Stream: implementers
Topic: Removing animal from Patient
Cooper Thompson (Jan 29 2018 at 23:25):
I'd like to solicit some input for gforge 14880 (move Patient.animal to standard extension). The challenges of Patient.animal are:
1. It is a modifier element, meaning implementers must inspect it. However I do not expect most implementers will do so. As I say in the tracker, doing the intuitive thing leads to being non-compliant with the spec. Which IMO means that the problem is in the spec.
2. One option is to make the animal property not a modifier, however it really should be, as it affects the interpretation of the rest of the resource. So that really leaves keeping it as a modifier, but moving it to a standard extension.
Grahame Grieve (Jan 30 2018 at 00:01):
then it would be a modifier extension. it would be less convenient for anyone that does use it, but make no difference otherwise
Brian Postlethwaite (Jan 30 2018 at 00:02):
Yes Cooper did mean a standard complex modifer extension
Lloyd McKenzie (Jan 30 2018 at 00:23):
The issue of people not checking modifier elements is a problem in general. We're quite clear that systems need to do so (or decide they don't care). Patient is only one of the places where this will occur - status and other things will suffer from the same problem. That's not, in itself, an argument to remove them.
Lloyd McKenzie (Jan 30 2018 at 00:24):
It being a modifier extension doesn't really make a difference - people can ignore that too. And it makes it much less clear that "yes, FHIR supports veterinary"
Eric Haas (Jan 30 2018 at 01:08):
"The issue of people not checking modifier elements is a problem in general. We're quite clear that systems need to do so (or decide they don't care)." To me this is evidence that the flag is useless because its so hard to interpret and apply. Maybe OK for profiles but I think doesn't add any value to the spec.
Eric Haas (Jan 30 2018 at 01:11):
BTW James and I suggested making an animal resource. Veterinary Practice software is about clients and patients and that might work better in the long run anyway
Eric Haas (Jan 30 2018 at 01:13):
James Case not James Herriott ;-)
Lloyd McKenzie (Jan 30 2018 at 02:07):
It's not "useless". It's important. If implementers ignore modifier elements, bad things will happen. The flag allows us to call out what implementers need to pay attention to so that it's clear for those who bother to look. The fact that not all implementers will adhere to the safety guidelines doesn't mean we should take away the safety guidelines or the elements they rely on.
Lloyd McKenzie (Jan 30 2018 at 02:08):
Not thrilled with adding an "animal" resource - animals need to appear everywhere Patient does (with the possible exception of a few "authored by/performed by situations). And they're not just relevant in veterinary. They matter in public health and research too.
Eric Haas (Jan 30 2018 at 02:33):
Even in research the animals don't need addresses and telephone numbers. just ids, species, breed, strain.
Lloyd McKenzie (Jan 30 2018 at 03:21):
Some animals actually do have URLs for telemetry purposes. They may have at least partial address information too, though agree that geographic boundaries are more typical.
Joginder Madra (Jan 30 2018 at 05:56):
I've struggled with this one...is there a definitive use case from an implementer that says we need to be able to deal with animal patients? Or are we designing for possibilities?
Richard Townley-O'Neill (Jan 30 2018 at 06:28):
How interested are we in implementers that are not part of the current FHIR community?
How interested should we be?
Richard Townley-O'Neill (Jan 30 2018 at 06:28):
I'm not troubled by including animal.
Joginder Madra (Jan 30 2018 at 06:46):
From the beginning, FHIR was viewed as a bottom-up, implementor driven activity so as to avoid some of the over-design pitfalls of v3.
Joginder Madra (Jan 30 2018 at 06:48):
...if that philosophy still holds, then I think this becomes a bit of a sanity check
Grahame Grieve (Jan 30 2018 at 09:48):
there's plenty of implementation of FHIR from contexts that includes animals. Most labs deal with animals too
Lloyd McKenzie (Jan 30 2018 at 13:07):
We don't have a lot of "veterinary EHR" systems at the table, but we certainly want FHIR to support them when/if they show up. Animals are relevant for many outpatient pharmacies. They're certainly relevant for clinical trials. They get monitored by public health agencies. It's not scope we want to give up - but it is reasonable to say "not yet well-exercised and thus this bit isn't normative yet".
