Could your smartwatch help detect the next coronavirus outbreak? Scripps Research scientists think so

A new study from Scripps Research indicates that data from wearable devices could help researchers identify coronavirus outbreaks.
(David McNew / Getty Images)

Scientists at Scripps Research in La Jolla found that data from wearable devices coupled with self-reported symptoms help predict whether someone has the virus.

A new study by scientists at Scripps Research in La Jolla describes a tool that could help public health officials spot and contain outbreaks of the coronavirus that causes COVID-19.

You might already be wearing it.

One in five Americans owns a wearable device such as a Fitbit or an Apple Watch. The gadgets monitor your heart rate, how many steps you take and your sleep patterns — measurements that often change when you’re sick.

Scripps scientists found that combining wearable device data with self-reported symptoms predicted whether a person had the virus better than either input on its own. That makes the popular devices a way to potentially track the scope and spread of the pandemic, according to Dr. Eric Topol, director and founder of the Scripps Research Translational Institute and executive vice president of Scripps Research.

“Everyone talks about ‘test, test, test.’ That isn’t working,” said Topol, one of the study’s authors. “We need other ways to track the toll of the virus and who might be affected.”

The study findings, published Oct. 29 in the journal Nature Medicine, are part of the ongoing DETECT study (Digital Engagement & Tracking for Early Control & Treatment). About 30,000 people across the United States enrolled between March 25 and June 7, sharing data from their wearable devices and reporting symptoms when they felt sick.

About 3,800 participants reported symptoms including stomachache, cough, difficulty breathing and loss of sense of taste and smell. Of those who felt sick, 333 were tested for the coronavirus; 54 tested positive and 279 tested negative.

The researchers then tried to predict who would test positive or negative with a statistical model based on self-reported symptoms; it performed about as well as a model based on wearable device data (heart rate, step count and sleep length). But combining the two predicted test results best.

“I see this approach as being more useful on a population level, in terms of seeing more activity in a population over time,” said Dr. Chip Schooley, a UC San Diego infectious-disease specialist who was not involved in the study.

Topol agreed, noting that researchers could regularly monitor wearable device data and self-reported symptoms to spot coronavirus outbreaks and tip off public health officials, who could then ramp up community testing and other measures to curtail the virus’ spread.

The DETECT study is ongoing, with researchers looking to enroll 100,000 participants. To learn more, visit ◆