Proposal: Additional field to the Drug Exposure table

[MelaniePhilofsky]

Issue: There isn't a concept_id field for the dose_unit_source_value field. The dose_unit_source_value field stores the unit of measurement for drugs. My specific use case (and those on the forum, need a way to represent the dosing of intravenous medications that are given per unit of time and/or per unit of body size (kg, bsa).

Proposed Solution: Add a dose_unit_concept_id field to the Drug Exposure table. This may require additions to the vocabulary.

Pros: The proposed solution will create a standardized way to represent dose units. Which will facilitate standardized analytics and reusable code for various analyses.

Cons: ?

[cgreich]

@MelaniePhilofsky: Wait. dose_concept_id is going away, or already went. folks wanted to add a field for doses taht are relative, not absolute.

[MelaniePhilofsky]

@cgreich Wait?!? You told me to move forward :)

We (see above link to forum post) want a way to put the 5 mcg/kg/min drug doses into the CDM with a concept_id for the mcg/kg/min or the like. We already have the quantity for the numerical representation and the dose_unit_source_value for the mcg/kg/min, now we need to standardize it with a concept_id.

I'm not sure about the dose_concept_id. It's not in there now and I don't remember it. Maybe it was prior to the Drug Strength table. What was the description for that field?

[cgreich]

@MelaniePhilofsky:

Oh, I am not against moving foward. I wanted to make sure we don't pick the same name of the field that we just abolished: drug_unit_concept_id. The unit should be take from the table DRUG_STRENGTH. It gives for each ingredient the unit we expect. The drug_unit_source_value is a remnant of that field, that's no longer there.

But somehow I am starting to think we won't be able to enforce that whole trick with the pre-specified unit. And we may just get the drug_unit_concept_id back, so you can use it. Even though you don't have absolute drug amount (mg, ml, µg) but relative (µg/kg/min). Let's discuss this on the WG next time.

[MelaniePhilofsky]

@cgreich ,

So, the dose_unit_source_value was the text/code representation for the drug_unit_concept_id? They had different names, but contained the same ideas? The description for each column in the CDM is something I still struggle with understanding. The descriptions are way too vague and ambiguous. dose_unit_source_value = The information about the dose unit as detailed in the source.

@clairblacketer Can we put this on the next CDM WG agenda?

[cgreich]

Argh. I misspelled. The duo was dose_unit_concept_id and dose_unit_source_value. I think. Need to look into old versions of the CDM doc.

Yes, we need to do an overhaul of the description.

[clairblacketer]

@MelaniePhilofsky unfortunately the agenda for November and December are set so we wouldn't be able to discuss this until January.

In addition, not every source_value in the CDM has or needs a source_concept_id. They are not meant for standard analytics so they would not be terribly useful. How about we add to the vocabulary standard concepts meant for mcg/kg/min, etc? Or, as suggested in the forum we could add a field called 'Dose_basis_concept_id' or something like that for the additional denominator

[DocEd]

Hi all. Could I ask if this issue was progressed at all? As far as I can tell, the "infusion" issue remains unresolved in OMOP. This remains the last piece of the puzzle before my team migrates to fully "OMOPify" our research database. I have some mitigation strategies in mind, that might entail extending OMOP for our particular case use, however it would be preferable to have a roadmap to be working with the core OMOP CDM.

[cgreich]

@DocEd:

Let's get done with it, I agree.

[DocEd]

Brilliant, how should we proceed?

[DocEd]

Is there any update here?

[cgreich]

Will bring up at the next CDM meeting.

[clairblacketer]

Hi @cgreich and @DocEd the CDM meetings are now shifting slightly to focus on documentation and ETL specifications through the end of next year. As Melanie says in the post above

The description for each column in the CDM is something I still struggle with understanding. The descriptions are way too vague and ambiguous. dose_unit_source_value = The information about the dose unit as detailed in the source.

This is the exact thing we are trying to rectify before continually adding more and more columns to the CDM. We will begin our new remit in November with a focus on the PERSON and OBSERVATION_PERIOD tables. I understand the infusion issue is an important one and it will be discussed once we move to the DRUG_EXPOSURE table in 2020.

@DocEd you mentioned you have some mitigation strategies in mind, I would be happy to meet with you separately to discuss if you like until we have a more permanent plan in place.

[DocEd]

Apologies for such a late response on this. @clairblacketer it would be great to get your insights here. I think my team will have to move forward without updates to the CDM. We don't want to extend OMOP unless we think our changes will be brought into master.

As such we will likely use the drug exposure table to record the drug at the RxNorm ingredient level. Then use the fact relationship table to link into either the measurement or observation table to include a semantic notion of the drug dose. We will probably use dose_unit_source_value to record the units, and forget about standardisation for the time being.

[cgreich]

@DocEd:

This is a long discussion. Are you waiting for the dose_unit_source_value? This is a additional field, it is very highly likely it will go in. In the mean time, you can just use it. Does that help?

[MelaniePhilofsky]

Closing this issue because University of Colorado doesn't have a network research use case for the data.