1

Call to order and role call

@Jim Case

This meeting is being recorded to ensure that important discussion points are not missed in the minutes.  The recording will be available to the SNOMED International community.  Joining the meeting by accepting the Zoom prompt declares that you have no objection to your comments being recorded

2

Conflicts of interest and agenda review

@Jim Case

None recorded

3

Update on previous discussions

@Jim Case

• Human gene name proposal - Following review on the BN forum, the proposal will move forward. The new attribute HAS ASSOCIATED GENE has been modeled and will be added to the MRCM in May or June.

• Deficiency and Excess - Following review on the BN forum, the proposal will move forward. The new attributes HAS DEFICIENCY OF and HAS EXCESS OF have been modeled and will be added to the MRCM in May or June.

• Endoscopy revision - Changes have been made and are awaiting review and promotion to the International release

• Replacement procedures - agreed approach to model with “removal” and “<<introduction” role groups to ensure clarity of meaning when the replacement does not involve the same entity.

4

Reattachment, reimplantation, replantation, refixation

@Victor Medina

Inconsistencies in representing reattachment, reimplantation, replantation and refixation surgical actions (See attached document)

A proposal to address clinical inconsistencies in the modeling of Reattachment, Replantation, and Refixation actions within SNOMED CT. Currently, "Replantation" is incorrectly used as a synonym for "Reattachment," despite significant clinical differences: reattachment is a broad term for reconnecting partially detached tissue, whereas replantation specifically refers to the microsurgical restoration of completely amputated parts requiring vascular anastomosis. This lack of precision has led to inconsistent classification and "semantic soup" across approximately 135 procedures involving limbs, digits, and organs.

To resolve these issues, the project recommends decoupling these terms by establishing Reattachment as a broad parent action and creating a new, dedicated Replantation action as a specific subtype. Determine whether it is correct to classify Reimplantation as a subtype of Reattachement. Additionally, the non-standard term Refixation will be retired, with its associated concepts—such as spinal refusion—reclassified under the Revision hierarchy to align with modern medical literature. These changes will ensure that surgical procedures are modeled with the anatomical and procedural rigor required for accurate clinical data retrieval.

See attached document.

Discussion:

Background

• Current terminology uses replantation as a synonym for reattachment, but definitions indicate these are distinct concepts

• Refixation action exists but is only used once (incorrectly) in the terminology

Proposed Changes

• Remove replantation as a synonym for reattachment and create it as a new, separate concept

• Make replantation a subtype of reattachment (not a sibling), given it is a more specialized form

• Evaluate reimplantation's relationship — likely a sibling to reattachment rather than a subtype, as it implies planned detachment and reattachment to a potentially different anatomical location

• Retire the refixation action entirely; update the sole associated concept (refusion of spine) to classify as a revision procedure

• ~35 existing replantation concepts to be reviewed for correct subsumption

• Dental reimplantation procedures to be reviewed case-by-case as they use these terms differently

Key Distinctions Agreed

• Reattachment — reconnection of a structure (partial or complete detachment) back to its original location; general term

• Replantation — complete traumatic severing, reattached using microsurgical techniques; subtype of reattachment

• Reimplantation — planned detachment with reattachment to the same or a different anatomical location; sibling to reattachment

Concerns Raised

• Risk that clinicians (particularly in primary care) will use the generic term reattachment regardless of the more precise options, creating inconsistent data across datasets — though this problem already exists with the current sibling structure

• Reference sets proposed as a practical solution: GPs use the generic reattachment concept; specialists have access to the more granular terms

Decision:

• Draft clear text definitions for reattachment, replantation, and reimplantation to make distinctions explicit

• Propose to extend scope to align the morphologic abnormality hierarchy (surgically re-implanted organ or tissue) with the agreed definitions

• Post BN to community forums to validate the proposed distinctions, assess use cases, and consult on proposed solution

5

Container model update

EAG

Container modeling revisions:

Want to model additives to containers. Original proposal discussed at the October 2025 meeting. Use of pharmaceutical products or substances as the value was tested.

Discussion:

The team discussed modeling container types with additives, focusing on whether to use a medicinal product hierarchy or a simpler substance-based approach. Feikje presented examples and explained that while the medicinal product approach would create more fully defined concepts, it would add unnecessary complexity for their use cases. The group agreed to proceed with the original proposal using substances as additives, as it better balanced simplicity with functionality. Jim outlined the next steps, including posting the proposal on the briefing notes page for community review.

