The Self Monitoring archetype (still in v0) has an open change request to get a widened scope to instead cover all kinds of self reported data.
Would that be a good idea or should we keep self monitoring as it is and create a new (wider) archetype for ‘Self reported data’? (E.g. with a concept description like ‘A composition to record data reported by the individual who is the subject of the record’.)
We need such a composition archetype farily soon at Karolinska and could create one locally, but would prefer to know where the international alternative is heading and rather contribute to that.
It seems like a good idea to define the scope for this composition archetype to be a container for all self reported data.
Self reported data will come in different flavours. One is data created by the user (ie forms). The other will be devices installed at home or even attached to the body of the patient. An open question is whether we need to be to distinguish those categories?
Perhaps that can be distinguished at at a lower SECTION or ENTRY level but, in many cases a COMPOSITION will contain a mix of “measurements” and “entry form elements” submitted togeteher, so I believe it would be good to have a composition that can carry a mix of both.
I would just rename the composition in the template to reflect that different usage, if needs be.
A more subtle difference is whether the information is regarded as curated and trusted and monitored and therefore clinical actionable.
I would think we should be regarding such patient-entered or direct-from-device measurements (e.g. BP, fetal heart rate, blood glucose) as being pretty reliable, and using the normal BP and other OBS archetypes. That means a priori querying will find them in the normal way and count them in the total samples (and the patient data might be the only samples), but if there was a reason to doubt or discount them, a query can be constructed to not include BPs etc inside this particular kind of Composition.
Is that what is being suggested here?
Unfortunately BP is one of the most error prone measurements. So assuming patient measured BP is reliable is risky. Blood glucose sensors are often unreliable unfortunately, so a professional filtering out the wrong measurements does add value to the recorded measurement.
But this also goes for these measurements when take on by a professional, it will depend on their qualifications, skill, knowledge etc, AND the risk tolerance of the reader how reliable a measurement is. I think AQL is the right level to solve this problem. But we should continue to spend serious thought on this issue.
If you want to explicitly mark measurements as potentially unreliable by putting them in a specific composition I think that’s an acceptable hack.
The change request has now been closed with this comment by @heather.leslie
Marit has worked up a COMPOSITION specifically focused on the recording ‘Self reported data’. It was proposed and imported to: https://ckm.openehr.org/ckm/archetypes/1013.1.6343.
The original scope and use case for this ‘Self monitoring’ archetype remains more on recording their own measurements and the source is unambiguous - BP monitoring, glucometer readings, PEFR readings from devices etc. The scope probably needs to be worked on for further clarity to reflect changes in understanding since the original authoring in 2013.
Great! Thanks for the new archetype (Marit as in @Marit_Alice_Venheim I supppose).
The above mentioned new Composition archetype “Self reported questionnaire” will work for Karolinska’s current immediate need that happens to be a classic questionnaire, but the future need of a more generic “Self reported data” that may contain a mix of…
- instrument readings (both manually entered and automated recordings e.g. from smartphone)
- questions (from questionnaires etc.) and
- (extracts from) PHR content
…still remains (but is not as urgent).
And the new archetype from @Marit_Alice_Venheim will be on review shortly. It has been worked on for a while, and the Change Request has just stayed there unattended