2024-05-06/07 SEC Face2face Meeting Berlin

Date

Jun 6, 2024 Jun 7, 2024

Location

Rahel Hirsch Center für Translationale Medizin (RHC)
Room 06 110 | Ada Lovelace (6th floor, AG Roland Eils)

Call Link

Discourse SEC calls, tbd

Participants

SEC Physical

SEC Remote

SEC members won’t attend or not responded

SEC Physical

SEC Remote

SEC members won’t attend or not responded

  • @Bostjan Lah

  • @Diego Bosca

  • @Ian McNicoll

  • @Jake Smolka

  • @Mattijs Kuhlmann

  • @Matija Polajnar

  • @Rong Chen

  • @Sebastian Garde

  • @Sebastian Iancu

  • @Severin Kohler

  • @Alex Vidrean (drop in/out as my schedule allows it)

  • @Erik Sundvall

  • @Joost Holslag

  • @Pablo Pazos

  • @Seref Arikan

  • @Bjørn Næss

  • @Birger Haarbrandt

  • @Thomas Beale

  • @Shinji Kobayashi

Goals

  • ADL2+

  • WG reports

  • Releases

  • openEHR & HL7

Discussion topics

4tarts

Duration

Item

Presenter/Suggested by

Issue

Notes

4tarts

Duration

Item

Presenter/Suggested by

Issue

Notes

Jun 6, 2024

09:00

10 min

Opening & Agenda

ALL

SEC Jira board.

 

 

30 min

Co-chair election

ALL

  •  

3 nominations for co-chair: Bostjan L, Rong C, Sebastian I

The SEC has no objection in relation with nominees, and agreed that they will be co-chairs for the next term up to June 2026.

 

Discussion about:

  • rotating co-chair task responsibilities - this needs to be further explored

  • continuity of co-chairs croup vs SEC member

  • @Severin Kohler needs onboarding docs for newcomers, process description

  • @Bostjan Lah we need to formerly ask inactive members to quite if they want so or delegate other persons

  • vote for @Severin Kohler to become a SEC member - all members agrees

    • todo: @Sebastian Iancu execute necessary changes

  • suggestion for two candidates for SEC expert: Borut Jures, Renaud Subigre

  •  

 

 

80 min

AM 2.4 (ADL2+)

@Bostjan Lah

Review

 

ADL2 - SWOT

  • @Bostjan Lah We are at the point where we could publish;

  • we had extensive review on the wording a branch

  • AD will have support for ADL 2.4 in June/July

  • Nedap is working hard on making support for ADL 2.4 grammar in Archie

    • @Bostjan Lah can we use (as a copy) Archie grammar as the official ADL 2 grammar ?

      • @Seref Arikan Archie should use the (copied) grammar from specs - consider also potential license issues

      • vote: all in favour

  • @Sebastian Garde migration for CKM archetypes is a major work item ; we will need first to agree on template formalism before moving to template

  • @Severin Kohler where/what is the change ticket for ADL 2.4

    • todo: @Sebastian Iancu make sure tickets are there processes correctly on the branch

  • @Seref Arikan we need a confluence page where we centralize all changes of from ADL 1.4 to 2.4

    • todo: @Ian McNicoll has a start (see SWOT page)

    • @Severin Kohler we should have an architecture doc explaining the changes and their reasoning

  • @Ian McNicoll we will need to assess if there is a need to backport some of the ADL2x changes that have no impact on CDRs to ADL1.4 (as 1.5)

    • @Bostjan Lah we are committed to ADL 2 also with AD

    • @Seref Arikan it is better to all be committed to 2.4

  • reconfirm that at-codes are the future for openEHR, we will have to make it clear in the specs so new openEHR RM implementers will implement at codes . There's a desire to deprecate id coding. but it will be left to decide for a next adl version. adl2.4 will support both at and ID coding.

  • @Joost Holslag will we carry on id codes in adl 2 archetypes ?

