Stream: implementers
Topic: Capturing patient-reported prescriptions
Morten Ernebjerg (Sep 17 2020 at 12:30):
Hi :wave: I am looking at a use case where patients can report medicines they were prescribed (independetly of what, if anything, was actually dispensed). That sounds like a MedicationReference MedicationRequest, but I saw the following in the documentation of that resource:
When reporting on the usage of a medication by a patient, the MedicationStatement resource should be used.
However, the reportedReference can point to a Patient to indicate how the information was reported.
My interpretation is that insofar the patient is reporting directly about a prescription, MedicationRequest with reportedReference pointing to the Patient is the right choice. Information about the medicine actually dispensed and taken would go in other resources, e.g. MedicationStatement. Is that understanding correct?
John Moehrke (Sep 17 2020 at 12:34):
@Jose Costa Teixeira @Rik Smithies @Lloyd McKenzie ?
John Moehrke (Sep 17 2020 at 12:36):
I am also very interested in the use-case, as it is one that I am working on as well. I have been using MedicatinoStatement only. Guided by the text on the medications page http://hl7.org/fhir/medications-module.html
Rik Smithies (Sep 17 2020 at 12:36):
I believe that is intended to be MedicationUsage, which is the new R5 name for what was called MedicationStatement in R4
John Moehrke (Sep 17 2020 at 12:37):
Rik, what about in R4?
Rik Smithies (Sep 17 2020 at 12:37):
-MedicationStatement in R4
John Moehrke (Sep 17 2020 at 12:42):
@Morten Ernebjerg what version of FHIR are you using? I looked around at R4, STU3, DSTU2, and current. I can't find a MedicationReference. did you mean Medication Resource?
Rik Smithies (Sep 17 2020 at 12:42):
both those are for anything that you want to record that the patient has taken, no matter who says it, but when you don't want to track any of the actual prescribe, dispense, administer cycle. But a system that does prescribe, would probably use MedicationUsage/Statement only (mostly) for the patient reported ones, because the clinical ones are covered by MedicationRequests.
Jose Costa Teixeira (Sep 17 2020 at 12:46):
MedicationStatement is for patients saying "i took this". That could be somewhat different from "i was prescribed this".
Morten Ernebjerg (Sep 17 2020 at 12:47):
@John Moehrke Oops typo, I meant _MedicationRequest_. I am mainly looking at R4 but would also be interested in the answer for STU3 (but R5 is not on the table).
Jose Costa Teixeira (Sep 17 2020 at 12:47):
MedicationRequest (with .reported=1) captures what was prescribed
Morten Ernebjerg (Sep 17 2020 at 12:49):
@Jose Costa Teixeira Right, exactly. I want to make it clear that the information does not necessaeily reflect what the patient was actually dispensed & is taking. I am not sure how onewould make that distinction clear in MedicationStatement.
Jose Costa Teixeira (Sep 17 2020 at 12:49):
So in your case I'd look at MedRequest
Melva Peters (Sep 17 2020 at 12:50):
You would use MedicationUsage (in R5) and MedicationStatement in previous versions. This can be used to report what the patient (or someone else) says they are taking (may be prescribed or not prescribed). MedicationRequest using the reported attribute can only be used if the intent is plan or proposal which doesn't really say "I was prescribed this", it says "I want to be prescribed this"
Jose Costa Teixeira (Sep 17 2020 at 12:50):
I had a use case @Melva Peters might remember: A patient says "i was prescribed this, dispensed this, but did not take it". -
Morten Ernebjerg (Sep 17 2020 at 12:50):
(deleted)
Jose Costa Teixeira (Sep 17 2020 at 12:51):
Melva Peters said:
MedicationRequest using the reported attribute can only be used if the intent is plan or proposal which doesn't really say "I was prescribed this", it says "I want to be prescribed this"
I was not aware of that constraint. Is is necessary (as it prevents this use case)?
Melva Peters (Sep 17 2020 at 12:52):
Check out medicationRequest.intent - in the comments
Jose Costa Teixeira (Sep 17 2020 at 12:52):
I think the use case is not
"i was taking, regardless of prescribed or not".
I think the use case is "it was prescribed, regardless of whether I took it or not"
Morten Ernebjerg (Sep 17 2020 at 12:53):
I am looking at the case where the patient says (1) I was prescribed X and (2) I was dispensed Y and am taking it - and I want to capture that as two pieces of information.
John Moehrke (Sep 17 2020 at 12:53):
but for normal patient reported medications... MedicationStatement is all that is needed.. right? This would be over-the-counter, prescribed elsewhere, self directed... right?
Melva Peters (Sep 17 2020 at 12:53):
Yes, @John Moehrke that is correct
John Moehrke (Sep 17 2020 at 12:54):
whew. I didn't want the basic flow to be overcome by an important but unusual case.. most patients will never notice that the drug they were dispensed was not what they were prescribed.
Jose Costa Teixeira (Sep 17 2020 at 12:54):
Thanks Melva. Can't find it MedRequest.intent comments:
"It is expected that the type of requester will be restricted for different stages of a MedicationRequest. For example, Proposals can be created by a patient, relatedPerson, Practitioner or Device. Plans can be created by Practitioners, Patients, RelatedPersons and Devices. Original orders can be created by a Practitioner only.
