Mapping from A.5.4 Results to Finding

[chgessner]

Discrimination between result and observation appears insufficient. More explanation needed.

[bvdh]

I not sure what you refer to. A Finding is an Observation, where do we discriminate between them?

[chgessner]

It is somewhat related to issue #20. I will work on a clarification before the next meeting. I am referring to result and observations in the eHN dataset and how they are mapped to the report.

[chgessner]

There is a quick solution and a longer discussion.

The quick solution is:
In Group 2 fix the TargetCode to section[findings].entry[finding].
This solves my problem of the reverse mapping being "insufficient". The reason was that there is another mapping in https://build.fhir.org/ig/hl7-eu/imaging/ConceptMap-A-5.5-ConclusionMapping.html Group 3, mapping between
conditionOrFinding (A.5.5.2.1 Condition or finding) and section[impression].entry[finding], competing with the current TargetCode section.entry[findings]

The longer discussion should add clarification to the hierarchies of Observations and Conditions in various parts of the specified structures: While the Guidelines have conditionOrFinding and stagingOrGrading under impression (A.5.5.1 Impression), the Composition provides section[impression].entry[finding] and section[impression].entry[impression]. And then there is a somewhat oblique mapping in Group 2 between both Guideline entries and Impression in Mapping of A.5.5 Conclusion. Moreover, the Report provides conclusion not as section or section.entry, but as markdown with CodeableConcept. So IMHO there is a need for more explanation and examples. This should include an explanation of which Findings go in section[results] and which go in section[impression].

[bvdh]

Christof, I'm not sure I understand your concern.

Do I understand you correctly that the source of your concern are:

• the fact that findings are referred to from the findings and impression sections
• that the appearance of findings in both sections makes it difficult the map the report back to the logical model.

[bvdh]

I totally agree we need to work on examples and show in what way elements of reports are reflected in this structure.