aehrc / spia-to-fhir Goto Github PK

It would be good if the transformation could also use the configured FHIR server to validate that the resources are valid according to the profiles that they assert.

There is a Log4J2 configuration file at src/main/resources, and Log4J dependencies are included in the POM.

However, this configuration does not seem to be active, as logging does not currently include times, originating class, etc.

Currently, we populate the title into the narrative. However, the narrative status is set to empty.

It should really be set to generated.

Currently, the date element on ValueSets and ConceptMaps are set to the date and time of the transformation. A more useful date for use within these elements will be the date of publication.

We need a way of passing the publication date in as a parameter, so that the dates within the output are stable and deterministic.

Currently there are some cases in which whitespace is trimmed from source content and omitted in the FHIR representation.

It would be good if anything like this is emitted as a warning to the logs, in the same way code validation results currently are.

This will allow for deficiencies in the source data to be fed back to the authors and addressed, so that the spreadsheet and the FHIR representation remain faithful to one another.

Currently, the plugin checks for the well-formedness of SNOMED and LOINC codes. It does not, however, check whether the codes are active.

This would help improve the quality of the source data, and prevent publishing ValueSets and ConceptMaps which refer to inactive codes.

There is a flag in the Chemical Pathology Reference Set which indicates whether it is safe to combine results for a particular test.

It would be great if we could somehow include this information in the FHIR bundle.