ERTRIAGE Outperforms Competing AI Triage Systems

Meta-Level Decision Fusion

• ERTRIAGE: Uses a multi-model fusion architecture (NEWS, SOFA, APACHE II) based on Dr. Gortzis' research for more accurate triage.
• Others: Typically rely on a single scoring system or fixed logic.

Real-Time Adaptive Learning

• ERTRIAGE: Continuously learns from real-time clinical input and outcomes.
• Others: Often run on static models, requiring manual retraining.

Explainable and Traceable AI

• ERTRIAGE: Knowledge graph-based traceability for regulatory compliance and clinician trust.
• Others: Black-box systems with low transparency.

Augmented Clinical Decision Support

• ERTRIAGE: Supports clinicians with override options and confidence levels.
• Others: May override clinician input without explanation.

Operational Context Awareness

• ERTRIAGE: Integrates with ED logistics—staff levels, ICU beds, diagnostic queues.
• Others: Typically disconnected from hospital operations.

Seamless Interoperability

• ERTRIAGE: FHIR and HL7 compatible for easy EHR integration.
• Others: Often closed or proprietary systems.

Summary Table

Comparison of Leading AI Triage Systems

• Accuracy: ~97%
• Protocols: ESI, HEART, NEWS, ROSIER
• Strength: Advanced decision logic, real-time learning, operational integration

ChatGPT-Based Triage Systems

• Accuracy: ~94.9%
• Strength: Natural language understanding, high diagnostic quality
• Limitations: Limited explainability and clinical protocol support

Aidoc

• Focus: Radiology and imaging (stroke, PE, etc.)
• Strength: Automated image analysis with high specificity
• Limitations: Not a comprehensive triage tool

NIB AI Symptom Checker

• Accuracy: ~97.2%
• Focus: Direct-to-consumer mobile triage
• Limitations: Lacks clinical integration or predictive logic

System Comparison Table