AI Aids DOD in Early Detection of COVID-19

 Sept. 22, 2020 | BY David Vergun , DOD News

The Defense Threat Reduction Agency and Defense Innovation Unit are applying commercial technology for early detection of the Severe Acute Respiratory Syndrome Coronavirus 2, which causes the COVID-19 coronavirus disease.

That technology is known as Rapid Analysis of Threat Exposure, or RATE, and it consists of non-invasive, wearable devices that provide early warning of infection up to 48 hours before a person becomes symptomatic, helping ensure military readiness, and protect against the threat of further spread of the disease, said DTRA Science and Technology Manager Ed Argenta.

Just like the ''check engine'' warning that comes on in a car, said DIU Human Systems Director Dr. Christian Whitchurch, the system is designed to alert service personnel that something is wrong, leading them to pursue early diagnostic testing with their physician  and then treatment.

RATE integrates consumer commercial-of-the-shelf wearables to measure key biomarkers. That data is then processed in the cloud to allow users to see their hourly RATE score through a secure website, he noted.

Data from 165 different biomarkers is collected by RATE and that data is then processed in the cloud to allow users to see their hourly RATE score through a secure website, he noted. 

The innovative technology uses artificial intelligence and machine learning to increase effectiveness of detection, he added

Using RATE, researchers discovered that exposure to infectious agents causes subtle changes in people's physiology before they experience symptoms. Identifying these changes early in the infection is critical to containing the spread, as asymptomatic and pre-symptomatic individuals don't yet show signs of infection, and can unwittingly spread the disease to others, Argenta said. An early warning solution could potentially alert people to their possible infection, and enable them to quarantine and change their behaviors sooner to avoid infecting others.

The RATE model was trained via AI and ML on 293,109 participants, including 256,320 controls and 36,782 with known hospital acquired infections and correlated to these common attributes: temperature, pulse oximeter and cardiac measures, he explained. 

Dr. Joe Frassica, the chief medical officer and head of Philips Research North America, the company assisting DTRA, said ''As we continue to get new data from monitored cases of COVID-19, we will be able to refine the RATE-COVID algorithm in the near future. We hope that this will not only allow us to protect people from contracting the disease, but to also intervene early and treat those who are infected.''

Whitchurch said developing RATE began in March 2018, and it did test for another coronavirus, SARS, as well as pneumonia in that DTRA/DIU development prototype. As the effort quickly pivoted in 2020 to identifying COVID-19, the CARES Act provided additional funding for a total of $7 million to refine the technology.

The Defense Department is set to oversee an extensive rollout of RATE devices to nearly 5,000 people in the coming weeks, Whitchurch said. The Navy, Office of the Secretary of Defense and DTRA and DIU personnel began receiving RATE devices in June. U.S. Northern Command is receiving RATE devices this month and the U.S. Military Academy is slated to receive theirs next month.