HEALTH-RECS
A digital healthcare record management system designed to help hospitals transition from manual patient record keeping to a secure, AI-powered digital platform using OCR, Vision AI, and full-stack web technologies.
Visit websiteOCR Data Extraction in Action
The core of HEALTH-RECS is its ability to automatically extract structured data from handwritten treatment cards and medical reports using advanced Vision AI models.
Handwritten patient treatment cards are uploaded to the system.
Patient Management Dashboard
The HEALTH-RECS dashboard provides healthcare staff with a comprehensive interface to manage patient records, view medical history, and access extracted data in real-time.
The Problem
Hospitals, particularly in underserved communities, continue to rely on manual paper-based patient record keeping. This approach creates significant operational challenges:
• Slow Data Retrieval: Finding patient history requires manual search through physical files
• Human Errors: Manual documentation is error-prone and inconsistent
• Risk of Data Loss: Records are vulnerable to damage, theft, or misplacement
• Administrative Burden: Healthcare staff spend excessive time on paperwork instead of patient care
• Limited Accessibility: Patient history is difficult to retrieve for treatment decisions
The Context: Anandwan
HEALTH-RECS was developed after visiting Anandwan, a rehabilitation center near Warora, Chandrapur, Maharashtra. Founded in 1949 by Baba Amte, Anandwan provides healthcare, education, and vocational support for marginalized communities including leprosy patients and differently-abled individuals.
The center faced significant challenges in managing patient records due to manual systems, making it the perfect real-world use case for developing a scalable, secure digital healthcare platform.
The Solution
HEALTH-RECS is a full-stack healthcare record management platform that digitizes and securely stores patient treatment cards, medical history, and health records using AI-powered OCR, Vision models, and cloud infrastructure.
The system transforms paper-based healthcare workflows into efficient digital processes, enabling instant record retrieval, automated data entry, and centralized patient management.
Core Features
OCR-Based Record Digitization: Upload treatment card images and automatically extract text. The system uses AI to process handwritten and printed data, auto-filling form fields and converting manual records into structured digital data.
AI Vision Processing: Leverages Llama-3.2-90B-Vision model via Groq API for intelligent image understanding, medical text extraction, and interpretation of handwritten treatment cards.
Patient Health Record Management: Comprehensive system for adding, updating, and retrieving patient records including medical history, treatment details, admission records, symptoms, and diagnoses.
Intelligent Form Autofill: OCR-extracted data automatically populates patient forms, reducing manual typing, minimizing errors, and accelerating the registration process.
Secure Database Management: Encrypted storage of medical records with centralized cloud-based management, controlled access, and comprehensive data protection measures.
Role-Based Access Control: Differentiated access levels for Admin, Doctors, and Staff, each with appropriate permissions for managing patient data.
System Workflow
Existing Patient Flow: Upload treatment card image → Vision Model/LLM processes image → JSON structured response → Frontend verification → Save to database
New Patient Flow: Add patient details manually → Upload required medical documents → Verify information → Save to database
The dual workflow accommodates both digitizing legacy paper records and managing new patients through the system.
Technical Architecture
Frontend: React.js with Tailwind CSS and Vite for a modern, responsive user interface with smooth interactions and fast rendering
Backend: Node.js + Express.js for API development and MongoDB integration, with Flask microservices for OCR processing
AI & OCR: Llama-3.2-90B-Vision model via Groq API for fast, accurate OCR processing and vision understanding of medical documents
Database: MongoDB for flexible, scalable storage of patient records and healthcare data
Deployment: Vercel for frontend hosting and Render for backend services, with planned AWS/Azure enterprise deployment
Key Technologies
• Llama-3.2-90B-Vision: Advanced vision model for OCR and document understanding
• Groq API: High-speed AI inference for real-time OCR processing
• React + Vite: Modern frontend development framework
• Node.js + Express: Backend API development
• MongoDB: Healthcare data persistence
Challenges & Solutions
OCR Accuracy: Handwritten treatment cards vary significantly in style and legibility. Addressed through AI vision models that handle diverse handwriting patterns.
Service Integration: Coordinating between frontend, backend, OCR service, and database required robust API synchronization and queue-based processing.
Data Security: Medical data privacy is paramount. Implemented encryption for sensitive patient information and secure authentication mechanisms.
Community Impact
Faster Healthcare Delivery: Doctors can instantly retrieve patient history and treatment details, enabling quicker medical decisions
Reduced Administrative Burden: Automated data entry and centralized management free up healthcare staff to focus on patient care
Enhanced Data Security: Digital backups and centralized storage significantly reduce the risk of record loss or misplacement
Improved Accessibility: Healthcare data becomes accessible to authorized personnel anytime, anywhere, improving coordination and continuity of care
The Anandwan team provided positive feedback, with doctors supporting the digitization initiative and staff appreciating the centralized database approach.
Current Implementation Status
• ✓ OCR Integration: Microservice fully integrated for text extraction from medical reports
• ✓ Frontend Development: React + Vite UI with upload and form management functionality
• ✓ Backend APIs: Node.js + Express APIs for record management and database operations
• ✓ Database Integration: MongoDB storage and retrieval operations fully functional
• ✓ Form Autofill System: OCR data successfully connected with frontend forms
Future Enhancements
• Enterprise Cloud Deployment: Scale to AWS and Azure for large-scale hospital networks
• Advanced OCR Models: Fine-tune OCR for medical-specific handwriting and multilingual support
• AI-Powered Insights: Add disease prediction, medical recommendations, and AI-assisted diagnostics
• Multi-Hospital Integration: Support multiple hospitals and healthcare networks with centralized oversight
• Performance Optimization: Handle larger patient volumes and concurrent users across distributed systems
Why This Project Matters
HEALTH-RECS demonstrates the power of technology applied to real-world healthcare challenges, particularly in underserved communities. It combines full-stack development expertise with AI integration to create a socially impactful solution that modernizes healthcare administration while maintaining security and patient privacy.
The project showcases practical implementation of OCR pipelines, AI vision models, and scalable architecture—creating a blueprint for digital healthcare transformation in hospitals worldwide.