Intelligent Medical Assistant MCP Server

🏥 Intelligent Medical Assistant

A comprehensive Patient-Centric Medical Intelligence System that combines MCP (Model Context Protocol) server capabilities with an intelligent web interface. This system provides natural language medical queries, patient data analysis, and professional medical insights powered by Gemini AI.

✨ Key Features

🤖 **Intelligent Medical Assistant**

• Natural Language Processing - Ask questions like "What is Ben's sleep summary?"
• Automatic Tool Routing - AI decides which patient data to access
• Conversational Interface - Single input for all medical queries
• Professional Medical Responses - Evidence-based information with disclaimers

👥 **Patient Data Management**

• Real Patient Data - CSV-based patient records and biometrics
• Comprehensive Tracking - Sleep patterns, vital signs, lab results, medications
• Trend Analysis - Historical data analysis and insights
• Multi-Patient Support - Manage multiple patient records

🔧 **Dual Architecture**

• MCP Server - Integration with Kiro IDE and other MCP clients
• Web Application - Professional web interface with authentication
• RESTful API - Programmatic access to medical data and insights

📊 **Available Patient Data**

• Ben Smith (34M) - Hypertension, Type 2 Diabetes
  • 15 days of detailed sleep data (duration, quality, efficiency)
  • Vital signs tracking (BP, heart rate, glucose, weight)
• Sarah Jones (28F) - Asthma
• Mike Wilson (45M) - High Cholesterol

🚀 Quick Start

1. **Installation**

# Clone and setup
git clone https://github.com/sp2learn/medical-mcp-server.git
cd medical-mcp-server

# Create virtual environment
python3 -m venv venv
source venv/bin/activate  # On Windows: venv\Scripts\activate

# Install dependencies
pip install -r requirements.txt

2. **Configuration**

# Copy environment template
cp .env.example .env

# Edit .env with your API keys
MEDICAL_MODEL=gemini
GOOGLE_API_KEY=your_gemini_api_key_here
DISPLAY_TIMEZONE=America/New_York

3. **Run the System**

Web Application:

python web_app.py
# Visit: http://localhost:8000
# Login: demo/password, doctor/secret123, admin/admin2024

MCP Server:

python server.py
# For integration with Kiro IDE or other MCP clients

🎯 Usage Examples

**Natural Language Queries**

The intelligent assistant understands natural language and automatically routes to appropriate tools:

🩺 "What is Ben's sleep summary for the past week?"
→ Analyzes sleep data, provides trends and insights

📊 "Show me Ben's blood pressure trends"
→ Reviews vital signs, identifies patterns

💊 "What medications is Ben taking?"
→ Lists current medications and adherence data

🔍 "What are the symptoms of diabetes?"
→ Provides general medical information

📈 "Compare Ben's glucose levels over time"
→ Analyzes lab data and trends

**MCP Integration**

Add to your Kiro IDE configuration (.kiro/settings/mcp.json):

{
  "mcpServers": {
    "medical-query": {
      "command": "/path/to/medical-mcp-server/venv/bin/python",
      "args": ["/path/to/medical-mcp-server/server.py"],
      "disabled": false,
      "autoApprove": [
        "medical_query",
        "symptom_checker", 
        "get_patient_sleep_pattern",
        "get_patient_vitals",
        "get_patient_labs",
        "get_medication_adherence",
        "get_patient_activity",
        "get_patient_overview"
      ]
    }
  }
}

🛠️ Available Tools

**MCP Tools (8 total)**

**Tool Management**

# List all tools and their status
python manage_tools.py list

# Show detailed tool information  
python manage_tools.py details get_patient_sleep_pattern

# Enable/disable tools
python manage_tools.py enable medical_query
python manage_tools.py disable symptom_checker

📁 Project Structure

medical-mcp-server/
├── 🤖 Core Intelligence
│   ├── intelligent_medical_assistant.py  # Natural language processing
│   ├── medical_client.py                 # AI model integration
│   └── patient_data_manager.py           # Patient data management
├── 🌐 Web Interface  
│   ├── web_app.py                        # FastAPI web application
│   ├── templates/