Hello everyone,
I am doing some sensitive analysis for the final model parameters, where I
want to increase/decrease the parameter estimation by 10% and look at its
impact on the model prediction.
My model is built in NONMEM 7. And I have WFN and PsN installed. I wonder
if anybody could suggest how to automatic carry out sensitive analysis. Is
PsN or WFN capable of doing so? Of course, the last resort would be to
manually change the parameters and run the model.
Thanks for your help.
Norman
Sensitive analysis
Hi Norman,
With what you're asking, is the real question, "How sensitive is parameter x to
changes in parameter y?"
If so, the best method would probably be bootstrapping. In PsN, that can be
done with the bootstrap command. An alternate could be likelihood profiling
(llp in PsN), though the summary of the changes in the "y" parameter would have
to be done manually as far as I know with that.
Thanks,
Bill
Quoted reply history
On Feb 18, 2012, at 6:46 PM, "Norman Z" <[email protected]> wrote:
> Hello everyone,
>
> I am doing some sensitive analysis for the final model parameters, where I
> want to increase/decrease the parameter estimation by 10% and look at its
> impact on the model prediction.
> My model is built in NONMEM 7. And I have WFN and PsN installed. I wonder if
> anybody could suggest how to automatic carry out sensitive analysis. Is PsN
> or WFN capable of doing so? Of course, the last resort would be to manually
> change the parameters and run the model.
>
> Thanks for your help.
>
> Norman
Norman,
Have a look at log-likelihood profiling in PsN. LLP documentation is on the
PsN website - I don't think it quite does what you want (it will look for
parameter values that give significant OFV changes rather than changing
parameters by certain proportions).
Best wishes,
Joe
Joseph F Standing
MRC Fellow, UCL Institute of Child Health
Antimicrobial Pharmacist, Great Ormond Street Hospital
Honorary Lecturer, UCL School of Pharmacy
Tel: +44(0)207 905 2370
Mobile: +44(0)7970 572435
Quoted reply history
________________________________
From: [email protected] [[email protected]] On Behalf Of
Norman Z [[email protected]]
Sent: 19 February 2012 02:16
To: [email protected]
Subject: [NMusers] Sensitive analysis
Hello everyone,
I am doing some sensitive analysis for the final model parameters, where I want
to increase/decrease the parameter estimation by 10% and look at its impact on
the model prediction.
My model is built in NONMEM 7. And I have WFN and PsN installed. I wonder if
anybody could suggest how to automatic carry out sensitive analysis. Is PsN or
WFN capable of doing so? Of course, the last resort would be to manually change
the parameters and run the model.
Thanks for your help.
Norman
********************************************************************************************************************
This message may contain confidential information. If you are not the intended
recipient please inform the
sender that you have received the message in error before deleting it.
Please do not disclose, copy or distribute information in this e-mail or take
any action in reliance on its contents:
to do so is strictly prohibited and may be unlawful.
Thank you for your co-operation.
NHSmail is the secure email and directory service available for all NHS staff
in England and Scotland
NHSmail is approved for exchanging patient data and other sensitive information
with NHSmail and GSi recipients
NHSmail provides an email address for your career in the NHS and can be
accessed anywhere
For more information and to find out how you can switch, visit
www.connectingforhealth.nhs.uk/nhsmail
********************************************************************************************************************
It seems to me that you have (at least) these two options:
- create a dense series of time points as a dummy data set for NONMEM, fix the
parameters, and write the predicted values into an output table (one per
parameter set)
- implement the model in Berkeley Madonna and export a table of predicted
values for the sets of parameters
For both cases I would calculate the AUC from the output table in R (or any
other suitable programming language).
Btw, a tool such as Pirana converts nonmem DES into BM code for you.
