Stream: social
Topic: Experian unique pt identifier??
Dave deBronkart (Jan 15 2020 at 14:01):
Not sure where this should go so I'll put it here for the moment: what do we make of this?
My friend @DrNic1 tweeted on Dec 24 that the patient can't know or see the identifier. What the heck?
So how exactly does this work…? @Experian has assigned everyone a "Universal Patient Identifier" but you can't see it or know it #Interoperability http://bit.ly/2ZaMVYI
- Nick van Terheyden, MD - "Dr Nick" (@drnic1) John Moehrke (Jan 15 2020 at 14:07):
That would be a cross-reference index value. Not unusual, but also not a "solution". Within the IHE XDS (Cross-Enterprise Document Sharing) there is an "Affinity Domain Identifier" that is a business identifier for each unique patient within the network, with a Patient Identity Cross-Reference (PIX) service that converts any local domain patient identity into that Affinity Domain Identifier. This value is just an identifier, not a set of attributes.
This same concept is often in the background of many centeralized, decenteralized, and distributed systems.
Dave deBronkart (Jan 15 2020 at 14:09):
This same concept is often in the background of many centeralized, decenteralized, and distributed systems.
So there's actually nothing innovative about this, except in the private little world of those two players?
David Pyke (Jan 15 2020 at 14:30):
Yep, they've created their own private "universal" MRN
Dave deBronkart (Jan 15 2020 at 14:41):
Is there a better place on Zulip to discuss this? I searched "patient identifier" and got nothing but this.
Dave deBronkart (Jan 15 2020 at 14:51):
Are Experian or NCPDP (specialty pharmaceuticals) involved with FHIR at all? I've done speeches for each of them ... might be able to reach out to them to align everyone's public messaging.
From the thread here, I have the impression (from @John Moehrke) that they've announced THEIR own unique identifiers, which are literally of no use to anyone else. Is that correct? If so, all they've announced is the equivalent of a patient number in their ecosystem, or a MRN (medical records number) in a hospital - not "universal" in any way. Correct?
Does HL7 have a formal definition of what a unique pt identifier would be? Seems to me this is a legit occasion to speak out, because healthcare Twitter's already confused by this.
John Moehrke (Jan 15 2020 at 15:02):
hard to fight marketing like Experian... It is likely best to let them stumble on their own feet
René Spronk (Jan 15 2020 at 15:17):
'uniqueness' of an ID is always within a certain context: a medical device, a provider organisation, a country. The latter is the best one can hope for right now. A truly universal unique Id would be something like a hash-code of your DNA.
Many countries that mandate the use of ID cards for absolutely everthing have country level unique citizen ids. The identifier itself is pretty useless without the ID card. In the US, where the system of common law doesn't mandate the use of official IDs (not to the same degree as in most napoleonic/roman law systems), it's difficult to create a nationwide patient identifier without backing it up with an official ID card. No ID cards means the number itself will have a value, much like a US SSN.
Disclaimer: I'm from a country with napoleonic law - so I'm biased in favor of unique national identifiers and the near universal use of ID cards.
Lloyd McKenzie (Jan 15 2020 at 16:47):
Even the DNA changes... :)
David Pyke (Jan 15 2020 at 16:58):
So, we've all agreed that a Universal ID is impossible, therefor we should stick with national IDs and create eMPIs for them with the WHO holding the eMPI servers.
John Moehrke (Jan 15 2020 at 17:01):
blockchain
Gino Canessa (Jan 15 2020 at 17:15):
...or keep pushing forward with Distributed Identities and make something that patients can use to maintain their own identity.
Dave deBronkart (Jan 15 2020 at 21:50):
Even the DNA changes... :slight_smile:
Yeah, it kinda fried my brain when I learned that a few years ago. What implications. So much for certainty! (yet again)
Dave deBronkart (Jan 15 2020 at 21:52):
So, we've all agreed that a Universal ID is impossible, therefor we should stick with national IDs and create eMPIs for them with the WHO holding the eMPI servers.
Does that "so" denote snark, or are you summarizing something that the FHIR community decided a few years ago, for real?
Grahame Grieve (Jan 15 2020 at 22:53):
HL7 has no authority for a decision like that nor could we come to agreement about that. So I think it's just a general observation about how people are acting
Dave deBronkart (Jan 16 2020 at 16:08):
it's just a general observation about how people are acting
Perfect - that's what I would have guessed. So, @David Pyke, iIs this an accurate paraphrase of your meaning?
"The world seems to have agreed ...
* ... that a universal / reliably unique patient ID is impossible,
* ... so let's stick with unique IDs within countries ...
* ... and create eMPIs between them (eMPI = mapping tables, right?)
* ... with WHO holding the mapping table servers
Or does "the world seems to have agreed" apply only to the first bullet, and the rest is YOUR view? :-)
David Pyke (Jan 16 2020 at 16:11):
It's just my (snarky) view. The problem with going past national borders would be huge and isn't in any way within HL7's ability to weigh in on. While some countries have national and/or regional IDs, there is no consensus (or discussion that I've seen) about going to an international level ID
Dave deBronkart (Jan 16 2020 at 16:20):
So, anyway, to wrap up this thread - it appears that what Experian and NCPDP announced is nothing "universal" at all - it really is just their own private "customer number," figuratively speaking, and is of no value to anyone else because only they can see it.
So it seems. If anyone finds out anything to the contrary, please let us/me know!
René Spronk (Jan 16 2020 at 16:49):
If we look at the EU, then the patient identifier used for cross border communications is (essentially) the tuple (country code, national [single country] patient id). Healthcare systems have a tendency to assign their own identifiers anyway, next to [hopefully] storing the [other country's] identifier.
In countries like Estonia and Norway a person is assigned a citizen identifier at birth - it's created by a national webservice, once a healthcare provider reports a new birth.
Jens Villadsen (Jan 16 2020 at 18:29):
Same story for Denmark as for Norway
Jens Villadsen (Jan 16 2020 at 18:36):
in EU eID is comming up
Jens Villadsen (Jan 16 2020 at 18:36):
https://ec.europa.eu/digital-single-market/en/e-identification
Jens Villadsen (Jan 16 2020 at 18:42):
@Dave deBronkart unless the Identifier from NCPDP & Experian Health is used and carried through interoperating systems, the identifier in itself is close to worthless.
Brendan Keeler (Jan 20 2020 at 17:26):
I expect Experian would productize such an ID for health systems and vendors to use. We're constantly asked for such a service by patient facing applications.
Dave deBronkart (Jan 20 2020 at 18:59):
Dave deBronkart unless the Identifier from NCPDP & Experian Health is used and carried through interoperating systems, the identifier in itself is close to worthless.
That was my point, exactly, higher in the thread. It's a bit disingenuous to brag about it being "universal" if it's not universally available - a market reality that's reflected by how many people responded on Twitter & LinkedIn saying "Wow, this will really change things!"
Last updated: Apr 12 2022 at 19:14 UTC
