search

nlmixrdevelopment / paws-2019

ASCEPT-PAGANZ joint meeting in Queenstown, New Zealand on 25 November 2019
2 stars 1 forks source link

question about confidence interval of prediction #1

Open dingying85 opened 4 years ago

dingying85 commented 4 years ago

Thanks for the nice course materials.

I can see the PAWS-2 is for parameter estimation with existing PK data and PAWS-1 is for prediction with hypothetical dosing schedule. I run through the PAWS-1, however, the curves are deterministic looks like, is that possible we could have some confidence interval of the prediction? Any suggestion or function can provide that?

Best Regards, Ying

mattfidler commented 4 years ago

Hi Ying,

I'm unsure what you wish to have. Do you mean the VPC?

Matt.

dingying85 commented 4 years ago

Dear Matt,

From the tutorial, seems vpc gives the confidence interval of the model built, but not the extrapolation of the model to a hypothetical new dosing schedule. For example, can we have a confidence interval of predicted PK profiles from the current biweekly dosing to monthly dosing. Are there any function good for that? Thanks.

Yours, Ying

mattfidler commented 4 years ago

Hi Ying,

An example like what you are describing is on slide 27 of the following slide-deck:

https://github.com/nlmixrdevelopment/ACoP9-2018/raw/master/nlmixr-tutorial_final.pdf

Let me know if you need any other help.

dingying85 commented 4 years ago

Hi Matt,

Really appreciated. This is really helpful.

Yours, Ying