ePCR Capture with Offline-First Sync

Overview#

Argus lets ambulance crews complete a full Patient Care Record on a tablet or mobile device with no network connection, then reconcile everything cleanly the moment connectivity returns.

A crew arriving at a rural scene with no signal still needs to capture the full Patient Care Record without compromise. Vitals, drug administrations, interventions, narratives, and witnessed signatures must all be recorded on the device exactly as they would be on a connected device, then reconcile cleanly the moment the radio or cellular link returns.

The ePCR Offline Capture module lets ambulance crews complete the entire Patient Care Record on the Argus responder tablet or Progressive Web App without network connectivity. Encounters are auto-created from the canonical incident the moment a medical call is dispatched, every field change queues locally, and updates replay with vector-clock and last-writer-wins conflict resolution per field once the tablet reconnects, so dispatchers see clinical activity on the incident timeline in near-real-time.

Key Features#

Auto-Created Encounters from Incidents: An encounter is opened the moment a medical incident is dispatched, so crews never start from a blank record and every encounter carries a verified link back to the originating incident.
True Offline-First Capture: Vitals, interventions, drug administrations, narratives, witnessed signatures, photos, and audio notes are all captured on the device without any network connectivity required.
Per-Field Conflict Resolution: Vector-clock and last-writer-wins reconciliation runs at the field level on reconnect, so two crew members editing different fields of the same record on separate devices never produce conflicting results.
Live Incident Timeline Updates: Every field update is published to the canonical incident, giving dispatchers near-real-time visibility into clinical activity in the field.
Narrative Dictation Assist: An optional Narrative Partner helps the crew structure free-text narrative from voice notes, while keeping the clinician in full control of the final wording.
Multi-Standard Export: A single captured encounter can be formatted as PHECC 2024 for Irish national reporting, NEMSIS v3.5 for cross-border or research export, or JRCALC 2024 for clinical guideline alignment.
Witnessed Signature Capture: Patient and crew signatures are captured locally as part of the encounter and travel with the record into the canonical store on sync.
Durable Replay Queue: The PWA service worker combined with IndexedDB on web, and equivalent persistent storage on the native responder app, guarantees that queued mutations survive app closure, device reboot, and intermittent reconnects.
Medical Device Streaming: The tablet agent accepts real-time vital-sign observations streamed over Bluetooth Low Energy from bedside monitors and wearables, buffering them locally when offline.
SMART on FHIR App Launch: Third-party clinical applications registered against the platform can be launched directly from within an active encounter using the SMART on FHIR 2.1 authorisation flow.

Use Cases#

Rural Scene with No Connectivity#

A crew at a remote incident captures the full Patient Care Record on the tablet, including vitals trends and drug administrations. Everything reconciles back to the canonical incident the moment the tablet reconnects, with no data loss and no double-entry.

Two-Crew Concurrent Editing#

One crew member updates vitals on the tablet while another adds the narrative on a second device. Per-field conflict resolution merges both contributions into a single coherent encounter.

Long-Distance Inter-Hospital Transfer#

An inter-hospital transfer crosses regions of weak or absent coverage. The ePCR keeps recording uninterrupted and replays in segments as connectivity comes and goes along the route.

Mass-Casualty Triage#

Multiple encounters are opened from a single incident, each one tied back to the canonical incident, so command sees clinical activity per patient on the same operational picture.

A clinical record captured once on scene flows to the PHECC 2024 dataset for Irish national submission and is also available as a NEMSIS v3.5 export for research or cross-jurisdictional sharing.

National Shared Care Record Integration#

When an encounter opens, the platform can automatically retrieve the patient's existing conditions, allergies, and procedures from the national shared care record and pre-populate the encounter, reducing manual transcription and improving clinical safety.

Integration#

Argus exposes ePCR capabilities over REST and GraphQL, secured by OAuth 2.0 bearer tokens (RS256-signed JWTs). All endpoints enforce tenant and crew scoping from the token claims, so multi-organisation deployments share no clinical data across boundaries.

