Existing Wording: conclusion, codedDiagnosis (attribute names)

Proposed Wording: finding, codedFinding (new attribute names)

Comment:

The term ?finding? is ambiguous across FHIR resources, as follows.

The DiagnosticReport description starts with ?The findings and interpretation of diagnostic tests? but it?s unclear whether findings refer to the result attribute?s referenced observations or to the conclusion & codedDiagnosis attributes. Furthermore, in the attribute name ?codedDiagnosis?, the included term ?diagnosis? seems more advanced on the clinical reasoning process than just finding or (interpretation), so it's somewhat inconsistent with the above description.

Diagnosis in turn is typically captured by the Condition resource, however, in the Condition description, finding is also mentioned in one of the examples where it is equated with diagnosis: ?It can be used to record information about a disease/illness identified from application of clinical reasoning over the pathologic and pathophysiologic findings (diagnosis)?. However, the essence of Condition seems to be more on problems, disorders, diseases, etc., per the resource description and most of the examples included there.

In the Observation resource, the finding concept is mentioned for the first time in section 10.1.4.4.3 (Interoperability Issues using code value pairs in FHIR), but is not defined nor clrearly distinguished from the concept of observation itself. In one of the patterns suggesting to bind the code attribute to the SNOMED CD terms with semantic type of clinical finding (a SNOMED top level hierarchy), the spec seems to be blurring the distinction between findings and observable information items like those defined by another SNOMED top level hierarchy of ?Observable entity? (Code 363787002).

Thus, finding seems to be an elusive concept and is so perhaps due to being the intermediate level of conclusion, somewhere between discrete observation?s interpretation and a more established diagnosis, captured in a Condition resource. For example, several abnormal ECG observations may be associated with a suspected atrial fibrillation finding, but to establish a paroxysmal atrial fibrillation diagnosis, there got to be a few of such findings appeared in specific certainty, order and timing, etc, as defined in clinical guidelines.

The DiagnosticReport.codedDiagnosis is bound to SNOMED Clinical finding hierarchy (the same as in one of the aforementioned Observation code-value pair patterns), but in DiagnosticReport the binding is at the spec level itself, and thus seems more substantial to this resource. Therefore, it is proposed that the DiagnosticReport resource will be described as the main placeholder to capture the concept of finding in a more explicit way. To this end, the ballot proposal is add description and calrity in this spirit, and to rename the conclusion and codedDiagnosis attributes to finding and codedFinding respectively, so that they reflect the finding ?nature? of DiagnosticReport backbone (i.e., code, results and conclusion).

In this way, the distinction between the terms of observation, finding and condition will be made clearer, at least within the scope of DiagnosticReport. To complete this change, there got to be more clarity in Observation and Condition about finding, and that might be submitted in separate ballot comments, however, if this ballot comment is found persuasive, then DiagnosticReport will become the focal point of findings.

Summary:

Make DiagnosticReport the finding placeholder