May 2020 - HL7 WGM - Wednesday 3pm Minutes

• HTA update on external URI process and open items

theExternal Terminology pages maintained by HTA: External Terminologies - Information

Process for new code systems is locked. Updates for published code system identifiers still TBD.

HTA Confluence Page: External Code System Owner Engagement 

Process defined and has been used successfully used. 

This is the page where all external code systems referenced in HL7 artifacts are documented. 

Reuben is drafting the FHIR pages that will reference the HTA pages. 

In some cases, IG authors have moved forward with their own process. As long as the authors share the information with HTA, HTA will update their material. HTA is sending the message that work groups/IG authors must coordinate with the HTA. 

Keep HTA and FHIR documentation in synch?

When HTA creates a new/changed code system pages, someone has to create the page for "Using Code System XXX" to include in the FHIR specification. Reuben does this now. Vocab owns the FHIR spec pages.

The next HTA communication is to UTG - new and changes. Different processes. 

Ted: V3 external code systems; some without any documentation (from CoreMIF). Does HTA plan to mine those entries to create HTA pages - even when there might not be any community pressure to create. 

Carol: good idea. Julie has a list from GG/FHIR.

Julie: the FHIR list isn't too bad. The CoreMIF list might contain items that would be of limited value to add.

Ted: maybe we should just have a page for those CoreMIF code systems with some text indicating we will add a URL if requested.

Julie: she suspects some of the FHIR information is obsolete or out of date. (GG pulled out all code systems in all IGs)

Ted: what should he do when questioned about a code system? Should he direct people to HTA, or remove a code system from the manifest. 

Rob M.: is there a way to be notified if someone expresses interest/someone is looking for a code system? Possibly go to a generic page that indicates the code system is old, please contact XYZ. 

Ted/Julie: OID was required to get into CodeMIF, but not all items in the OID registry ended up in the CoreMIF. 

Julie: Why were code systems in the CoreMIF? 

Ted: Because it was referenced in a V3 artifact, and someone asked to add it to the CoreMIF

Julie: Not all code systems used in V3 artifacts are in CoreMIF. 

Julie: What was the motivation? 

Ted: Because the ballot was generated from the CoreMIF, if referenced in a ballot, it had to be in the CoreMIF (in some cases a ballot referenced a Code System and not the Value Set).  There isn't a clear, precise reason why a Code System is in the CoreMIF - no consistency. The question is whether we want to lose the list? 

Carol: What is the harm is leaving them? 

Ted: None. Its just that the HTA list and FHIR pages will not align. 

Carol: She proposes that anything actively maintained in the HTA site MUST be in UTG with aligned data values. One direction. 

Ted: If something is not on HTA, but in UTG - what do we do?

Julie: Not sure that leaving data in UTG where we cannot define its provenance, etc. does no harm. 

Carol: If a Code System were never published, then it wouldn't be a big deal to remove them. But some were published. HTA will not maintain them nor advertise them as vetted content. 

Motion:

Ted: Historical code systems brought in from the CoreMIF will be documented in the UTG entries as historical, in use 20 or more years ago and no further information is available at this time.

Second: Susan Matney

Further discussion: don't mention number of years. 

Add this text for any code system that falls into this category. 

This is a historical code system brought in from the V3 CoreMIF.  No further information is available at this time. This is not recommended for use. If you have a need to use this Code System, please contact the HTA. 

(These code systems might have been brought in from UMLS)

Ted: Accepts changes. Susan: continues to second

Against: 0 Abstain: 0 For: 12    

Clarification: this is a vocab decision (thanks Davera)

Ted - Susan there a bunch of nursing items in the UTG list. Would she review them? 

Susan: these will be part of the HTA cross check

• Process for identifying and implementing URI changes in normative HL7 products

Draft policy: Policy on Maintenance of External Code System Identifiers (URIs)

We know that non-authoritative URI was used for ICD-10-PCS. The implementer community recognized this, contacted HTA. HTA reached out and things were corrected.  

ICD-10-CM, FHIR declared a code system URI. The URI has been used however, the HTA is getting an authoritative URI that will most likely not match what has been in use. 

It has been proposed that via UTG and using NamingSystem resource, Until a proposed element is added to NamingSystem to note an instance as authoritative, Ted has proposed that we use the comment element to provide the user with the authoritative information.  

The HTA and Vocab need to come to an agreement on the process for triggering and managing a change to a URI that has been in use. Should the authoritative identifier be used going forward? The use of an authoritative identifier is preferred.

The issue is new work vs: existing work.  Is the existing work normative/STU? 

Julie: the authoritative identifier should be used in all cases. Identifiers in themselves should be meaningless. If we have built something that cannot change, then we're in big trouble. 

We need to separate out the what from the how. 

Rob H.: For any new work/artifact, the authoritative URI MUST be used. For any update to existing artifacts, the authoritative URI MUST be updated. This includes updates to specifications that must be updated on a time cycle. At that time, the authoritative URI MUST be updated.

Rob M.: This (Code System URI changes ) must be tracked. Its very easy for something like this to slip through a ballot cycle without comment.  The same issue applies to Value Sets. Tooling is required to identify where changes are necessary, and to make the change. Instance data will have varying identifiers for the same Code System.

Ted: He disagrees with the premise - Code System and Value Set identifiers aren't the same animal. VSD declares that if there is a change to a Value Set identifier, it is a new Value Set. Rob M. disagrees. Either way, the Code System issue remains.

The tooling needs to be supported, and tested at connectathons. 

Davera moves, Julie seconds that HTA and Vocab adopt this policy:

Identifiers provided by the code system owner are authoritative and are viewed by both the HTA and Vocab to provide, to the extent possible, a stable identifier. Thus, the use of an authoritative identifier is preferred. For any new use of the code system, the authoritative identifier MUST be used. For existing HL7 balloted artifacts, either at the time of update or renewal, the authoritative identifier MUST be used. 

Discussion: update or renewal covers all 

Reuben: make it explicit that the identifers are endorsed in HTA, and published in HTA. 

Carol: the motion is only regarding the policy, not the execution. The execution will be worked out.

Abstain: 0 Against: 0 For: 10 (one person dropped before this motion)

• Request from Vocab for HTA to seek clarification on use of pre-release codes in FHIR
  • It has been proposed that they can be represented as a code system fragment, however this declares the pre-release codes as members of the code system - which may be in violation of the code system authorities intent.