This AI tool helps hospitals predict COVID-19 bed and ventilator demand

NHS Digital said the initiative marks the first time a project of such scale involving machine learning has been rolled out.

Decades of biology go into Google’s work on COVID-19 coronavirus
3:34

The NHS has started trials of a machine-learning system designed to help hospitals in England anticipate the demand on resources caused by COVID-19.

The COVID-19 Capacity Planning and System (CPAS) is being piloted at four acute hospitals in England to demonstrate whether it can help the NHS predict the demand for equipment like ICU beds and ventilators. If successful, CPAS will be rolled out nationally.

latest developments

Coronavirus: Business and technology in a pandemic

From cancelled conferences to disrupted supply chains, not a corner of the global economy is immune to the spread of COVID-19.

Read More

NHS Digital told ZDNet the initiative marked the "first time any project of this scale and scope using machine learning has been rolled out in the NHS."

SEE: How to implement AI and machine learning (ZDNet special report) | Download the report as a PDF (TechRepublic)

Trials are expected to last a week and will take place at NHS trusts Cambridge University Hospitals, Kings College Hospital, Lancashire Teaching Hospitals and University Hospital Southampton.

If initial deployments prove successful, the tool will be rolled out nationally "over a period of days" through a secure online portal where each hospital can access their own data reports, an NHS Digital spokesperson told ZDNet.

CPAS is based on a machine-learning system called Cambridge Adjutorium, which has been developed by a team of researchers at Cambridge University led by Professor Mihaela van der Schaar.

Cambridge Adjutorium has already been used to develop insights into cardiovascular disease and cystic fibrosis.

The system being used by NHS Digital has been trained with depersonalised data from 4,000 patients provided by Public Health England to provide near-term projections of demand on hospital resources.

Hospitals using the CPAS tool in the trial can run "simulation environments" allowing them to see how different scenarios would play out, such as by increasing the number of available beds or changes in the demographics of admitted patients.

Professor van der Schaar told ZDNet that CPAS had not been designed to take decisions on capacity planning, but rather help hospitals make better decisions based on the available data.

"It is important to note that this is not a medical device," she said. "It is a planning tool, to help predict the resources that are needed, and capacity over time."

The quality of the predictions generated by CPAS will depend on the quality of the data submitted by each healthcare provider, NHS Digital added. "We recognize that data quality can always be improved so we are working with Public Health England and NHS England & Improvement to help units improve their data returns," the spokesperson said.

NHS Digital is also identifying other sources of readily available data that can be used in the model "so that the burden of data collection can be reduced for the very busy teams," they added.

"We know that once the data is useful and provides rapid feedback to submitters, the data quality will likely improve."

Beyond COVID-19, there is scope to develop the machine-learning system into a broader framework for helping the NHS manage stretched resources.

SEE: Coronavirus: Business and technology in a pandemic  

Professor van der Schaar said that there has also been interest in the technology from the Netherlands, Italy, Israel and the United States.

"There is an opportunity here to empower this large-scale system to plan more effective delivery of healthcare services. This is not just about optimising disease response, but ensuring the NHS has the right resources at the right time," she said.

"You often here about AI models being apps. Here, we are talking about something more exciting – artificial intelligence to help empower national health systems."

More on using technology to tackle coronavirus