1 reply to this topic

[Teodora]

Hi,

 

I have a QT dataset with ECG data from a predose day and from the treatment day, specified in a column "DAY" where DAY=0 for the predose or DAY=1 for the treatment day. What I want is to model QT interval as a function of the RR interval using only the predose data. Is there a way I can subset the dataset within the PML script to do this? I am thinking something like an if statement, but I cannot figure out how to implement it. This is what I have so far:

 

test(){

    covariate©

    E = BSL*(C/1000)**(alpha)

    error(EEps = 0.1)

    observe(EObs© = E + EEps)

    stparm(BSL = tvBSL + nBSL)

    stparm(alpha = tvalpha + nalpha)

    fixef(tvBSL = c(300, 350, 800))

    fixef(tvalpha = c(0.1, 0.3, 1))

    ranef(diag(nBSL, nalpha) = c(0.1, 0.1))

}

 

where EObs are the observed QT intervals, C is mapped to the "RR" column in the dataset, BSL and alpha are model parameters. how do I bring the "DAY" into the model?

 

Any suggestion would be greately appreciated!

 

Thanks,

Dora

 

PS. I know I can filter the dataset using the "DataWizard" object, but I am trying to avoid doing this.

[serge guzy]

Dear Dora

Here is what I suggest.

Add an MDV column with 1 for all data you do not want to use (all Day 1 data).

In the user interface, you have input options with MDV as an option. Map it to the MDV column you added in the data set (should map automatically).

If you run the mdoel, only the Day 0 data will be used as oyu put MDV=1 for the treatment Day (Day=1).

This seems to me the easiest way to handle your issue but there should be other ways too.

best

Serge