NHS Rolls Out AI Tool That Predicts Falls and Hospital Admissions

A NEW artificial intelligence (AI) tool is now being used across the NHS to predict falls and health deterioration in patients, with the potential to prevent up to 2,000 falls and hospital admissions daily. The tool is proven to predict falls with 97% accuracy.

The AI tool works by monitoring vital health signs during over 2 million home care visits each month. It tracks parameters like blood pressure, heart rate, and temperature, using real-time data to predict risks, such as falls or the onset of winter viruses like flu, COVID-19, and norovirus. Healthcare professionals are then alerted to intervene and reduce the need for hospitalisation. The software operates across more than two-thirds of NHS integrated care systems, generating approximately 5,000 high-risk alerts each day.

Initial results show significant benefits for both patients and the healthcare system. Falls, which are the leading cause of emergency hospital admissions for older people, affect around 30% of those aged over 65 years. In the first months of its roll-out, the AI tool has already reduced hospital admissions by up to 70%, saving the NHS over £1 million daily. It also helps free up hospital beds by keeping patients safe and well at home, thus contributing to overall efficiency in care delivery.

This technology is part of the NHS’s ongoing efforts to embrace digital solutions, ensuring better care for patients while alleviating pressures on emergency services. As part of the upcoming 10-year Health Plan, the integration of AI in home care signifies a shift towards preventative care and greater reliance on community-based treatment, ultimately improving quality of life for older and vulnerable individuals across the UK.

Ada Enesco, EMJ

Reference

NHS England. Nationwide roll out of artificial intelligence tool that predicts falls and viruses. 2025. Available at: https://www.england.nhs.uk/2025/03/nationwide-roll-out-of-artificial-intelligence-tool-that-predicts-falls-and-viruses/. Last accessed: 5 March 2025.