FHIR Chat · Group vs List. Should a Group be able to point to a List? · implementers

Stream: implementers

Topic: Group vs List. Should a Group be able to point to a List?


view this post on Zulip Isaac Vetter (Jun 14 2018 at 17:29):

Hi!

Could a characteristic of a group be the existence in a specific List?

I (think) I understand the subtle difference between a Group of patients and a List of patients -- a List is a "flat" collection of the same type of resource, a Group can either be a flat collection or a set of characteristics that the member of the group shares. A Group is treated as a whole, a List is a collection. Further, Group.type restricts the members to only a few types of resources; whereas, List. entry.item is Reference(Any).

The bulk data api uses Group and the quality MeasureReport resource uses List, -- which both seem pretty reasonable. Is it possible to unify these, such that a bulk data client could query for the information of patients in a quality measure ?

Isaac

p.s. I'm trying to incorporate @Kevin Weaver 's feedback on this bulk data PR.

view this post on Zulip Lloyd McKenzie (Jun 14 2018 at 17:50):

The other difference is that List is a collector but not an actor. Groups, on the other hand, can be the subject of events. (e.g. observations, medication administrations, research studies, etc.) So a Group is an enumerated and/or characteristic derived collection of entities that are intended to be the subject of one or more actions. A List is an enumerated currated collection of arbitrary entities that cannot directly be a subject of an action. I think MeasureReport is incorrect. It should be using Group because it's talking about the 'subject' of the report. The only reason I can think they might have gone with List is because they wanted to force enumeration. @Bryn Rhodes , any thoughts?

view this post on Zulip Bryn Rhodes (Jun 14 2018 at 17:56):

Yes, we wanted to force enumeration, and we didn't want the added complexity of having to represent the result with a resource that included mechanisms to describe the Group. Having said that, I'd be amenable to using the Group resource with language that in that context, it would be expected to be a strict enumeration Group (no definitional components). I think it would also make sense to extend that back to the Measure to say that Group could be used in a strictly definitional role to define the populations. Perhaps profiles of Group for GroupDefinition and GroupEnumeration?

view this post on Zulip Bryn Rhodes (Jun 14 2018 at 17:58):

Or even more strict, separate Group into GroupDefinition and Group?

view this post on Zulip Isaac Vetter (Jun 14 2018 at 18:00):

Either of these two approaches would work great for my specific question / use-case.

view this post on Zulip Isaac Vetter (Jun 14 2018 at 18:01):

I can see the appeal of moving Group.characteristic to a GroupDefinition resource, but wouldn't this further muddy the distinction between List and Group?

view this post on Zulip Isaac Vetter (Jun 14 2018 at 18:07):

Bryn, to ensure that this is more formally considered, should I submit a gforge issue against MeasureReport ?

view this post on Zulip Bryn Rhodes (Jun 14 2018 at 18:10):

Yes please, thank you!

view this post on Zulip Bryn Rhodes (Jun 14 2018 at 18:11):

As far as the distinction between List and Group, I don't think so, I agree with @Lloyd McKenzie there, Groups are intended to be groups of subjects, and it's such a ubiquitous concept, it makes sense that we'd have more to say about Groups of subjects, than we do about just Lists.

view this post on Zulip Isaac Vetter (Jun 14 2018 at 18:22):

https://gforge.hl7.org/gf/project/fhir/tracker/?action=TrackerItemEdit&tracker_item_id=17355

view this post on Zulip Lloyd McKenzie (Jun 14 2018 at 18:39):

I don't think Bryn was proposing a distinct resource, just a profile that constrained whether the enumeration had to be present or not

view this post on Zulip Bryn Rhodes (Jun 14 2018 at 19:51):

I did suggest it as an option. Seems worth considering to me since the distinction comes up in several different places already.

view this post on Zulip Bryn Rhodes (Jun 14 2018 at 19:52):

Especially since Group is partly definitional and I can't reference it in the same way I do other definitional resources.

view this post on Zulip Bryn Rhodes (Jun 14 2018 at 19:52):

Not a huge deal, but worth considering I think.

view this post on Zulip Lloyd McKenzie (Jun 14 2018 at 20:01):

It's actually an instance, even if you only describe the criteria. If you spray all the chickens in barn 5, you don't have to enumerate the chickens. It's still a specific group of individuals.

view this post on Zulip Grahame Grieve (Jun 14 2018 at 23:39):

I am not in favour of splitting Group into GroupList and GroupDefinition since you often mix both, and the general case, it's valid to have either. That's true for MeasureReport too. I understand that you might want to simplify for a particular case, but I don't see that as a reason for forcing a separation into 2 resources


Last updated: Apr 12 2022 at 19:14 UTC