Lloyd McKenzie (Jan 30 2018 at 13:09):
Most of the incentives for exchange are certainly on the human medicine side of the table. Being able to move Fluffy or Bessie's records from system to system is unlikely to have government incentives behind it in most jurisdictions. However, some degree of spill-over is certainly possible. And where we have intersection of human and animal health, they'll definitely need ot be supported.
Simone Heckmann (Jan 30 2018 at 13:32):
I don't think that moving "animal" to a standard extension supports veterinary any less that the way it is now.
Lloyd McKenzie (Jan 30 2018 at 14:13):
No, but it is a declaration of whether veterinary is considered to be a "core" part of FHIR. That was why we consiously included it in Patient - even though it's clear that 80% of systems in the healthcare space don't support animals
Cooper Thompson (Jan 30 2018 at 14:32):
@Grahame Grieve regarding lab workflows for animals, I was just looking at Specimen and realized that we may need to improve that a bit. There are a lot of specimen-based lab workflows that do not have a known patient. We have to maybe add Account links so the lab charges know where to go, but beside animals, there are a lot of other lab flows that should not go through patient (or Group/Substance/Device or anything else in Specimen.subject). So if this is a lab system specifically for veterinary, a Patient might make sense. But if it is a "human" centered lab, then they likely will want to treat animal specimens as non-patient samples.
Grahame Grieve (Jan 30 2018 at 14:34):
the lab I used to work in registered animals as patients, and also handled non-patient samples
Grahame Grieve (Jan 30 2018 at 14:34):
I'd have to revisit specimen with non-patient specimen in mind
Eric Haas (Jan 30 2018 at 16:06):
I think the pharmacy use case is very strong argument for keeping Fluffy in Patient - Thanks @Lloyd. YOu have me convinced it should stay...
Eric Haas (Jan 30 2018 at 16:08):
But then I think we need to look at is_mod ... is the juice worth the squeeze
Lloyd McKenzie (Jan 30 2018 at 20:50):
So - modifiers are a given in FHIR. Systems are expected to pay attention to them. When we get into certification, we'll absolutely beat the heck out of that point. If a system doesn't check to see if "animal" exists and accidentally end up including Fluffy in their opiod reporting statistics, it's their own fault. This particular modifier is a lot less risky to ignore than others. The biggest risk is if some system "corrects" the dose of a medication for Fluffy to be within normal human range, but I'm not aware of that happening. There's much more risk associated with "doNotPerform" and "status" on other resources.
John Silva (Jan 30 2018 at 22:13):
For many people the 'animal' (pet) is part of the family and may even be named in wills! ;-)
(yeh, my thoughts were "why is this in the 80% rule" -- and the same thoughts others expressed about not dealing with this. Are systems that do not deal with animals allowed to create their own profile that removes the animal property or is that non-compliant because of the isModifier flag?)
Brian Postlethwaite (Jan 30 2018 at 22:15):
My issue with is-modifier on patient.animal is that its not modifying the patient resource, its modifying how the other resources that reference is are used/interpreted. So its a level of indirection away, which is different to the other uses of isModifier i can think of.
Grahame Grieve (Jan 30 2018 at 22:16):
making an extension doesn't make any difference to that
Brian Postlethwaite (Jan 30 2018 at 22:17):
true, which kinda points if was extension, to make it a non-modifier.
Lloyd McKenzie (Jan 30 2018 at 22:21):
I don't think "animal" actually is a modifier. It doesn't change the meaning of any other element - with the possible exception of the contact information - which could shift in meaning from "how to contact the patient" to "how to contact those responsible for the patient". But frankly, for an infant human, the use wouldn't be any different from that of a cat.
Drew Torres (Jan 30 2018 at 22:22):
I concur with this assessment. The animal modifier doesn't make sense. It doesn't change any of the other attributes of patients. I brought this up in the meeting where PA felt moving animal to an extension would be appropriate.
