Title: UCLH Utilizes Machine Learning to Address Emergency Bed Demands
University College London Hospitals NHS Foundation Trust (UCLH) is revolutionizing its approach to bed capacity management by leveraging a machine learning tool. Developed in collaboration with analysts from University College London (UCL), this tool accurately predicts the number of emergency beds required in various departments, including pediatrics, oncology, surgical, medical, and hematology, for the upcoming eight-hour period.
Craig Wood, the clinical operations manager at UCLH, emphasized the significance of managing patient flow within the hospital. Balancing elective treatments with emergency cases is a continuous and delicate process. Since introducing the machine learning tool, targeted actions have been taken to optimize bed availability in specific areas of the hospital, significantly improving the management of hospital capacity.
The NHS as a whole is currently grappling with demand, with many hospitals operating at full or over capacity. UCLH aims to empower its operational staff with this tool, ensuring an adequate number of beds are available across different departments.
When patients arrive at the Accident & Emergency (A&E) department and undergo vital observations and tests, the machine learning tool analyzes data such as vital signs, blood test results, and the need for specialist consultations. The forecasts are updated every 30 minutes. Additionally, the tool predicts the number of patients likely to arrive at A&E over the next eight hours, drawing upon electronic patient records from August 2021 to August 2022.
Zella King, one of the co-developers of this technology from UCL’s clinical operational research unit, highlighted the potential of machine learning and data to enhance day-to-day operations in hospitals. However, she emphasized that hospital capacity is influenced by numerous systemic factors, and machine learning should be seen as a valuable support tool rather than a comprehensive solution.
The current research builds upon previous work by the team in 2022, during which the tool could predict the total number of required beds but lacked the ability to differentiate between departments.
In 2020, amidst the Covid-19 pandemic, Faculty, an artificial intelligence firm, developed a tool employed by the government to provide the NHS with early warnings of any potential surges in Covid-19 cases. This tool significantly aided the national and local levels of the NHS in understanding bed capacity and availability.
UCLH’s adoption of machine learning to forecast and manage emergency bed demands marks a significant milestone in optimizing operational efficiency. While there is no doubt that machine learning has the potential to make a meaningful impact, it must be used with actionable and aspirational outputs in order to effectively support hospital operational staff.
With ongoing advancements in technology, UCLH’s innovative approach serves as a prime example of harnessing the power of machine learning to support crucial decision-making in healthcare. By developing highly accurate bed demand predictions, hospitals can efficiently allocate resources and provide optimal care to patients in need.
