AI Triage Engine

Overview

Intelligent decision support for rapid patient prioritization in mass casualty incidents

The AI Triage Engine is a machine learning-powered decision support system that automatically categorizes trauma patients based on vital signs data from the Smart Pad. Using a weighted KPI (Key Performance Indicator) model trained on the START (Simple Triage and Rapid Treatment) protocol, it provides consistent, evidence-based triage recommendations in seconds.

🎯

Consistent

Eliminates inter-rater variability in triage decisions

⚡

Rapid

Assessment in under 2 seconds from vital signs input

📊

Evidence-Based

Trained on ATLS and START protocol guidelines

Clinical Foundation

START Protocol Integration

Simple Triage and Rapid Treatment - the gold standard for mass casualty incidents

🔴 RED - Immediate (Priority 1)

Life-threatening injuries requiring immediate intervention

SpO₂ <90%
Systolic BP <90 mmHg
Unable to follow commands
Unconscious or unresponsive
Severe active bleeding

Action: Immediate intervention required

🟡 YELLOW - Delayed (Priority 2)

Serious injuries but stable, can wait for treatment

SpO₂ 90-94%
HR 80-120 BPM
Follows commands
Non-ambulatory
Moderate injuries

Action: Monitor, treat when resources available

🟢 GREEN - Minor (Priority 3)

Walking wounded with minor injuries

Ambulatory
Normal vital signs
SpO₂ ≥95%
Minor injuries only
Can follow instructions

Action: Self-care area, delayed treatment

⚫ BLACK - Expectant (Priority 4)

Deceased or injuries incompatible with survival given available resources.

Algorithm

Weighted KPI Model

Multi-parameter scoring system with clinical weights

30% Weight

🧠 Consciousness Score

Assessment of patient alertness and ability to follow commands.

Alert & Responsive	100 pts
Confused/Drowsy	50-85 pts
Unconscious	0-40 pts

25% Weight

🫁 Oxygenation Score

SpO₂ and oxygen saturation level assessment.

SpO₂ ≥95%	100 pts
SpO₂ 90-94%	55-75 pts
SpO₂ <90%	0-50 pts

25% Weight

❤️ Cardiovascular Score

Heart rate, blood pressure, and perfusion status assessment.

HR 60-100, BP normal	100 pts
HR >100, BP borderline	50-75 pts
HR >120, BP <90	0-40 pts

20% Weight

🩸 Perfusion Score

Skin color, temperature, and perfusion quality assessment.

Normal perfusion	100 pts
Mild compromise	60-80 pts
Severe compromise	0-50 pts

Final Score Calculation

// Weighted composite score calculation
float finalScore = 
    (consciousnessScore * 0.30) + 
    (oxygenationScore * 0.25) + 
    (cardiovascularScore * 0.25) + 
    (perfusionScore * 0.20);

// Triage category determination
if (finalScore < 40 || criticalAlerts > 0) 
    return \"RED\";
else if (finalScore < 70) 
    return \"YELLOW\";
else 
    return \"GREEN\";
                

Integration

Hospital Dashboard Integration

Real-time patient data for receiving hospitals

📡

Real-time Data

Continuous vital sign streaming via BLE

🏥

Pre-arrival Triage

ER prepared before patient arrives

📊

Trend Analysis

Historical vital sign patterns

🚨

Alert System

Automated critical alerts

Dashboard Features

Patient Overview

Real-time vital signs display
Triage category with confidence
Estimated time of arrival
Patient identification

Clinical Decision Support

AI-generated recommendations
Intervention priority list
Resource allocation guidance
Clinical reasoning display

Analytics

Trend visualization
Pattern recognition alerts
Outcome tracking
Performance metrics

Interactive Demo

Try Smart Pad Triage Demo

Experience simulated vital signs from Smart Pad and real-time AI triage assessment

Data Flow: Smart Pad → BLE → Dashboard → AI Triage

Smart Pad Sensors

→

BLE 5.0 Transmit

→

Dashboard Display

→

AI Triage Assessment

78

Heart Rate

BPM

98

SpO₂

%

120/80

Blood Pressure

mmHg (via PTT)

36.8

Temperature

°C

Normal Sinus

ECG Rhythm

Rhythm

Select Patient Profile

GREEN Minor - Walking Wounded

All vital signs within normal limits. Patient is stable and can wait for routine assessment.

Try the Triage Demo

Experience the AI Triage Engine with interactive trauma scenarios based on real clinical cases.

View on GitHub → View Effectiveness Data

Smart Trauma Assessment Engine