Jenni Syed (Jan 30 2018 at 22:33):
I wondered about age interpretation/DOB. A 14 yo dog vs a 14 yo human are very different stages of life. HOWEVER that's a very different type of "modifier" than what I see on other resources
Jenni Syed (Jan 30 2018 at 22:35):
I think it feels weird because Patient is so core. If you interpret this patient as a person instead of an animal, all the regulation and clinical data around the patient probably doesn't mean the same thing...
Grahame Grieve (Jan 30 2018 at 22:36):
it's labelled as a modifier for this reason: if you care about whether the patient is human, you can't claim to be caught by surprise if it's populated
Grahame Grieve (Feb 09 2018 at 05:25):
Full write up from @Brian Postlethwaite (I subsetted this in my formal report):
Grahame Grieve (Feb 09 2018 at 05:25):
We have received a few items of feedback regarding the inclusion of the animal part of the Patient resource which the Patient Administration workgroup would like some additional feedback from the FHIR community prior to making a final decision for the normative definition of Patient.
There is no argument that the animal component is not part of the 80% (as per normal FHIR guidelines) and has been an intentional exception to the rule, the desire being to make a definitive statement that veterinary usage is a part of the scope of FHIR.
Most systems will either be all human, or all animal (veterinary), and not both. However systems like lab and pharmacy do handle both human and animal content.
So what’s the issue then?
The Patient resource is going normative, and this part of the resource has not been adequately tested/verified with real systems. Which is a part of the FMM progression scales.
The current backbone element is marked as a modifier element, which isn’t a modifier like anywhere else in FHIR. Other properties that are modifiers impact how to interpret the resource itself. For Patient, it impacts how you might interpret other resources.
Patient Administration is missing feedback from the many sides of veterinary – domestic and agricultural
The content we currently have appears to satisfy the domestic side of things (pets etc) but not clear how this would work with herds or exotic species.
The 2 trackers in question are:
14880 Move Patient.animal to a standard extension (Jan 2018 ballot)
14393 Patient.animal needs a meaning when absent
Discussion is already underway on chat.fhir.org:
https://chat.fhir.org/#narrow/stream/implementers/topic/Removing.20animal.20from.20Patient
Alternatives considered:
Leave alone
Change to a non-modifier element
Remove backbone element
Move backbone element to a complex extension
New Alternative suggested (at FMG):
Just include the species in the patient (no backbone)
The Patient Administration workgroup currently supports moving the existing backbone element into a complex extension (with same properties), and include a new section on the page describing the veterinary usage (attempting to satisfy the intent to declare its inclusion in fhir).
This would also then permit different forms of veterinary to define their own extensions for handling herds … exactly as they desire it to be.
At a minimum the workgroup proposes to remove the modifier from the backbone element.
Also note that PractitionerRole is to be included as an optional reference type in the Patient.generalPractitioner property (tracker #14224)
Simone Heckmann (Feb 09 2018 at 17:53):
+1 for removing animal and adding species. That'll make FHIR future-proof for the Post-First-Contact-Aera :vulcan_salute: :alien:
Brian Postlethwaite (Feb 09 2018 at 22:57):
If you've implemented handling of the animal property(s) please let us know, as this change will break things for you!
Eric Haas (Feb 09 2018 at 23:35):
That make
+1 for removing animal and adding species. That'll make FHIR future-proof for the Post-First-Contact-Aera :vulcan_salute: :alien:
That makes no sense for veterinary practice management software.
- species
- breed
- neutered
are key attributes to a patient at a minimum and extension for strain for research and production medicine. If we want to support Veterinary need to support it 100%.
Aaron Seib (Feb 11 2018 at 00:23):
These seems like a silly debate to me but I do have to add my 2 cents and suggest life form.
Live long and prosper.
Simone Heckmann (Feb 13 2018 at 18:48):
...as boolean?
Last updated: Apr 12 2022 at 19:14 UTC
