SIGL=12 on $EST with LAPLACEAN
For pre-dose samples, do you mean pre-first dose? Then those should be ignored,
removed from the data set, or use EVID=2 MDV=1, then they will be ignored at
estimation. After washout, if all samples are BQLs, I would ignore them in a
similar way, at least initially.
For washout setting, all versions are fine, You can use TIME=0 EVID=4 for the
first dose after washout.
To check the code, I would first remove BQLs (use MDV=1 EVID=2 for those, to
get IPRED), use FOCEI rather than LAPLACEAN, and make sure that the model fit
is good. Then check that IPRED at the points of BQLs is small. If not, check
whether those BQLs are reasonable or could be data errors. Then switch to
LAPLACEAN with initial values of all parameters set at the final values of the
previous FOCEI run. Make sure you use INTERACTION options for all runs.
Good luck!
Leonid
On 5/29/2023 8:02 PM, Hiba Sliem wrote:
Hi
Thank you for your assistance, unfortunately I still haven't found a solution
to my issue, and I keep running into either one of these error messages
depending on my dataset codification:
#PROGRAM TERMINATED BY OBJ
ERROR IN NCONTR WITH INDIVIDUAL 6 ID= 6.00000000000000E+00
NUMERICAL HESSIAN OF OBJ. FUNC. FOR COMPUTING CONDITIONAL ESTIMATE
IS NON POSITIVE DEFINITE
#R MATRIX ALGORITHMICALLY SINGULAR
AND ALGORITHMICALLY NON-POSITIVE-SEMIDEFINITE 0R MATRIX IS OUTPUT
0COVARIANCE STEP ABORTED
I finally managed to have both estimation and covariance not fail with the
following error code :
$ERROR
LOQ=0.1
IPRED = F
W1 = THETA(8)*IPRED
W2 = THETA(9)
IRES = DV - IPRED
IWRES = IRES/(W1 + W2)
DUM = (LOQ -IPRED)/(W1 + W2)
CUMD = PHI(DUM)
IF(BLQ.EQ.0) THEN
F_FLAG=0
Y= IPRED +W1*ERR(1) + W2*ERR(2)
ELSE
F_FLAG=1
Y=CUMD
MDVRES = 1
ENDIF
$THETA
(0.01, 0.38) ; [w1]
(0.01, 0.1) ; [w2]
$SIGMA
1 FIX ;[P] sigma(1,1)
1 FIX ;[P] sigma(2,2)
However my results were accompanied by the following message:
MINIMIZATION SUCCESSFUL
HOWEVER, PROBLEMS OCCURRED WITH THE MINIMIZATION.
REGARD THE RESULTS OF THE ESTIMATION STEP CAREFULLY, AND ACCEPT THEM ONLY
AFTER CHECKING THAT THE COVARIANCE STEP PRODUCES REASONABLE OUTPUT.
I think a part of the issue is the way I've been formatting my dataset, since I
get different results depending on the way I've set it up, so I'd like to have
your opinion on the best way to proceed in the following situations:
>predose samples that are BLQ or higher at Time of dosing:
ID TIME DV AMT BLQ
1 0 0.1 0 1
1 0 0 10 0
Do I keep the original times or do something like that?:
ID TIME DV AMT BLQ
1 -0.01 0.1 0 1
1 0 0 10 0
And can I ignore these observations altogether (using something like MDV=100
for example)?
Washout period followed by a predose sample and a new administration of the
same drug (different dose):
ID TIME DV AMT BLQ EVID
1 0 0 10 0
1
1 1 0.5 0 0
0
1 .... ... ....
.... ....
After washout > 1 1000 0.1 0 1 0
1 1000 0 20 0
1
Do I leave it inchanged, or use EVID4 either this way:
ID TIME DV AMT BLQ EVID
1 0 0 10 0
1
1 1 0.5 0 0
0
1 .... ... ....
.... ....
1 1000 0.1 0 1
0
1 0 0 20 0
4
Or that way?:
ID TIME DV AMT BLQ EVID
1 0 0 10 0
1
1 1 0.5 0 0
0
1 .... ... ....
.... ....
1 0 0 20 0
4
1 0 0.1 0 1
0
Sorry if these all seem like obvious questions, but I've been struggling to get
satisfying results over the last few days and I'd like to understand what I've
been doing wrong.
Kind regards,
-----Original Message-----
From: Philip Harder Delff <phi...@delff.dk>
Sent: Friday, 26 May 2023 21:42
To: Leonid Gibiansky <lgibian...@quantpharm.com>
Cc: Hiba Sliem <hiba.sl...@pharmalex.com>; nmusers
<nmusers@globomaxnm.com>
Subject: Re: [NMusers] Problem with estimating sigma when using M3
method
[You don't often get email from phi...@delff.dk. Learn why this is
important at https://aka.ms/LearnAboutSenderIdentification ]
Hi Hiba,
I agree that often the issues should be found in the data rather than the
model. I recommend checking the data with NMcheckData from the R package called
NMdata. It scans for a long list of potential issues, some that will make
Nonmem fail, some that won't. If it finds issues, they will be returned in a
data.frame with reference to row numbers and ID's so you can easily identify
the root cause. If you look at ?NMcheckData you may identify arguments you can
specify to add to the list of checks the function can run.
Having said this, a data/model issue can also be that your data poorly supports
estimation of parts of your model (practical identifiability).
NMcheckData won't help you identify such issues.
