Inspiration
Canada welcomed almost 170,000 refugees in 2023, but for the average displaced person, our healthcare system is broken, disjointed, and failing those that need it the most.
More than 60% of refugee patients receive healthcare from two different countries or providers, yet most have no way to carry their medical history across borders.
We wanted to build a simple platform that addresses the core pain points for today's refugees.
What it does
Medi is a patient-owned AI-powered passport that securely stores all medical documents and provides a concise overview of a patient’s health history.
After a clinic visit, users can upload documents such as immunization records, prescriptions, and doctor’s notes, along with basic details like clinic location and visit type (e.g., routine checkup, emergency visit). Medi automatically summarizes the contents of these documents into brief, easy-to-understand descriptions and offers translations into any language by calling on Google's Gemini API. This is displayed as a table on the dashboard and on the backend, in a database called Supabase.
At the top of the dashboard, users can access a high-level summary of their medical profile including potential critical conditions, ensuring that the most essential health information is always organized and readily available. With Medi, important medical records are never misplaced. It empowers patients—especially refugees and those frequently on the move—to receive faster, more informed care through accessible summaries of their health history.
How we built it
We built Medi using React and JavaScript for the frontend, creating an intuitive and accessible user interface.
For backend storage, we chose Supabase, which manages user authentication and securely stores medical documents. Supabase is HIPAA-compliant, meaning it meets strict healthcare security standards. We selected Supabase because it is trusted by thousands of developers for building and deploying secure applications. Under Supabase’s shared responsibility model, Protected Health Information (PHI) can be securely stored once a Business Associate Agreement (BAA) is established.
To enable document summarization and translation, we integrated Google's Gemini API. This allows Medi to generate AI-powered medical summaries and translate them into multiple languages, helping bridge communication gaps between patients and doctors.
Challenges we ran into
Database Design - Structuring the database to accommodate various types of medical records while ensuring fast and secure access was a challenge. We spent time refining our SQL schema to support different document types and relationships efficiently.
AI Summarization & Translation - Generating clear and medically accurate summaries was initially tricky. Medical documents contain complex terminology, so we iterated on prompt engineering to improve our results.
Authentication & Security - Ensuring secure access control so that only patients and authorized individuals can view medical records requires extra effort. We implemented Supabase authentication to manage user sessions safely.
Accomplishments that we're proud of
Making a fully dynamic, non-hardcoded project!
Successfully integrating AI-powered medical summaries and translation to improve accessibility for non-English-speaking refugees.
Building a secure, HIPAA-compliant platform that keeps user data private and encrypted.
Designing an intuitive dashboard UI that allows users to manage their medical history effortlessly.
Overcoming technical challenges in structuring a robust SQL schema to handle diverse medical record types.
What we learned
Structured data in healthcare applications is essential. A well-organized schema makes querying and displaying information much more manageable.
How to leverage AI for medical applications, including the challenges of generating accurate and concise summaries.
Healthcare security and compliance are complexities, mainly when dealing with sensitive patient data.
How to integrate Supabase authentication to ensure secure user access while maintaining a smooth login experience.
What's next for Medi
Clinic Recommendations - We aim to integrate location-based clinic recommendations, allowing users to find nearby healthcare providers easily.
Expanded AI Summaries - We plan to enhance AI-generated medical summaries with condition-specific alerts (e.g., flagging potential allergies or medication conflicts).
Offline Access - Developing an offline mode so users can retrieve vital medical summaries even in areas with limited internet access.
Blockchain Encryption - Implementing blockchain-based encryption for secure, tamper-proof medical record storage. This ensures that data integrity is maintained and prevents unauthorized alterations.
Decentralized Login - Enabling blockchain-based authentication where users can log in with a secure wallet or decentralized identity (DID), eliminating reliance on traditional authentication systems and improving privacy.
Built With
- css
- gemini
- html5
- javascript
- react
- supabase
- tailwindcss
Log in or sign up for Devpost to join the conversation.