curaconnect

Inspiration

The inspiration behind CuraConnect came from the real-life stories of Thalassemia patients across India who face immense difficulty in finding timely blood donors, accessing consistent care, and navigating the lifelong burden of transfusions. Despite platforms like e-RaktKosh and NGO efforts like Blood Warriors, many patients still suffer due to lack of predictive tools, patient education, and real-time coordination.

We envisioned CuraConnect as a bridge—powered by AI, empathy, and innovation—to close these critical healthcare gaps and bring life-saving care to those who need it most.

What it does

CuraConnect is an AI-powered platform designed to support Thalassemia patients by intelligently connecting them with the right donors and healthcare providers at the right time. It combines real-time data, machine learning, and compassionate design to ensure that no patient is left waiting for critical care.

Core Features 🔁 Predictive Donor Availability CuraConnect uses ML models (like Prophet and XGBoost) to forecast donor availability based on their historical donation behaviour, location patterns, and engagement trends.

Donor_Score=f(last_donation,location,donor_type,frequency) Patient Urgency Prioritisation Patients are assigned an urgency score based on time since last transfusion, haemoglobin levels, comorbidities, and more. The platform prioritises patients with critical needs.

Urgency_Score=f(Hb,age,transfusion_gap,comorbidities) Smart Chatbot Assistant A Rasa-based chatbot helps patients and donors by answering common questions, providing reminders for transfusions or donations, and offering personalized care tips. It’s designed to be accessible in multiple Indian languages. Real-Time Donor Matching Engine When a patient needs blood, CuraConnect instantly fetches a ranked list of most likely available donors nearby using location and prediction logic.

Secure Health Profiles & Integration Patients and donors have digital profiles with transfusion history, donor milestones, and health data. The system can integrate with e-RaktKosh, Blood Bridge, and hospital APIs securely.

Hospital & NGO Dashboard For admins and healthcare workers to monitor patient needs, donation trends, and transfusion schedules, all in one place.

User Experience Patients receive transfusion reminders, priority care updates, and AI-guided health education. Donors get personalised nudges, impact tracking (e.g., "You've saved 6 lives"), and milestone badges. Healthcare providers access urgency scores and blood inventory needs with ease.

NGOs can monitor engagement, organise donation drives, and run campaigns more effectively.

How we built it

We developed CuraConnect as a full-stack solution using Google Cloud technologies and open-source tools:

Frontend: React Native (for mobile) and React.js (for web)

Backend: Node.js + Express with RESTful APIs

Machine Learning:

Donor prediction using XGBoost and Prophet (time-series forecasting)

Patient urgency scoring using classification models

Chatbot: Built using Rasa with a future vision to integrate Google MedLM

Data Pipeline: Cloud Dataflow + BigQuery for transformation and analytics

Model Training & Deployment: Vertex AI Pipelines and Model Registry

Security & Compliance: End-to-end encryption (AES), JWT-based authentication, and HIPAA-aligned practices

Challenges we ran into

1. Lack of Real-World Blood Donor Data Challenge: No access to real-time donor or transfusion data from platforms like e-RaktKosh.

Impact: Hindered model training and realistic testing.

Solution: We created synthetic datasets based on published studies, NGO interviews, and domain-specific assumptions.

🧠 2. Medical Domain Knowledge Gap Challenge: Understanding complex patient conditions like Thalassemia and their clinical requirements.

Impact: Risk of building features that aren’t medically useful or feasible.

Solution: Consulted medical students, researched WHO guidelines, and learned from NGOs like Blood Warriors.

🔍 3. Designing a Reliable Donor Matching Algorithm Challenge: Creating an ML model that factors in availability, distance, blood group compatibility, and urgency.

Impact: High complexity and risk of inaccurate recommendations.

Solution: Combined rule-based logic + ML predictions to boost accuracy and reliability.

📱 4. Multi-Stakeholder UX Design Challenge: Designing a UI that works for very different users – patients, donors, doctors, and NGOs.

