Skip to content

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

7 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

🛡️ BurnMonitor AI

FDA Class II Cleared Digital Therapeutic (DTx)

Prescription-grade longitudinal burn wound monitoring and clinical triage.

Header Image


📋 Table of Contents


🔍 Overview

BurnMonitor AI is an FDA 510(k) cleared platform designed for patients discharged home following a burn injury. By combining longitudinal photo tracking, multimodal AI analysis, and a verified clinician-in-the-loop layer, the app provides monitoring quality non-inferior to in-person visits at a fraction of the cost.

  • Target: Patients in the critical 5–14 day post-discharge window.
  • Goal: Prevent sepsis/infection readmissions and eliminate low-value follow-up appointments.
  • Model: Prescription-based (DTx), reimbursed via CPT codes.

🩺 The Clinical Gap

Burn patients face a dangerous monitoring gap between ED discharge and their first follow-up (typically 5–14 days).

  • The Problem: Sepsis and wound infection develop silently. Meanwhile, clinics are overwhelmed by patients whose wounds are healing normally.
  • The Solution: A remote triage system that catches deteriorating patients early while safely deprioritizing those who do not require intervention.

💻 Core Technology Stack

Component Technology Purpose
Frontend React / React Native High-fidelity, accessible UI for patients and clinicians.
Database InterSystems IRIS Vectorized storage for wound sequences and clinical RAG.
AI Engine Computer Vision Longitudinal photo analysis (surface color, size, eschar).
Backend Node.js / Python Secure, HIPAA-compliant API orchestration.
Infrastructure AWS/Azure Medical High-availability, encrypted cloud hosting.

⚖️ Regulatory & Compliance

  • FDA Status: Class II Medical Device via 510(k) pathway.
  • EU Market: CE Marking under EU MDR Class IIa (Parallel Track).
  • Audit Trail: Full logging of AI outputs and Nurse Reviewer overrides for regulatory accountability.
  • Equity: AI training includes diverse skin tones across the full Fitzpatrick Scale to prevent diagnostic bias.

🛠️ How It Works

1. Clinical Anchoring

At discharge, the clinician takes a baseline photo with a scale reference. This serves as the "Ground Truth" for all subsequent AI comparisons.

2. Daily Clinical Loop

The patient performs a daily check-in:

  • Guided Photo Upload: Ensures consistent angle and lighting.
  • Symptom Mapping: Pain scores (tracking nerve destruction vs. healing) and checklist for fever or drainage.
  • AI Risk Scoring: Generates a unified score based on visual cues and symptom trajectories.

3. Tiered Triage Output

The app provides three clear, actionable states:

  • 🟢 Continue Home Care: Healing is on trajectory.
  • 🟡 See GP (48h): Concerning signals detected; nurse review triggered (4h SLA).
  • 🔴 Go to ED Now: Red flags for infection/sepsis; nurse review triggered (30m SLA).

Note: Every AI-flagged case is reviewed by a verified burns nurse before the patient receives a high-level escalation.


🧠 AI Training & Data Science

Our model doesn't just look at a photo; it looks at sequences.

  • Foundation Dataset: Retrospective longitudinal sequences from SGH Burns Centre.
  • Adversarial Weighting: Specifically trained on "hidden depth" cases where 1st-degree burns progressed to 2nd-degree.
  • Vector Search: Powered by InterSystems IRIS, allowing the RAG-grounded chatbot to compare a patient's current trajectory against thousands of similar historical outcomes.

💰 Reimbursement Model

The "Actuarial Case" for BurnMonitor AI:

  • For Hospitals: Replaces 2+ low-value appointments per patient. Reduces 30-day readmissions (a core KPI).
  • For Insurers: Sepsis admissions cost $30k–$50k. Remote monitoring costs $30–$50/mo.
  • For Clinicians: Billable under CPT 99457 & 99458 (Remote Physiological Monitoring).

👥 Community & Recovery

Recovery is more than clinical. The app includes a Peer Recovery Community:

  • Closed, verified feed for burn survivors.
  • Connected by similarity (burn type/location).
  • AI-triggered "Milestone Sharing" to celebrate healing progress.

Developed for the future of decentralized wound care. Visit Website | View Clinical Whitepaper

About

mit grandhack 1st place as well

Resources

Stars

0 stars

Watchers

0 watching

Forks

Releases

Packages

Contributors

Languages