Stream: implementers
Topic: Meaning of MedicationStatement/-Usage about future
Morten Ernebjerg (Feb 16 2021 at 13:15):
I know that MedicationStatement/-Usage and MedicationRequest have a complex relationship and some overlaps. One thing, however, that I used to think was a key difference was this: a MedicationStatement is alway information (albeit possibly imprecise) about someone actually taking/having taken a concrete medical product, whereas MedicationRequest is about ordered medication that is not necessarily yet "instantiated" as a concrete medical product (and may be instantiated by a pharmacist as a different medical product than is on the prescription). But I just realized that the spec explicitly says that a MedicationStatement/-Usage can also be used to say that "the patient [...] will be taking the medication in the future" (not sure how I missed that :see_no_evil: ). In which case, of course, the concrete medical product is unknown.
My question is then: How do we interpret a MedicationStatement about the future & what use cases are targeted? In particular,
- what is supposed to be covered by the phrasing "will be taking" (i.e. what kind of certainty/evidence)?
- how can we retrospectively know this was a statement about the future? E.g. must
dateAssertedthen always be filled with a date beforeeffectiveDateTime/effectivePeriod.end? The latter would not allow encoding "the patient will be taking this at some unknown point in the future".
A concrete use case is this: The International Patient Summary (IPS) lists current medications using MedicationStatements. Suppose now that a patient is given a prescription, e.g. on being discharged from a hospital. One could perhaps argue that when the corresponding MedicationRequests is generated, one could also quite automatically generate a future MedicationStatement by interpreting the prescriptions as a statement about medications that patient will (presumably) be taking later. Would that be legal, assuming the "futureness" can be recognized? I know such an automatic conversion is not possible retrospectively because the patient may have gotten a different medication than what was prescribed (or may not have filled the prescription at all).
Lloyd McKenzie (Feb 16 2021 at 15:07):
Even if known, the Medication Statement could still be at a high level and say something like "Patient is taking an ACE Inhibitor".
I'm a bit surprised to see the bit about "will be taking the medication in the future" because that seems to be a clear overlap with MedicationRequest (which is about intent) while MedicationStatement is supposed to be 'event'. @Melva Peters @John Hatem ?
Peter Jordan (Feb 16 2021 at 19:40):
This (future) use case appears to be covered by the MedicationStatement.status value of 'intended' ('The medication may be taken at some time in the future.'). From a practical viewpoint, I would be reluctant to include a medication with that status unless supported by a MedicationDIspense event, but much will depend on the completeness of the system used to generate a MedicationStatement.
Currently in NZ, hospital discharge prescriptions, and related dispensing, will only be recorded on our National ePrescription Service, if the dispensing is done at a community pharmacy. Otherwise, what occurs in secondary care facilities is only available on a patient's primary (master) record in the form of a discharge summary PDF and won't be included on a Medication Statement.
Melva Peters (Feb 17 2021 at 20:57):
@Lloyd McKenzie when a patient tells a provider that "I'm not taking this now, but I intend to take it in the future" - for example, antihistamines.
Lloyd McKenzie (Feb 17 2021 at 23:58):
It's messy. Because they could put that same thing in a MedicationRequest with intent of 'plan' and it would mean exactly the same thing. What guidance do you have on when to use one approach vs. the other?
Jean Duteau (Feb 18 2021 at 00:00):
we are working on the guidance of the overlap. we have a Jira issue exactly about this
Lloyd McKenzie (Feb 18 2021 at 00:03):
It would help if the guidance reflected that MedicationUsage reflects an "event" (what is happening) while MedicationRequest emphasizes "intent".
Peter Jordan (Feb 18 2021 at 03:32):
Melva Peters said:
when a patient tells a provider that "I'm not taking this now, but I intend to take it in the future" - for example, antihistamines.
I'd only include that in a MedicationStatement if the medication has actually been supplied to the patient. Countless patients tell their doctors one thing and do another...not that I'm one of them. :smiley:
Lloyd McKenzie (Feb 18 2021 at 03:59):
Depends whether it's prescription med...
Jean Duteau (Feb 18 2021 at 04:03):
In one jurisdiction in Canada (maybe more), we use MedicationStatement for recording over-the-counter meds and supplements that the patient tells us they are taking. "I am taking and will continue taking Vitamin D"
Lloyd McKenzie (Feb 18 2021 at 04:16):
Currently taking and expectations for continuation is fine. It's the "not yet taking but expect to" that gives two different ways to say the same thing.
Peter Jordan (Feb 18 2021 at 05:37):
Lloyd McKenzie said:
Depends whether it's prescription med...
I'm not so sure. Sometimes the only difference is the dose strength, e.g. you can purchase lower strength diclofenac sodium tablets OTC in NZ. Take 2 at once and you're effectively self-administering a 'prescription' strength dose.
Melva Peters (Feb 18 2021 at 14:39):
There are medications that may be prescribed or may be purchased over the counter that the patient only takes once a year. For example seasonal allergy medications. They may not be taking now, but they will be taking in a few months when it is allergy season. The use of MedicationRequest for recording these isn't the typical workflow in my experience. They are recorded as MedicationStatements.
Lloyd McKenzie (Feb 18 2021 at 15:14):
Ok, so capturing in MedicationStatement isn't for "I've never taken or been prescribed but expect to start" but rather "It's prescribed and I'm supposed to be on it, but that hasn't quite begun yet" or "I take it sometimes, not at the moment, but I will again soon-ish". Is that accurate?
Melva Peters (Feb 18 2021 at 16:02):
Yes
Melva Peters (Feb 18 2021 at 16:03):
Although I expect some would argue for "I have never taken", but I wouldn't use it for that
Lloyd McKenzie (Feb 18 2021 at 16:26):
"I have never taken" is fine. That's also 'event' :)
Morten Ernebjerg (Feb 23 2021 at 08:16):
[Slightly delayed here, was caught up elsewhere!] Thanks you for all the input! I can see there are definitely sensible future use cases, but the exact distinctions are not easy. E.g. in the table listing the differences between the medication resources, the grammatical tense seems to suggests only backward-looking usage (my emphasis):
A medication statement is not a part of the prescribe->dispense->administer sequence but is a report that such a sequence (or at least a part of it) did take place, resulting in a belief that the patient has received a particular medication.
I think a further point that makes the overlap confusing is that a MedicationRequest can be reported by a patient (informationSource can point to a Patient). @Melva Peters What is the use case covered by a patient-reported MedicationRequest that is not covered by MedicationUsage/-Statement?
Picking up @Lloyd McKenzie's example of a possible future use of MedicationStatement:
"It's prescribed and I'm supposed to be on it, but that hasn't quite begun yet",
It seems to me that although this is not equivalent to a MedicationRequest, this statement can be derived from a MedicationRequest (with intent e.g. "active"), insofar as the latter expresses the intent (on part of the prescriber) that the patient will/should be taking a medication. Specifically, suppose I have a MedicationRequest with the following properties
status= "active" (or "completed")intent= "order"requester= Doctor XauthoredOn= date Dmedication= medication M
Then it seems one can automatically derive a MediationStatement/-Usage with the following properties:
status= "intended"basedOn= [medication request above]dateAsserted= date DinformationSource= Doctor Xmedication= medication M
Or would the latter incorrectly imply something that is not implied by the MedicationRequest? (Again, for usage context, the latter fits right into IPS, the former doesn't).
Last updated: Apr 12 2022 at 19:14 UTC