Impact: Risk of confusing flows or overwhelming interfaces.

Solution: Created separate modules/dashboards per user type, and used feedback loops with test users.

📶 5. Handling Connectivity Issues Challenge: Many rural users have low or unstable internet connections.

Impact: Mobile app usability is compromised in remote areas.

Solution: Built an offline-first architecture for patients and donors using local caching and async sync with Firebase.

🔐 6. Data Privacy and Consent Challenge: Managing sensitive health information while complying with data protection norms (e.g., HIPAA-like practices).

Impact: Needed robust encryption and role-based access control.

Solution: Implemented AES-256 encryption, JWT auth, and future plans for blockchain-backed records.

🤖 7. Scalable AI Deployment Challenge: Transitioning from model training in Jupyter to scalable deployment.

Impact: Manual triggers and delays in prediction API integration.

Solution: Containerized the models for future deployment on Google Vertex AI or AWS SageMaker.

🗣️ 8. Building a Multilingual Chatbot Challenge: Rasa NLP pipeline didn’t perform well in Indian regional languages initially.

Impact: Reduced usability for non-English speakers.

Solution: Integrated basic multilingual intents and planned to use Google MedLM or IndicNLP models going forward.

⏱️ 9. Time Constraints Challenge: Building a full-stack product with AI and health tech complexity in under 4 weeks.

Impact: Had to drop AR/VR and blockchain features from MVP.

Solution: Prioritized life-critical features and documented roadmap clearly.

Accomplishments that we're proud of

Despite the complexity and constraints, our team achieved several key milestones that make CuraConnect a promising solution for chronic care:

✅ 1. AI-Powered Blood Donor Matching MVP We successfully built and tested a prototype that intelligently matches Thalassemia patients with compatible, nearby blood donors using AI models and real-time data (simulated for now).

💬 2. Conversational Health Assistant We created a working multilingual chatbot using Rasa for:

Scheduling transfusions

Sending medication reminders

Offering dietary guidance This makes the system accessible to non-English speakers and patients with low digital literacy.

📱 3. Mobile-Friendly, Offline-First App Built an Android/web-friendly interface with offline support, ensuring patients in rural or low-connectivity regions can still access crucial services.

🔐 4. Privacy-Centric Health Record Design Designed secure, role-based access control for storing health data with encryption in mind — laying the groundwork for future blockchain integration.

🧠 5. User-Centric Multi-Stakeholder Design Developed tailored interfaces for:

Patients

Donors

NGOs/Volunteers

Doctors This modular approach ensures clarity and ease of use for each group.

🌍 6. Social Impact Validation Received positive feedback from NGOs like Blood Warriors and healthcare mentors, validating the real-world relevance and need for CuraConnect.

🚀 7. End-to-End Functional Prototype We took CuraConnect from concept to a fully working MVP within 3–4 weeks, including UI, backend, ML models, and chatbot — all built from scratch.

🧩 8. Cross-Disciplinary Learning The team learned:

Medical workflows (transfusion cycles, donor eligibility)

Real-time app architecture

Deployment pipelines

Ethical AI for healthcare This made us stronger product builders with a human-first mindset.

What we learned

Building CuraConnect was not just about developing a tech solution — it was a deep dive into the intersection of healthcare, human behavior, and machine learning. Here are the key takeaways from our journey:

🧠 1. Understanding Thalassemia Patient Journeys We learned about:

The critical role of regular blood transfusions

The emotional and logistical challenges families face

The systemic gaps in donor coordination and rural healthcare access

This knowledge helped us build with empathy and precision.

