Introducing a Learning Project Integrating FHIR and HL7v2
fhir-interop
Summary
A developer built a pipeline that processes both clinical data formats, FHIR R4 bundles and raw HL7v2 messages, normalizing and validating them into a single internal model. It is implemented in Go and serves the data via GraphQL API and FHIR REST API.
Details
This project is a learning-focused pipeline designed to ingest and integrate clinical data from two different formats: FHIR R4 bundles and raw HL7v2 messages, mapping them onto a common internal model. The process involves distinct stages of 'Parse,' 'Normalize,' and 'Validate.' Technically, it features custom parsing for HL7v2 segments and mapping those structures onto the same Go structs used by JSON-based FHIR data. For quality assurance, terminology checks are performed against LOINC and SNOMED CT using a public server. The processed data is persisted in Postgres and exposed via a GraphQL API (the primary read method) and a dedicated FHIR REST API for Patient records. This implementation demonstrates a practical mechanism for connecting modern standards with legacy formats.
Original content copyright by respective publishers