Andreas
-----
Andreas Krause, PhD
Director, Lead Scientist Modeling and Simulation
Dept. of Clinical Pharmacology
Actelion Pharmaceuticals Ltd / Gewerbestrasse 16 / CH-4123 Allschwil /
Switzerland Phone +41 61 565 6891 / Email [email protected]
[email protected] wrote: -----
To: "Denney, William S." <[email protected]>, "Standing Joseph (GREAT
ORMOND STREET HOSPITAL FOR CHILDREN NHS TRUST)" <[email protected]>
From: Norman Z <[email protected]>
Sent by: [email protected]
Date: 2012-02-19 18:16
Cc: [email protected]
Subject: Re: [NMusers] Sensitive analysis
Hi Bill and Joe,
Thank you very much for your suggestions.
What I am asking is "How sensitive is the simulated PK profiles (summerized by
AUC or Cmax) to changes in parameter y?".
So instead of looking at the OFV change with different parameters, I want to
summarize the result of the sensitive analysis. My goal is to obtain the AUC
and Cmax for the predicted PK profiles (the model have multiple compartments,
and I need to extract the AUC and Cmax for several compartments) with different
parameter values.
Does bootstrap or LLP in PsN out put all the intermediate simulation results?
Kind regards,
Norman
The information of this email and in any file transmitted with it is strictly
confidential and may be legally privileged.
It is intended solely for the addressee. If you are not the intended recipient,
any copying, distribution or any other use of this email is prohibited and may
be unlawful. In such case, you should please notify the sender immediately and
destroy this email.
The content of this email is not legally binding unless confirmed by letter.
Any views expressed in this message are those of the individual sender, except
where the message states otherwise and the sender is authorised to state them
to be the views of the sender's company. For further information about Actelion
please see our website at http://www.actelion.com
Hi Norman,
If you have the most recent PsN version, then there is functionality to
simulate with uncertainty in population parameters. PsN can use either the
nonmem cov matrix or (non-parametric) bootstrap for that purpose. Each
replicate dataset that is simulated would be simulated with a different set of
population parameters.
Unless you want to do a sensitivity analysis on parameters that you have fixed
in your analysis, I think this is exactly what you need.
Best regards
Jakob
Sent from my iPhone
Quoted reply history
On 19 Feb 2012, at 18:42, "Norman Z" <[email protected]> wrote:
> Hi Bill and Joe,
>
> Thank you very much for your suggestions.
>
> What I am asking is "How sensitive is the simulated PK profiles (summerized
> by AUC or Cmax) to changes in parameter y?".
>
> So instead of looking at the OFV change with different parameters, I want to
> summarize the result of the sensitive analysis. My goal is to obtain the AUC
> and Cmax for the predicted PK profiles (the model have multiple compartments,
> and I need to extract the AUC and Cmax for several compartments) with
> different parameter values.
> Does bootstrap or LLP in PsN out put all the intermediate simulation results?
>
> Kind regards,
>
> Norman
>
>
Norman,
I am intrigued by your question because I cannot think of any good reason to want to understand this 'sensitivity'.
Note for AUC the 'sensitivity' is trivial to calculate because it is inversely proportional to CL (and proportional to F if you have F in our model).
Can you give me some idea what you want to do with this 'sensitivity'?
Best wishes,
Nick
Quoted reply history
On 20/02/2012 1:16 a.m., Norman Z wrote:
> Hi Bill and Joe,
> Thank you very much for your suggestions.
>
> What I am asking is "How sensitive is the simulated PK profiles (summerized by AUC or Cmax) to changes in parameter y?". So instead of looking at the OFV change with different parameters, I want to summarize the result of the sensitive analysis. My goal is to obtain the AUC and Cmax for the predicted PK profiles (the model have multiple compartments, and I need to extract the AUC and Cmax for several compartments) with different parameter values. Does bootstrap or LLP in PsN out put all the intermediate simulation results?
