Anatomical location

There will definitely be cases where bilateral is used as a qualifier, without a specific organ location e.g ‘bilateral mastectomy’. And I do appreciate that conceptually this is different from an actual physical location. My worry is that it is going to be a very fine distinction for most people

I guess the use-case which is making me try to stick with the single way of ‘doing anatomy’ is @bna DIPS Opthalmology example where they wanted to identify ‘all things about the left eye’, using SNOMEDCT irrespective of the containing Entry i.e to include orders procedures, diagnoses etc. Current AQL implementations let us do that on CLUSTERS, but not on ELEMENTS.

In that case I would ensure that a statement like ‘bilateral cataract’ did have the organ location ‘Eyes’ held in anatomical location as well as ‘bilateral’. or the unilateral equivalent.

this is horrible!!

This is definitely an ambitious use case. But it seems more sustainable to refine AQL or the algorithm/logic rather than munge the archetypes. For the use case I’d suggest again that you specialise for your purpose. In the meantime we have identified a need to extend the archetypes with ‘Side’, including ‘bilateral’ as a value and unless there are objections will make CRs with the intent to update them accordingly.

1 Like

I’m happy with that approach. It is a messy area and we may just have to live with a bit of variation to suit differing circumstances.

In some ways, Silje’s earlier suggestion about using SNOMEDCT and pre-corrdinated terms or even post-coordination limited to ‘laterality + xyz’ would be the best solution, but that would mean SNOMED relaxing their license.

Perhaps it is worth discussion (with SNOMED ? FHIR)

Seems like we have a plan for the archetypes, at least. Thanks for the robust discussion everyone :star_struck:

1 Like

Nothing better than a decent brawl, underpinned by Marquis of Queensbury Rules for Gentlefolk :slight_smile:

I strongly object to rule #2. :joy:

Yes - I see your point - perhaps I used the wrong sport.

Rule 2 is what we do.

Ok - do we need a new archetype based on Rule #2, or can we blend it alongside the other rules (and constrain out) or are we back to specialisation as a compromise? :space_invader: :space_invader:

1 Like

A wrestling archetype? Or am I totally lost? :woozy_face:

My entirely arbitrary use of the space invaders emoji may have over-loaded an already convoluted and over-stretched ‘meta-joke’.

So apologies, my bad, but yes you are lost :wink: Quite probably along with everyone else - it made sense to me at the time :joy:

It’s a familiar feeling… :confounded:

Shouldn’t bilateral just be a user interface artefact to select both left and right to be recorded?

1 Like

In some use cases this could well be a good solution! :grinning:

Hi,

Good point, Colin. You can get around the problem that way, but I am afraid that doesn’t solve the connundrum about expressing localization in human sorts, while being precise as we must be.

Please don’t take my thoughts as more than a go in keeping the discussion alive. I am not sure at all about the right way to express this.

For laterality, you can assume it after picking left and right sides. That still means one more click, or engineering an UI-wise way of recording both in a go.
For most cases, this is a useful way of summarizing a bilateral problem, albeit it doesn’t seem to accommodate for assimetry. ‘Bilateral’ may be seen as a short expression of a kind of laterality. Maybe it should be viewed (and represented in the IM) as a coarser, though informative, way of localizing something. Other examples would be ‘diffuse muscle pain’, ‘centrifugal progression of an exanthema’, etc.

IMHO, UI artifacts can help to keep accordance to a consistent information model, but the IM must express those human-readable forms in a logical way.

Best regards,

A segunda, 10/08/2020, 09:03, Silje Ljosland Bakke via openEHR <discourse@openehr.org> escreveu:

Coming late to the party (sorry).

I think that bilateral eye is not a anatomical location. It’s actually two anatomical locations. The human body has in fact two eyes if you didn’t loose one.

When it comes to defining anatomical location within the body we should be specific enough.

I approve the suggestion from Colin about leaving bilateral to the user-interface. It’s a situation when you create data which are equal for two different parts of the body.

2 Likes

Sorry for opening this thread again. Just want to share some thoughts on the bilateral issue.

First a metaphor.
Life is like a box of chocolates, you never know what you’re going to get. ” –Forrest Gump.

Bilateral is not like Forrest Gump describes it. Bilateral is a box where you know what you get. You get two similar chocolates - or a more clinical example is two identical attributes of both eyes, knees or any other pair-organ. For INSTRUCTION and ACTION you get a box with two identical requests or procedures performed.

Depending on the level-of-details the two items in the box might be identical or have minor difference. I.e. you might have a box defining decreased bilateral knee flexion . Looking in the details the right knee might be worse then the left.

Bilateral is therefore a complex and hard semantic term. The interpretation of it will depend.

Anatomical location should be used to define a “locatable point” in the human body. Bilateral can not be used for this. This will be more related to some “organ classifier”. Like the eye, knee or hip. In this context we are not talking about a “point of interest” - but more about a function of the body.

2 Likes

Have you thought about co-opting the Foundational Model of Anatomy. In it Ears has a different number to Left Ear and Right Ear. It is already a fully specified ontology.

A

Hello everyone. Apologies for re-opening this discussion. I replied to an old Change request and Silje directed me to this thread, which I have read and I hope to have understood.

I am very new to the openEHR world, so there are some things I don’t fully understand, and my biggest doubt in this is what is the harm of having ‘bilateral’ as an option.

Whenever I have come across laterality, the options tend to be left, right, bilateral, midline, unknow. I understand that it could be a UI implementation and store in the background left+right, however, to be honest, we are pointing to an external reference data set with those values.

I believe there are use-cases in which describing something as ‘bilateral’ (or ‘left and right’ as SNOMED describes it) is relevant. For example:

  • bilateral symptoms and signs may point to one disease and discard others which are typically unilateral (ie Guillain Barré typically has bilateral and symmetrical muscle weakness or CHF has bilateral oedema in the legs and crackles in chest examination).
  • bilateral presentation may also trigger different treatments (ie a bilateral inguinal hernia may prompt a laparoscopic repair rather than conventional hernia repair or unilateral pain in a leg may be a DVT but it would be rare for it to be bilateral and a Doppler US may be requested in one and not the other).
  • metastasis can be bilateral as a simple description with no further requirements to clarify on them (ie bilateral lung metastasis on a rectal cancer, or bilateral node involvement).
  • COVID 19 was typically described as bilateral and the difference was established depending on how much of the lungs were affected.
  • CLIS as a risk factor for breast cancer can be bilateral.

I understand the theory of ‘things’ being only in one place at a time, but given that we tend to group in the descriptions, wouldn’t that justify having bilateral as an option and disabling it depending on the use-case?

Thanks!