Dear group,
Recently I was running a sparse sampling animal data (one data per animal)
using NONMEM v6 & WFN and ran into this trouble:
"PARAMETER ESTIMATE IS NEAR ITS BOUNDARY
THIS MUST BE ADDRESSED BEFORE THE COVARIANCE STEP CAN BE IMPLEMENTED".
I searched the forum for any solution and found the message from Mark Sale
(Next Level Solutions).
My question: a) Is there any option to overcome this issue without suppressing
the message as suggested earlier?
b) What causes the above problem? Do I need to avoid specifying any boundary
for the thetas?
c) What does this message indicate? a poor selection of model? poor initial
estimates? narrow boundary? or any correlation issues with the estimates?
I highly appreciate your input. Thanks in advance
Regards,
Chandra Durairaj,
Post-doctoral Fellow,
Univ of Colorado Denver
PARAMETER ESTIMATE IS NEAR ITS BOUNDARY with NONMEM v6
Chandra,
Please post your control stream that causes the warning. I wonder how you are dealing with the one obs/animal problem.
My advice on how to handle the warning is to look at the parameter estimates and ask yourself if they seem sensible. Interpretation of boundaries is up to the user and his/her knowledge of the underlying problem.
Nick
Durairaj Chandrasekar wrote:
> Dear group,
>
> Recently I was running a sparse sampling animal data (one data per animal) using
> NONMEM v6 & WFN and ran into this trouble:
>
> "PARAMETER ESTIMATE IS NEAR ITS BOUNDARY
> THIS MUST BE ADDRESSED BEFORE THE COVARIANCE STEP CAN BE IMPLEMENTED".
>
> I searched the forum for any solution and found the message from Mark Sale
> (Next Level Solutions).
>
> My question: a) Is there any option to overcome this issue without suppressing
> the message as suggested earlier?
> b) What causes the above problem? Do I need to avoid specifying any boundary
> for the thetas?
> c) What does this message indicate? a poor selection of model? poor initial
> estimates? narrow boundary? or any correlation issues with the estimates?
>
> I highly appreciate your input. Thanks in advance
>
> Regards,
> Chandra Durairaj,
> Post-doctoral Fellow,
> Univ of Colorado Denver
--
Nick Holford, Professor Clinical Pharmacology
Dept Pharmacology & Clinical Pharmacology
University of Auckland, 85 Park Rd, Private Bag 92019, Auckland, New Zealand
[email protected] tel:+64(9)923-6730 fax:+64(9)373-7090
mobile: +64 21 46 23 53
http://www.fmhs.auckland.ac.nz/sms/pharmacology/holford
Durairaj,
My comments are as follows:
a) Is there any option to overcome this issue without suppressing the message as suggested earlier?
In my experience, changing initial estimates to estimates nearer the final estimate may cause this problem to disappear.
Also changing the units on your clearance, volume of distribution, etc., can cause this problem to disappear.
b) What causes the above problem? Do I need to avoid specifying any boundary for the thetas?
Your scaled parameter estimate is near the boundary of what was specified. When you expect the estimate or large to be small, this problem can occur.
c) What does this message indicate? a poor selection of model? poor initial estimates? narrow boundary? or any correlation issues with the estimates?
This message indicates that one of your parameters is either too small or too large, which sometimes may be taken care of with changing initial estimates. The rest of the questions are more easily answered in context of your model, what is expected.
Matt.
Quoted reply history
-----Original Message-----
From: owner-nmusers
On Behalf Of Durairaj Chandrasekar
Sent: Thursday, September 24, 2009 3:19 PM
To: nmusers
Subject: [NMusers] PARAMETER ESTIMATE IS NEAR ITS BOUNDARY with NONMEM v6
Dear group,
Recently I was running a sparse sampling animal data (one data per animal) using NONMEM v6 & WFN and ran into this trouble:
"PARAMETER ESTIMATE IS NEAR ITS BOUNDARY
THIS MUST BE ADDRESSED BEFORE THE COVARIANCE STEP CAN BE IMPLEMENTED".
I searched the forum for any solution and found the message from Mark Sale (Next Level Solutions).
My question: a) Is there any option to overcome this issue without suppressing the message as suggested earlier?
b) What causes the above problem? Do I need to avoid specifying any boundary for the thetas?
c) What does this message indicate? a poor selection of model? poor initial estimates? narrow boundary? or any correlation issues with the estimates?
I highly appreciate your input. Thanks in advance
Regards,
Chandra Durairaj,
Post-doctoral Fellow,
Univ of Colorado Denver
This e-mail (including any attachments) is confidential and may be legally privileged. If you are not an intended recipient or an authorized representative of an intended recipient, you are prohibited from using, copying or distributing the information in this e-mail or its attachments. If you have received this e-mail in error, please notify the sender immediately by return e-mail and delete all copies of this message and any attachments.
Thank you.
Durairaj,
My comments are as follows:
a) Is there any option to overcome this issue without suppressing the message
as suggested earlier?
In my experience, changing initial estimates to estimates nearer the final
estimate may cause this problem to disappear.
Also changing the units on your clearance, volume of distribution, etc., can
cause this problem to disappear.
b) What causes the above problem? Do I need to avoid specifying any boundary
for the thetas?
Your scaled parameter estimate is near the boundary of what was specified.
When you expect the estimate or large to be small, this problem can occur.
c) What does this message indicate? a poor selection of model? poor initial
estimates? narrow boundary? or any correlation issues with the estimates?
This message indicates that one of your parameters is either too small or too
large, which sometimes may be taken care of with changing initial estimates.
The rest of the questions are more easily answered in context of your model,
what is expected.
Matt.
Quoted reply history
-----Original Message-----
From: [email protected] [mailto:[email protected]] On
Behalf Of Durairaj Chandrasekar
Sent: Thursday, September 24, 2009 3:19 PM
To: [email protected]
Subject: [NMusers] PARAMETER ESTIMATE IS NEAR ITS BOUNDARY with NONMEM v6
Dear group,
Recently I was running a sparse sampling animal data (one data per animal)
using NONMEM v6 & WFN and ran into this trouble:
"PARAMETER ESTIMATE IS NEAR ITS BOUNDARY
THIS MUST BE ADDRESSED BEFORE THE COVARIANCE STEP CAN BE IMPLEMENTED".
I searched the forum for any solution and found the message from Mark Sale
(Next Level Solutions).
My question: a) Is there any option to overcome this issue without suppressing
the message as suggested earlier?
b) What causes the above problem? Do I need to avoid specifying any boundary
for the thetas?
c) What does this message indicate? a poor selection of model? poor initial
estimates? narrow boundary? or any correlation issues with the estimates?
I highly appreciate your input. Thanks in advance
Regards,
Chandra Durairaj,
Post-doctoral Fellow,
Univ of Colorado Denver
This e-mail (including any attachments) is confidential and may be legally
privileged. If you are not an intended recipient or an authorized
representative of an intended recipient, you are prohibited from using, copying
or distributing the information in this e-mail or its attachments. If you have
received this e-mail in error, please notify the sender immediately by return
e-mail and delete all copies of this message and any attachments.
Thank you.