Preliminary analysis used wearables data to identify more than 50 percent of COVID-19 cases by third day of symptoms
Following promising initial results, the National Institutes of Health (NIH) and Evidation Health have entered Phase II of a study to identify early warning signs of COVID-19 using data from commercial wearable sensors. Now underway, Phase II will continue through the summer of 2021.
The ongoing study is supported by Evidation’s Achievement platform and network, which includes nearly 4 million participants in 9 out of 10 American zip codes. The study uses machine learning models to differentiate the earliest signals of COVID-19 from other infections. In Phase I, machine learning models based on data collected from commercial wearables correctly identified more than 50 percent of COVID-19 cases by the third day of symptoms with high specificity.
“Early detection of COVID-19 is one important way we can mitigate the spread of the virus, and our early results show wearables are a useful tool in making early detection possible on a large scale,” said Luca Foschini, Ph.D., the project leader and Evidation’s co-founder and chief data scientist. “In the next stage of this NIH-supported project, Evidation will examine an expanded set of commercial wearable sensors. Our goal is to develop an early warning system for COVID-19 that is sensor agnostic.”
The model developed in Phase I attained similar detection performance even when validated on data collected from a completely different study from the one it was developed on, showing promise that performance can be maintained in real-world settings. In Phase II, consenting members of the Achievement population will continue to share wearable data containing steps, sleep, and heart rate information to help researchers understand COVID-19 symptom onset.
This study is a one of a broad set of initiatives Evidation is engaged in to better understand and characterize COVID-19 and its impacts. The company recently shared a preliminary analysis of an ongoing, multi-part research study on vaccine sentiment and COVID-19. In June, Evidation and BARDA launched a study to track symptoms of COVID-19 in those at particularly high risk, including health care workers and other first responders, in order to better understand susceptibility to SARS-CoV-2 infection and enable models for earlier detection. Evidation also partnered with NYC Health and Mt. Sinai, to better understand COVID-19 and its impact on mental health.
The National Cancer Institute (NCI) and the National Institute of Biomedical Imaging and Bioengineering (NIBIB), both part of NIH, announced in September 2020 Evidation’s selection as one of seven digital health solutions as part of the agencies’ congressionally supported response to COVID-19.