And there is OCT, which is - if I’m not mistaken - not a specialisation of the generic Archetype, even though it is a different type of imaging examination.
Which of these should I use for an OCT Examination? If I understand correctly, then the best practise for all types of imaging examinations would be to specialise the generic imaging archetype - is there a reason why the OCT archetype deviates from this practice?
The OCT archetype is old and unpublished so probably no one has made the effort to bring it into line with the design pattern of the imaging exam.
Feel free to do so! We’re here to help:)
Yes the current preferred approach is to use the imaging exam archetype along with Cluster archetype specific to the modaility and site eg OCT. In ybe details slot
We woud be interested in collaborating as we have been doing some work on opthalmogy.
One challenge is tbat this kind of imaging reporting has been pretty non standardised, other than DICOM which is much more about backend RIS.
The OBSERVATION.imaging_exam_result provides the starting framework for all imaging results reporting, no matter what the modality.
DICOM documentation is supported, although limited to context of a report only
There is a growing group of CLUSTER archetypes related to imaging of regions or structures that are modality agnostic, where possible.
The best way to model other very specific archetypes such as Cobb angle and Fetal biometry is being investigated. As are reusable components across regions/structures, such as US vascularization findings.
Published archetypes have been through a peer review process (green tick icon) so are regarded as relatively stable & any changes will be tightly governed.
Draft archetypes within a Project are probably still under development, could change significantly through the peer review process & should be used understanding that they are not stable. Here be medium-sized dragons.
Any other archetypes found in incubators should be considered with caution. They may be historical, or for a specific use case or a very early draft. Here be the largest dragons.
We are currently working with opthalmogy. Like screening for diabetic rethinopaty. There are different techniques for imaging. The imaging examination archetype seems to be a good information model for any imaging report. For opthamolgy the image is the primary source for diagnostic, findings and other descriptions. This seems to match the intention of the archetype.
Did you also use the imaging exam archetype in your project?
Would this be a good choice?
@heather.leslie - I like the quoted item above. And as mentioned in the post above we are exploring the usage of this archetype for imaging examination of the eye. These are not radiology based. There are different techniques applied for the imaging procedure. The data we are modelling are the doctors evaluation/assessment of the findings in the image. This is why the imaging_exam_result archetypes seems like a good fit.
The only issue is the use-definition of the archetype: Use to record the findings on imaging examination of a specified body structure or region and their interpretation , using radiological techniques.
If we skip “using radiological techniques” everything will be ok ?
Let me know if this intepretation is acceptable or if anyone has other opinions on this.
PS!
The term “we” above is a cooperative project with Helse Nord and DIPS. @lekkim is also involved in this project - and we have an ongoing discussion with the editorial group for archetypes in Norway - arketyper.no
There are ongoing editorial discussions about the concept OBSERVATION.imaging_exam_result with regards to OCT and similar ophthalmology-related imaging. Will update this thread when discussions have concluded.
The Snomed-CT tree does indeed indicate OCT as a child of [Radiographic imaging procedure (procedure)], so there is that.
I completely agree with this suggestion. I’m sure we will come across exceptions to the rule, but I think there is should be a good presumption that the start point for imaging results of any modality should be handled by the OBSERVATION.imaging_exam_result archetype with detailed clusters.
I’m less sure about the current approach to then specialise this archetype to broad anatomical locations, as it feels to me as if we can handle any commonality just as effectively by using terminology.
Specifying and quering anatomical locations can certainly be done with only the base CLUSTER.imaging_exam, which is the same pattern as for CLUSTER.exam. Indeed the main common difference between the parent and children archetypes is the constraining of the ‘Body structure’ using SNOMED CT terms. But when more specific data elements need to be recorded, an additional archetype is required in any case.
Whether this additional archetype is a specialisation of the CLUSTER.imaging_exam or an extra CLUSTER containing only the relevant elements is to me less important, other than the fact that we chose the specialisation route back when and have by now published several archetypes following this pattern.
If we’re to change our choice of pattern now, IMO there has to be a better reason than being less sure about it
