Unilateral Concepts

JCA presented the slide seen on the left.

KCA said that the approach seemed to be that the negation would live in the observation result model, so to represent unilateral, one would say that it is present on the right and absent on the left within that observation result model. Unilateral within SNOMED was unnecessary, he said, and he did not think it was a patient safety issue.

JCA asked, given that it could be represented in alternative ways, did having “unilateral” potentially cause confusion?

The AG members felt that it did. KCA recommended removing the references to unilateral. GRE said for clinical use it was useful, but from the terminology perspective it was a nightmare. Having them split between situations and original hierarchy was not helpful, he said. User feedback showed confusion.

There was a discussion about when unilateral was useful. KCA said he thought some of the use cases came from ICD codes being converted to SNOMED, when left or right was unknown. He said he preferred moving them to a refset and advising users that using such concepts in modeling was problematic and not best practice.

JCA summarized the AG conclusions: we need to do something in the short term about unilateral. It might take a long time to reach consensus, but the group felt that the inconsistency needed to be resolved, regardless of whether it was retired or not.

GRE suggested removing them from the situation hierarchy, providing a clear message to users, and eventually getting rid of them. He mentioned primitive groups, which could not be aggregated. He said navigational concepts could not have children. JCA asked if GRE was suggesting putting them in the navigational hierarchy. GRE replied that it was one option. JCA said that they would be useful, so they would have to be put into the legitimate hierarchy and not just the navigational hierarchy. GRE said that that was an intermediate compromise.

Limited Status Concepts - WAS A

JCA introduced the topic (see screen shot to the right). There was a discussion about different options and solutions tried in the past. 

JCA put forward a motion: if one wants to do a run into a limited status concept, which is the source for a target concept that one wants to retire, then one can retire the WAS A relationship and then go on and retire the concept that one originally planned to retire. KCA seconded the motion. It was passed unanimously. 

Decision 6: The AG recommended the following: if one wants to do a run into a limited status concept, which is the source for a target concept that one wants to retire, then one can retire the WAS A relationship and then go on and retire the concept that one originally planned to retire.

SNOMED CT Interface or Reference Terminology?

KCA explained that many years ago, he and Kent Spackman came up with the interface vs. reference terminology differentiation in order to convince a particular vendor to adopt SNOMED. Then, in the conversion from SNOMED RT to SNOMED CT, the interface terminology was deemed to be unacceptable to the British NHS, so as part of the SNOMED CT design, the interface terminology was part was brought into SNOMED CT. So SNOMED CT wholly supports interface terminology, and the distinction should be made as a dialect. The way one should implement an interface terminology, KCA said, was by implementing a dialect of SNOMED CT, not by doing an ongoing mapping that would never be semantically sound and that would cause a number of other problems.

GRE and KCA discussed the issue and agreed that the dialect would be an extension.

Linda Parisien said she really like the idea of a dialect, and that would help resolve a number of issues in Canada, enabling people to benefit from the full power of SNOMED.

After further discussion, KCA said one could say an interface terminology = content extension + dialect refset. The content extension would have the descriptions and the dialect would indicate “this is preferred.” 

Decision 7 There was no vote on the issue, but the AG agreed that the public message on interface terminology should be that there are different ways to implement an interface terminology, but IHTSDO recommends using a content extension plus a dialect refset.

Bullet Points for Closing Plenary 

Adjournment

There was a discussion about future meetings. JCA then adjourned the meeting. Including the half-hour break, it was 4 hours.