MedMonitor: Open-Source Medical IoT Telemetry Gateway (SaMD)

MedMonitor is an open-source, pre-compliant Medical IoT Gateway and real-time vital signs dashboard, designed to demonstrate feasibility for MIC@Home (Mobile Inpatient Care at Home) and hospital step-down wards in the ASEAN region.

📁 GitHub Repository · 📄 View Docs

---

⚠️ IMPORTANT DISCLAIMER – NOT A MEDICAL DEVICE

This software is provided for RESEARCH, DEVELOPMENT, INVESTIGATIONAL/PROTOTYPE, AND EDUCATIONAL PURPOSES ONLY.
It is NOT a medical device, NOT cleared or approved by any regulatory agency (such as Singapore HSA, Malaysia MDA, US FDA, or EMA), and NOT intended for clinical use.

🔬 Data Source & Synthetic Nature

MedMonitor is a functional prototype. All physiological data processed by this system (Heart Rate, SpO2, BP) is derived from public Kaggle ICU datasets or generated via synthetic simulation. This project does not collect, process, or store real-world Patient Identifiable Information (PII) or Protected Health Information (PHI).

⚖️ Investigational Use Only (IUO)

This software is intended to demonstrate technical feasibility for medical IoT telemetry. Any use of this software in a clinical setting or with real patients is strictly prohibited unless conducted under a formal Investigational Testing Exemption (ITE) or within a sanctioned Regulatory Sandbox (e.g., Singapore MIC@Home or Malaysia MOH Sandbox) under the supervision of qualified medical professionals.

❌ What this software is NOT

• Not a primary monitoring system for life-critical situations.
• Not a substitute for professional medical judgment or hospital-grade hardware.
• Not for generating clinical alerts that lead to immediate medical intervention without human verification.

✅ What this software IS

• A technical demonstration of IEC 62304 Class B software architectural patterns.
• A research tool for testing PDPA-compliant data isolation and audit-trail integrity.
• An open-source reference for Modified Early Warning Score (MEWS) algorithm implementation.

🔒 Liability Waiver

THE SOFTWARE IS PROVIDED “AS IS”, WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.

By using this software, you confirm that you understand and agree to these terms.

---

🛑 The Problem & 💡 The Solution

Hospital IT architects and researchers frequently encounter the same roadblocks when deploying medical telemetry. MedMonitor explicitly solves these core industry challenges:

1. Alarm Fatigue in Step-Down Wards

• The Problem: Up to 99% of clinical alarms in non-ICU environments are false or clinically insignificant, leading to severe alarm fatigue (Cvach, 2012) and missed deterioration events.
• The Solution: MedMonitor implements a 5-minute rolling suppression window for duplicate alert types and utilises the Modified Early Warning Score (MEWS) composite algorithm. Instead of alerting on transient single-parameter spikes, it evaluates HR, RR, BP, and Temp holistically to generate high-fidelity CRITICAL_DETERIORATION alerts.

2. Network Instability in MIC@Home

• The Problem: Remote patient monitoring (Hospital at Home) relies on residential Wi-Fi or 4G/5G, which is prone to dropouts, resulting in lost clinical telemetry.
• The Solution: MedMonitor features Edge Buffering. Python-based edge nodes queue physiological payloads in local memory during network disconnects and perform a rapid “catch-up flush” to the .NET REST API the moment connectivity is restored.

3. Strict Regulatory & Privacy Compliance

• The Problem: Medical data sovereignty (Malaysia PDPA, Singapore HSA CLS-MD) and software safety (IEC 62304) make building custom IoT dashboards prohibitively expensive and legally risky.
• The Solution: MedMonitor bakes compliance into the lowest layer:
  • PDPA Isolation: PostgreSQL Row-Level Security (RLS) restricts data access via DepartmentId session variables.
  • Audit Integrity: An immutable, HMAC-SHA256 cryptographically chained audit_log ensures non-repudiation of clinical actions.
  • Right to be Forgotten: Automated Hangfire jobs physically purge telemetry older than 30 days.

