Skip to content

Document Retrieval

Pull patient documents that Unomed has prepared (e.g. from the inbox) into your primary system, using the standard FHIR DocumentReference API.

This guide covers the case where the documents already exist in Unomed. It does not cover how they get there (email ingestion, upload, patient-matching).

The model

  • You pull, we don't push. Poll the DocumentReference endpoint on your own schedule and file the documents on your side.
  • Standard FHIR. Search returns a FHIR searchset Bundle (application/fhir+json); each entry is a DocumentReference whose attachment carries a short-lived download URL. Any FHIR-aware client works.
  • Org-scoped. Every endpoint is scoped to your organization UUID and only returns that organization's documents.

Flow

sequenceDiagram
    participant PS as Your primary system
    participant U as Unomed PACS
    PS->>U: GET /{org}/fhir/r4/DocumentReference?date=ge…&_count=100
    U-->>PS: Bundle (searchset) of DocumentReference
    loop each entry
        PS->>U: GET attachment.url (presigned) — or Binary/{id}
        U-->>PS: file bytes
    end
    PS->>PS: file document under the patient, advance the poll cursor

Base URL and auth

Sandbox:

https://pacs.dev.unomed.ch

The production URL (pacs.unomed.ch) is shared during rollout. Every call needs an OAuth2 access token as a bearer token, scoped to your organization:

Authorization: Bearer YOUR_ACCESS_TOKEN

Contact info@unomed.ch to receive credentials for your organization. Replace YOUR_ORG_UUID with your organization's UUID throughout.

See the API Documentation for full endpoint details.

Next steps