An instance-order is an instantiation of a request or order and may be used to populate Medication Administration Record.
This element is labeled as a modifier because the intent alters when and how the resource is actually applicable."
Melva Peters (Sep 17 2020 at 12:56):
What intent would you see being used for this use case? @Jose Costa Teixeira
Jose Costa Teixeira (Sep 17 2020 at 12:56):
John Moehrke said:
whew. I didn't want the basic flow to be overcome by an important but unusual case.. most patients will never notice that the drug they were dispensed was not what they were prescribed.
the case is not "i was prescribed X but I took Y". the case (at least how I read it) is : Iwas prescribed my antiepileptic / insulin but I did not take it.
John Moehrke (Sep 17 2020 at 12:57):
Im not arguing with you. I don't disagree. I just wanted the basic flow to be agreed before we go down a targeted use-case
Morten Ernebjerg (Sep 17 2020 at 12:57):
For me, prescribed X, got/taking Y is one case I am looking at.
Melva Peters (Sep 17 2020 at 12:58):
You could have a MedicationUsage with the the "takenAsOrdered" attribute = False
Morten Ernebjerg (Sep 17 2020 at 12:59):
But unfotrtunately not in R4, correct?
Jose Costa Teixeira (Sep 17 2020 at 12:59):
sorry @Morten.
For that, I'd say you use MedStatement for drug Y
Morten Ernebjerg (Sep 17 2020 at 13:01):
And MedRequest for X (R4/STU3)?
Jose Costa Teixeira (Sep 17 2020 at 13:01):
you can use something to say "but watch out, that was not was was prescribed" - I don't know if you'd use medStatement.derivedFrom? @Melva Peters
Jose Costa Teixeira (Sep 17 2020 at 13:03):
Melva Peters said:
What intent would you see being used for this use case? Jose Costa Teixeira
not crucial anymore but for "my" use case, I'd expect that the only intent is "order" because I am asserting it was ordered.
Morten Ernebjerg (Sep 17 2020 at 13:03):
But what about the other direction? I.e. saying here is smt. the patient said s/he was prescrided, but it may not be what they got in the end?
Morten Ernebjerg (Sep 17 2020 at 13:05):
(Will need to drop out for a while, btw)
Jose Costa Teixeira (Sep 17 2020 at 13:12):
Morten Ernebjerg said:
But what about the other direction? I.e. saying here is smt. the patient said s/he was prescrided, but it may not be what they got in the end?
that is a boundary case and is hard, especially if we need to map to the current resources 1:1.
Jose Costa Teixeira (Sep 17 2020 at 13:12):
IHE Pharmacy had some discussions on this, and there seems to be a pattern on how (good) systems address this mess.
Melva Peters (Sep 17 2020 at 14:01):
@Morten Ernebjerg I would suggest you add a questions or change request and describe your use case. The Pharmacy Work Group can discuss and see if we can make clarifications in the Current Build for this use case. That doesn't help with previous releases, but at least we'd add clarification for the current build. We have a discussion about MedicationUsage on Tuesday from 2-4pm Eastern at the upcoming HL7 Working Group Meeting. It would be great to include your use case in this as well.
Morten Ernebjerg (Sep 17 2020 at 15:30):
Hi @Melva Peters (I think you mentioned the wrong Morten, BTW) I will open a JIRA ticket describing my use case in some detail. Will that be enough information for the discussion? (not sure if I would be able to make the call)
Morten Ernebjerg (Sep 17 2020 at 15:39):
Thanks for all the input! As @John Moehrke pointed out, patients may often overlook differences between prescribed and dispensed medicines. For exactly that reason, my primary concern was to enforce a very clear distinction between the two (in patient-reported data) to avoid misunderstandings downstream. What I take away from this discussion is that correctly capturing patient-reported prescription is (currently) hard and should probably be avoided in favour of just collection information about the medicine being taken (in the form of MedicationStatement/MedicationUsage). Would you say that is a fair summary?
(And one bonus question, about partial capture of prescription-related info: Is there some standard way of attaching information along the lines of "This is what Dr. Johnson always prescribes to me" to MedicationStatement? - it' didn't see any)
Melva Peters (Sep 17 2020 at 15:54):
@Morten Ernebjerg yes, that will be great
Tom Stanis (Sep 17 2020 at 16:22):
Just throwing this out there, but another option would be to use Questionnaire.
Lloyd McKenzie (Sep 17 2020 at 16:42):
Questionnaire data wouldn't generally be queriable. I.e. You can't search for QuestionnaireResponses that indicate a patient has reported being on a particular med.
Morten Ernebjerg (Sep 17 2020 at 19:38):
@Melva Peters Here is ticket I opened - ended up a bit long but I wanted to outline the various lines of reasoning: FHIR-28575
Morten Ernebjerg (Sep 18 2020 at 06:11):
Would be very interested in hearing the outcome of the discussion!
Melva Peters (Sep 18 2020 at 13:26):
Thanks for creating the issue. Pharmacy will be reviewing issues next week during the Working Group Meeting - the issue will be updated based on that discussion.
Last updated: Apr 12 2022 at 19:14 UTC
