Date
10 January 2017
Location
GoToMeeting
Attendees and Observers
See agenda page. KCA joined during the ECE discussion.
Topics
Welcome, call to order, attendance, conflicts of interest, minutes
JCA welcomed everyone. No changes to the agenda. Attendance (see agenda). No Conflicts declared. Minutes approved.
Guidelines surrounding use of numbers or roman numerals
GRE said he had been looking at examples and he was most comfortable with the last proposal, doing it as in the literature, because there was no consistency. He said they should go with the diversity, no way to standardize. On whether to represent the alternative number as a synonym, GRE said search engines would not expect to do the transformation, so if users might do it both ways, then the team would have to record both ways. If they are usually represented one way, there was no reason to write the synonym.
BGO said if there is no preference in common usage, then just choose one. JCA suggested using arabic if there is no preference in the literature. The AG members agreed.
As noted on the agenda, "Recommendation: use the most common use in the literature for the FSN. The alternative representation will be included as a synonym if also represented in the literature. If no preference, use the arabic representation."
Clinical genomics
Presentation by Ian Green (IGR). See powerpoint presentation linked to agenda.
Discussion (27:30):
PAM said from UK perspective this is quite important and there were strong political imperatives involved, so fast action was needed so that people would not come up with their own, independent solutions. Not mentioned was how tests - LOINC - would interact, which was key due to the IHTSDO/LOINC agreement.
JCA said he realized there were some possible problems with the LOINC agreement but SNOMED International and LOINC would be meeting with soon to discuss the agreement. He said another big topic not mentioned was how to reference external terminologies as values for SNOMED, and whether that was possible or even feasible in the existing environment. GRE suggested looking at linkage tables, which would have to be referenced but he was not sure whether they would be defining or not. He said the strategy and the role SNOMED would play was key. It was a nice start to define what SNOMED CT's role and then could try to align with that.
JCA said it would be a continuing topic on the AG's agenda as the organization tried to take the strategy and realize it as tactical objectives. He expected other groups like modeling to be involved as well.
PAM said he was due to become involved in the topic next month. NHS Digital would have discussions with SNOMED International in early February. IGR said NHS Digital was focused on rare diseases at the moment, so incredibly interested in Orphanet.
Drug Model
38:00-45:20 TMO gave a presentation on the drug concept model and products as attributes (see Powerpoint on the agenda page).
JCA asked a question about reviews of the document with the proposed model and linkage of terms to other standards like IDMP. Do we want to document the separate classes we have with those represented in IDMP? TMO agreed and said she and the drug group would do that. JCA said he just wanted to note the relation between SNOMED and other standards.
JCA also noted that the group was focusing on manufactured form, not administrable form as the focus for this particular model.
47:00-49:00 TMO went on with her presentation on extension of causative agent to include product concepts (same powerpoint).
JCA noted that once product level concepts were cleaned up, they could provide clearer editorial guidance on use of products for values of relationships. TMO agreed. She said this was a subproduct of the drug work. JCA said it involved the work of cleaning up the foundational hierarchies and it would be easier to address these hierarchies once foundational hierarchies cleaned up.
GRE: do you have a set of examples of the type of products you would like to reference? Sometimes we have been using substance when we should have been using product b/c it was manufactured. in case of disorders may get more complicated. Could get in the same position as laterality. Someone has allergy to asperin products and somoen else allergy to asperin substance, so some examples would help so if you could post them after this? TMO: absolutely.
- Toni Morrison to post some examples of causative agents for AG review.
JCA: one of the actions is if we want to ask Tech Services to possibly extend the range of the causative agent relationship to include products prior to the cleanup of the hierarchy. He asked TMO if that was premature. TMO said she thought it was. JCA said the action should be deferred until the cleanup of the product hierarchy. The AG agreed. He noted that BGO would be impacted by ECE work. BGO said it was okay - there was a lot of work to do in the area.
- Jim Case To adjust prior action regarding Technical Services extending the range of the causative agent relationship to include products to delay such action until after the cleanup of the hierarchy.
ECE Update (Event, Condition and Episode)
55:00 BGO gave a presentation (see powerpoint on agenda page or recording).
