Wound EHR Analyzer

An AI-powered tool for analyzing and interpreting wound care management data, providing healthcare professionals with advanced insights and recommendations.

Overview

This application leverages large language models (LLMs) to analyze wound care data, generating comprehensive insights and evidence-based recommendations for healthcare providers. The system includes both an interactive web dashboard and command-line tools for efficient data processing and analysis.

Key Features

• Interactive Analysis Dashboard: Streamlit-based interface for real-time data visualization and AI-powered insights
• Multi-Model LLM Support: Compatible with various LLM platforms including OpenAI and custom endpoints
• Advanced Statistical Analysis: Comprehensive wound healing trend analysis and progression tracking
• Flexible Data Handling: Support for diverse data types including images, time-series data, and clinical notes
• Robust Error Handling: Graceful recovery from API interruptions and connection issues
• Containerized Deployment: Docker support for consistent deployment across environments

Dashboard Components

The interactive dashboard provides comprehensive wound analysis through specialized tabs:

• Overview: Patient demographics, wound summary statistics, and population-level trends
• Impedance Analysis: Electrical measurements visualization with clinical interpretations
• Temperature: Thermal gradient analysis for wound assessment
• Oxygenation: Tissue oxygen saturation monitoring and analysis
• Exudate: Characterization and trending of wound drainage
• Risk Factors: Patient-specific risk factor evaluation and impact analysis
• LLM Analysis: AI-powered natural language processing for comprehensive wound assessment

Quick Start

Prerequisites

• Python 3.12+
• Docker (for containerized deployment)
• OpenAI API key or compatible service

Installation & Setup

We provide convenient scripts for all setup operations. Choose the deployment method that best fits your needs:

Option 0: Pip Installation (Simplest Method)

# Install directly from PyPI (once published)
pip install wound-analysis

# Or install the latest version from GitHub
pip install git+https://github.com/artinmajdi/Wound_management_interpreter_LLM.git

# Run the dashboard
wound-dashboard

# Or run analysis from command line
wound-analysis --record-id 41

See documentation/INSTALL.md for detailed pip installation instructions.

Option 1: Docker Deployment (Recommended for Production)

# 1. Set up environment variables (API keys, etc.)
./scripts/setup_env_variables.sh

# 2. Start the application in Docker
./scripts/run_docker.sh start

# 3. Access the dashboard at http://localhost:8501

# 4. Run CLI analysis for a specific patient record
./scripts/run_docker.sh cli 41

# 5. Verify dataset structure and integrity
./scripts/run_docker.sh verify

Option 2: Conda Environment (Recommended for Development)

# 1. Create and configure the conda environment
./scripts/install.sh

# 2. Activate the environment
conda activate wound_analysis

# 3. Run the dashboard
streamlit run wound_analysis/dashboard.py

Option 3: Python Virtual Environment

# 1. Create and activate a virtual environment
python -m venv .venv
source .venv/bin/activate  # On Windows: .venv\Scripts\activate

# 2. Install dependencies and set up environment
pip install -r setup_config/requirements.txt
pip install -e .
./scripts/setup_env_variables.sh

# 3. Run the dashboard
streamlit run wound_analysis/dashboard.py

Documentation

• Configuration Guide: Environment variables and configuration options
• Docker Usage Guide: Detailed containerization instructions
• API Documentation: API reference and component documentation
• Installation Guide: Pip installation instructions
• Data Processing: Information on supported data formats and processing pipelines
• LLM Integration: Guide to configuring and using different LLM models

Project Structure

wound_management_interpreter_LLM/
├── setup.py                  # Package configuration
├── setup_config/                   # Configuration files
│   ├── .env.example          # Template for environment variables
│   ├── environment.yml       # Conda environment specification
│   ├── MANIFEST.in           # Package manifest file
│   ├── pyproject.toml        # Modern Python project metadata
│   ├── pytest.ini            # PyTest configuration
│   └── requirements.txt      # Python dependencies
├── documentation/            # Documentation files
│   ├── INSTALL.md            # Installation instructions
│   ├── LICENSE               # License file
│   ├── configuration.md      # Configuration guide
│   ├── docker_usage.md       # Docker deployment instructions
│   ├── data_processing.md    # Data processing guide
│   ├── llm_integration.md    # LLM integration guide
│   └── index.md              # Documentation index
├── docker/                   # Docker configuration
│   ├── Dockerfile            # Container definition
│   ├── docker-compose.yml    # Service orchestration
│   └── .dockerignore         # Build exclusions
├── scripts/                  # Utility scripts
│   ├── run_docker.sh         # Docker management script
│   ├── install.sh        # Conda environment setup
│   └── setup_env_variables.sh          # Environment configuration
├── tests/                    # Test suite
├── wound_analysis/           # Core application code
│   ├── dashboard.py          # Streamlit interface
│   ├── main.py               # CLI entry point
│   ├── cli.py                # Command line interface
│   ├── utils/                # Utility modules
│   └── dashboard_components/ # Dashboard components
│       ├── overview.py       # Overview tab component
│       ├── impedance.py      # Impedance analysis component
│       ├── temperature.py    # Temperature analysis component
│       ├── oxygenation.py    # Oxygenation analysis component
│       ├── exudate.py        # Exudate analysis component
│       └── risk_factors.py   # Risk factors analysis component
├── dataset/                  # Data directory (mounted at runtime)
├── .env                      # Environment variables
└── ide_config/               # IDE configuration
    └── Wound_management_interpreter_LLM.code-workspace  # VSCode workspace file

License

This project is licensed under the Creative Commons Attribution-NonCommercial 4.0 International License (CC BY-NC 4.0), which permits non-commercial use with attribution. See the documentation/LICENSE file for details.

Acknowledgments

• This project was developed as part of advanced research in AI-assisted healthcare
• Special thanks to the healthcare professionals who provided domain expertise