Page tree
Skip to end of metadata
Go to start of metadata

From: Robert McClure MD <>
Sent: Wednesday, April 29, 2020 12:23 PM
To: Brett Marquard <>; Davera Gabriel <>
Cc: Austin Kreisler <>; Calvin E. Beebe <>; Eric Haas DVM, MS <>;;; Gay Dolin MSN RN <>;; Julie James <>; Jean Duteau <>; Floyd Eisenberg MD, MPH, FACP <>
Subject: Re: Utilization of ISO codes for Preferred Language: BCP 47 in USCDI FHIR IG


Nope, still not sorted out. I will note that the language codes and use of the syntax and process defined in BCP 47 (RFC 5646) has open access IP (I’d have to track down the reference for that) meaning we don’t need to get ISO permission to use it. I don’t know what Davera means by “nested reference” but I’ll pick up the “syntax” reference because this (and UCUM) are two of the code systems that are problem children in that they are not a typical code system and instead are a set of grammatical elements plus the syntax (process) to be used to construct valid codes. As such we can certainly define specific code sets (value sets) when a restricted known set is needed. But we have a problem when we simply want an element bound to any valid code. FHIR accomplishes this in a backhand way via the implicit ValueSet that leaves implementation unclear. C-CDA and all other HL7 standards can do essentially the same thing by doing an “all codes” CDL. Problem is actually implementing this using the standard approach of having a “current static list” creates an incomplete solution that is known to break. 

I’m not sure this is an HTA issue, but I’d like to see us formally clarify that for code systems that function like this, when a value set that is “all codes” is bound to an element, the best-practice implementation would be to not create a static value set member list but instead should be implemented as validation operation, presumably with a known common set of valid codes for selection and display. I’m not sure if that means we don’t bind to a value set, or (perhaps better) somehow we clarify that implementing that “all codes” CLD is done differently in these situations. 

. . .

All content represents my view only

Robert McClure MD : President, MD Partners, Inc.

303.926.6771 :

On Apr 29, 2020, at 5:39 AM, Brett Marquard <> wrote:

Good morning Davera, 

Thanks for reaching out.

A few quick items:

  • The US FHIR Core supports USCDI, but isn’t restricted to USCDI. There is no HL7 IG named USCDI FHIR IG.
  • US FHIR core is now ‘owned’ by the cross projects work group (co-chair Jean/Floyd added to cc), although emailing Structured Documents leadership is never a bad idea since they were involved in much of the original development
  • This value set is in use in both C-CDA and US Core

USCDI requires:

When you dig into this you end up at ISO. This has always been a bit of a mess, I have added Rob to see if he had ever untangled this during his efforts to load into VSAC.



Not ideal, but  always reference this site when looking for language codes.

From: Davera Gabriel <> 
Sent: Tuesday, April 28, 2020 2:05 PM
To:;;; Brett Marquard <>;;;;
Subject: Utilization of ISO codes for Preferred Language: BCP 47 in USCDI FHIR IG

Hello All,

As a representative of the HL7 Terminology Authority, I am sending this message to members of the Structured Documents Leadership team (Austin Kreisler, Calvin E. Beebe, Sean McIlvenna, Benjamin Flessner, Gay Dolin, Andrew Statler) as well as the Argonaut SMEs (Eric Haas, Brett Marquard) per information available on the and confluence sites regarding the groups with ‘ownership” of the USCDI FHIR IG.  If this message was sent to you inappropriately, if you could please let us know who can address our question, that would be most appreciated.

In reviewing the utilization of code systems within HL7 specifications, we are taking a look at those terminologies, vocabularies, and value sets that require a license for use.  Among these are the ISO standards.  In the HL7 FHIR® US Core Implementation Guide STU3 Release 3.1.0, Preferred Language utilizes an ISO standard <> .  This is a reflection of the current USCDI for Preferred Language <>  which links to BCP 47 <> .  Our questions is whether this is actually a "nested reference" and if the IG is actually a syntax using ISO codes.

Can you help us make that determination? 

Many thanks in advance for your time and or direction in resolving this issue.



Davera Gabriel, RN

Senior Research Terminologist

Johns Hopkins University

Institute for Clinical and Translational Research

2024 E Monument St, Suite I-200

Baltimore, MD 21287

+1 (530) 723-8356

  • No labels

1 Comment

  1. prior Meeting minutes / notes:

    It still unclear what the implications of the difference are. The applicable section to review is "2.2 Language Subtag Sources and Interpretation" in Davera will look at the US Core FHIR IG to determine if they are referencing the ISO set for Language Codes and/or another set or a syntax

    Previous commentary is shown below:

    Process comment: Meeting minutes from meeting in which these items were originally discussed were updated after the fact. Requesting that assignments are tagged with a user name so users can be notified. 

    • FHIR also uses the IANA Time Zones
    • Davera Gabriel will look up more information about BCP-47 (a document authored by the IETF trust) for language codes in the US Core FHIR IG.  Davera will check if this is actually a "nested reference" and is actually a syntax using ISO codes
    • (follow-up from Davera) Appologies for needing to step away & then not getting back to this in the meeting.  In the document linked above is the answer to the question.  On Page 8, under the section: "2.2. Language Subtag Sources and Interpretation" there is an explanation of the name spacing used to construct the BCP-47 "tags."  It is my read that the answer to the question as posited is a hybrid of the two options you propose in your question.  There is structure within the tags, but Im not sure that includes your posited definition of "syntax."  One would have to parse the tags to derive syntax, but as this is standard it could be done.  As such, what I intended to do in the meeting is discuss what was meant by question created the amended meeting minutes and NOT in the original meeting minutes: "Davera will check if this is actually a "nested reference" and is actually a syntax using ISO codes" to see if the tag / subtag structure fit your intended question definition of "syntax."  In practice, the subtags are intended to designate corollaries to portions of ISO 639.x designating spoken or written languages. 
    • I would further suggest we consult with the folks that suggested the use of BCP-47 as to their intentions.  I'd be very interested to learn whether they are interested in capturing a written versus spoken language.  My hunch is that there's a much simpler explanation than that, in which case splitting the BCP-47 tags is not implied, nor necessary for their use case, and if so:  we are overthinking a bit

    No progress just yet