fhir-bulk-data

FHIR bulk data access uses the $export operation to initiate large dataset retrievals. You trigger the export via a specific endpoint, then implement an asynchronous polling workflow to check status and retrieve generated NDJSON files when ready.

NDJSON, or Newline Delimited JSON, is the file format used in FHIR bulk data exports to stream large volumes of resources. Each line contains a single FHIR resource, allowing efficient processing of large datasets without loading entire files into memory.

SMART Backend Services authorization secures FHIR bulk data access by using asymmetric cryptography. Clients authenticate via a signed JWT token to obtain access tokens, enabling machine-to-machine data retrieval without interactive user consent.

Building a FHIR bulk data client requires initiating the $export request, receiving a polling URL, and periodically checking its status. Once complete, the client downloads and processes the output NDJSON files from the provided links.

FHIR bulk data access supports exporting specific resources like Observation and Condition. You can use type-level $export operations or apply parameters to filter data by resource type and date ranges.

Asynchronous polling in FHIR bulk data workflows requires handling potential timeouts and large file downloads. Servers may delete output files quickly after generation, so clients must download NDJSON files immediately upon completion.