CareRoute

landing page

Inspiration:

Access to the right medical care at the right time is still a major challenge, especially in emergency or uncertain health situations. Many patients struggle to decide whether their condition is serious, which specialist to consult, or which hospital is best suited for their case.

We were inspired by this gap between symptoms and actionable medical guidance. We wanted to build a system that doesn’t just list doctors but intelligently analyses symptoms and medical reports to guide patients toward the most appropriate care.

What it does:

CareRoute is an AI-powered medical triage and hospital recommendation system.

It allows users to:

Enter symptoms
Upload medical reports (PDF/image)
Receive an AI-based urgency assessment
Get recommendations for the most suitable hospital and department nearby

The system evaluates severity (Normal / Consultation Needed / Emergency) and routes the patient accordingly. CareRoute acts as a decision-support layer between patients and healthcare providers.

How we built it:

CareRoute consists of three main components:

Symptom Intelligence Engine -> NLP-based symptom analysis -> Keyword extraction and severity mapping
Medical Report Analyzer -> PDF parsing and text extraction -> Identification of critical values (e.g., WBC, glucose levels) -> Rule-based + ML-assisted risk scoring
Triage & Routing Engine -> Urgency classification logic -> Disease-to-department mapping -> Location-based hospital recommendation

The frontend provides a clean interface for symptom input and report upload, while the backend processes medical data and returns structured recommendations in real time.

Challenges we ran into:

-> Extracting structured insights from unstructured medical reports -> Designing a reliable urgency scoring mechanism -> Avoiding overprediction in emergency classification -> Ensuring the system remains supportive, not diagnostic

Balancing accuracy and safety was one of our biggest technical and ethical challenges.

Accomplishments that we're proud of:

-> Successfully built a working AI triage prototype -> Integrated symptom analysis and report intelligence in one system -> Designed a scalable decision-support architecture -> Created a real-world healthcare-focused solution rather than a generic booking app

What we learned:

-> Healthcare AI requires careful logic design and responsible output framing -> Even simple rule-based layers can significantly improve decision clarity -> Building impactful systems is not just about ML models but also about workflow design

What's next for CareRoute:

-> Improve triage accuracy with larger medical datasets -> Integrate real-time hospital availability data -> AI-based X-ray report analysis using computer vision models to detect potential abnormalities such as fractures, infections, or early radiological indicators -> AI-powered skin disease assessment through image-based deep learning models to support preliminary screening of dermatological conditions -> Pilot testing in tier-2 and tier-3 healthcare environments