This also relates to what nowadays FHIR's LOCATION https://www.hl7.org/fhir/location.html and DEVICE https://www.hl7.org/fhir/device.html

Have you looked into the CONSTYS definitions related to this topic?

• PLACE seems to be something like Point of Care ; https://contsys.org/concept/point_of_care

• RESOURCE is another important entity when doing resource planning to execute healthcare services: https://contsys.org/concept/resource

For our Surgery module we created an Archetype to model healthcare resources: https://github.com/bjornna/dips-ckm/blob/nymaster/archetypes/cluster/openEHR-EHR-CLUSTER.healthcare_resource_dips.v1.adl (I should have it translated)

@Sebastian Iancu concretely that will work, but only if you know you have to do it; it's something that would be completely unexpected because Party-Party relationships don't normally include 'relationships' to passive objects or physical places, in organisational theory.

I completely agree that the information is important, I just think we should model it in a clear way. I have not yet tried to build a model proposal in UML, but I would suggest that PLACE (or whatever we want to call it) is a top level object type, like PARTY. PLACE would be able to contain PLACEs (not the same instances, obviously), and a PLACE would have various identifying information and other details ('purpose' for example).

BTW: in ontologies like BFO, there is quite a developed notion of 'Site', which is more or less a 'Place' as we are describing here, and which occupies a 'spatial region'. I'll look into the correct relationships between sites and sub-sites.

PARTY would have some generic optional relationships with PLACE, e.g. 'owns', 'uses' etc - we would need to think these through.

ORGANISATION, and sub-Organisation units would have more specific optional relationships with PLACE, e.g. 'facilities'.

PERSON would probably have some special relationships as well, including to represent 'being at' during a certain period of time.

I have some good modelling publications that address some of this - when I get a moment, I'll provide some pictures of possible models.

Thomas, running a query like that is not really a problem, just filter out based on relationship type.
I also see no reason for confusion, as you said a PLACE is a passive entity. Very often what is more important than whether or not somebody is admitted to an ORGANISATION is where is he/she physically located, or what physical (passive) resource does he/she consumes.
By volatile information do you refer to the burden having versioning/audit?

But fine, I see you see these differently - so, what is your conclusion for now? What should be done on this issue?

@Sebastian Iancu I don't agree that an ACTOR would have a contractual relationship with a PLACE - an ACTOR can just be in a PLACE, and that is his/her location for some time. A PLACE isn't an active Party, it's a passive entity. Again, I think we need to follow the ontological semantics here - the relationships are clearly different in that sense. Consider if you ran a query looking for all current relationships of a PERSON xyx, you would not expect to get back also connections to physical places, like their physical office or department building, only organisational / personal relationships.

In any case, I'm not sure that representing Actor current location inside the demographic system is a good idea, since this is likely to be very volatile information.

Also if PLACE is modelled as an ACTOR subtype, it's going to be confusing distinguishing between the ORGANISATION that a PERSON is admitted to, and the physical Location at the organisation site where they will stay (some PLACE, e.g. Ward 2, Building B).