The Accelerated Capability Environment (ACE) has worked with NHS AI Lab to develop a pioneering artificial intelligence tool to identify those patients who are at risk of prolonged hospital stays, which is helping to transform patient outcomes and reduce health costs.

ACE – a Home Office capability tackling security challenges arising from digital and data technology – has co-developed the new tool that assesses hospital data and uses AI technology to predict which patients are at risk of extended hospital stays. By spotting these patients early, more can be done to avoid the negative implications of lengthy hospital stays by adjusting treatment plans.

Prolonged hospital stays can lead to higher rates of mortality, higher risks of readmission as well as physical decline, particularly in elderly patients. Gloucestershire Hospitals NHS Foundation Trust, who were involved in the project, found that 4% of all admissions to the trust resulted in a stay of 21 days or longer. And this group of long-stayers accounts for over a third (34%) of all bed stays.

Professor Peter Brindle, strategic engagement lead, ACE, said: “The ability to not only identify individuals at high risk of an extended hospital stay before their admission, but also pinpoint the specific factors contributing to this likelihood is instrumental. It enables swift action to explore alternatives to admission. If hospitalisation remains necessary, it allows for the early mitigation of risk factors associated with a prolonged stay.

“Ultimately, this translates to reduced time spent in the hospital, which is excellent news for patients and a significant relief for busy healthcare facilities.”

ACE worked with Polygeist to develop a long stay stratification tool. The AI model was trained on 460,000 anonymised records, and used information available from initial patient data collection to make its analysis.

ACA and NHS AI Lab delivered a proof of concept in just 12 weeks. When trialled by Gloucestershire Hospitals NHS Foundation Trust, the tool was able to detect 66% of long stayers within the highest risk categories. As well as bringing tremendous health benefits to patients, it also delivered enormous financial benefits. A reduction of just a single day on an average hospital stay can lead to £1.7m in savings for Gloucestershire Hospitals alone.

Following the proof of concept, the tool was integrated with the trust’s electronic health record system via application programming interfaces (APIs).

Similarly, a scheme led by NHS Cheshire and Merseyside integrated care board, with technology partner C2-Ai is also using AI to help identify patients on waiting lists who are at high risk of deterioration.