1 reply to this topic

[Shu-Pei Wu]

Hi All,

 

Recently, I am setting up a PKPD model.

In the dataset I have 0 (I put them as blank) for PK concentration and some values for PD for vehicle arm.

 

I would like to model PKPD simultaneously.

However, what I found is that the PK estimation will be biased by those missing value in concentration.

I try to use the following code to avoid that but without total success.

--

observe(CObs = (Nominal_dose==0?0:C) * (1 + CEps))

--

I will get different parameter estimates and a warning saying:

Warning: Cannot determine error variance - 1.0 assumed
 

Any suggestion?

Thank you very much,

Shu-Pei

[Simon Davis]

Hi Shu-Pei if you can post (or send us directly) the project file I think this would be easier to investigate.

 

 thanks, Simon.