FHIR Chat · Discharge medication list · implementers

Stream: implementers

Topic: Discharge medication list


view this post on Zulip Floyd Eisenberg (Jul 10 2019 at 20:52):

I'm trying to determine the best resource to use to request a list of all medications on the hospital medication discharge list provided to the patient - i.e., some include new prescriptions, still others are to continue at home (either over-the-counter or based on prescriptions previously filled).

view this post on Zulip Jose Costa Teixeira (Jul 10 2019 at 22:27):

best resource to ask for the list, or the best resource to contain that list?

view this post on Zulip Floyd Eisenberg (Jul 11 2019 at 00:48):

Right - so how do I model that in FHIR R4? Is the list a MedicationStatement with a status = community [note that MedicationStatement.category references Medication status codes (one of which is "community" - http://hl7.org/fhir/medicationstatement-definitions.html#MedicationStatement.category] and assure the expression requests that MedicationStatement with dateAsserted as the time of discharge from the hospital?

view this post on Zulip Floyd Eisenberg (Jul 11 2019 at 00:48):

So what list would I request?

view this post on Zulip Floyd Eisenberg (Jul 11 2019 at 00:49):

I guess I would expect to ask for the "list" referencing the best resource to contain it - or am I missing something?

view this post on Zulip Melva Peters (Jul 11 2019 at 01:11):

If you are looking for orders for medications, then you should use the MedicationRequest resource. This is a list on discharge of all of the medications a patient is supposed to be taking.

view this post on Zulip Floyd Eisenberg (Jul 11 2019 at 01:17):

Agree - but the list I'm seeking contains meds that have been ordered to be taken as outpatient after discharge AND meds that are to be continued at home for which new orders are not being written (some of them are over-the-counter and don't require orders). So my first approach was to use MedicationRequest with Medication.Request.intent = plan or order - i.e., the "plan" indicates the meds on the list recommended to the patient and the "order" includes those with actual prescriptions - hence a full list of all meds with those criteria (and with category = community). Today I presented that option to our quality measure group and received considerable feedback that there has to be an easier way and that hospitals aren't necessarily implementing that way. The group wants something comparable to a CDA discharge medication list. Hence, my question seeking an alternative.

view this post on Zulip Richard Townley-O'Neill (Jul 11 2019 at 03:32):

HL7 AU has a profile for a medication list. http://build.fhir.org/ig/hl7au/au-fhir-base-stu3/StructureDefinition-au-medlist.html
Your list could be populated with MedicationRequests. The list resource could be created, curated and stored with a link to the discharge encounter.

view this post on Zulip Jose Costa Teixeira (Jul 11 2019 at 09:50):

I was thinking of it as a list but now I see this as a composition which contains
- medRequests (if that is what your requirements say you should keep),
- MedDispenses (if that is what your requirements say you should keep, and keeping the link to the Requests),
- MedAdmins(if your requirements say it is desirable and/or you can link those to requests/dispenses),
- MedStatements (to complement the others)
We should be looking at an IHE profile on this - later. This is my current preference from prior research and discussions

view this post on Zulip Jose Costa Teixeira (Jul 11 2019 at 09:51):

. I do have one open question: How would we capture the notion of a "treatment" as a glue between the different elements?
some treatments are known from the existence of a prescription, others are from the existence of a dispense, others from a statement.. if we do not have a "treatment" object, the reconciliation is harder.

view this post on Zulip Jose Costa Teixeira (Jul 11 2019 at 09:53):

this notion of "treatment" is not a common requirement in pharmacy workflows. Or rather, it gets confused with the prescription, this is why we abusively say "status of this prescription is complete" when we should say "status of the treatment initiated by this prescription is complete".

view this post on Zulip Jose Costa Teixeira (Jul 11 2019 at 09:55):

going back to the list: Normally any Medication List would have some need for authorship / context / coherence so it is more in the space of Composition than List IMO.

view this post on Zulip Jose Costa Teixeira (Jul 11 2019 at 09:55):

Discharge Medication list sounds like an example of something that has an author and clinical context and validity.

view this post on Zulip Richard Townley-O'Neill (Jul 12 2019 at 05:52):

@Jose Costa Teixeira
Sometimes people want a list as a document and sometimes as a list. We are working on a design for a medication list resource that uses Medication.source as author and can be included in various documents, including a medication list document. Using List allows recording information about changes.
See https://github.com/AuDigitalHealth/ci-fhir-stu3/tree/master/output/SharedMedicinesList

view this post on Zulip Jose Costa Teixeira (Jul 12 2019 at 07:28):

in IHE we had some discussions, perhaps complementing this. here are some points:
1. We expect there to be a process for reconciliation and producing a "formal medication list". if you are performing that process, your input is more likely a list, output more likely a document/composition.

view this post on Zulip Jose Costa Teixeira (Jul 12 2019 at 07:28):

2. "Medication.source"? You mean "List.source"?
in the cases where we want composition, composition.attester should be equivalent. I would like to see device as a possible attester in composition.

view this post on Zulip Jose Costa Teixeira (Jul 12 2019 at 07:34):

3. Indeed it seems List is richer in terms of change tracking. And Composition is more about the "finished" document. i think there may be a middle ground there to explore

view this post on Zulip Jose Costa Teixeira (Jul 12 2019 at 08:54):

btw, @Richard Townley-O'Neill I like that IG. which template / framework did you use? Lloyd's ? Eric's?

view this post on Zulip Lloyd McKenzie (Jul 14 2019 at 03:13):

Composite is always an option for sharing data, but should never be a requirement

view this post on Zulip Richard Townley-O'Neill (Jul 15 2019 at 00:14):

@Jose Costa Teixeira

2. "Medication.source"? You mean "List.source"?

Yes I meant List.source :embarrassed:

which template / framework did you use?

We started from Eric's Healthdata1 IG-Template.

view this post on Zulip Jose Costa Teixeira (Jul 17 2019 at 05:40):

Back to @Floyd Eisenberg :
1. if you want something like a "CDA discharge medication list" then you are picking a flavour of a list - and flavours are context-dependent and may not be that interoperable. The International Patient summary may have another way to form such a list, and many countries have their own rules for that. For example if you make a list based on prescribed meds, and I make a list with actually dispensed meds, they may match or not.
2. For this reason, in IHE we have a Medication List transaction which is just "get the raw data, I will filter and reconcile on my side".
As long as there is no universal flavour of "what should be in a medicatino list" you must prepare for the fact the only semantically solid ground is the "ingredients" - prescription, dispense... and these have different meanings.
If you want to see how different those medication lists are, here is one overview:
https://www.nist.gov/sites/default/files/nist_ncpdp_medication_rec_and_stds_overview.pdf

view this post on Zulip Jose Costa Teixeira (Jul 17 2019 at 05:40):

If you want to make a list, perhaps start checking the International Patient Summary.

view this post on Zulip Jose Costa Teixeira (Jul 17 2019 at 05:41):

IHE will be implementing this, once the FHIR resources are a bit more stable


Last updated: Apr 12 2022 at 19:14 UTC