1:05 discussion: JCA said there are there substantial number of late effect concepts - are they sequelae of disorder? BGO said no, many of those were under the complication hierarchy, events classified as sequelae of disorder would be classified as complications too. Many would need to be remodeled to be classified under sequelae, so a lot of clean up work would be needed to fully implement these models. JCA asked if that might be done through a batch process? BGO replied to an extent, but there were specific issues too that would require attention.
GRE said he was generally content with the model, but existing content is scattered around. What is the plan for that content, and the scope? So it's not only the model, it's the plan in applying the model.
BGO replied that it started out focusing on modeling perioperative complications, so starting out with that hierarchy would be the first step. He invited GRE to work with him on it. GRE agreed. JCA said he was also planning on working on it. we can use some of the tooling Peter Williams has developed and see if batch processing is an option to put it into more consistent modeling. BGO said he would be interested in working on that. JCA said they were close to being happy with the model, then it could be put into a batch queue.
1:11:16 BGO then gave a presentation on the allergy model.
1:24 discussion: JCA asked how if fit with the proposal years ago for the condition class. BGO replied that it fit very well with that. This made it more consistent, he said. There was a discussion of high-level primitive concepts in relation to the model.
Ed Cheetham (ECH) offered some comments on the model. He said it systematized what had been sloppy.
At this point, KCA joined the meeting.
GRE said all models have problems, he said, but what is the appropriate representation of reality? Over the years there had been some good ideas, but when they were modeled it turned out that reality was much more complex. We may feel that we need some small changes to improve it, but we will not have universal application, but we can improve some specific cases. He felt that pathological process was necessary, and simplifying the model to only the really necessary attributes would help. He would accept allergy concepts to be fully defined. He had not seen examples where they needed to role group things.
GRE noted that BGO's understanding of the model is much more detailed and complex than the average user and that could be a problem. GRE agreed that they could not make work arounds for years and years. But he felt they needed to get back to what is possible to consistently model.
KCA said he felt similarly to GRE. We need to evolve the models slowly over time, he said, and focus on adding new value bit by bit rather than adding a lot of complexity all at once. He said he could not comment on the allergies model because he had missed the presentation, but the observables model was similar.
BGO said he thought the model he presented was simpler than the current model. But they needed to make decisions due to the allergy to substances being dependent on those decisions.
PAM said any solution chosen, the Community of Practice could use a user's guide because a significant number of GPs wouldn't know what to record. What impact would this have on systems checking on allergies and offering decision support? BGO replied to the second part.
JCA noted that the results were exactly the same, it was just a difference of making one of the concepts primitive. What is benefit of making allergic sensitization primitive? BGO: disposition occurs after the sensitization and sensitization is caused by the substance. Other one after the sensitization, causative agent due to some unspecified process. JCA said he agreed with GRE and KCA about not making something more complicated if there is no clear benefit. BGO said he still felt his proposal actually simplified things. But he felt that that needed to be discussed more in ECE.
JCA asked what the AG would need to make a recommendation. GRE replied maybe slides or short document (GRE) get back to requirements and what are things we have agreed and particularly so many batch sand changes and work arounds, get back to what problem is and what want to achieve, whether still feasible or if there are simpler solutions. GRE said maybe he needed to go through the slides and maybe have a one-on-one call with BGO, but he was not comfortable recommending a change unless they were certain it was a better way, even though the current state was problematic.
KCA said the solution would have to be simplifying, he would defer.
JCA asked BGO if he had a paper on this. BGO said he could add to the paper this latest proposal.
PAM wondered if this might be tested in a system to see how people would actually use it. Would the tinkering with the model affect the way it's used? It might have no impact, in which case simplest is probably best. But we should test it out. BGO asked how it might be done. JCA said he did not think there was a mechanism to do that in advance. KCA said real experience would be interesting. Most of the systems are just using the strings and the codes, so changing this model would probably have 0 impact from the clinically-facing perspective. there could be some smaller systems if they are trying to implement the models. So he did not think it would have much of a positive or negative impact at this point. He cautioned to make sure the changes were something they wanted to live with for a long time. But impact would be zero right now, so we need to think about business case, including industry's perspective on the value of SNOMED. JCA said there were also complaints about SNOMED's lack of internal consistency, and areas like this are mess, so cleaning up the mess would be an advantage from a reputational perspective.
