Objective
Develop a backend solution for the
"Primary Doctor Recommendation System," designed to intelligently match patients with primary care physicians based on various criteria such as patient preferences, doctor specialties, and collaborative filtering algorithms. This system aims to enhance the patient experience by providing personalized doctor recommendations, taking into account both user profiles and historical match data.
Tools & Technologies
Flask (Python), MySQL (Hosted on Google Cloud SQL), ReactJS, Node.js, FastAPI, TensorFlow, Custom models for recommendations (Item-based collaborative filtering)
About the Model
The recommendation model in this system is built on an item-based collaborative filtering algorithm, designed to analyze patterns in patient preferences and past interactions to suggest the most suitable primary care physicians. The model takes into account various factors such as:
- Patient Preferences: Age, gender, preferred communication style, and medical history.
- Doctor Specialties: Specific areas of expertise, patient reviews, and availability.
- Collaborative Filtering: Leveraging past user data to identify patterns and similarities among patients for personalized doctor recommendations.
Currently, as real-world data is not readily available, we propose collecting this data through surveys targeting patients, doctors, and healthcare providers. We also aim to consider health insurance information to enhance the matching criteria and provide even more tailored recommendations.
For the prototype, we have utilized dummy data to test and validate our model's initial functionality and effectiveness.
This project was developed during the
TFC Civic Tech Hackathon 2024, an event focused on building innovative solutions to improve civic engagement and public services. The hackathon gathered developers, designers, and innovators to create impactful projects addressing real-world challenges in our communities.