FHIRcast | Medplum

Overview

FHIRcast is a lightweight, topic-based publish/subscribe protocol that enables healthcare applications to synchronize context in real-time. Built on HTTP and websockets, it facilitates seamless communication between a "context source" (the application initiating the context change, e.g., an EHR) and "context subscribers" (applications that need to react to that change, e.g., a clinical decision support app).

Key Concepts

• Context: Refers to the current patient, encounter, or other relevant clinical information that needs to be shared between applications.
• Context Source: The application that publishes context changes. Typically, this is the primary clinical application, like an Electronic Health Record (EHR) system or Picture Archiving Communication System (PACS).
• Context Subscriber: An application that listens for and reacts to context changes published by a source. These can be various SMART on FHIR apps, decision support tools, or other integrated systems.
• Topic: A specific channel for context updates, often identified by a UUID. Applications subscribe to topics to receive relevant context changes.
• Event: A specific type of context change, such as Patient-open, Patient-close, Encounter-open, or Encounter-close. FHIRcast defines a set of standard events.
• WebSockets: FHIRcast primarily uses WebSockets for real-time, bidirectional communication, allowing subscribers to receive updates as they happen.

How FHIRcast Works

1. Handshake: A subscriber initiates a connection with a context source, typically by establishing a websocket connection.
2. Subscription: The subscriber expresses interest in specific context topics (e.g., patient context).
3. Context Change: When a user in the context source application (e.g., an EHR) changes the active patient, the source publishes a FHIRcast event to the relevant topic.
4. Notification: Subscribers to that topic receive the event, which includes the new context (e.g., the Patient FHIR resource, which contains information like the patient's ID, name, date of birth, etc.).
5. Synchronization: Upon receiving the event, the subscriber application can then update its own display or internal state to reflect the new context.

Use Cases

• EHR to SMART App Integration: An EHR system publishes Patient-open events, and a SMART on FHIR application automatically loads the newly selected patient's data.
• Clinical Decision Support: A CDS application subscribes to Encounter-open events, and when a new encounter is opened, it automatically retrieves relevant patient data to provide real-time recommendations.
• Multi-Application Workflows: In a clinic setting, as a clinician moves from patient to patient in one application, other connected applications (e.g., a billing system, a lab order entry system) automatically switch to the same patient context.
• PACS Synchronization: A Picture Archiving and Communication System (PACS) can subscribe to patient or encounter context changes, automatically loading the relevant imaging studies when a patient or encounter is selected in another clinical application.
• Streamlined Radiology Workflows: As a radiologist creates imaging measurements in a PACS, the PACS creates DiagnosticReport-update events containing ImagingSelection resources, which informs the report creator of the new measurements. As the radiologist is dictating and highlighting relevant measurements in the PACS, the PACS sends DiagnosticReport-select events, letting the report creator know which measurement is being referenced at different parts without any extra input from the radiologist.