Summary
Information is being sought to support a review of concepts that represent <disorder> without a <disorder>
- Examples:
- Blister without infection (disorder)
- Measles without complication (disorder)
- Brain injury without open intracranial wound (disorder)
- Esophageal varices without bleeding (disorder)
- Seen as a classification construct.
- Could potentially be represented:
- using 2 concepts:
- A disorder concept for the disorder that is present
- A situation concept representing that disorder that is not present
- A situation concept with two relationship groups.
- using 2 concepts:
- Could potentially be represented:
- Looking to gather information of the use of <disorder> without <disorder> concepts by clinicians in member countries.
- Topic to be taken to the EAG once input has been provided. EAG are meeting 26 Jan 2018. Information to be provided to EAG for that meeting.
| Date | Requested action | Requester(s) | Response required by: | Comments |
|---|---|---|---|---|
Relevant documents
Actions:
| Date | Requested action | Requester(s) | Response required by: | Comments |
|---|---|---|---|---|
| 13-12-17 | National input requested on use by clinicians within your country of <disorder> without <disorder> concepts. |
|
| Please post your final responses in the Country response table below. Discussion comments can be made as comments. |
Links
PCP-59 - Getting issue details... STATUS
2018-01-26 Editorial Advisory Group Conference Call
Country response
| Country | Date | Response |
|---|---|---|
| Member countries without a CMAG rep |
CMAG response
| Date | CMAG Response | Next steps |
|---|---|---|
Final outcome:
Date:
11 Comments
Daniel Karlsson
In the Swedish extension there is one (primitive) instance of <disorder> without <disorder>, 171000052102 | Urinary tract infection without fever (disorder) |. This is used for collecting information about antibiotics use nationally. These could also alternatively be represented as a single situation with two role groups, one with 408729009 |Finding context| = 410515003 |Known present (qualifier value)| and one with 408729009 |Finding context| = 410516002 |Known absent (qualifier value)|.
What kind of response is expected?
Jim Case
Daniel,
The pattern you propose was proposed a few years ago and was essentially ignored, but it might be good to look at the implications again. From an implementation standpoint it causes problems when one half of a with/without pair is in findings and the other is in situations. That, I think was one of the primary oppositions. I have used this pattern in the US extension for testing and it certainly works out logically, but again, places it in the situation hierarchy, which confounds folks.
Elaine Wooler
I've added the extrapolated UK Primary Care use based on Read codes mapped to SNOMED CT. We have a number of reference sets which may also contain this type of content. This includes UK extension content as well as International.
Jim Case
Elaine,
Not sure how to read this attachment. What do the values "0-50" mean in the columns?
Elaine Wooler
Hi Jim
0-50 means the true absolute usage value - when standardized to a constant population of 65 million people over one year – would be between zero and 49 inclusive
50-250 means the true value would be between 50 to 249 inclusive
So 0-50 is a rarely used code and 50-250 not much better! - though this is for Primary Care.
Elze de Groot
We have several of these concepts in the Dutch extension for example 12171000146107 |Penetrating wound of eye without intraocular prolapse (disorder)|. But most of our concepts are 'without spinal cord lesion', and these are children of 269061001 |Fracture of vertebra without spinal cord lesion (disorder)|. I partly agree with Daniel about the modelling of these concepts, a situation with two role groups can be a solution We also considered this solution for the modelling of some concepts some time ago, but it wasn't allowed.. I just am not sure about the context part of the solution. You are not able to use the disorder for an initial diagnosis that isn't sure yet, because of the 'known present' part. And another withdrawal is that it does not show up in the same hierarchy as the disorder without the ' without'. So it might be dependent of the kind of concept you want to model which solution you want to choose.
John Fountain
There has been no response about the development/use of this type of concept in New Zealand. However, there is a post-coordination approach to deal with drug indications involving "without" type concepts e.g. panic disorder without agoraphobia, pernicious anemia without neurological involvement, atrial fibrillation without left ventricular dysfunction, etc
Matt Cordell
We've only got two extension concepts matching this pattern. "Ulcerative colitis without dysplasia (disorder)" "Crohn's disease without dysplasia (disorder)".
Both primitive, with just the 'present disorder' attributes stated. I suspect a single concept solution is needed/expected by the community of practice.
Agree with the Jim Case's comment about having the content split across heirarchies.
A typical retreival of "All patients with Ulcerative colitis" might be done as simply as:
<< 64766004 | Ulcerative colitis |And one could exclude the patients with (asserted) dysplasia
<< 64766004 | Ulcerative colitis | minus << 308870004 | Dysplasia of colon |But neither of these queries is going to include a concept from the SWEC hierarchy, if a situation concept was created.
Of course the queries could be modified to include these, but it's not going to be pretty... Something like the following is really needed to get all the patients with a given finding.
<< 64766004 | Ulcerative colitis | OR (<< 413350009 | Finding with exlicit context| : 246090004 |associated finding | = << 125605004, 408729009 |finding context | = 410515003 |known present |)It's seems cumbersome though? But probably needed already regardless of the "without <disorder>" use case. A patient with 51436006|Unilateral traumatic amputation below elbow without complication (situation)| isn't going to come back for a patient with Traumatic amputation (<< 262595009).
Then you look at concepts like 225551005|Wound discharge finding (finding)|. It has a single subtype 239164002|Wound discharge (finding)|.
What's this second concept mean if not the same as the first? Then see 698752003|Absence of wound discharge (situation)|.
I know there's implications around negation etc. But perhaps the problem is with the SWEC hierarchy... The historical concepts make sense to have separate. But having unilateral and 'known absent' concepts treated as "not findings" does complicate implementation. Both clinical finding and SWEC heirarchies are usually valid for a "presenting problem" type field.
Apologies for the ramble... I went down a rabbit hole....
Linda Parisien
CA stakeholders were asked but no feedback was received. As a result, we don't know if these are use and the frequency. We don,t have these types of concepts in our CA Edition.
Olivier Bodenreider
There is abundant literature arguing for considering A without B a situation involving disorder A. I understand Jim's reservations about having disorders and situations for similar things and how it might confuse users. On the other hand, we also want to model this in a way that is logically sound (e.g., to support analytics and clinical decision).
Jim Case
So in light of the fact that we do not know who uses these terms and how frequently, we can "safely" redevelop these into situations that more appropriately represent the specific absence of an aspect of the disease.