README.md

metadata

title: PNA Assistant
emoji: 👨🏾‍⚕️
colorFrom: indigo
colorTo: purple
sdk: gradio
sdk_version: 5.12.0
app_file: app.py
pinned: false

Professional Nurse Advocate Assistant

AI-Powered Support for the A-EQUIP Model & Restorative Supervision

👨🏾‍⚕️ 👩🏽‍⚕️ 👨🏿‍⚕️ 👩🏻‍⚕️ 👩‍⚕️

---

📋 Table of Contents

• About
• Key Features
• The A-EQUIP Model
• Technology Stack
• Repository Structure
• Roadmap
• Local Development
• Contributing
• Important Disclaimers
• Citation
• Acknowledgements

---

🏥 About

The Professional Nurse Advocate (PNA) Assistant is an AI-powered educational tool designed to support nursing professionals in understanding and applying the A-EQUIP model (Advocating and Educating for Quality Improvement).

This tool focuses on:

• Restorative Clinical Supervision - Supporting staff wellbeing
• Person-Centred Communication - Compassionate, reflective dialogue
• Quality Improvement - Guiding continuous professional development

---

🚀 Key Features

Feature	Description
RAG-Powered Knowledge	Retrieval-Augmented Generation using the official A-EQUIP Model Guide
Foundation Model	Powered by Gemma 2 2B IT, instruction-tuned for clear, safe, and helpful dialogue
Premium UI	Modern, accessible interface with NHS-inspired design
Diversity-First	Inclusive design with diverse healthcare worker representation

---

📚 The A-EQUIP Model

The A-EQUIP model provides a framework for Professional Nurse Advocates with four key functions:

1. Normative - Monitoring quality and standards
2. Formative - Education and skill development
3. Restorative - Clinical supervision and staff wellbeing (primary focus)
4. Personal Action for Quality Improvement - Driving positive change

---

🧠 Technology Stack

• Base Model: google/gemma-2-2b-it
• Architecture: RAG (Context-Aware) + Instruction Tuned LLM
• Framework: Gradio + Hugging Face Spaces (ZeroGPU)
• Knowledge Base: RAG with Sentence Transformers + FAISS

---

� Repository Structure

├── app.py                 # Main application file (Gradio UI & Logic)
├── pna_client.py          # LLM Client handling generation
├── knowledge_base.py      # RAG implementation (FAISS + Embeddings)
├── requirements.txt       # Project dependencies
├── Professional...Guide.md# Source knowledge base document
└── README.md              # Documentation

---

🗺️ Roadmap

• Phase 1: Initial Deployment with RAG & Fine-tuned Model
• Phase 2: TranslateGemma Live (Real-Time WebRTC Audio/Video Translation)
• Phase 3: Voice Integration for Spoken Restorative Supervision
• Phase 4: Integration with NHS e-Learning Platforms

---

🛠️ Local Development

# Clone the repository
git clone https://github.com/ClinyQAi/Professional-Nurse-Advocate-Assistant.git
cd Professional-Nurse-Advocate-Assistant

# Install dependencies
pip install -r requirements.txt

# Run locally
python app.py

---

🤝 Contributing

Contributions are always welcome! Please see CONTRIBUTING.md for details on how to get started.

---

⚠️ Important Disclaimers

For Educational & Research Purposes Only

This AI assistant is designed exclusively for educational and research purposes. It does NOT provide:

• Clinical advice or diagnosis
• Treatment recommendations
• Mental health crisis support

Important Note on AI Limitations: While this assistant may provide general information about accessing services or resources, we cannot guarantee the accuracy of this information. AI models can "hallucinate" or provide outdated details. Always verify outputs with authoritative sources.

Not a Substitute for Professional Support

• This tool is not a replacement for qualified Professional Nurse Advocates, clinical supervisors, or mental health professionals.
• If you or someone you know is experiencing a mental health crisis, please contact a qualified healthcare professional immediately or call your local emergency services.

---

🎓 Citation

If you use this tool or model in your research, please cite:

@software{gombedza_pna_assistant_2024,
  author = {Gombedza, Lincoln},
  title = {The Professional Nurse Advocate AI: A specialized language model for supporting nursing workforce resilience research},
  year = {2024},
  publisher = {Hugging Face},
  url = {https://huggingface.co/spaces/NurseCitizenDeveloper/PNA-Assistant},
  version = {1.0.0}
}

📜 System Card & Methodology

For a detailed technical breakdown of the system architecture, RAG implementation, and ethical considerations, please see the System Card.

---

👤 Author

Lincoln Gombedza
Registered Learning Disability Nurse | Practice Educator | AI Researcher | Nurse Citizen Developer

• 🐙 GitHub: @NurseCitizenDeveloper
• 🤗 Hugging Face: NurseCitizenDeveloper
• 🏥 Organisation: ClinyQAi

---

🙏 Acknowledgements

This project builds upon the work of many individuals and organisations:

• Foundation of Nursing Studies (FONS) - For person-centred nursing principles
• NHS England - For the A-EQUIP Model and PNA Framework
• Google - For the Gemma 2 open model family
• Hugging Face - For hosting and ZeroGPU infrastructure

---

📄 License

This project is licensed under the Apache License 2.0 - see the LICENSE file for details.

_{Made with ❤️ for the nursing profession}