    • vote: no, we will not carry them. TODO: @Bostjan Lah edit out the line in the specs. Will reconsider for adl3.

  • agree to start from at0000, keep it the same in ADL1.4; TODO @Bostjan Lah make it clear in specs

  • From @Ian McNicoll 's list, only the numbering scheme for at codes (first point) needs to be specified. The other points can be solved after adl2.4 vote.

    • agree to localize it to at9000 and above - todo: @Bostjan Lah add it to specs of ADL 2.4

  • @Joost Holslag we need to discuss common expected formats (op2/adls/json/yaml) between AD, CKM and CDR. But only after publishing adl2.4.

  • AOM changes are not necessary because we decided that we don’t need to carry id-codes

  • review implications of latest BASE required by AOM2 and propose changes if needed so that it can be correctly used by ADL2.4 see for example comments on https://openehr.atlassian.net/browse/SPECTERM-30

  • the group agrees that the notes about at-codes choice as openEHR vendors should go in ADL 2.4 as well as conformance specs

11:00

10 min

Break

 

 

 

80 min

AM 2.4 (ADL2+)

@Ian McNicoll

A few outstanding crucial ADL2 issues...

13:00

45 min

Lunch Break

 

 

 

13:45

60 min

ADL 2.4

 

 

  • @Joost Holslag @Seref Arikan we will need ADL2/OPT2 → OPT 1.4 conversion

    • @Bostjan Lah we have that in AD, we will still going to support it, we might consider opensource part of that conversion code

  • source formalism for template, something as alternative to .oet files (from v1.4 world)

    • .adlt file could be a an alternative, based on openEHR Template specs (AOM2)

    • (.adls is the archetype version of the …)

  • re: decision to keep the MD5 hash (optional field) in archetypes, but don't use it as identifier of the archetype

    • @Joost Holslag the hashes are useful in some situations, it would be better to keep a place to keep the hash; @Sebastian Garde maybe we should just use md5 over the whole file (thus not the canonical hash)

  • @Severin Kohler the annotations is a inconsistently specified; it should be free to use annotations but recommendation is to use properly

  • on the question of keeping historical version of archetypes in CKM, we will keep the history (in ADL1.4)

  • @Ian McNicoll we should have a new attribute to capture the template title

    • @Joost Holslag we think it is already there in the root; need further investigation todo @Ian McNicoll

  •  

 

 

RM & BASE

@Sebastian Iancu

 

not discussed

but there was ad-hoc discussion about demographic virtual values

  • todo @Severin Kohler will start a WG to work further on this

 

15 min

REST API WG

@Sebastian Iancu

 

not discussed

 

15 min

Federation WG, Authorization

@Joost Holslag

 

  • do we have an agreement to consider the entry the lowest level?

    • @Matija Polajnar we limit on the composition level

    • yes, agreement authorisation scope should be no smaller than an entry. Unsolved: how to handle conflicts between entry level and composition level acces policies.

  • @Seref Arikan we may disagree with legal/social requirements, but we should be able to facilitate ACL policies

  • @Sebastian Iancu @Seref Arikan UK and Turkey have broad acces policy by default, unless specific data (model) is marked as sensitive. Warns against defining fine grained access policies, due to complexities.

  • @Joost Holslag NL law requires a no acces policy by default

  • @Severin Kohler @Joost Holslag political realities require decentralising CDRs

  • @Joost Holslag There’s a requirement to blind specific elements in a specific composition for a specific user/role. E.g. patient has access to all data in the EHR, except for specific information that could be harmful for the patient to know, e.g. psychiatric patients. This requires blinding on the element level by the commiter of the information. And shouldn't be done at the model/archetype level as not to conflict with the entry as unit of information rule.