Responder Mobile and PWA: Field screens on the native responder app and browser-based PWA connect to the same encounter endpoints. The PWA service worker hosts the mutation replay queue, providing identical offline behaviour on both surfaces.
Encounter Lifecycle Events: Encounter opened, field updated, and encounter signed events are published using the CloudEvents 1.0 envelope, allowing downstream consumers such as dispatch dashboards and reporting pipelines to subscribe without coupling to the clinical data model.
FHIR Export Endpoint: A dedicated endpoint assembles a complete HL7 FHIR R4 transaction Bundle from the captured encounter, including Encounter, Observation, Procedure, MedicationAdministration, and Composition resources, ready for submission to receiving hospital systems.
Standards-Aligned Formatters: The platform produces PHECC 2024, NEMSIS v3.5, and JRCALC 2024 exports from the same captured record without requiring re-entry.
Medical Device Streaming: An ingest endpoint accepts batches of IEEE 11073 observations from a Bluetooth LE tablet agent, associating them with the active encounter and queuing them locally when offline.
SMART on FHIR App Registry: Organisations can register third-party clinical applications. The platform manages the authorisation code flow and issues short-lived access tokens scoped to the encounter, enabling clinicians to launch registered apps in context.
Normalised Data Model: All clinical events use a consistent, normalised model internally; outbound formatters translate to the target schema without loss of fidelity.

Open Standards#

HL7 FHIR R4: encounters, observations, procedures, medication administrations, and handover compositions are modelled as FHIR R4 resources; the platform builds and submits conformant transaction Bundles.
SMART on FHIR 2.1: third-party clinical app launch uses the SMART on FHIR 2.1 authorisation flow, including the PKCE-secured authorisation code exchange and RS256-signed access tokens.
NEMSIS v3.5: the National EMS Information System dataset version 3.5 is supported as a cross-export format for research, cross-jurisdictional sharing, and alignment with US EMS data structures.
PHECC 2024 Dataset: the Pre-Hospital Emergency Care Council clinical record schema is the primary national submission target for Irish ambulance operations.
JRCALC 2024: the Joint Royal Colleges Ambulance Liaison Committee clinical practice guidelines are reflected in intervention and medication structures; narrative sections follow the JRCALC 2024 ABCDE structure.
SNOMED CT: clinical findings, chief complaints, and triage decisions are coded using SNOMED CT terminology to keep records interoperable across clinical systems.
LOINC: vital-sign observations (systolic and diastolic blood pressure, pulse oximetry SpO2, heart rate, and others) use standard LOINC observation codes for consistent identification.
ICD-10-AM: patient conditions imported from national shared care records are coded in ICD-10-AM and mapped into the encounter's condition list.
IEEE 11073: vital-sign observations streamed over Bluetooth Low Energy from bedside monitors and wearables use the IEEE 11073 personal health device communication standard.
CloudEvents 1.0: encounter lifecycle notifications use the CloudEvents 1.0 envelope, enabling standard event routing and filtering by downstream consumers.
W3C Service Worker: PWA offline replay on the web client is built on the standard Service Worker API, providing background sync without requiring a native application.
W3C IndexedDB: queued mutations on the web client are stored in IndexedDB, the standard browser-side structured storage layer, ensuring durability across page reloads and browser restarts.
OAuth 2.0 and JWT (RS256): all API access is governed by OAuth 2.0 bearer tokens; access tokens are RS256-signed JWTs carrying tenant, organisation, and crew claims.
EU GDPR Article 9: patient consent for special-category health data is captured as a short-lived, QR-coded consent token before record creation, meeting the explicit consent requirement for health data processing.
HTTPS / TLS: all reconnect-time synchronisation and API communication uses standard secure transport.

Security and Compliance#

All clinical data in transit is protected by TLS. At rest, encounter records are tenant-isolated and crew-scoped; no cross-tenant access is possible at the API layer. Witnessed signatures are cryptographically bound to the encounter at capture time and cannot be detached from the record.

Patient consent for health-data processing is collected in compliance with GDPR Article 9 before the encounter record is created. Encounter lifecycle events are published to an immutable audit log, providing a complete chain of custody from dispatch to final export.

Access tokens follow short-lived JWT best practices (maximum one-hour lifetime) and carry the minimum necessary claims. The SMART on FHIR 2.1 flow adds PKCE to prevent authorisation code interception by third-party applications.

Last Reviewed: 2026-05-05 / Last Updated: 2026-05-05

ePCR Capture with Offline-First Sync

Ready to Build?