Covariates and rxSolve: iCov vs params

[john-harrold]

I believe you can specify covariates using either iCov or params. Is this correct? If so what are the benefits/trade offs of using one vs the other. The documentation says that iCov is for non-time-varying covariates.

[mattfidler]

Both iCov and params are supported for non time-varying covariates. At this point there is likely little difference between using either approach.

[john-harrold]

And what is the method for time-varying parameters?

[mattfidler]

You add it to the data frame / event table you are solving. In fact the non time-varying covariate can be in there as well. In that case, rxode2 will:

• Identify it as a non time-varying covariate
• Add it to the non-time varying covariate pool

So, it is actually marginally better to put non time-varying covariates in either iCov or the params data frame.

[john-harrold]

I assume that the event table can also contain the parameters as well (like in NONMEM and I think mrgsolve)?

[mattfidler]

Yes, this can contain parameters as well.