MII FHIR Flattener - Overview

Flattening is the act of transforming FHIR Data from an object format into a flattened format - typically csv or parquet.

To flatten data the fhir-flattener webservice is used, which based on the pathling library takes a FHIR viewDefinition and FHIR resources and flattens them.

Use in the DUP pipeline

In the DUP pipeline aether takes the extracted resources and based on a CRTDL generates ViewDefinitions for the specific DUP using a flatteningLookup table, which is provided as part of the FDPG ontology.

The lookup table is still in development, but a first version can be downloaded from our example deployment to try out the flattening in conjunction with aether.

The flattening rules for the Core Data Set (CDS)

The flatteningLookup table is generated based on specific rules and the CDS profiles as described here.

Use of the flattener without aether

The fhir-flattener can also be used without aether and the viewDefinition lookup table. In this case the the users have to create their own view definition and use the fhir-flattener api to flatten their resources.