Quartz 5

Concept — SMART-on-FHIR

Properties2
descriptionSMART-on-FHIR: the open EHR integration standard that enables third-party apps like Abridge to read schedules and write clinical notes.
tagsconcept, ai-internship, neurology, interoperability-standard

3 min read

SMART-on-FHIR — EHR Integration Standard

What SMART-on-FHIR Is

SMART-on-FHIR is an open standard that allows third-party healthcare software applications to securely access and interact with electronic health record (EHR) systems. It combines two standards:

  • SMART — Substitutable Medical Applications, Reusable Technologies
  • FHIR — Fast Healthcare Interoperability Resources (a modern data format standard from HL7)

Think of it as “the app store model for healthcare software” — SMART-on-FHIR creates a standardized way for apps to read from and write to EHRs without requiring custom integration work for each system.

Why It Matters for AI Medical Scribes

Before SMART-on-FHIR, integrating with an EHR required:

  • A bespoke integration with each EHR vendor (Epic, Cerner, athenahealth, etc.)
  • Hospital IT teams to build and maintain custom connections
  • Significant cost and time (often 6–18 months of integration work)

SMART-on-FHIR changes this by providing a standardized API layer that works across EHR systems. Once an app is certified as SMART-on-FHIR compliant, any hospital running a compatible EHR can install it without custom integration work.

For AI medical scribes, SMART-on-FHIR enables:

  • Reading patient schedules from the EHR (to know which patient is being seen)
  • Writing clinical notes back into the EHR (to deliver the AI-drafted note)
  • Reading patient context (relevant history, medications, problem list)

SMART-on-FHIR in the Abridge Workflow

1. Clinician opens Abridge app
        ↓
2. Abridge reads patient schedule from Epic (via SMART-on-FHIR)
        ↓
3. Clinician records the patient encounter
        ↓
4. Abridge generates the clinical note
        ↓
5. Abridge writes the note back to Epic (via SMART-on-FHIR)
        ↓
6. Clinician uses dot-phrases (.hpisec, .meds, etc.) to pull
   sections into the Epic note template

This is the integration architecture behind Abridge’s EHR write-back capability.

FHIR Resources — What Gets Exchanged

FHIR organizes healthcare data into “resources” — standardized data structures for things like:

  • Patient — demographics
  • Encounter — a clinical visit
  • Observation — a clinical finding or measurement
  • Condition — a diagnosis or health concern
  • MedicationStatement — a medication being taken
  • DiagnosticReport — lab or imaging results
  • DocumentReference — a clinical document (like an AI-drafted note)

A SMART-on-FHIR app can read and write any of these resources via a standard REST API.

The Competitive Moat of Deep EHR Integration

The depth of SMART-on-FHIR integration varies significantly by vendor:

VendorSMART-on-FHIR depthNotes
AbridgeDeep — reads schedule, writes notes, supports dot-phrasesConsidered deepest ambient AI integration
Nuance DAXDeep within Epic (via Microsoft/Epic partnership)Uses Dragon Medical One + GPT-4
DeepScribeModerateFocused on outpatient/ambulatory
SukiModerateLightweight, faster implementation
FreedLightSelf-serve, minimal EHR integration

The deeper the SMART-on-FHIR integration, the higher the switching cost for a health system — because the dot-phrase library, training, and clinical workflows are all built around that specific vendor’s integration.

Graph View

Backlinks