Stream: implementers
Topic: Encoding counts in `SimpleQuantity`
Morten Ernebjerg (Nov 02 2018 at 10:15):
I am working on MedicationStatementand looking at how to indicate how much of the medication needs to be taken (field doseQuantity, of type SimpleQuantity). In the simplest case, one gives a count, e.g. the medication comes in the form of pills and one wants to say "take 3 of those (pills)". "In the wild" I have seen the following 4 ways of encoding "One pill:
1. Only value (form of medication in Medication resource)
{ "value": 1, }
2. Only UCUM code for count (form of medication in Medication resource)
{ "unit": "1", "system": "http://unitsofmeasure.org", "code": "1" }
3. UCUM code for count and value (form of medication in Medication resource)
{ "value": 1, "unit": "1", "system": "http://unitsofmeasure.org", "code": "1" }
4. Units that explicitly give the form of the medication (and value on top)
{ "value": 1, "unit": "TAB", "system": "http://hl7.org/fhir/v3/orderableDrugForm", "code": "TAB" }
My question is: Is one way preferable? Or do they actually indicate different things? I would expect Medicationto contain information about the form, in which no. 4 seems redundant. On a more conceptual level, I am wondering how to interpret valueif code & system are also given.
Stefan Lang (Nov 02 2018 at 10:55):
In my personal opinion, there should always be a unit (i.e. code+system) to keep the resource data self contained, though FHIR does not force you to do so.
Also, there are real world cases where e.g. the package contains 20 tablets, but the statement says "150 mg". Or a bottle of 100 ml, take 20 drops daily. In these cases you need code+system anyway, so I would use it consistently in all instances.
Also note: "If the unit can be coded in UCUM and a code is provided, it SHOULD be a UCUM code.". Again no hard enforcement, but a preference on UCUM over other code systems, including the v3 orderableDrugForm. But that is dependant on your use case.
I probably would go with UCUM, in the following form:
{
"value": "1",
"unit": "tablet",
"system": "http://unitsofmeasure.org",
"code": "{tbl}"
}
John Silva (Nov 02 2018 at 12:22):
In V2 we had the dosage form (RXO-5, RXE-6 , RXA-8) that helped clarify that the dosage amount (RXO-2, RXE-3 ,RXA-6) and units. Is there an equivalent of "dosage form" in FHIR?
Morten Ernebjerg (Nov 02 2018 at 14:32):
@Stefan Lang Thanks, that makes sense - I didn't consider the fact that UCUM also has codes for some "entity types".
Stefan Lang (Nov 02 2018 at 14:37):
Well, you can write anything in curly braces and it will become a valid UCUM code ;-) But some more common ones are explicitely mentioned in the UCUM spec
Lloyd McKenzie (Nov 02 2018 at 14:56):
Using the curly braces for the UCUM unit is likely to cause you trouble in some implementations. It's probably best to leave the UCUM unit as "1" - which is the equivalent of anything in bare curly braces. But it's fine to have the unit element contain a useful human description. (The UCUM code is for the computer - and the computer has no use for the stuff in curly braces . It just makes the computer's work harder.)
John Silva (Nov 02 2018 at 14:59):
However, just knowing the "dosage form" (in UCUM or other) it doesn't tell you the amount (dose amount) in the dosage form, e.g. a tablet could be 250mg or 500mg. Wouldn't it still be necessary to encode that somewhere, I assume in the dosage property?
Lloyd McKenzie (Nov 02 2018 at 15:00):
Only if it's not implicit in the selected drug code
Stefan Lang (Nov 02 2018 at 15:02):
And then you would put it into Medication.ingredient.amount
Stefan Lang (Nov 02 2018 at 15:03):
i.e. 500 mg per 1 tablet
Jose Costa Teixeira (Nov 08 2018 at 18:41):
I noticed that you want to use medicationStatement to indicate "what should be taken". Is this an actual treatment that will be acted upon i.e. having prescriptions, dispenses...?
Jose Costa Teixeira (Nov 08 2018 at 18:44):
(won't bother you with yet another discussions on why medicationStatement is not a treatment plan).
Jose Costa Teixeira (Nov 08 2018 at 18:44):
@Morten Ernebjerg
Morten Ernebjerg (Nov 09 2018 at 07:38):
@Jose Costa Teixeira This from the context of a German medication plan. These plans are "patient-centric" and serve as reminders for the patient on what medication to take, and for showing others what they are taking (e.g. to avoid problematic interactions between drugs). They are distinct from the prescriptions, but will probably often be given in parallel. E.g. when a doctor prescribes new medicine, she might also give the patient such a medication plan as a reminder, along with the prescriptions (the plan need not be limited to what was just prescribe, it could also contain other medicine the patient is taking). Plans can also be issued by pharmacies when they fill prescriptions. There is standard format for this in Germany, which has also been implemented in FHIR (_Medikationsplan PLUS_ project, on Simplifier here - German only, though) and here, MedicationStatement is used to record instruction about when and how much of a given medication to take (as part of a FHIR document).
