Stream: implementers
Topic: Representing non-medication administrations
Lloyd McKenzie (Sep 14 2020 at 17:07):
There are a number of clinical processes that involve the notion of dosage, rate, maxDose, etc. that don't involve administering actual medications. For example, radiation therapy, photo-therapy, blood transfusions, etc. We will need to be able to represent both orders as well as events. There are three possibilities for representing these using FHIR:
- leverage the existing MedicationRequest and MedicationAdministration resources and revamp the scope to allow capture of administration of 'substances' other than medications. This would allow leveraging the existing elements that deal with dose, rate, etc.
- leverage the ServiceRequest and Procedure resources and add new core elements or extensions to support capturing concept such as 'dose', 'rate', etc.
- define a new set of resources
In general, we usually drive such decisions based on "What do systems currently use to represent these concepts?" - so for systems that handle ordering and retaining records of radiation, photo-therapy, transfusions, etc, would you generally use the same structures/interfaces you use for ordering and resulting meds, for ordering and resulting imaging/labs, or something else?
NOTE: Landing this is important to allowing ServiceRequest, Procedure and the Medication resources to progress to normative.
Hans Buitendijk (Sep 23 2020 at 15:11):
We should consider in this context the full flow of request (what dose/rate), actual dose (now it can involve x-ray, thus not specifically ordered, vs. therapy), and dose registry. IHE is working on an effort addressing dose summary and registry.
Jose Costa Teixeira (Sep 23 2020 at 15:26):
@Lloyd McKenzie can you please change the title of this stream by removing the last '? It messes up linking to this stream
Hans Buitendijk (Sep 23 2020 at 15:40):
Can you do that?
Jose Costa Teixeira (Sep 23 2020 at 15:44):
I think you can change the name of the topics you started
Lloyd McKenzie (Sep 23 2020 at 15:46):
Done
Rob Dingwell (Oct 28 2020 at 19:38):
@Lloyd McKenzie is there any intent on providing guidance on this subject for previous/current versions of FHIR? Coming up with a path forward is great and something I am interested in from a Radiation Therapy perspective but not having a means to deal with this for previous and current versions of FHIR is problematic because systems do not necessarily keep in step with the latest. There are still some significantly sized EHRs out there that are still DSTU2. Going back that far may be a stretch but guidance for R4 would certainly be useful.
Lloyd McKenzie (Oct 29 2020 at 03:44):
In general, we don't backport guidance into historical releases as it's easy enough for implementers to look for guidance in the current release. By and large we can't add capabilities to past releases, though there's a standard extension mechanism that allows you to adopt data elements introduced in future releases in historical ones. In any event, the first step is to land on what our guidance is for the R5 release - and I think we've got a good way to go there yet
Lin Zhang (Oct 29 2020 at 04:34):
Substance administration would be broader than medication administration.
Lin Zhang (Oct 29 2020 at 04:35):
X-ray is a kind of energy/force...
Lloyd McKenzie (Oct 29 2020 at 04:42):
Agree. We need to decide what resource to use for it though. And our options are listed at the top of the thread. The decision of which of those options is most appropriate should be guided by implementers sharing what would work best for how they currently represent such information in their systems. Unfortunately, no one's done that yet. So badgering time :)
@Jenni Syed @Michael Donnelly @Jeffrey Danford - not sure if you folks are the right individuals from your respective organizations - feel free to draw others into the conversation as appropriate. Given how you currently capture such things as radiation therapy, photo therapy, shock therapy and other 'administrations' where the notion of 'dose' and 'rate' come into play, but where we're not talking about medications, what would make the most sense from a resource perspective - broaden the Medication[x] resources to encompass these concepts; extend Procedure to capture dose & rate-related elements; define an additional set of resources to cover the space; something else?? (The use-case would need to cover both the Request side and the Administration side - presume we don't really need to worry about 'summary' or 'dispense' concepts here.