Apr 1, 2026

The Briefing note has been available for review. Only one comment has been received. Are there needed changes before a proposal to move forward? Do we need a formal review by the MAG due to the need for new attributes?

Apr 13, 2026

Proposed MRCM Changes

• Three new attributes proposed: has additive, has separator, and intended content

• The evacuated/non-evacuated distinction will be removed as it is not relevant to lab analysis

Substance vs. Medicinal Product Modeling (Resolved)

• A medicinal product hierarchy approach was explored but rejected — too many new concepts required and conflicted with existing modeling guidelines

• Agreed to link additives directly to the substance hierarchy; concepts with specific concentrations will remain primitive leaf nodes

Scope

• Remodeling will cover all existing content below 706046003 |Specimen receptacle (physical object)| (~50–60 containers)

• No additional attributes beyond the three proposed are anticipated

Administrative Notes

• Attribute names to use verb form (e.g., has intended content)

• Acronyms in FSNs must be spelled out or universally unambiguous

Decision:

• Formalize the attribute proposal and submit to SNOMED

• Remodel all existing content below specimen receptacle using the new attributes

• Add new eHealth-required concepts

• Likely proceed as a collaborative authoring project — proposing team does modeling, SNOMED does review

6

Recanalization of blood vessel

@Feikje Hielkema

Recanalization of Blood Vessel – Modeling Proposal (See attached document)

• Proposal to introduce and standardize modeling of recanalization procedures (restoration of blood flow in occluded vessels)

• Current content is inconsistent and undermodelled across International and national extensions, limiting queryability and aggregation

• Key issue: recanalization represents an intended outcome (reopening a vessel) rather than a single, consistently modeled procedure

• Existing methods (e.g., angioplasty, stenting, thrombolysis) are modeled differently, making subsumption under a single concept difficult

• Open questions:

  • Should this be modeled as a subtype of repair or as an outcome without a defined method?

  • How should site and morphology be represented?

  • Can a concept be created that meaningfully groups relevant procedures?

• Broader implication: raises whether SNOMED CT should support procedure aggregation by intent vs. method

Discussion:

Background

• A Dutch organization (with ~15,000 procedure concepts) had requested 50+ recanalization of blood vessel concepts, with 30 more pending — none classifying consistently

• A request was submitted to SNOMED to create a parent recanalization concept; flagged as complex by the reviewer

Core Modeling Problem

• Recanalization can be achieved via multiple methods (stenting, thrombectomy, angioplasty) which share no common modeling attributes in SNOMED — only body site is shared

• It is unclear whether recanalization is a procedure method or an outcome, complicating placement in the hierarchy

• Additional complexity: recanalization applies beyond blood vessels (fallopian tubes, ureters, gastrostomy tubes), and the term may describe removing an obstruction or recreating a passage (e.g., fallopian tubes), which are fundamentally different

Options Discussed

• Using repair as a parent — rejected as recanalization is more of an outcome than an action

• Using has intent or has focus attributes — considered but would broaden scope of those attributes beyond their intended use

• Using GCIs (General Concept Inclusions) — acknowledged as theoretically possible but would require an extensive and complex set

• Creating a primitive grouper concept — possible, but offers limited analytical value beyond vocabulary grouping

Key Tension

• Clinicians (e.g., cardiology societies) want to record that a recanalization occurred without specifying how, but SNOMED ideally requires all methods to subsume under any such grouper concept — which is currently not achievable

Decision:

• EAG members to consider possible solutions before next meeting

• Proposer to post the problem to SNOMED community discussion forums, including the background document, to seek community input

• Check whether any national extensions have already addressed this

7

Update/decision on Nontraumatic vs. spontaneous injury

All

See previous discussion at:

Nontraumatic vs. Spontaneous injuryPreview

Summary of prior discussion:

Nontraumatic vs. Spontaneous Injury – Terminology Discussion

• Ongoing inconsistency in SNOMED CT:

  • “Nontraumatic” and “spontaneous” are sometimes treated as distinct concepts, other times as synonymous

• Proposed distinction:

  • Nontraumatic = no external cause (may have premonitory signs)

  • Spontaneous = no external cause and no premonitory signs (more specific)

• Key issue:

  • Most participants agree the distinction is not reliably reproducible or clinically meaningful

  • Terms are often used interchangeably with others (e.g., idiopathic, cryptogenic)

• Modeling concerns:

  • Duplicate concepts exist (spontaneous vs nontraumatic variants)

  • Use of “due to spontaneous event” is questioned as unclear and potentially low value

• Emerging consensus (pragmatic approach):

  • Merge duplicate concepts where both forms exist

  • Choose a single preferred term (often “nontraumatic”) for FSN

  • Avoid systematically adding both terms as synonyms

  • Reassess or potentially remove “due to spontaneous event” modeling

  • Conduct broader review of injury hierarchy and definitions

• Status:

  • No formal resolution, but general agreement that strict differentiation is not sustainable

The questions that were driven by the prior discussion:

1. Where there are both "non-traumatic" and "spontaneous" concepts for the same disorder, combine them, inactivating the "younger" of the two, adding the inactivated concept description to the retained concept?

2. Where there are both "non-traumatic" and "spontaneous" descriptions on the same concept, evaluate the clinical literature for the predominant term and make that the FSN?

3. Where both spontaneous and traumatic descriptions exist on a concept, leave them both?

4. Do not automatically add "non-traumatic' or "spontaneous" synonyms on concepts where they do not exist?

5. Limit the hierarchy to only those terms that specify “spontaneous” or “nontraumatic” in the FSN? (Many disorder concepts “could be” classified as nontraumatic)?

6. Review the need for "due to spontaneous event".  - Argument that only a proposed parent of 1119219007 |Nontraumatic injury (disorder)| would be primitive and subtypes would not need the DUE TO event.

7. Review the model for the entire injury hierarchy and develop a more refined definition for what constitutes an injury. Forum discussion on Injury vs. lesion: https://forums.snomed.org/t/lesion-vs-injury/440/12

Discussion:

Background

• Long-standing inconsistency in SNOMED: "non-traumatic" and "spontaneous" injury terms are sometimes treated as synonyms, sometimes as supertype/subtype — with no consistent pattern

• Issue originally surfaced from Kaiser (US) content requests; resurfaced via Netherlands translation work

• Approximately 240 spontaneous disorder concepts and 70 spontaneous finding concepts exist; ~40 are descendants of non-traumatic injury, of which only ~10 have both terms on the same concept

Proposed Distinction (Historical)

• Non-traumatic: no external physical force as cause

• Spontaneous: no external physical force and no premonitory signs — intended as more specific

• However, the distinction was deemed clinically non-actionable and not reproducibly applied in the terminology

Key Discussion Points

• A non-traumatic injury can have an identifiable internal cause (e.g., tumor compression), whereas spontaneous implies no discernible cause — so they are not always interchangeable

• Overuse/stress injuries raised as a grey area, but noted these are modeled separately in SNOMED with their own morphology and would not be affected

• General consensus: for practical purposes the two terms are synonymous in the vast majority of cases, and the distinction is not worth maintaining

Emerging Consensus

• Treat the terms as synonymous; retain non-traumatic as the preferred FSN and PT for consistency, as the top-level hierarchy already uses this term

• Where both a non-traumatic and spontaneous concept exist for the same disorder, combine them — retaining the non-traumatic concept and inactivating the spontaneous one

• Retain spontaneous as an alternative description on the merged concept for continuity of existing searches

• The due-to spontaneous event attribute relationship would likely be retired as part of this cleanup

• Do not automatically add spontaneous synonyms to concepts that currently lack them

Decision:

• Compile a full list of affected concepts (~40 candidates) and distribute to the EAG forum for review before final decisions are made

• Confirm whether any cases exist where the distinction is clinically meaningful, and handle those as exceptions

• Review whether the spontaneous event concept used as an attribute value also needs to be retired

8

Inconsistencies and proposed revision of the Situation hierarchy

@jeremy rogers

EAG Review of SNOMED CT Situations – Key Issues See attached documnet)

• Identification of widespread inconsistencies in modelling “Situation” concepts, indicating need for clearer or new editorial policy

• Major problem areas include:

  • “Treatment with/for X”: inconsistent or overly abstract associated procedures; lack of guidance on when to create matching procedure concepts

  • Recommendations/advice: multiple competing modelling patterns with incomplete representation of what is recommended and by whom

  • Performer/requestor (“actor”) modelling: proliferation of codes embedding actors in terminology vs. handling in information models

  • Communication modality and recipient: duplication with information model capabilities and unclear scope for terminology

  • Location (“where”) of care: overlap with information models and inconsistent representation

  • Multi-actor scenarios (e.g., maternal–fetal, donor–recipient): ambiguity in subject context and relationships

• Core theme:

  • Need to clarify what belongs in SNOMED vs. the information model

  • Current approach leads to inconsistency, redundancy, and limited semantic clarity

• Recommendation:

  • Undertake policy review and standardization of modelling patterns across Situation concepts

Discussion:

Jeremy presented a set fo slides (attached above) outlining the many issues with the current Situation with explicit context hierarchy. This will be used as a basis for future discussions on how context shouldbe represented in SNOMED.

