Are ADDL (additional doses) and II (interval) nlme columns supported in Simbiology datasets for model fitting?

I have a dataset with 4 IDs, each ID receives a different dose level, and doses are given daily during 4 weeks.

I would like to fit the data to a model we already have, but I am not sure how to make the mapping between the IDs and the doses. Do we need to create as many rows as doses administered in the dataset (so doses from dataset can be used) or is there another way through variants, each variant receiving a different dose defined in the model?

Thanks a lot for your help!

Best,

Zinnia

Answer 1

[Arthur Goldsipe]

When you actually perform the fit, you typically want one row in your data that describes each distinct dose event. (It is also possible to set up separate doses and/or variants to use during fitting, but I don't think you need to do that here.)

SimBiology can understand the ADDL and II columns when importing data in the NONMEM format and convert that into a table that's appropriate for fitting. (You can read more about NONMEM support here.) As shown here, you can either rely on the default names for NONMEM columns, or you can explicitly tell SimBiology that ADDL is your DoseRepeatLabel and II is your DoseIntervalLable. When you import the data, SimBiology will convert this one row with ADDL and II information into one row per dose event.

I hope that makes sense. If you need more details, just let me know whether you're working at the command line or in the SimBiology Model Analyzer app (or both!), and what version of MATLAB you're using.

-Arthur

[Zinnia Parra-Guillen]

Thanks a lot Arthur! It does make sense and works perfectly well.

I have one more question related to this dataset, hope you can help. Is it possible to specify a different variant (defined in the model, I am using simbiology app) and map each variant to each group in the dataset?

The reason is that in the PK model we are using, dose is a covariate in Q, so we have defined one variant for each dose level with different Q values. The only way I can think of doing it right now is to include Q in the dataset and map that as a variant to our Q component in the model...does it make sense?

Thanks again!

Zinnia

[Arthur Goldsipe]

Hi, as long as you are using R2022a or later, then you should be able to accomplish this in the app. Your variant information can either be stored in the data or as variants on the model. You just need to configure the table of variant and dose setup to use this information. We have an example that shows how to do this.

[Zinnia Parra-Guillen]

Sorry! I was using the wrong dataset, as soon as I had several groups I saw how you can specify a different variant for each of them when using simulation group :)