- JCA said they could not come up with a recommendation at the moment but he asked BGO to but these suggested modifications in the discussion paper around allergies and then share the link to that with the group, then move apace at trying to get a recommendation from the ECE.
Moving on, BGO said there had not been much progress since Wellington on the final approval on the editorial updates but provided a short update.
End of recording #1.
ICD-11 External Causes chapter
Start of recording #2.
JCA said there was a request from the contractors harmonizing SCT and ICD-11, and review of external causes of morbidity and mortality, found a number of concepts and they made some suggestions. See document on agenda page.
JCA explained that there would be a need for about 5000 new primitive concepts and would require a lot of rewording. Do people feel that they should be added to the core?
GRE asked why there would be a need to represent them in SNOMED CT? JCA said they were new concepts with increased level of detail. The agreement with the WHO is to align SNOMED with the new chapter in ICD-11 as much as possible.
KCA said if we have the rights to bring them in and they're necessary for interoperability with ICD, I think they should be brought in but put in a module. This is also ICD-11 international version, not ICD-11 CM, being considered in the US. I would have different feelings if it were the latter. GRE said he was not against including content as they are, or changing the wording, but concerns with significant overlap with what was in ICD10. If you recreate the meaning of an extension, then the model will not say this is a concept from one or the other. UKTC would have to deactivate its concepts. That is a potential side effect. He was not sure UKTC would be okay with inactivating their content. JCA said that was a risk, and this would be a relatively massive scale. It would be many thousands of concepts. The problem with referring to the near identical term is that ICD often makes small changes so they are not identical. There was a fair amount of this content that was in ICD9 and inactivated, so now do we want to go back and activate it again? GRE said the same would happen when ICD moves to ICD12. GRE said if they were really necessary, he had no objections to adding them. JCA said he did not think there was a contractual obligation to enter them and had not entered some other ICD concepts in the past.
PAM said most countries had an obligation to return information to WHO about accidents, but he was not sure about the level of detail required. He said the UK military recorded accidents in quite a lot of detail so he could understand why some people might want the concepts. There could be a causes of accidents properly modeled in SNOMED. It would be nice and more user friendly than what's in ICD. Also ICD usually provided mapping tables between ICD9 to ICD10 and ICD10 to ICD11. So those mapping tables could be used.
JCA said the first part of ICD11MMS is supposed to come out in the first quarter of 2018. So it would be a few years probably until first implementation of it, far longer in the US. He did not dispute the value of injury or accident representation in SNOMED, but was it worth while having it to the level of granularity in this chapter? PAM said individuals might use individual concepts, but not sure they use it widely. Space craft are quite rare but car accidents aren't.
KCA said if you need that level of detail in SNOMED to meet regulatory requirements for reporting in ICD, then SNOMED needed that level of granularity. PAM agreed but did not know what the regulatory requirements were.
BGO said those concepts were in CMT, but he had never used them. They were very granular. The military section might be useful in a module or refset, but in general medical practice would have almost no use. If could identify a subset of concepts that are clinical, then those could be included. The others could be in a module or not included.
JCA said David Robinson's options had not included a fourth option discussed by the AG. He went through David's 3 options (see paper). JCA noted that adding the concepts could be relatively easy since tooling could be used since they are primitive. But filtering out non-clinical concepts could be more work. Some words, like "pavement," meant different things in different countries. Fourth option, note mentioned in the paper: identify the clinically useful concepts, add those, then take the others and put them in a module.
PAM said there was antoher option. If ICD11 may not be used for a number of years, reintroduce the ICD10 ones then use the ICD10 to 11 map. Then later introduce something that is useful and modeled properly in SNOMED. Wouldn't that be the least intensive resource option? JCA said he was not sure it was the least work intensive due to the shift of concepts with the UKTC. KCA asked if they might get rid of the process of reassigning identifiers when moving from one place to another? JCA said it was done only one way. If moved to an extension, then it needs to have an extension identifier. But if that were to be changed it would have to go to the modeling group. KCA said it should be changed. Identifiers should be permanent. JCA agreed, and said now that the org is no longer supporting RF1, then that solution may be possible. David Markwell had some thoughts on that. JCA said that was a sixth option, involving a change on the policy of identifiers, working with the material moved to UKTC.
JCA said they could recommend to the contractors the postponement of working with the content until they decide what to do with it. The AG agreed.
