Harry Abel

• Enhancing the National Early Warning Score 2 Using Observational Patient Data and Machine Learning
• MBBS

The National Early Warning Score 2 (NEWS2) is widely used across the NHS to identify patients at risk of deterioration. While NEWS2 is effective for 24-hour prediction, its accuracy declines over longer periods and may be less reliable in certain groups, such as older adults.

This project investigated the potential of a machine learning model to enhance NEWS2 by integrating patient observations and additional data, such as age, body mass index (BMI), diastolic blood pressure, and type of oxygen delivery device. The model, trained on anonymised historical patient data from Newcastle Hospitals NHS Foundation Trust, demonstrated promising predictive accuracy.

The main challenge was a high-class imbalance in patient data, which was addressed using categorisation techniques based on clinical relevance. The work highlights the importance of data diversity and collaboration between key stakeholders to ensure both clinical relevance and technical readiness of the model. Further refinement and validation on additional datasets are required to improve model performance and its generalisability across NHS trusts.