NMdata:
https://eur02.safelinks.protection.outlook.com/?url=https%3A%2F%2Fphil
ipdelff.github.io%2FNMdata%2F&data=05%7C01%7CHiba.Sliem%40pharmalex.co
m%7Ca65c013926fc4c8b3b9c08db60b3fcdd%7Cff9ac3ce3c4141c3b556e1b32a662fe
d%7C0%7C0%7C638210098390223373%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjA
wMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sd
ata=wjIRuNgQJ9jEvmHCKVJx94bH0IEEY9hwVoUXVf2PSjA%3D&reserved=0
NMcheckData manual:
https://eur02.safelinks.protection.outlook.com/?url=https%3A%2F%2Fphil
ipdelff.github.io%2FNMdata%2Freference%2FNMcheckData.html&data=05%7C01
%7CHiba.Sliem%40pharmalex.com%7Ca65c013926fc4c8b3b9c08db60b3fcdd%7Cff9
ac3ce3c4141c3b556e1b32a662fed%7C0%7C0%7C638210098390223373%7CUnknown%7
CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXV
CI6Mn0%3D%7C3000%7C%7C%7C&sdata=uUA1xwU0xQw6XVhuOZD0C59TmvpbutubpINW7s
TwGjU%3D&reserved=0
An example with a few arguments that activate additional checks:
res.checks <-
NMcheckData(mydata,covs="WEIGHTBL",cols.num="WEIGHT",col.usubjid="USUB
JID") Here, NMcheckData will (in addition to a bunch of other checks)
see if WEIGHTBL exists numeric, non-na and unique within subjects, WEIGHT
exists and is numeric and non-NA, and that ID is unique against USUBJID and
vice versa. (Obviously, Nonmem can't read USUBJID if it contains characters,
but you could still keep it to the right in the dataset for reference). See the
manual above for more options.
Best,
Philip
On 2023-05-26 11:06 AM, Leonid Gibiansky wrote:
Yes, SIGMA should be fixed to 1 (do not try anything else, it has to
be done correctly in the code first, and then we should worry about
how to make it work)
For combined error, expression is
W = SQRT(W1**2 + W2**2) (squares in both terms)
Do not worry about error 134, this is harmless, and you can fix it
any time after you get your model right. Add UNCONDITIONAL MATRIX=S
to the $COV step.
For PARAMETER ESTIMATE IS NEAR ITS BOUNDARY try to add
NOSIGMABOUNDTEST NOOMEGABOUNDTEST NOTHETABOUNDTEST to $est record
Most of the time, numerical difficulties come from the problems with
the data, so it makes sense to clean the data set first as much as
possible.
Leonid
On 5/26/2023 9:30 AM, Hiba Sliem wrote:
Hi
I already tried fixing the value of sigma to 1, the covariance step
isn't implemented when I do that.
If I try fixing it to 0.144 the minimization isn't successful.
I also tried a combined error model like this:
LOQ=0.1
IPRED = F
W1 = THETA(8)*IPRED
W2 = THETA(9)
W = SQRT(W1**1 + W2**2)
DEL = 0
IF(W.EQ.0) DEL = 1
IRES = DV - IPRED
IWRES = IRES/(W + DEL)
DUM = (LOQ -IPRED)/(W + DEL)
CUMD = PHI(DUM)
IF(BLQ.EQ.0) THEN
F_FLAG=0
Y= IPRED +W*ERR(1)
ELSE
F_FLAG=1
Y=CUMD
MDVRES = 1
ENDIF
In which case I get a PARAMETER ESTIMATE IS NEAR ITS BOUNDARY error
message When trying to fix Sigma in the combined model I have a
MINIMIZATION TERMINATED
DUE TO ROUNDING ERRORS (ERROR=134) message.
My dataset has a lot of predose samples and washouts between
different periods, is it possible the issue comes from my dataset?
Regards
-----Original Message-----
From: Leonid Gibiansky <lgibian...@quantpharm.com>
Sent: Friday, 26 May 2023 14:51
To: Hiba Sliem <hiba.sl...@pharmalex.com>; nmusers@globomaxnm.com
Subject: Re: [NMusers] Problem with estimating sigma when using M3
method
[You don't often get email from lgibian...@quantpharm.com. Learn why
this is important at https://aka.ms/LearnAboutSenderIdentification ]
you should fix
$SIGMA
1 FIX
as you are already estimating the SD using THETA(8).
Leonid
On 5/26/2023 4:57 AM, Hiba Sliem wrote:
Hello
I'm fairly new to nonmem, I'm currently trying to model a phase 1
study with BLQ values, while the run was successful with no error
message, my
residual error has a %rse >70 and a confidence interval that
includes
zero.
Here's my code:
$ERROR
LOQ=0.1
IPRED = F
SD = THETA(8)*IPRED
DEL = 0
IF(SD.EQ.0) DEL = 1
IRES = DV - IPRED
IWRES = IRES / (SD + DEL)
DUM = (LOQ -IPRED) / (SD + DEL)
CUMD = PHI(DUM) + DEL
IF(BLQ.EQ.0) THEN
F_FLAG=0
Y= IPRED +SD*ERR(1)
ELSE
F_FLAG=1
Y=CUMD
MDVRES = 1
ENDIF
$EST METHOD=1 INTERACTION LAPLACIAN PRINT=5 MAX=9999 SIG=3 SLOW
NUMERICAL MSFO=*.msf
$SIGMA
0.38 ;[P] sigma(1,1) (estimated in a previous model)
Furthermore, when trying to fit this model to my phase 2 dataset,
covariance step fails when I implement it.
Any suggestions are welcome
Thank you