>
> Kind regards,
> Norman
--
Nick Holford, Professor Clinical Pharmacology
Dept Pharmacology& Clinical Pharmacology, Bldg 505 Room 202D
University of Auckland,85 Park Rd,Private Bag 92019,Auckland,New Zealand
tel:+64(9)923-6730 fax:+64(9)373-7090 mobile:+64(21)46 23 53
email: [email protected]
http://www.fmhs.auckland.ac.nz/sms/pharmacology/holford
Dear Norman,
If you are an R user, I have had some success using the FME package for this
sort of sensitivity analysis:
http://cran.r-project.org/web/packages/FME/index.html
www.jstatsoft.org/v33/i03/paper
Hope that might be of some help.
Joanna
Joanna Lewis
PhD student
UCL Institute of Child Health,
30 Guilford Street,
London WC1N 1EH
0207 905 2307
Quoted reply history
From: Norman Z <[email protected]<mailto:[email protected]>>
Date: Sat, 18 Feb 2012 21:16:15 -0500
To: <[email protected]<mailto:[email protected]>>
Subject: [NMusers] Sensitive analysis
Hello everyone,
I am doing some sensitive analysis for the final model parameters, where I want
to increase/decrease the parameter estimation by 10% and look at its impact on
the model prediction.
My model is built in NONMEM 7. And I have WFN and PsN installed. I wonder if
anybody could suggest how to automatic carry out sensitive analysis. Is PsN or
WFN capable of doing so? Of course, the last resort would be to manually change
the parameters and run the model.
Thanks for your help.
Norman
Hi NMUsers,
A lot of very constructive suggestions were provided since I posted my
question. I sincerely appreciate everybody for your help.
Norman
Quoted reply history
On Mon, Feb 20, 2012 at 5:22 AM, Lewis, Joanna <[email protected]>wrote:
> Dear Norman,
>
> If you are an R user, I have had some success using the FME package for
> this sort of sensitivity analysis:
>
> http://cran.r-project.org/web/packages/FME/index.html
> www.jstatsoft.org/v33/i03/paper
>
> Hope that might be of some help.
>
> Joanna
>
> Joanna Lewis
> PhD student
> UCL Institute of Child Health,
> 30 Guilford Street,
> London WC1N 1EH
> 0207 905 2307
>
> From: Norman Z <[email protected]>
> Date: Sat, 18 Feb 2012 21:16:15 -0500
>
> To: <[email protected]>
> Subject: [NMusers] Sensitive analysis
>
> Hello everyone,
>
>
>
> I am doing some sensitive analysis for the final model parameters, where I
> want to increase/decrease the parameter estimation by 10% and look at its
> impact on the model prediction.
>
> My model is built in NONMEM 7. And I have WFN and PsN installed. I wonder
> if anybody could suggest how to automatic carry out sensitive analysis. Is
> PsN or WFN capable of doing so? Of course, the last resort would be to
> manually change the parameters and run the model.
>
>
>
> Thanks for your help.
>
>
>
> Norman
>
>
>
Sorry to join the conversation late but an alternative would be Phoenix
WinNonlin which can do this type of sensitivity analysis quite easily.
Simply append copies of the dataset for each scenario (to set sampling times
for each subject) and set the parameters for each scenario. A simulation will
then produce the predicted values for each case.
These can then be passed to the NCA tool in WinNonlin can be used to get Cmax
and AUC (which are sensitive to sampling times) and you can build a nice table
of the results. All within one software package and the workflow can be easily
exported as a template for re-use on other projects, or to share with
colleagues.
Best regards,
Simon.
___________________________________
[email protected]<mailto:[email protected]>
Senior Scientific Consultant
Pharsight- A Certara™ Company
Mobile : +44 7980 832 666
Facsimile : +1 801 991 7145
Home Office : +44 113 274 1198
July Phoenix 6.2.1, including AutoPilot 2.0 released,
March Phoenix 6.2, including the IVIVC toolkit released, get tips and discuss
Pharsight products with other users; www.pharsight.com/extranet
___________________________________
http://www.pharsight.com/training/training_upcoming.php
Introduction to Phoenix WinNonlin 21-23 Feb. 2012 Milan, Italy
Population Modeling Methodology using Phoenix NLME 19-21 Mar. 2012 Paris, France
Introduction to Phoenix WinNonlin 17-19 Apr. 2012 Berlin, Germany
Introduction to Phoenix WinNonlin 15-17 May 2012 Amsterdam, The Netherlands
Satellite workshops at PAGE, 5 June 2012, Venice, Italy.