---

🏗️ System Architecture

MedMonitor uses a modern, decoupled architecture designed for high-throughput sensor telemetry.

graph TD
    subgraph Edge["MIC@Home / Ward Edge"]
        D1[Patient Sensor 1] -->|Raw Vitals| Sim[Edge Buffer Gateway]
        D2[Patient Sensor 2] -->|Raw Vitals| Sim
        Sim -- "REST (JSON) + X-Api-Key" --> API
    end
    subgraph Backend [".NET 8 Application Server"]
        API[Ingestion API] --> RS[Reading Service / MEWS Logic]
        RS --> |Save| ORM[EF Core]
        RS --> |Broadcast| SigR[SignalR WebSocket Hub]
    end
    subgraph Database ["PostgreSQL (Supabase/AWS)"]
        ORM -- "Session Pooling (5432) + RLS" --> DB[(PostgreSQL)]
        DB --> Aud[HMAC-SHA256 Audit Log]
    end
    subgraph Frontend["React 19 SPA"]
        SigR -- "Real-time Telemetry" --> UI[Clinical Dashboard]
        UI -- "JWT Authentication" --> API
    end

---

⚙️ Tech Stack & Regulatory Mapping

Component	Technology	Regulatory / Security Purpose
Backend API	.NET 8 (C#)	High-performance async ingestion; handles EF Core execution strategies.
Real-time Engine	SignalR (WebSockets)	Sub-second telemetry propagation to clinical dashboards.
Database	PostgreSQL (Supabase)	Managed JSONB datastore; Port 5432 Session Pooling for RLS enforcement.
Frontend	React 19 + Vite + Recharts	Append-only UI rendering to prevent DOM blocking under high data loads.
Authentication	JWT + TOTP (2FA)	Secures clinical API endpoints; bakes dynamic RBAC capabilities into claims.
Observability	VictoriaMetrics + Loki	15-day system metric retention (PMS evidence for regulatory audits).
PDF Reporting	QuestPDF (.NET)	Generates end-of-shift clinical handover reports offline without external dependencies.

---

⚖️ MedMonitor vs. Traditional Medical IoT

Feature / Capability	Traditional IoT Gateways	MedMonitor
Audit Log Integrity	Standard text/DB logs (editable by DBAs)	HMAC-SHA256 Hash Chain — cryptographically tamper-proof
Device Authentication	Static API keys or IP whitelisting	Mutual TLS (mTLS) using X.509 client certificates
Cross-Ward Data Leakage	Application-level filtering only	PostgreSQL Row-Level Security (RLS) injected into DB session pools
Alarm Fatigue Mitigation	Triggers on every threshold breach	IEC 60601-1-8 Compliant — 5-min rolling suppression + MEWS scoring
Regulatory Alignment	Black-box compliance	Pre-mapped for IEC 62304 Class B & HSA CLS-MD Level 2
Deployment Cost	High licensing fees, vendor lock-in	Open-source, deployable on PaaS (Render / AWS / Supabase)

---

🛡️ Built for Compliance Reviewers & Healthcare IT

MedMonitor implements a strict Dynamic RBAC (Role-Based Access Control) where the API role (medmon_api) has its UPDATE and DELETE privileges explicitly revoked for clinical telemetry and audit logs — ensuring full compliance with data immutability requirements under zero-trust architecture guidelines for medical devices.

---

🚀 Getting Started

# 1. Clone the repository
git clone https://github.com/Calaweh/medmonitor-iot-gateway.git
cd medmonitor-iot-gateway

# 2. Copy environment variables
cp .env.example .env

# 3. Start all services with Docker
docker compose up --build

See the full setup guide and configuration options in the repository README.

---

MedMonitor is an open-source project by Calaweh · Licensed under MIT