How artificial intelligence can cut your wait for a hospital bed by an hour when you go to A&E
Artificial intelligence can significantly reduce the length of time an A&E patient needs to wait for a hospital bed – and reduce the chance of them having to move beds once they have one, a new study has found.
Researchers have developed an AI tool that is able to predict much more accurately whether any given visitor to a casualty ward is likely to need to stay overnight and which can be swiftly discharged by looking at range of personal information and live data.
By knowing the likely demand for hospital beds four to eight hours in advance managers can configure their wards more efficiently. This could see a more accurate ratio of male to female bays – which are unisex and typically house 4 to 5 beds – and of medical, surgical or elderly wards.
And it could help staff to prepare the right type of beds, which can have quite different requirements – such as a ‘cardiac monitor’ equipped to keep an eye on heart activity or an ‘airway’ bed, with special protective mattresses designed for greater comfort and enhanced circulation.
The AI model was developed by University College London researchers and has been trialled at the University’s hospital, UCLH.
“Our AI models provide a much richer picture about the likely demand on beds throughout the course of the day. They make use of patient data the instant this data is recorded and show it’s possible to improve quality and patient care without having to spend millions on highly sophisticated AI technologies,” said Dr Zella King, of UCL.
“We hope this can help planners to manage patient flow – a complex task that involves balancing planned-for patients with emergency admissions. This is important in reducing the number of cancelled surgeries and in ensuring high-quality care.”
During the pilot, the new AI system for bed predictions outperformed the conventional method, with daily forecasts, on average four admissions, or beds, off the actual figure – compared to 6.5 for the conventional method. This means inaccuracies were reduced by nearly a third.
“This AI tool will be hugely valuable in helping us manage admissions and patient flow,” said Alison Clements, Head of Operations, Patient Flow & Emergency Preparedness, Resilience & Response at UCLH.
Based on preliminary observations and anecdotal evidence she estimates that, on average, the superior forecasting abilities of the AI system shaved about an hour off the wait for the average A&E patient for a bed – although there can be considerable variation from one patient to the next and this estimate needs to be confirmed with further research, she said.
The researchers trained machine learning models using patient data recorded at UCLH between May 2019 and July 2021. These models assessed each patient’s probability of being admitted to the hospital from the emergency department based on data ranging from age and how the patient arrived in hospital, to test results and number of consultations, and combined these probabilities for an overall estimate of the number of beds needed.
They then compared the models’ predictions to actual admissions between May 2019 to March 2020.
After Covid hit, the researchers were able to adapt the models to take account of significant variations both in the numbers of people arriving and the amount of time they spent in the emergency department.
Professor Sonya Crowe, Director of the UCL Clinical Operational Research Unit, said: “Most applications of AI in healthcare so far have focused on clinical questions whereas this tool aims to help the operational side of healthcare – that is, how it is run and managed.”
UCL and UCLH will continue to refine the tool in the hope of it being used more extensive in their hospital – and potentially rolled out to other emergency hospital departments in the NHS.
“We see this project as the thin end of the wedge,” said Dr King.
The trial is detailed in the journal Nature Digital Medicine.
This work was funded by grants from the UCL and Partner Hospitals Wellcome Institutional Strategic Support Fund (ISSF) and the NIHR UCLH Biomedical Research Centre. Some of the contributing authors were funded by the National Institute for Health Research and NHSX.