08:30 – 12:30 “Phoenix Connect: Linking Modeling and Simulation Tools,
including NONMEM, R, Xpose & PsN, in Phoenix Workflows”
https://pharsight.securesites.com/training/course_display_new.php?details_id=98.
13:00 – 17:00 “An Overview of the QRPEM Algorithm in Phoenix NLME”
https://pharsight.securesites.com/training/course_display_new.php?details_id=99.
Introduction to Phoenix WinNonlin 11-13 Sep. 2012 Amsterdam, The Netherlands
IVIVC Toolkit for Phoenix WinNonlin - In Vivo-In Vitro Correlations 13-14 Sep.
Amsterdam, The Netherlands
Quoted reply history
-----Original Message-----
From: [email protected]<mailto:[email protected]>
[mailto:[email protected]] On Behalf Of
[email protected]<mailto:[email protected]>
Sent: Sunday, February 19, 2012 5:55 PM
To: Norman Z
Cc: Denney, William S.; Standing Joseph (GREAT ORMOND STREET HOSPITAL FOR
CHILDREN NHS TRUST); [email protected]<mailto:[email protected]>
Subject: Re: [NMusers] Sensitive analysis
It seems to me that you have (at least) these two options:
- create a dense series of time points as a dummy data set for NONMEM, fix the
parameters, and write the predicted values into an output table (one per
parameter set)
- implement the model in Berkeley Madonna and export a table of predicted
values for the sets of parameters
For both cases I would calculate the AUC from the output table in R (or any
other suitable programming language).
Btw, a tool such as Pirana converts nonmem DES into BM code for you.
Andreas
-----
Andreas Krause, PhD
Director, Lead Scientist Modeling and Simulation Dept. of Clinical Pharmacology
Actelion Pharmaceuticals Ltd / Gewerbestrasse 16 / CH-4123 Allschwil /
Switzerland Phone +41 61 565 6891 / Email
[email protected]<mailto:[email protected]>
[email protected]<mailto:[email protected]>
wrote: -----
To: "Denney, William S."
<[email protected]<mailto:[email protected]>>, "Standing
Joseph (GREAT ORMOND STREET HOSPITAL FOR CHILDREN NHS TRUST)"
<[email protected]<mailto:[email protected]>>
From: Norman Z <[email protected]<mailto:[email protected]>>
Sent by: [email protected]<mailto:[email protected]>
Date: 2012-02-19 18:16
Cc: [email protected]<mailto:[email protected]>
Subject: Re: [NMusers] Sensitive analysis
Hi Bill and Joe,
Thank you very much for your suggestions.
What I am asking is "How sensitive is the simulated PK profiles (summerized by
AUC or Cmax) to changes in parameter y?".
So instead of looking at the OFV change with different parameters, I want to
summarize the result of the sensitive analysis. My goal is to obtain the AUC
and Cmax for the predicted PK profiles (the model have multiple compartments,
and I need to extract the AUC and Cmax for several compartments) with different
parameter values.
Does bootstrap or LLP in PsN out put all the intermediate simulation results?
Kind regards,
Norman
The information of this email and in any file transmitted with it is strictly
confidential and may be legally privileged.
It is intended solely for the addressee. If you are not the intended recipient,
any copying, distribution or any other use of this email is prohibited and may
be unlawful. In such case, you should please notify the sender immediately and
destroy this email.
The content of this email is not legally binding unless confirmed by letter.
Any views expressed in this message are those of the individual sender, except
where the message states otherwise and the sender is authorised to state them
to be the views of the sender's company. For further information about Actelion
please see our website at http://www.actelion.com
_________________________________________________________________