shyamvaidhyanathan/PatientAppointmentMCP

🏥 Patient Health Classification Agent

This project demonstrates how to use FastAPI, Model Context Protocol (MCP), and CrewAI to classify a patient’s health status based on their number of appointments in the past year. The system includes:

• A REST API for patient classification
• An MCP server exposing the API as a callable tool. The tools available in this MCP Server can be listed.
• An CrewAI Agent that invokes the tool given to it using the MCP Server to classify a patient.

---

📁 Project Structure

patient_classifier_agent_project/
├── agent/
│ └── crewai_agent.py # ADK Agent using CrewAI
├── api/
│ └── patient_classifier_api.py # FastAPI classification endpoint
├── mcp/
│ └── mcp_server.py # MCP server wrapping the API
├── tools/
│ └── classify_tool.py # Tool wrapper to interface API with agent
├── data/
│ └── appointment_db.py # Dummy appointment data
├── requirements.txt # Python dependencies
└── README.md # This file
└── .env

---

⚙️ Setup Instructions

git clone https://github.com/shyamvaidhyanathan/patient-health-agent.git cd patient-health-agent

Create a python virtual environment

python -m venv env source env/bin/activate # On Windows: env\Scripts\activate

Install Dependencies

pip install -r requirements.txt

⚙️ Execution Instructions

Running the Patient Classifier API found inside api directory

python api/main.py

Running the MCP Server

uvicorn mcp.mcp_server:app --reload --port 8001

Running the Agent

python agent/crewai_agent.py

🧠 Health Classification Logic

The patient is classified based on the number of appointments in the past year:

CHRONIC: More than 10 appointments

MODERATE: 5 to 10 appointments

HEALTHY: Less than 5 appointments

🛠️ Tech Stack

FastAPI – for building REST API

CrewAI – for autonomous agent management

LangChain – for tool abstraction

Model Context Protocol (MCP) – for wrapping APIs as tools

Uvicorn – for serving ASGI apps

📄 License

MIT License

🙋‍♀️ Maintainer

Shyam Vaidhyanathan