Stream: implementers
Topic: Reuse of Extensions from the FHIR Registry
Stephen Royce (Dec 14 2016 at 23:50):
All, I have a need for a couple of extensions:
- a general one for notes/comments in (typically clinical) resources (that don't already have such, obviously!); and
- one for
StructureDefinitionsto indicate that they have been deprecated.
There exist a few "standard" extensions which meet my needs:questionnaireresponse-noteandstructuredefinition-annotationfor the first andcodesystem-deprecatedfor the the second. However, all are tied to resources that are not the same as my actual requirements. Is it okay to use them anyway? Should I just create my own? Should I be prompting someone to make more generic extensions available as standard?
Stephen Royce (Dec 14 2016 at 23:53):
I should point out that the use of a Boolean to indicate that a (version of a) StructureDefinition is deprecated is a simplistic, short-term solution to the more general problem of supersession of StructureDefinitions, either by other new StructureDefinitions or new versions of the same StructureDefinition.
Grahame Grieve (Dec 15 2016 at 00:32):
have you looked at StructureDefinition.status, at the extension http://hl7.org/fhir/StructureDefinition/structuredefinition-template-status? but deprecated applies to more than just CodeSystem - can generalise that
Grahame Grieve (Dec 15 2016 at 00:32):
comments is a bit different -where do you want to add comments, and why?
Lloyd McKenzie (Dec 15 2016 at 15:59):
@Stephen Royce You can't use an extension anywhere but the paths it's defined to be allowed. However, you can submit a change request for the scope of an extension to be broadened.
Stephen Royce (Dec 16 2016 at 02:08):
I have looked at the http://hl7.org/fhir/StructureDefinition/structuredefinition-template-status extension and we are using it in fact. However, I'm really cheating with wanting to use the deprecated extension; what I actually want to do is add a boolean to StructureDefinitions to indicate that they've been superseded so that I can write an XPath like fn:collection('...')/fhir:StructureDefinition[fhir:id/@value = '...' and not(<superseded>)] to make it easy to always get the most current iteration of any given StructureDefinition. So the status of theStructureDefinition would not actually be deprecated, it would merely indicate that it not the most recent version.
Stephen Royce (Dec 16 2016 at 02:11):
As far as comments go, our current models have an element for notes or comments that is often lacking in the equivalent FHIR resource, e.g. within Observation.component to allow a comment to be made about an individual component of an Observation that does not apply to the whole.
Stephen Royce (Dec 16 2016 at 02:14):
@Lloyd McKenzie By "the paths it's defined to be allowed," do you mean the Context of Use written in the documentation of the extension?
Lloyd McKenzie (Dec 16 2016 at 03:30):
yes
Eric Haas (Dec 20 2016 at 04:36):
Re comments on a component vs the whole obs is by design. Components are inseparable and looked at together so you would not expect a comment to apply to one part but to the whole. Consider using related or diagnostic report if they are separable individual observations. Otherwise you can make your own extensions.
Last updated: Apr 12 2022 at 19:14 UTC
