README.md

metadata

title: Nursing Language Translator
emoji: 🏥
colorFrom: blue
colorTo: indigo
sdk: gradio
sdk_version: 4.44.0
app_file: app.py
pinned: false
license: apache-2.0
short_description: Translate NHS clinical shorthand to formal language

🏥 Nursing Language Translator

Translate NHS clinical shorthand to formal language using AI.

Powered by NurseEmbed-300M, a clinical embedding model fine-tuned on NHS nursing terminology.

🚀 Try It Now

Live Demo: https://huggingface.co/spaces/NurseCitizenDeveloper/Nursing-Language-Translator

✨ Features

Feature	Description
🔤 170+ NHS Abbreviations	SOB, NEWS2, NOF, UTI, COPD, AF, LMWH, etc.
⚠️ NEWS2 Interpretation	Automatic risk scoring with clinical actions
🧠 Semantic Matching	AI-powered translation, not just string matching
📊 Confidence Scores	See how confident the model is in each translation
📚 Reference Guide	Browse all abbreviations by category

📝 Example

Input:

72M c/o SOB, NEWS2 score is 7, PMH: COPD, AF. Started on Salbutamol NEB and LMWH.

Output:

Term	Translation	Category
72M	72-year-old Male	Demographics
c/o	complaining of	Assessment
SOB	Shortness of Breath	Respiratory
NEWS2 7	🔴 High risk - Emergency response required	Assessment
PMH	Past Medical History	History
COPD	Chronic Obstructive Pulmonary Disease	Respiratory
AF	Atrial Fibrillation	Cardiovascular
NEB	Nebuliser	Route
LMWH	Low Molecular Weight Heparin	Medication

🛠️ Local Installation

# Clone the repository
git clone https://github.com/Clinical-Quality-Artifical-Intelligence/nursing-language-translator.git
cd nursing-language-translator

# Install dependencies
pip install -r requirements.txt

# Run the app
python app.py

Open http://localhost:7860 in your browser.

📁 Project Structure

nursing-language-translator/
├── app.py                  # Main Gradio application
├── knowledge_base.json     # 170+ NHS abbreviations database
├── requirements.txt        # Python dependencies
└── README.md               # This file

🧠 How It Works

1. Input Parsing: Extracts individual terms and phrases from clinical text
2. Semantic Embedding: Uses NurseEmbed-300M to create vector representations
3. Knowledge Base Matching: Finds the closest matching abbreviation in the database
4. NEWS2 Detection: Automatically identifies and interprets Early Warning Scores
5. Translation Generation: Produces formal clinical language with confidence scores

📊 Model Performance

NurseEmbed-300M was trained using a two-stage hybrid approach:

Stage	Dataset	Samples	Performance
Stage 1 (Medical)	miriad-4.4M-split	10,000	81.3% Acc@1
Stage 2 (Nursing)	Custom NHS dataset	200	95.4% Acc@10

🏷️ Abbreviation Categories

• Assessment: NEWS2, MUST, Waterlow, AVPU, GCS
• Respiratory: SOB, COPD, URTI, LRTI, O2 sats
• Cardiovascular: AF, MI, PE, DVT, CCF
• Medications: LMWH, GTN, PRN, STAT, OD/BD/TDS
• Routes: IV, IM, SC, PO, PR, NEB
• Investigations: CXR, ECG, FBC, U&E, CT, MRI
• Locations: A&E, AMU, ITU, HDU, CCU
• And 10+ more categories...

🤝 Contributing

Contributions are welcome! To add new abbreviations:

1. Fork the repository
2. Edit knowledge_base.json to add new entries
3. Submit a pull request

📄 License

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

👨‍⚕️ Author

Lincoln Gombedza (@NurseCitizenDeveloper)

• 🏥 Registered Learning Disability Nurse
• 🎓 Practice Educator
• 💻 Co-Chair, Digital & Technology Working Group NHS Professional Strategy for Nursing and Midwifery
• 🚀 Founder, Nursing Citizen Development Movement

🔗 Related Projects

• NurseEmbed-300M - The embedding model
• NurseSim-Triage - AI triage training
• The Virtual Ward - RL environment for ward management

---

Built with ❤️ for NHS Nurses