Jose Costa Teixeira (Nov 09 2018 at 07:52):
Ok, I do not feel comfortable with this use of medicationStatement.
what you describe is a medicationRequest, not a statement
Jose Costa Teixeira (Nov 09 2018 at 07:53):
But thanks for pointing it out. I was unaware of this and will bring it up in HL7 DE.
Jose Costa Teixeira (Nov 09 2018 at 07:55):
i think we have not been clear enough that medicationRequest is not only a "Rezept"
Jose Costa Teixeira (Nov 09 2018 at 08:39):
"for showing others what they are taking (e.g. to avoid problematic interactions between drugs)" is a good use of medicationStatement. perhaps I read "medicationplan" differently from what it is intended.
Where in the IG do we read that "MedicationStatement is used to record instruction about when and how much of a given medication to take"?
Morten Ernebjerg (Nov 09 2018 at 09:01):
Maybe I'm out of my depth here w.r.t. exactly how to interpret the plan, I am by no means an expert on this. In Germany, the intake instruction are also on the prescription slips, so the medication plan is by no means required to instruct the patient (in fact, you only have the right to get one printed if you are taking 3+ medications). @Stefan Lang ?
All the IG stuff I've seen is in German, but here is an example of a full German medication plan using this standard (there are no other Medication[x] resources involved):
http://fhir.hl7.de:8080/baseDstu3/Composition/266/$document?persist=false
Jose Costa Teixeira (Nov 09 2018 at 09:16):
i was looking for the german documentation. The document I see seems ok - this seems just an informative list of what the patient is taking or scheduled to take
Jose Costa Teixeira (Nov 09 2018 at 09:18):
that is why i asked to see where in the guide there is something that could be misread as recording instructions to take medication.
Jose Costa Teixeira (Nov 09 2018 at 09:21):
btw, the trigger for this was when you wrote you wanted to express "you take 3 of those (pills)" - which is a request, not a statement
Jose Costa Teixeira (Nov 09 2018 at 09:21):
statement is "I will be taking 3 of those pills"
Morten Ernebjerg (Nov 09 2018 at 11:48):
The formulations I used above are mine, I did not take them from any IG, so any imprecisions are mine (as mentioned, I am not an expert on the details). Here is the German IG: http://wiki.hl7.de/index.php?title=IG:Patientenbezogener_Medikationsplan_Plus. The concept for such a medication plan is independent of the FHIR implementation, see e.g. http://www.kbv.de/media/sp/Praxisinformation_Medikationsplan.pdf
Lloyd McKenzie (Nov 09 2018 at 12:07):
Jose is correct that MedicationStatement reflects what's actually being done (or at least is believed to be happening), while MedicationRequest reflects what's recommended/planned/ordered to be done.
Morten Ernebjerg (Nov 12 2018 at 09:11):
My feeling is that the German medication plan may actually balance on the border between (or mix) those two categories, e.g. it could be used to show to other docs what's actually being done but also to remind the patient what s/he should be taking. @Simone Heckmann @Stefan Lang, do you know how this was settled in the context of the MPP/Bundeseinheitliche Medikationsplan? (BTW, my apologies if I am accidentally repeating a discussion that already happened!)
Stefan Lang (Nov 12 2018 at 12:58):
Right, MPPlus is an informative list, based on the Bundesmedikationsplan. The latter one gives every patient with 3+ medications the right to receive a standardized list of their medication (continuous or as needed), as @Morten Ernebjerg mentioned above. This may include information from the patient, like "I take a pill of Magnesium and a pill of Vitamin B12 daily", if the author (i.e. practitioner or pharmacist) thinks it's relevant, which is definitely not a prescription.
On the deeper background probably @dr Kai U. Heitmann can eleborate.
We had a similar lengthy discussion in the German stream on MStatement vs. MRequest which ended with the conclusion that MStatement is the right resource to use here.
If you want to express details in the context of a prescription, you would use MedicationRequest, what the pharmacy does is a MedicationDispense, and to express things like "today at 8:00 I took a pill of Ibuoprofene 400 mg" (or: "today from 14:00 to 17:00 the patient received an infusion of Taxol") you use MedicationAdministration.
Lloyd McKenzie (Nov 12 2018 at 13:00):
I hope the discussion recognized that MedicationRequest isn't only for prescriptions, but also for medication "plans" and that MedicationStatement is an assertion of what's believed to actually be happening - not what "should" be happening. I.e. If the patient has been prescribed a medication but has chosen to not take it, MedicationRequest (as a plan) would show they should be taking it, while MedicationStatement would show them as not taking it.
Last updated: Apr 12 2022 at 19:14 UTC
