AI-Enabled Digital Health

Improving the ability of the National Early Warning Score (NEWS2) system to predict critical outcomes through additional patient data or amendments to the scoring process (Video)

The NEWS2 scoring system is widely used for systematically documenting and identifying clinical deterioration, with proven accuracy in predicting mortality within 24 hours, facilitating timely interventions and improving clinician communication. However, its accuracy in predicting adverse events beyond 24 hours is limited, particularly in vulnerable populations such as older adults, who were not adequately represented in its development. Enhancing the system’s ability to identify early deterioration in patients with low NEWS scores who later die in hospital is critical, along with incorporating additional routinely collected variables and leveraging digital technologies for longitudinal monitoring.

In collaboration with Dr Chris Plummer, Consultant Cardiologist and Chief Clinical Informatics Officer at Newcastle upon Tyne Hospitals NHS Foundation Trust, this project examines demographic, observational, and outcomes data to refine the NEWS2 scoring system. By exploring new variables and weightings, we aim to improve NEWS2's accuracy in predicting clinical deterioration, particularly in older adults. This work will ensure the tool remains clinically valuable in an ageing population increasingly susceptible to sudden physiological changes, ultimately contributing to better patient outcomes.

Relevant publications: 

Evaluating autonomous telemedicine for cataract surgery follow-up 

Ageing populations increase demand for many healthcare services, which can exceed the capacity of the clinical workforce, causing longer waits and staff burnout. Digital technology has the potential to help automate routine tasks, freeing up healthcare professionals to work on higher-skill tasks and spend more time on meaningful patient interactions. One example of this is cataract surgery, the UK’s most common operation. Although few patients experience complications, many Trusts still require face-to-face postoperative follow-up assessments. This creates additional pressure on healthcare services and inconvenience for patients, while not resulting in care changes for most patients. 

Our lab co-led a project with Ufonia Ltd., funded by the NIHR AI in Health and Care Award. We investigated whether an autonomous conversational agent, delivered via telephone, could safely and accurately perform a post-cataract surgery follow-up assessment in two NHS Trusts and detect patients in need of further review. The performance of the conversational agent (Dora, version R1) was compared to ophthalmic specialists. Our analysis found good preliminary evidence of Dora’s safety and accuracy. It was also generally considered easy to use and acceptable but highlighted the importance of maintaining a ‘human element’ in the follow-up process, particularly for patients with concerns or complications. 

Relevant publications: