1

Call to order and role call

JCA

2

Conflicts of interest

Approval of minutes Jan 2018 conference call

JCA

No conflicts reported (GRE has his usual ones, contracts with SI and supports multiple NRCs)

3

ECE Update

BGO

• Reactive arthritis and modeling extraarticular manifestations using simple co-occurrence

• Test of allergic disorder model as pathologic structure

• Sepsis

  • Sepsis
    1. Agreement to add dysregulated host response as a pathological process.
    2. Agreement not to add pathological process infectious process which would result in sepsis being defined as organ dysfunction syndrome co-occurrent and due to infectious disease.

• Test of device complications model

  • Device complications
    Problems with the device itself should be a finding and not a disorder. This would allow some rearrangement of the current device problem findings. The modleing structure would be to use the INTERPRETS/HAS INTERPRETATION pair to define the findings.
    Should also create a more specific "device failure" to segregate from general external equipment failure.
    desire to see more examples for each of the three patterns.

See Events, Conditions, Episodes Project Group meeting agenda 2-12-2018

4

Drug Model Update

TMO

@Former user (Deleted) to provide an update on the status of the drug project

Open

5

Observables Model Update

DKA

Links to two Observables discussion items:

https://snomed.atlassian.net/wiki/display/OBSERVABLE/LOINC+Property+-+Likelihood+and+associated+LOINC+terms

https://snomed.atlassian.net/wiki/display/OBSERVABLE/LOINC+component+-+Mapping+of+LOINC+component+for+Susceptibility+LOINC+terms

6

Revision of Amyloidosis

JCA

On the ICD-11 MSAC call recently, a discussion ensured about the recent changes to nomenclature for Amyloidosis, with an emphasis on etiology (i.e. the type of amyloid) with secondary interest in the anatomical location of the amyloid deposition. This is contrary to the current way SNOMED classifies amyloidosis, which focuses on the site of deposition as opposed to detailed information about the type of amyloid.

WHO description:

Amyloidosis is a vast group of diseases defined by the presence of insoluble protein deposits in tissues. Its diagnosis is based on histological findings. Amyloidoses are classified according to clinical signs and biochemical type of amyloid protein involved. Most amyloidoses are multisystemic, 'generalized' or 'diffuse'. There are a few forms of localized amylosis. The most frequent forms are AL amyloidosis (immunoglobulins), AA (inflammatory), and ATTR (transthyretin accumulation).

Reference for amyloidosis nomenclature is here.

Discussion points:

1) The current subhierarchy <<17602002 |Amyloidosis (disorder)| is mostly primitive and does not go into much detail about the biochemical characteristics of the amyloid protein. Do we want to add the specific subtypes of amyloid?

2) Do we want to align the SNOMED CT amyloidosis content to the ICD-11 approach? What is the clinical importance? Most general clinical references emphasize the underlying origin of the amyloid (i.e. genetic, inflammation, dialysis).

7

Expand range of "USING DEVICE"

JCA

Currently, the entire subhierarchy under 445536008 |Assessment using assessment scale (procedure)| is primitive. There is no linkage between the hierarchy of scales 273249006 |Assessment scales (assessment scale)| and procedures using those scales.

Proposal for consideration is to expand the range of USING DEVICE to include 273249006 |Assessment scales (assessment scale)| <<

The ability to SD the concepts using 445536008 |Assessment using assessment scale (procedure)| will improve the ability to organize and locate procedures using assessment scales of the same type.

8

Disorder without disorder redux

JCA

Comments from the CMAG suggest the following:

• while there is limited use of existing concepts with this pattern, consensus was that the construct involving the specific absence of a particular clinical manifestation provides additional valuable clinical information.

• consensus that this type of concept is a situation (currently there are only 15 concepts with this pattern in the situation hierarchy.

• consensus that creation of a query to find all concepts with or without a particular manifestation would be extremely challenging if they were placed into different hierarchies

• Most of the commonly used concepts by members exist in their national extensions

Another comment questioned whether alternative mechanisms to record this specific absence without the use of a precoordinated construct would be acceptable to the SNOMED CT CoP.

Q: Have we sufficiently considered the clinical usefulness or are these really just useful as billing codes to represent condition complexity? (Note: Current usage does not support widespread usefulness)

Q: Which of these solutions provides the least impact (users vs. editors):

• Inactivation

• Movement to the situation hierarchy

• Leave primitive in Findings

• Other?

9

SNOMED International Quality Improvement Initiative

JCA

The SNOMED quality initiative has been kicked off and a pilot project to test the proposed strategy has been completed.

Information about the project can be found at:

Quality Improvements 2018

A number of templates are being created to support the automated reconstruction of subhierarchies.

10

Specimen from subjects other than the patient

 JCA

Update since Bratislava:Currently we have many concepts in the specimen hierarchy that include “from patient”as well as those that do not include it as an ancestor.  Current editorial guidance simply states "The | Specimen | hierarchy contains concepts representing entities that are obtained (usually from a patient) for examination or analysis." 

Thus, the editorial guidance does not specify the implied context of "from patient". In the current Specimen concept model, there is only a single attribute that allows for identification of the "source" of the specimen, i.e SPECIMEN SOURCE IDENTITY. The allowable range for this attribute is extremely broad encompassing:

Person 125676002 (<<)

Family 35359004 (<<)

Community 133928008 (<<)

Device 49062001 (<<)

Environment 276339004 (<<)

Further review of the issue suggests that this is not a significant problem and that while exceptional conditions exist in the terminology, the number of affected concepts is very small.

Number of concepts with SPECIMEN SOURCE IDENTITY as a stated relationship: 100/1,695 = 5.9%

Number of concepts with SPECIMEN SOURCE IDENTITY as a stated relationship: 118/1,695 = 7.0%

Number of concepts with "from patient" in the FSN: 8/1695 = 0.5%

The primary issue is that because of the very broad range of values that can be used for SPECIMEN SOURCE IDENTITY, the inferences that result cause substantial fragmentation of the content (for example, Specimen from blood bag from patient is not a subtype of Specimen from patient).

It is possible that if the problem were more pervasive, there would be justification to look at revamping the concept model for specimens; however, the small number of affected content should at this time be handled as exceptions.

Currently editorial guidance does not specify a soft default, so the context is provided by the information model in which the concept is used.

Recommend: tabling this topic.

11

Quality improvement project update

JCA

Review of the proposed approaches to the structural improvements that are being applied to the Clinical findings hierarchy.

Pilot project

12

Discussion topics for London

JCA

A draft discussion paper related to prior "naked kernel" discussions has been posted for review by the EAG members. This will be the sole/primary discussion at the EAG sessions in London in April.

13

Future meeting

JCA

SNOMED International Business meeting April 10-11