Continuous covariate with zero value

Phoenix Certara Forums Modelling and Simulation
1

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 ?

2

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

3

Hi,

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

Thanks

4

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,))

5

Thanks Kniefort, Cradhakr and Sameer for your responses.