fetching-fhir-resources

To extract clinical narratives from FHIR R4 resources, fetch Patient, DocumentReference, DiagnosticReport, Observation, and Condition resources, then decode base64-encoded text attachments to prepare the clinical content for de-identification and analysis workflows.

Yes, FHIR R4 resource retrieval supports Epic, Cerner, HAPI, Medplum, HealthLake, and other US Core servers for EHR chart retrieval, including handling paginated searches and patient or encounter rejoining across these platforms.

When fetching FHIR resources, the skill automatically handles Bundle pagination to retrieve complete result sets and decodes base64 attachments found in DocumentReference and DiagnosticReport resources to extract the underlying clinical text narratives.

SMART-on-FHIR authentication considerations are accounted for during patient records retrieval, requiring standard FHIR REST access to securely fetch and route clinical narratives through OpenMed privacy processing while preserving patient and encounter rejoin keys.

The best way to de-identify EHR clinical notes from FHIR APIs is to route the extracted narratives through OpenMed de-identification workflows, which removes sensitive patient data while preserving the necessary patient and encounter rejoin keys for grouping.

FHIR DocumentReference retrieval may return non-text content if resources lack base64 text attachments, and this non-text routing is handled automatically by separating clinical narratives from unsupported content during the extraction and de-identification process.