Pak-Medic
A Mobile-First AI-driven Centralized Healthcare Platform which is an amalgam of AI and NLP for the creation and maintenance of electronic medical records. Diagnosis and prognosis for doctors, support communities for interactive communication and patient chatbot for seamless healthcare.
View ProductThe Problem
Lack of medical knowledge in the majority of people leads to health issues. Minor problems prompt self-medication, as visiting doctors is inconvenient. Limited awareness about specialists results in misguided advice. Doctors' misinterpretation of conditions hinders disease diagnosis due to a lack of history tracking. Poorly managed health records pose challenges. Social taboos impede open discussions on medical issues. Lack of information and the greater threat of quackery have irreversible consequences. Illegible prescriptions risk misuse. Overall, improving medical education, awareness, and record-keeping is crucial for better public health in Pakistan.
The Process
We spun out our ideating phase with a Ph.D. AI Professor who helped us list down features and the potential solutions to the problems. We began brainstorming after conducting user research and crafted a rough user journey map. We then began to design the wireframes and test the user flow before going High Fidelity as It would distract us from the flow and emphasize the design elements.
User Journey
Low Fidelity Wireframes
High Fidelity Mockups
AI-Powered Platform
Features like Disease Diagnosis and Prognosis for Doctors, real-time doctor Specialty Recommendation, Medicine Recommendation, Admin Bert Question Answering through textual Data, support communities and Telemedicine for easier healthcare providence. The inspiration for this platform came from the need for a centralized platform to aid Humans and Save Lives.
Dashboard for Analytics and Bert
The admin dashboard system was built from the ground up to be flexible and extensible. It’s built to have a single source of truth for the admin to monitor the platform and have an entire overview of things. Furthermore, encompassing the powers of AI, A Bert Agent was introduced to reduce the workload of understanding textual data and the ability to ask up to 5 questions from it.
Technology Stack
The tech stack was chosen to handle each specific use case efficiently. The front end is built with React Native, React, and CSS. The backend is built with Node.js (Express.js), Python for Machine Learning, Material UI for Admin Portal and Docker for Containerization. The backend is deployed on AWS EC2 in a dockerized Container. MongoDB Atlas is used for the databases. The front end and back end are hosted on separate servers to allow for horizontal scaling.