MedAssist – Real-Time Emergency Decision Intelligence System

A real-time emergency decision system that transforms symptoms into explainable, actionable life-saving steps within seconds.

Comment

🚨 About the Project :

💡 Inspiration

In emergency situations, the biggest failure is not the absence of help , it is the delay in making the right decision.

People panic. They hesitate. They are unsure what to do next.

Most existing solutions act as information providers , they list symptoms or possibilities, but do not guide immediate action.

This gap between symptom → decision → action inspired MedAssist.

«The goal was to build a system that guides users clearly and instantly when every second matters.»

⚙️ How We Built It

MedAssist is designed as a real-time decision support system that evaluates symptoms through a structured and explainable process.

The system follows a clear pipeline:

User Input → Structured Questions → Risk Evaluation → Decision → Action

🧠 Decision Engine

At its core, MedAssist uses a weighted risk evaluation model:

  • Each response contributes to a cumulative risk score
  • Higher risk symptoms carry greater weight
  • The final score is mapped to severity levels:
    • LOW
    • MEDIUM
    • HIGH && CRITICAL

Mathematical Representation (Conceptual)

Risk Score = sum of (weight × response) across all questions

Where:

  • weight = importance of each symptom
  • response = user input (yes/no mapped to 1/0)

This ensures that severity is quantified rather than guessed, making the system consistent and interpretable.

🔄 Structured Question Flow

The system uses a guided sequence of targeted questions:

  • Each question captures clinically relevant signals
  • The flow is consistent to ensure reliability
  • Users are guided step by step without requiring medical knowledge

📊 Explainable Output

MedAssist emphasizes transparency and clarity:

  • Displays risk level clearly
  • Provides a confidence estimate
  • Explains reasoning behind the result
  • Suggests immediate next steps

🎨 Emergency-Focused Interface

The UI is designed for high stress scenarios:

  • High-contrast severity colors (Green / Yellow / Red)
  • Minimal interaction steps
  • Clear layout and typography
  • Smooth transitions to simulate real-time analysis

🌐 Deployment & Integration

  • Deployed using Vercel
  • Built with Next.js + TypeScript
  • Integrated real-world actions:
    • Emergency call guidance (108)
    • Nearby hospital discovery via Google Maps

🧗 Challenges We Faced

  1. Designing a Balanced Risk Model Ensuring meaningful scoring without oversimplification required careful tuning of weights.

  2. Clarity Under Stress The system had to be usable in panic conditions — simplicity became critical.

  3. Maintaining Explainability Avoiding black-box logic while still producing useful results was a key challenge.

  4. Real-Time Experience Design Balancing responsiveness with perceived intelligence required careful UI timing.

📚 What We Learned

  • Structured decision systems are more reliable than generic symptom checkers
  • Explainability increases user trust significantly
  • UX must adapt to cognitive load during emergencies
  • Simplicity, when engineered correctly, is powerful

🚀 Final Outcome

MedAssist is not just a symptom checker.

It is a system that:

  • Evaluates symptoms systematically
  • Quantifies severity clearly
  • Guides users with actionable steps

All within seconds.

«It transforms uncertainty into clear, reliable decisions , when every second matters.»

Note: MedAssist is designed for mobile-first emergency use , voice input is supported on both mobile and desktop via the Web Speech API. For the most reliable speech recognition, a smartphone is recommended, as mobile browsers handle microphone permissions more consistently than desktop. Manual symptom selection is always available as a fallback and works instantly on all devices.

Built With

Share this project:

Updates