Dear informatics colleagues,
A tricky semantic conundrum for you.
The Anatomical location archetype was published initially in 2015 and then republished in 2017 as 1.1.0 - https://ckm.openehr.org/ckm/archetypes/1013.1.587/18. The intent was to be able to identify a specific place on the body, at a surface anatomy/macroscopic level, where we observed tenderness or a lump, lesion, burn, laceration, bruise or similar as part of physical examination. In that context it only made sense that a lump could only be in one place, and if there were multiple lumps then we need to be able to describe the location of each lump in sufficient detail so that we could tell one lump from another on physical examination or accurately record which lump we had excised.
Confusion starts to arise when people want to express clinical findings the other way around in some circumstances. We wouldn’t usually say bilateral lumps or burns, but describe the findings specific to each lump. But it is not unreasonable to want to describe that there was bilateral pneumonia as part of an assessment on an Xray or as a diagnosis. Or to schedule a ‘Bilateral myringotomy’ procedure to be put on the surgical wait list - both ears to have grommets inserted. Or a problem list might record bilateral hip replacements as being done, even though they were done a year apart and would have been recorded as separate operations at the time. There have been occasions where modellers have wanted to use the anatomical location archetype to support these kinds of use cases, especially in the messy areas of secondary use of data eg in a registry, but they are subtly (and importantly) different.
In Mar 2017, a change request was made - CR-149 ‘Left and right are not sufficient for laterality’. You can see in the notes that I pushed back quite strongly, and we eventually agreed that the data was messy in registries. I think that this resulted in the creation of the ‘Anatomical side’ archetype in the Apperta CKM as a means to resolve the local registry data representation.
In July 2019, there was an additional request to add ‘a body part’ with the ability to specify if it is ‘left / right / both sides / middle / unknown’ - see Discussion thread entitled ‘Laterality’. The clinical context was another data registry requiring recording of the part of the body irradiated - for example ‘Irradiated target: palatine tonsils; Position of the target area: bilateral.’
In 2019, we decided to try to embrace this repeatedly stated need, despite the fact that it broke our original intent. We added ‘left & right’ as a value to ‘Laterality’ and added the ‘Occurrence’ data element with values ‘Unilateral’ and ‘Bilateral’. This archetype is currently our latest revision, in the ‘Reassess’ state meaning that it needs to go through a review to ratify or amend the new content before republishing.
However as I’ve considered starting a new review, I’m increasingly uncomfortable with the new content. It is hard to express it precisely, but changing the scope/intent of the archetype has compromised it’s original integrity. It is hard to explain/justify the changes to outsiders, which raises a red flag that this is not a semantically sound model.
It is interesting to note that we have created a modelling pattern that has been used successfully in many archetypes now, but started in the Blood Pressure model where we added both a data element for ‘Location of measurement’ and then also a SLOT to carry the ‘Anatomical location’ archetype if we needed to describe the site in more precise details. For 99% of situations we anticipate that the simple coded value set has been adequate, and I certainly haven’t ever had the need to fill the SLOT. IMO the key point here is that we identified that while we were more recording the method of taking the BP using a general body site value set, recorded as part of Protocol rather than a precise anatomical location within the Data itself - this is an important, but subtle difference.
So when we consider how to record a target for irradiation, it is also more a method rather than purely describing where something was on the body. So in an INSTRUCTION for radiotherapy it seems appropriate to record the ‘Target area’ as ‘bilateral palatine tonsils’ rather than the ‘Anatomical location/site’ of bilateral tonsils. If a precise area needed to be described then a SLOT for use of Anatomical location might be useful to describe each site accurately (as per the BP pattern).
Similarly ‘bilateral pneumonia’ or ‘bilateral hip replacements’ are fine to be recorded as ‘Problem/diagnosis name’ or ‘Procedure name’.
So, currently the archetype’s reassess status needs resolution, and our usual way to do that is to send it out for another review, but in a review it is not easy to support reviewers to understand the history, context and specific issues needing resolution.
This is a long post, but the point of it is to request feedback on a proposal to revert the archetype back to the last published revision, focused on describing a single place on the surface of the body.
I’d appreciate your thoughts.