Long before the start of the Covid-19 pandemic, the U.S. Department of Defense and Royal Philips had been working on a project to detect potential outbreaks of infectious disease before they happen. The idea was to use wearable devices to catch the early signs of a bacterial or viral infection before symptoms appear — potentially useful in preventing the spread of disease in naval ships, dorms, and other close groups.
The project, called Rapid Analysis of Threat Exposure (RATE), wasn’t just intended to pick up on common infections, such as the flu or strep. It was also designed with unknown pathogens in mind.
“It was also aimed at a pandemic. What would happen if we don’t know the agent?” Dr. Joe Frassica, chief medical officer and head of research for Philips North America, said in a phone interview.
As Covid-19 began to spread across the U.S. in the spring, Frassica and leaders at the Defense Innovation Unit (DIU) and U.S. Defense Threat Reduction Agency (DTRA) quickly came up with a plan to apply this system to the current pandemic.
“We all recognized that the background work we had been doing for the past few years could be leveraged for a new purpose that was so needed and immediate,” he said.
A version of the system designed for the Covid-19 pandemic was rolled out in June, and the study is expected to grow to several thousand participants.
Here’s how it works: Philips developed a risk algorithm from a large dataset of more than 41,000 cases of hospital acquired infection. It identified 163 biomarkers that go into it, such as heart rate, respiratory rate and oxygen saturation.
The system is designed to work with everyday wearables, such as smartwatches or rings. Philips partnered with Texas A&M University to set up a secure way of transferring de-identified data from the devices to the cloud. Then, individuals could sign in anonymously to check their scores hourly.
This early alert system could clearly be useful in a military context, as bases in the U.S. and overseas work to slow the spread of the novel coronavirus. In an outbreak earlier in the year on the USS Theodore Roosevelt, one in five sailors who tested positive were asymptomatic.
“RATE would allow us to non-invasively monitor a service member’s health and provide early alerts to potential infection that will help us to ensure troop readiness, better support their health and protect against the threat of further spread of the disease,” Edward Argenta, science and technology manager at DTRA, said in a news release “Unlike other more narrow approaches, this solution is designed to recognize a wide variety of infections and can help identify future novel threats.”
It could also be helpful in a civilian context, to monitor hospital patients for infection, and in a workforce setting, to identify individuals who should self-isolate and work from home.
While data collected in a hospital setting is generally more accurate than using wearables at home, Frassica said with multiple biomarkers combined, the algorithm can become a good predictor of infection.
“We’ve made a lot of progress in the last few months,” he said. “My major takeaway from this, having worked for a long time in AI and the collection of large scale data, this is an example of one of the reasons that we do it.”