Jump to content


Photo

Time varying covariate in NONMEM


  • Please log in to reply
3 replies to this topic

#1 Nyein Hsu Maung

Nyein Hsu Maung

    Newbie

  • Members
  • Pip
  • 2 posts

Posted 27 June 2019 - 04:33 AM

Hello everyone, could you please suggest me some books or materials about how to build population pharmacokinetic models with time varying covariates in NONMEM. Thank you.



#2 Simon Davis

Simon Davis

    Advanced Member

  • Administrators
  • 1,316 posts

Posted 27 June 2019 - 05:57 AM

Dear Dr Maung,

 

   This is a forum primarily for Phoenix and therefore we mainly support building population models in NLME.  There are a few discussions already around this if you search for "time varying" e.g.

 

https://support.cert...ual-error-model

 

and using reset

 

https://support.cert...osing-modelling

 

   Try having a look at those and begin building your model, remember you can probably get much of your structure first using built-in or graphical mode and then perform further edits if required in textual mode.  I make sure I do that in a copy of my model so I can more easily roll back if I want to.

 

   If you get stuck then please reply here with your project and we'll take it from there.

If you definitely want to construct your model in NONMEM then perhaps try the NM users mail list.

 

  Simon.


Edited by Simon Davis, 27 June 2019 - 05:57 AM.

  • Nyein Hsu Maung likes this

#3 Nyein Hsu Maung

Nyein Hsu Maung

    Newbie

  • Members
  • Pip
  • 2 posts

Posted 27 June 2019 - 06:36 AM

Thank you very much!



#4 smouksassi1

smouksassi1

    Advanced Member

  • Members
  • PipPipPip
  • 231 posts
  • LocationMontreal

Posted 27 June 2019 - 07:45 AM

regardless of what tool you are using you need to understand the underlying assumptions and with time varying covariate you can  carry the observation backward or forward

 

ID TIME DV COVTIME

1      0                 10

1     0.5   2        

 1      1    5

1     10    5.5      12

...

 

so above you had the covariate evaluated at time 0 and 10  how do you want it to fill in the values between 0 and 10

is it 10 up to time 10 or you move the 12 back to zero.

 

that is why in phoenix you have the covariate and fcovariate ( forward move covariate)

 

in NONMEM you have the $BIND

 

 

in some situations you might want to build a time varying model for your time varying covariate ( joint modeling)

 

Samer






0 user(s) are reading this topic

0 members, 0 guests, 0 anonymous users