Stream: Covid-19 Response
Topic: What's the official HSLOC URL
Keith Boone (Sep 04 2020 at 21:09):
VSAC has it as http://cts.nlm.nih.gov/fhir/cs/hsloc, VADS as http://cdc.gov/nhsn, I think it's actually http://cdc.gov/hsloc, and the actual URL where it is defined is https://www.cdc.gov/nhsn/cdaportal/terminology/codesystem/hsloc.html (which would also work for me).
Keith Boone (Sep 04 2020 at 21:10):
@Austin Kreisler Seems like this should be NHSN's call as to what it actually is. Until they make a clear statement, the only thing everyone agrees on is the OID.
Austin Kreisler (Sep 04 2020 at 21:25):
NHSN went through a lot of effort to make the last URL (https://www.cdc.gov/nhsn/cdaportal/terminology/codesystem/hsloc.html) our official source of truth for the HSLOC code system. After the fact we found out that HL7 Vocab doesn't like it for a FHIR vocab URL.
Keith Boone (Sep 04 2020 at 23:11):
I honestly don't see what's wrong with it, NOR do I believe it's their choice. It meets the specification for a URL.
Josh Mandel (Sep 04 2020 at 23:13):
incidentally what does it mean for them not to like it? Is this an aesthetic judgment or does it violate some rule of thumb?
Keith Boone (Sep 04 2020 at 23:48):
And it does exactly what a CodeSystem URL should do.
Keith Boone (Sep 04 2020 at 23:50):
And given that's the intent of the publisher of the code system for the FHIR URL, that's the ONE I'm going to use for connectathon. I would also suggest that be documented in the page that defines the code system so that nobody else (HL7, VADS, VSAC or otherwise) is ever confused.
Lloyd McKenzie (Sep 05 2020 at 02:12):
I expect one of our issues is the ".html" on the end - because the content that's retrieved when you use that resource as a canonical could be a variety of things - XML, JSON, RDF, HTML or possibly other stuff.
Keith Boone (Sep 05 2020 at 02:37):
Not everyone has the luxury of controlling their corporate server's redirects. I think we need to live with the limitations of the real world and not impose our hoped for ideals on organizations whose IT departments don't necessarily see eye to eye with us.
Keith Boone (Sep 05 2020 at 02:40):
And using that value would sure beat the heck out of having three different published values which AREN'T what the publisher of the vocabulary intended being published by systems which are supposed to be able to be relied upon by standards implementers in the US (HL7, VSAC and PHIN VADS).
Robert McClure (Sep 09 2020 at 22:57):
@Keith Boone So you don't think @Lloyd McKenzie point is worth pushing for? The rule, as noted in HTA's requirements are clear. And yes, if .html for a jason artifact is not worthy of discussion for you I'm a bit surprised. All said, you are right, we have to take what they force us to take.
Keith Boone (Sep 10 2020 at 00:24):
I think HL7's Advice and guidance is perfectly fine, and it's recomendation that It should also not be specific to the format of the page. Having said that, these ARE guidelines, not requirements, and having to navigate through the IT bowels of an organization that employs more than 20,000 people to get them to change the processes that they use for their organizational web presence is NOT something that you want to wade through to address a 6 month or more delay in publication of a FHIR endpoint URL. Yeah, this is all easy web stuff, but finding the right person in IT who UNDERSTANDS THAT for an organization that large isn't. Been there, done that. It took a team of 5 many years to get that fixed at a former employer.
At the present time, the "URL" for HSLOC can at best be represented by an OID, even though CDC staff and contractors did everything they thought they needed to in order to create an appropriate URL. It meets the general goals of FHIR, and this seems to me to be a case of perfection being the enemy of the good. At this point, both HL7 AND VSAC have published TWO DIFFERENT FHIR URLS for this vocabulary, and NEITHER of them is the publisher of it, nor comes anywhere close to the publisher's intent. That situation is even worse.
John Moehrke (Sep 10 2020 at 15:07):
long live the OID
Last updated: Apr 12 2022 at 19:14 UTC
