Inspiration
Emergency transport decisions are often made under extreme time pressure with incomplete information. Ambulances may travel to familiar hospitals rather than the most suitable ones, while critical resources like ICU beds remain unevenly utilized. We wanted to explore how structured digital triage and live capacity visibility could support faster and more consistent decisions.
What it does
The system converts patient vitals and symptoms into a severity score and priority class. It then evaluates nearby hospitals using travel time and reported bed or ICU availability to recommend the most appropriate destination.
The goal is not to replace medical professionals, but to assist them with transparent, repeatable decision support.
How we built it
We designed a rule-based scoring engine where each clinical parameter contributes weighted risk points. The total determines whether the case falls into RED, YELLOW, or GREEN priority.
The workflow includes:
- Data entry by paramedics or operators
- Automated severity calculation
- Real-time comparison of hospital capacity
- Ranked recommendation
- Explainable decision trace
The architecture is modular so machine learning models can be introduced later using historical outcomes.
Challenges we ran into
Obtaining real-time hospital data is difficult in practice, so we designed the system assuming updates through an administrator portal or future API integrations.
Another challenge was balancing simplicity with realism. The model must be understandable in emergencies while still reflecting clinical reasoning.
What we learned
We learned how critical explainability is in healthcare systems. Users must understand why a recommendation is made. We also realized interoperability and data reliability are as important as algorithms.
What's next
Future work includes predictive demand modeling, traffic-aware routing, automated device integration from ambulances, and learning from treatment outcomes to refine scoring accuracy.
Log in or sign up for Devpost to join the conversation.