- JCA to tell David Robinson that the AG needed to think some more about the options, so the team should postpone taking action on those concepts.
Need to address semantics and placement of "Care plan for X" or "X care plan"
35:34 JCA introduced the topic. He said it was a difficult topic that had been around for a while. It had arisen again due to issues raised by the nursing community.
JCA asked, at the highest level, what does the AG think a care plan is? Process of planning, info in the plan, execution of the plan, etc?
The AG discussed the topic.
Monica Harry (MHA) described the Nursing SIG's position, and noted that the UK was developing a national care plan library. The Nursing SIG wanted them as record artifacts. JCA noted that the SIG was talking about a general template of a plan. MHA described what how the SIG was defining it and provided some examples.
KCA said the VA had care plans that were put into the patient record as text, without coding.
JCA asked the MHA how the Nursing SIG wanted to use it, given that care plan is just a name. GRE said the SIG presumably had a particular purpose in mind. But there were only a limited number of care plan concepts in SNOMED, so moving them might affect people who are using them. So could inactivate them as ambiguous, or listing as artifacts. But would be useful to have feedback on the use of terms, like cancer care plan, perhaps using cancer care plan document. Need a clear use case.
GRE said he would accept them as inclusions to artifacts, ask them to add documents to all the FSNs to make clear the meaning, then define what to do with the ambiguous concept, but figure out what to do with them. JCA agreed they needed clarity of vision amongst other realms. He suggested taking it to Susan Matney. KCA agreed. JCA asked if BGO could get some info on whether it would impact the use of care plan in Kaiser. BGO said he did not think it would impact but he could get more information. JCA said he agreed with GRE's suggestion. No one knew what to do with the concepts at the moment so this would clarify and help people use them. Adding "Document" to the FSN would help provide that clarity.
JCA recommended writing up a fast track to provide clarity on current ambiguous content would be retired, renamed and replaced in the record artifact to provide structure around the existing content, then bounce off the Content Managers AG. The AG agreed.
- Jim Case to arrange for a fast track to provide clarity on the current ambiguous content on care plan and a proposal for relocating "Care plan" concepts under a single representation as record artifacts.
Arteriography vs. Arteriogram
1:04 MHA introduced the topic. She said -gram and -graphy are used interchangeably, but there is also guidance that says only use one. So editorial guidance was needed. Continue using both on every concept? Penni Hernandez (PHE) said she also wanted clarity on the finding sites for those concepts. They were all confused in the current modeling.
JCA said it goes back to common clinical usage. BGO said at Kaiser orders are used inconsistently, clinicians seemed to equate the two. So he would continue using both. Use the verb (-graphy) for the FSN and the noun (-gram) for the synonym. JCA asked the group if the AG agreed with that, the noun not necessarily as the preferred term. GRE noted that the advancement of technology would change the meaning of the terms in the future.
There was a long discussion about which was used by different organizations.
Consistency across procedures, JCA and MHA noted, was a problem. MHA offered a suggestion to reduce the inconsistency by using synonyms. There needed to be some clean up.
JCA asked MHA if she had enough guidance? MHA do we want to always add angiography? JCA said yes, stick to the current editorial guidance. MHA: indicating the modality in the synonym? JCA said "correct."
PHE provided a specific example regarding angiogram and arteries. Is angiogram can it be done on something other than a blood vessel? JCA said no. PAM said if it was something else it would be called ventriculargram (of the brain) so you would know from the name. There was further discussion. JCA told PHE that to be accurate they would want to use vascular structure of X organ, the finding site for angiography, what are the inferences that come out of using that? PHE said she could do a test on that.
- Penni Hernandez to do a test on angiography and come back to the group with the results.
JCA told PHE that they should look at the organ on a number of test cases and vascular on a number of tests. Then go back and look at the test results.
Adjournment
JCA said the last topic was KCA's, but he had dropped off the call.
- JCA said he would ask KCA to put together a high level document on "Aligning SNOMED CT with clinical use cases".
- JCA said he would put the action items on the next agenda.
The AG agreed to the date of its next call. JCA thanked everyone for getting through so much material in near record time. The AG agreed that 4 hour calls allowed them to get through a lot more work.
The call took around 3 hours and 45 minutes, not including the 45 minute break.