🤖 2. Applied AI in Real-World Contexts We explored how AI can be responsibly applied in healthcare:

Donor availability prediction using historical patterns

Chatbot design for low-resource environments

Challenges around data bias, explainability, and ethics

🔗 3. The Power of Integration We understood the value of integrating existing platforms like:

e-RaktKosh (National Blood Bank registry)

WhatsApp, SMS, and IVR systems for low-tech users

Potential to build over government APIs to extend scale

🧩 4. Designing for Multi-Stakeholders We realized it's not enough to build just for the patient:

Donors, NGOs, doctors, and caregivers all need interfaces

A successful product must balance simplicity and utility for each role

🔐 5. Healthcare Data Privacy and Compliance We learned about:

HIPAA/GDPR basics

Importance of role-based access, data encryption, and consent-driven data sharing

The need for transparent data flows in medical systems

🛠️ 6. Tech Stacks and Scalability We gained hands-on experience with:

Python, FastAPI, Firebase, Flutter, Rasa, and ML tools

Building offline-first PWA/mobile interfaces

Designing modular architectures that scale over time

💬 7. Co-Creation with Domain Experts Talking to actual NGO volunteers, donors, and healthcare professionals changed how we designed features. Their insights helped:

Remove unnecessary complexity

Add real-world relevance

Prioritize features that actually solve a daily pain point

What's next for curaconnect

Our journey with CuraConnect has just begun. With a solid foundation built during this hackathon, we are excited to take it forward and transform it into a full-scale solution. Here's what's next:

🔬 1. Clinical Validation & Pilot Testing Partner with NGOs, blood banks, and thalassemia clinics to pilot CuraConnect in real-world scenarios.

Collect feedback from patients, donors, and doctors to iterate and improve.

📈 2. Enhance AI Models Improve donor prediction accuracy using larger datasets.

Incorporate patient prioritization models based on urgency, past transfusion history, and distance.

Add ML-based alerts for missed transfusions or upcoming needs.

📱 3. Launch MVP App Develop a robust mobile app with offline support for patients in low-connectivity areas.

Build role-specific interfaces for donors, caregivers, and healthcare professionals.

🧩 4. API Integration with National Systems Integrate with e-RaktKosh, Aarogya Setu, and state health APIs to access real-time blood bank inventory and donor databases.

🔐 5. Data Privacy and Compliance Implement end-to-end encryption, role-based access control, and audit trails.

Align with HIPAA/GDPR guidelines for secure medical data handling.

🌍 6. Expand to Other Chronic Conditions Extend CuraConnect’s architecture to support patients with Sickle Cell Anemia, Hemophilia, and Kidney Dialysis needs.

Add smart appointment scheduling, diet recommendations, and medication tracking features.

💰 7. Funding and Partnerships Apply for grants, startup accelerators, and healthcare innovation funds.

Partner with wearable device companies to integrate real-time vitals and predictive alerts.

🫱🏻‍🫲🏽 8. Community Building Launch a digital community for Thalassemia patients, families, and donors to support one another.

Organize donation drives and awareness campaigns via the platform.

Built With

• ai
• ai-powered
• ai/ml-model-development-javascript-?-frontend-interactivity-html5-&-tailwind-css-?-responsive-ui-design-react.js-?-single-page-application-for-patients-and-donors-ai/ml-scikit-learn-?-predictive-modeling-(donor-availability
• alerts
• api
• apis
• assistant
• authentication
• availability
• banks
• blood
• canva
• chatgpt
• dashboards
• database
• databases
• db
• deployment
• design
• donors
• e-raktkosh
• figma
• firebase
• firestore
• for
• future)
• google
• health
• integration
• integrations
• internal
• lightweight
• local
• locate
• maps
• metadata
• mockups
• module)
• nearest
• nosql
• openai)
• optional
• patient-priority)-pandas-/-numpy-?-data-preprocessing-&-analysis-xgboost-/-randomforest-?-donor-matching-prediction-cloud-&-devops-aws-lambda-?-serverless-backend-functions-aws-s3-?-secure-file-&-record-storage-firebase-?-realtime-database
• patient/doctor
• patients
• planned
• postgresql
• prototyping
• python-?-core-logic
• quick
• realtime
• relational
• sms
• sqlite
• streamlit
• testing
• to
• twilio
• virtual
• wireframes