15:30

15 min

Break

 

 

 

 

75 min

CDS WG

 

  • Try to kick-start WG decision/process support (GDL2/TP etc) @Rong Chen

    • 1h - 2h

    • BPM+ in FHIR

  • consider deprecating/parking TP, try to shrink more of the specs/components, reduce complexity

  • @Ian McNicoll check if BPM+ might be something to be adopted

  • @Seref Arikan can we adopt e.g. typescript base scripting? check ‘wasi’

  • @Ian McNicoll we should have a really basic way of expressing rules in archetypes, the way Nedap is doing; rules should “stay” with archetype

  • WG: Rong, Seref, Ian, Matija,

17:30

Ending

Jun 7, 2024

09:00

5 min

Opening & Agenda

 

 

WG: virtual demographic archetype

  • Severin, Sebastian I, Ian, Seref, Chunlan, Rong, Bostjan, Jake(?), Diego

  • 2-3 meetings; review usecase

 

10 min

Terminology v3.1.0

@Sebastian Iancu

 

not discussed

 

60 min

Terminology

@Severin Kohler @Ian McNicoll

  • simplifying terminology (for ADL2 mainly)

  • also terminology binding mappings @Severin Kohler

  • Severin presents issues he find using term mapping on model and data side

  • Ian presents issues with terminology bindings, comparison with valuesets

    • question: what if we would use or express archetype ontologies (local at-codes) as FHIR values sets inside archetypes ?

    • AQL syntactic sugar for querying across mappings and defining_code

    • @Ian McNicoll let's take this as a question for HL7 collaboration project context

    • should we consider allowing mixed valuesets including null_flavour codes?

    • DV_TEXT/DV_CODED_TEXT => why not use CodeableConcept

      • @Bostjan Lah why not add a new type DV_CODEABLE_TEXT, which looks like DV_CODED_TEXT but the code is optional, similar to FHIR CodeableConcept ?

      • @Seref Arikan good idea, but what would be the impact on existing CDR data

      •  

 

60 min

interop FHIR+OMOP

@Severin Kohler

  • structural mapping

  • @Severin Kohler presents proposal to have common mapping language based on FHIR connect specifications

  • FHIR and OMOCL mappings are going to be supported by CKM on a mapping tab

  • @Severin Kohler and @Diego Bosca takes indicative's on taking this further and formalize it, perhaps propose a WG

13:00

45 min

Lunch

 

 

 

13:45

90 min

openEHR & HL7

@rachel dunscombe

  • FHIR / HL7 collaboration plan & impact

  • SEC: produce a doc on what would be our vision

 

15:15

15 min

Break

 

 

 

15:30

90 min

SEC general matters

 

  • todo: review what should be deprecated

  • strategical planning of spec (long term)

  • UML > BMM > asciidoctor

  • consensus from present members to initiate procedure on Retire following specs:

    • CNF tests SUT

    • SM entirely

    • PROC entirely, overview merges into CDS

    • RM: EHR extract

    • LANG: P_BMM (our move it to ITS), EL

  • needs review

    • CNF

  • We should remake the main landing page of specifications and menu:

    • todo: @Sebastian Iancu hide all D specs on release baseline

    • todo: @Sebastian Iancu @Bostjan Lah ask feedback in our companies and on discourse about how the landing page should look like, how to reorganize the way we present specs

      • menu too complex, too much text on main page

    • todo: @Sebastian Iancu generate new website (test)

  • Jira: continue with jira-adoption; if a tickwet is not assigned to a person we will have to add a label so that we need to discuss it in the group

    • @Sebastian Iancu consider @Ian McNicoll suggestion to label tickets based on action that has to be takken on the ticket

  • We need to review the ToR doc so that:

    • members that are de facto inactive for a specific period of time (6months) should be asked to resign their role or propose a replacement, or to join the SEC expert group

    • add requirements that members should be present at least twice a year on SEC meetings

    • todo: @Rong Chen execute this

  • Propose: url of stable specs should be now ‘latest’, and the dev-baseline goes to ‘development’

    • vote all in favor

  • Propose: move all amendment sections of specs to be moved by the end of docs

    • vote: all in favor

 

 

 

 

  • Jira tickets (overview)

not discussed

17:00

Ending

Action items

 

 

 

Related content