Body as biosensor: Faster COVID detection boosts military readiness
Keeping the military at peak readiness is job #1, so COVID-19 presents a particularly thorny emerging threat. There is no natural immunity, medical diagnostics or environmental detection for a virus whose victims may be utterly asymptomatic or might land in intensive care and then die from infection. It’s the kind of problem “that we think about all the time … in this space,” said John Hannan, chief of the Digital Battlespace Management Division at the Defense Threat Reduction Agency (DTRA).
Speaking at a Dec. 2 FCW Health IT workshop, Hannan said the “pandemic has presented a real-world scenario in which current detection and counter-measures are ineffective.” That pushed DTRA to develop a way to diagnose infected service members before observable symptoms appeared and get them screened for COVID before their illness impacted readiness. The agency was looking for a kind of “check engine light” that would signal that “something going wrong in the physiology of the human being or the soldier,” he said. “That allows you to put that screening assay upfront, quicker, and then everything that follows behind that is quicker.”
The effort actually predates COVID-19. In partnership with the Defense Innovation Unit and Philips, DTRA in 2018 developed wearables – a wrist-worn device and a ring -- that tracked temperature trends, respiration and heart rates and blood oxygen levels to screen for hospital-acquired infections before symptoms developed. When a machine-learning algorithm called the Rapid Analysis of Threat Exposure (RATE) was applied to the health data, it warned of infection up to 48 hours before the onset of symptoms.
Wearable technology uses the body as its own biosensor, Hannan said. It offers a non-invasive, continuous monitoring that can give health officials a heads up that something was wrong -- a “threat agnostic” approach – that could be widely applied.
Based on that earlier work, DTRA was able to take the technology out of the hospital environment, modify it and test it with volunteers to see if RATE could provide pre-symptomatic infection alerts an operational environment.
When modifying the technology to spot subclinical COVID infections, DTRA faced challenges adapting low-cost consumer wearables to detect infections outside a hospital setting, where conditions are pristine and the quality of both the wearable and the patient data is high. Other hurdles were transmitting data in unconnected environments and analyzing the results to determine if the algorithm could actually alert pre-symptomatic wearers that they may have the virus and unknowingly infect others.
The technology is being tested by volunteers from West Point and the Naval Academy who wear the devices 24/7. Participants connect their wearable to a smartphone via Bluetooth, take a brief symptom survey every 24 hours and ensure the wearable is uploading data. The app evaluates the symptom survey and the data from the wearable in the cloud and assigns a RATE score that indicates variance from baseline readings. Coordinators log into the secure cloud database and monitor individual scores and trends and advise commanders of health alerts.
The technology is not quite ready for roll out. “This generalized algorithm that was developed for the hospital acquired infections,” Hannan said, so it has to be evaluated outside health care facilities for COVID. This testing is really looking at the data collection and analysis, he said. We need to see “how the algorithm does and how that might need to be retrained. Or maybe we need to be looking at other metrics -- that's the back end challenge.”
Short term, RATE will allow commanders to identify infections earlier, potentially reducing community spread and increasing survival. On the front lines, it would improve readiness levels and increase personnel redundancies, Hannan said. Longer term, it could become a tool for leadership to monitor well-being and readiness of personnel, from commanders down to individuals.
Connect with the GCN staff on Twitter @GCNtech.