In every FHIR resources being copied from A to B, is it good enough to add capability to overwrite `meta.source` as such it tells where does the FHIR resource actually sourced from?

[muhammad-levi]

Proposed scenario

GIVEN a federated architecture setting of multiple FHIR Servers, assigned one of the FHIR servers as the "central", WHEN google/fhir-data-pipes sync the FHIR servers other than the "central" to the "central" FHIR server, THEN all the FHIR resources copied to the "central" FHIR server has meta.source value equals to ${SOURCE_FHIR_SERVER_URL}/${RESOURCE_TYPE}/${RESOURCE_ID}

Implications

• Ability to trace back to the original source
• Some analytics grouped by different original sources

References

• FHIR Resource Metadata
• https://www.hl7.org/fhir/R4/search.html#uri

[bashir2]

Thanks for the suggestion. Yes, if we decide to make the data consolidation features of the pipelines production-ready, we need to add some sort of resource tracking feature too. This is in addition to other requirements that I mentioned under #825. Using the meta.source field is one idea as you suggested; there are also other ideas like creating Provenance resources which we need to consider. Adding @fredhersch and @pmanko as we have discussed these FHIR-server to FHIR-server scenarios recently.

[muhammad-levi]

Thank you. Yes, as the FHIR spec described, meta.source provides a minimal amount of Provenance information and maybe making use of the Provenance resources might enhance the experience of for example data quality investigations. I will try my best to help contribute to these as well adding "FHIR-server to FHIR-server" to the controller.

[bashir2]

This is related to both #825 and #923. Depending on how/if we support those features and if we end up with supporting multiple FHIR-sources, we may want to implement this feature too as an option.