News

Researchers Detail Use of Machine Learning Algorithms for Pandemic Predictions

• By John K. Waters
• 04/22/2020

Two projects that use machine learning algorithms to enhance our understanding of what works and what doesn't in the fight against COVID-19 have come to fruition.

In one, University of Cambridge researchers , working with the NHS Digital, have begun trials of a system that uses machine learning to make accurate predictions about the demand for ICU beds and ventilators to treat patients with COVID-19 at individual hospitals and across regions in England.

The COVID-19 Capacity Planning and Analysis System (CPAS) uses data from Public Health England to help hospitals plan more accurately and deploy resources strategically.

"With the pressure being placed on intensive care by the current coronavirus pandemic, it is essential to be able to predict demand for critical care beds, equipment, and staff," said NHS Digital chief medical officer Jonathan Benger in a statement.

"CPAS allows individual hospitals to plan ahead, ensuring they can give the best care to every patient. At the same time, the wider NHS can ensure that the ventilators, other equipment, and drugs that each intensive care unit will need are in place at exactly the time they are required. In the longer term, it is hoped that CPAS can be used to predict hospital length of hospital stay, discharge planning and wider intensive care demand in the time that will come after the pandemic."

CPAS is built around a machine learning engine called Cambridge Adjutorium, developed by Mihaela van der Schaar and her team at the University of Cambridge. A highly flexible machine learning system created for medical researchers, Cambridge Adjutorium has been used to develop insights into cardiovascular disease and cystic fibrosis.

Meanwhile, a graduate student and a professor at MIT have deployed a machine learning algorithm to answer the question, "What has been the impact of quarantine measures on the spread of COVID-19?"

"Our model is the first which uses data from the coronavirus itself and integrates two fields: machine learning and standard epidemiology," said Raj Dandekar, a PhD candidate studying civil and environmental engineering, in an article on the MIT campus news site. Dandekar collaborated with George Barbastathis, professor of mechanical engineering, to develop the model as part of the final project in the "Learning Machines" class.

Dandekar and Barbastathis focused their analysis on four regions struck by the pandemic: Wuhan, Italy, South Korea and the United States. Their goal was to capture the number of infected people under quarantine, and therefore no longer spreading the infection to others. They used an enhancement of a well-known disease prediction model, the SEIR, which categorizes people into four groups: susceptible, exposed, infected and recovered. They used that enhanced model to compare the role played by the quarantine and isolation measures in each of the four regions in controlling the effective reproduction number of the virus.

The MIT duo published the results of their study in a paper ("Quantifying the effect of quarantine control in Covid-19 infectious spread using machine learning"), currently in a pre-print edition available online only. "Our results unequivocally indicate that the countries in which rapid government interventions and strict public health measures for quarantine and isolation were implemented were successful in halting the spread of infection and prevent it from exploding exponentially," they wrote.

"Our model shows that quarantine restrictions are successful in getting the effective reproduction number from larger than one to smaller than one," Barbastathis told MIT News. "That corresponds to the point where we can flatten the curve and start seeing fewer infections."