Philip’s healthcare, with its Department of Defense (DoD) partners, Defense Innovation Unit, Defense Threat Reduction Agency and DARPA, just published results in Nature Scientific Reports highlighting how researchers have been able to use data from Oura and Garmin to build the “RATE algorithm”, a tool that generates real-time risk scores for identifying COVID-19 cases within their workforce.
Results show that Rapid Analysis of Threat Exposure, or RATE, risk scores begin to rise as early as 6 days prior to testing positive for COVID-19 and, on average, appear elevated 2.3 days in advance.
These scores successfully serve as a “check engine light” that can alert organizations and individuals of possible early cases of infection in an active workforce that is “committed to being ‘on location’ and ‘in person’ during public health crises.”
The RATE predictive algorithm is actively being improved via machine learning and already has an 82% chance of distinguishing between positive COVID-19 cases from the rest of the population (AUC = 0.82). Oura’s engineers are actively collaborating to make changes that support rapid incremental improvements.
Check out the full paper.