4 replies to this topic

[Davidq8]

Hello, I'm working on a model and trying to determine how a continuous covariate would affect the Pk parameters of a drug.

 

This covariate has a value of zero for some subjects and an integer value "above zero" for other subjects.

 

Even though eta plots demonstrate a significant correlation between the value of this covariate and Pk parameters, I don't get any iterations when I conduct stepwise analysis to evaluate the impact of this covariate on the drug parameters.

 

Does that imply that the continuous covariates must always have a value greater than zero ?

[kniefort]

Can you share a snippet of your model code with how you have the covariate influence coded on your parameter?  For example, the default syntax:

 

Cl = tvCl * COV^dCldCOV * exp(nCl)) 

 

Would result in Cl taking a value of 0 when the covariate value is 0. --  This run would fail.

 

The "1+" option would be 

 

Cl = tvCl * (1+(COV)*dCldCOV) * exp(nCl))

 

This syntax would still allow Cl to take a positive value when COV=0

[cradhakr]

Hi,

 

If you don't want to share your project file in this domain - please send it support@certara.com.

 

Thanks

[smouksassi1]

you can also go into textual mode and go free style: 

stparm(Clind = tvCl * exp(nCl))
Clcoveffect= CRCL > 0 ? (CRCL/112)^dCldCRCL : 1
Cl = Clind *Clcoveffect
fixef(dCldCRCL = c(,0.5,))

[Davidq8]

Thanks Kniefort, Cradhakr and Sameer for your responses.