DISTRIBUTED SYSTEMS · EMERGENCY CARE

Emergency transit,
coordinated as a system.

A closed-loop research prototype connecting EMS events, route review, facility acceptance, and clinical context through an organization-isolated Supabase backend.

RESEARCH BUILDHUMAN-APPROVED OPERATIONSNO SYNTHETIC CASES

01 / THE PROBLEM

One transport. Many systems.
One traceable timeline.

Emergency transport depends on updates arriving from different actors at different times. The research question is how to preserve a trustworthy case timeline across those boundaries—while exposing only the information each role needs.

Core principle

Clinical telemetry informs qualified human review. It never diagnoses or dispatches by itself.

02 / SYSTEM MAP

From provider event
to authorized live view.

The proposed event path is explicit about validation, tenancy, and measurement. Each arrow is a system boundary that can be instrumented and evaluated.

OBSERVE THE SYSTEMp50 / p95 / p99 latencyEvent freshnessFailure recoveryDuplicate side effectsCross-tenant access

Architecture and metric map · evaluation measures are planned; no operational values are represented here.

03 / CLOSED LOOP

A route stays a proposal
until a person approves it.

Acceptance, arrival, and handoff remain separate state transitions. Rejections return to review as a new route version.

  1. 01IncidentReceived & triaged
  2. 02Route proposalVersioned & timestamped
  3. 03Dispatcher reviewAccept or request revision
  4. 04TransportEn route & arrived
  5. 05Facility acceptanceExplicit confirmation
  6. 06HandoffResponsibility recorded

04 / RESEARCH CONTRIBUTION

Integration is the contribution.
Claims stay bounded.

The project focuses on a verifiable workflow across event ingestion, multi-tenant persistence, human review, and clinical-data separation—not on claiming a new routing algorithm.

01

Replay-safe event history

Repeated deliveries are deduplicated, payload conflicts are rejected, and case state updates respect source timestamps.

02

Role-aware data boundaries

Organization membership controls case access; clinical observations are limited to clinical and privacy roles.

03

Human approval boundary

Route suggestions and facility capacity are not treated as approval, acceptance, or a clinical decision.

04

Measurable system behavior

The evaluation plan covers event delay, freshness, duplicate handling, recovery, and cross-tenant access.

Novelty statement: the defensible contribution is systems integration and its evaluation method. Priority over prior work is not claimed; a complete literature review and measured comparison remain necessary.

05 / EVALUATION

Designed to measure.
Not dressed up as results.

The planned experiment compares a centralized polling baseline with the event-driven Supabase configuration under equal workload traces. Both configurations must be measured before any performance conclusion is reported.

Read the IEEE methodology
PLANNED MEASURESNOT YET RUN
End-to-end event latencyp50 · p95 · p99
Case / route freshnesssource-to-view age
Recovery under interruptiontime · acknowledged writes
Duplicate and ordering safetyside effects · stale overwrites
Tenant authorizationunauthorized reads target: 0

Chart values are intentionally absent until controlled experiments produce them.

06 / DATA & EVIDENCE

Historical data for research.
Live feeds for operations.

These are different evidence sources and must never be presented as interchangeable.

RETROSPECTIVE EMS DATA

NEMSIS research dataset

Potential source for aggregate EMS case-mix characterization after access approval and review of dataset terms. It is not a live incident stream.

NEMSIS access information ↗
HISTORICAL CLINICAL DATA

MIMIC-IV

Credentialed, governed retrospective EHR data. It does not validate an EMS provider connection or represent real-time telemetry.

PhysioNet dataset page ↗
PLATFORM CONTROLS

Supabase RLS & Realtime

RLS policies and explicit grants form the tenant boundary; Realtime refreshes authorized views for a bounded pilot design.

RLS documentation ↗
!

Research prototype only. The Supabase schema and webhook function are deployed, but no authorized EMS feed is connected and no measured results have been collected. This system is not for dispatch, diagnosis, or clinical care. Do not enter real patient data until institutional, privacy, security, and provider approvals are complete.