Background

• A review of the ~5,000 concepts in the SNOMED situations hierarchy revealed widespread inconsistency, poor modeling quality, and many concepts that remain as primitives without the expected associated finding or procedure relationships

• This creates a significant interoperability problem with FHIR, which has its own independent mechanisms for representing context (recommended, canceled, declined, etc.), resulting in at least 5 non-interoperable ways to represent the same thing

Key Problem Areas Identified (4 of 7 presented)

• Treatment with X concepts (e.g., metformin not tolerated, glitazones contraindicated) — often lack a properly modeled associated procedure; some link to therapeutic drug classes that the pharmacy working group has already removed from the drug hierarchy

• Treatment for X concepts (e.g., diabetes treatment changed) — associated procedures often modeled only at a very generic level, which is clearly inadequate

• Procedure performed by actor concepts (e.g., counseling by physiotherapist, vaccination hesitancy by parent) — ~1,000 such codes with ~400 different actor types, grown incrementally without clear purpose; some contain the word "other" which loses meaning when records move between contexts

• Recommendation concepts — 4 mutually non-interoperable SNOMED patterns for saying something has been recommended, plus a 5th via FHIR, with poor or absent relationships to the underlying procedure being recommended

Broader Issues Raised

• The concept of "default context" was clarified as a misnomer — SNOMED concepts are context-agnostic, with context supplied by the information model (e.g., FHIR resource type or problem list location), not the terminology itself

• Using SNOMED to represent negated findings (e.g., "no headache") produces logically incorrect classifier inferences — the editorial guide already advises against this

• The situations hierarchy has been discussed as problematic since at least 2003 (Alan Rector) with no resolution; a 2018 effort and a more recent project both stalled

Diverging Views on the Path Forward

• Some argued SNOMED should progressively hand context representation off to FHIR/information models and stop expanding the situations hierarchy

• Others cautioned that FHIR adoption remains very low (estimated ~1-2% of US data flow), and non-FHIR implementers still need a solution

• General agreement that some pre-coordinated context concepts remain necessary (e.g., family history, fetal context in maternal records) and cannot easily be replaced by post-coordination

• Suggestion that a selective approach — retiring the worst-behaved branches while retaining those that work — may be preferable to abolishing the hierarchy entirely

Decision:

• Reinstate the previously proposed situations hierarchy review group (previously led by Alejandro) to assess which branches should be retained, improved, or retired

• Develop clearer editorial authoring guidance on minimum modeling requirements for situation concepts — both new and existing

• Evaluate whether SNOMED should formally declare certain types of context out of scope and defer to information models

9

Formation of a group to discuss Observable Entity vs. Evaluation procedures

@Jim Case

The lack of precision regarding SI editorial policy on Evaluation procedures vs. Observable entity needs further discussion. The scope of Evaluation procedures is larger than Observable entity, yet there is substantial overlap between the two hierarchies. The impact of the LOINC ontology on this issue has yet to be determined. It is proposed to form a Member group to identify the ongoing issues, impact on terminology modeling and classification.

Evaluation Procedure vs Observable Entity – Ongoing Debate

• Longstanding overlap between Evaluation procedure and Observable entity hierarchies in SNOMED CT

• Two competing perspectives:

  • Procedure-focused view: ordering a test is a clinical action (procedure), distinct from the result

  • Observation-focused view (LOINC/NPU): the order is fundamentally for an observation, with the procedure being incidental

• Key issues:

  • Whether SNOMED should create parallel laboratory tests with evaluation procedures and observables?

    • Latest LOINC ontology contains over 2500 order level concepts: (e.g. 732441010000108 |Measurement of chromium in urine (observable entity)|)

  • To what level of detail should evaluation procedures be modeled?

    • Some existing evaluation procedures mimic Observable entity content: (e.g. 443934001 |Quantitative measurement of mass concentration of chromium in peripheral blood specimen (procedure)|)

• Implication:

  • Impacts model consistency, classification, interoperability with LOINC/NPU, and future content strategy

Discussion:

Continued to later call due to time constraints.

Decision:

10

Next meeting

@Jim Case 

TBD