Elliot Silver (Oct 29 2020 at 05:27):
Please loop @Abderrazek Boufahja and @Jonathan Whitby in for discussion of ionizing radiation dose. From my limited understanding, this is a complex subject. @Kevin O'Donnell may have an opinion (my guess is, it will be "get the medical physicists involved"). There may also be overlap between this discussion and an earlier one about maximum patient dosage ("how much more radiation therapy can we give the patient?")
Sharon Sebastian (Oct 29 2020 at 14:01):
The CodeX Radiation Therapy Treatment Data for Cancer (RTTD) use case is working to express a summary of a patient's radiation oncology (RO) treatments. As we consider how to model radiation-related concepts in FHIR, it will be important to consider the entire RO workflow and how each step in the RO process will fit together.
Lin Zhang (Oct 29 2020 at 23:56):
FYI, physical force may be used to model/cover such “non-materials”.
Daniel Rutz (Nov 23 2020 at 22:01):
@Lloyd McKenzie , we do actually need to model the 'dispense' actions for sure in the blood use cases. The far more common (currently/v2) interfaced flows we see are between EMR and Blood Bank LIS, where the ordering provider requests X units of various blood products and there's a status progression of (more or less) Selected, Issued, Dispensed, Received, Transfusing, Transfused for each individual & discretely represented product (with its own identifiers).
Daniel Rutz (Nov 24 2020 at 16:00):
Also, I'm reminded that we have talked about this several times in PC discussions for several reasons, but ensuring that AdverseEvent can correctly reference it (for blood transfusion reactions, or presumably similar events with radiation and other therapies) is another component of the use case. Tagging @Michelle (Moseman) Miller , @Rob Hausam , @Jay Lyle , @Stephen Chu , @Emma Jones (I'm sure there were others too, sorry) for that part.
Emma Jones (Nov 25 2020 at 17:28):
Back to Lloyds original comment about defining a new set of resource as an option.
Nutrition is another entity that can be ordered/requested and administered. NutritionOrder is its own resource. What determines if a resource is needed vs use of an existing resource?
Lloyd McKenzie (Nov 25 2020 at 21:00):
@Daniel Rutz My point is that the 'dispense' wouldn't typically capture administration information. You don't 'dispense' radiation or blood with specific administration instructions. As such, the simple 'supply' resource should presumably work fine.
Lloyd McKenzie (Nov 26 2020 at 00:09):
The decision to use a new resource is driven by a few things:
- how similar are the elements in the different use-cases
- do some existing systems typically capture the data from both use-cases in the same database table and/or with the same user interfaces
- can existing systems differentiate between the use-cases or would they be unable to decide which resource to use to represent the data
- when referencing the resource instances, is there commonly going to be a need to include one set and exclude the other set?
Rob Dingwell (Dec 15 2020 at 12:54):
@Lloyd McKenzie Has been any additional thought on this? Asking because I noticed there was a request in JIRA for a new project to related radiation dosing. https://jira.hl7.org/browse/PSS-1678 If there is going to be a larger effort to support other types of administrations that required dosing type information having 2 efforts that overlap may produce competing ways to represent the same thing.
Lloyd McKenzie (Dec 15 2020 at 14:04):
I was actually happy to see that project as I was hoping they might land this question. I'm not aware of another HL7 project?
Rob Dingwell (Dec 15 2020 at 14:11):
Wasn't thinking of another specific project, just the general notion that you put out with regard to thinking about this more generally. Was not sure if you had started on something or not but if that is a direction that the overall FHIR spec was thinking of taking then the radiation dosing project may be a means to start defining that
Rob Dingwell (Dec 15 2020 at 14:16):
Actually, looking at their confluence page they are not even looking to define dosing at an administration level. It looks like they want to create an Observation profile that is a summary of information about the radiation therapy that has been given.
Last updated: Apr 12 2022 at 19:14 UTC
