💧 KidneyCare: AI-Powered CKD Awareness & Early Detection

✨ Inspiration

Chronic Kidney Disease (CKD) is a growing health concern in Sri Lanka, especially in rural communities where early detection and awareness are limited. Many patients are diagnosed at advanced stages due to lack of accessible screening and education. KidneyCare was inspired by the idea of using AI to empower individuals with knowledge, early detection tools, and actionable prevention tips—all through a mobile platform that runs offline on Arm-based devices.

🚀 What it does

KidneyCare is a mobile health platform that combines education, risk assessment, and early detection tools for CKD:

🧠 Educational Modules: Interactive lessons about kidney health, CKD risk factors, diet, and lifestyle.
🩺 Risk Assessment: AI-driven questionnaire analyzes personal risk factors (age, lifestyle, blood pressure, family history).
📊 Early Detection: Allows users to log simple health metrics (e.g., urine color, blood pressure, creatinine levels) to receive AI-guided feedback.
🔔 Prevention Tips: Personalized recommendations for diet, hydration, exercise, and routine checkups.
🏡 Offline Mode: Runs fully offline on iPads and other Arm-based devices to reach remote areas.

🛠 How we built it

KidneyCare is built with a hybrid AI-mobile approach optimized for Arm devices:

🤖 AI Risk Engine: Lightweight decision tree and transformer model to predict CKD risk from user input.
📱 Mobile App: React Native + SwiftUI hybrid for cross-platform interactive UI.
🖼 Illustrations: Local image generation for educational diagrams using Core ML + Metal acceleration.
💾 Data Storage: SQLite database for offline tracking of user health metrics.
⚡ Arm Optimization: int8 quantized AI models, thread-balanced execution, and Apple Neural Engine support for speed and efficiency.

🧗 Challenges we ran into

🐢 Offline performance: Ensuring AI models run efficiently on iPads without cloud dependency.
📊 Accurate detection: Designing an AI system that gives reliable risk feedback without requiring invasive tests.
🌍 Local relevance: Adapting content and risk factors specifically to Sri Lankan population and dietary habits.

🏆 Accomplishments

Developed a fully offline CKD awareness and detection app for Arm-based tablets.
Created interactive educational modules that simplify CKD understanding for general users.
AI-powered risk scoring gives actionable prevention guidance.
Optimized app for smooth performance on iPads, ensuring accessibility in rural areas.

📚 What we learned

Mobile AI can bring health education and early detection to underserved populations.
Localization of content (language, diet, and lifestyle) is critical for adoption.
Offline AI ensures accessibility in regions with limited internet connectivity.

🌟 What's next

Integrate urine and blood test image detection using the device camera.
Add community tracking and aggregated insights for Sri Lankan health authorities.
Expand the educational library with dietary guidance and CKD-friendly recipes.
Support multiple languages and dialects spoken in Sri Lanka.