Wearable Device Uses AI to Detect COVID-19 Symptoms

Continuous monitoring and machine learning in a patch-like wearable device enables early detection of COVID-19 symptoms. Remote monitoring capabilities extend the reach of doctors and the safety of patients.


Article Key

Image source: Sonica Health

If necessity is the mother of invention, COVID-19 is the necessity that has spawned invention at an amazing rate. New products are emerging to help combat the spread of disease and to monitor those at risk of catching it.

Sometimes it requires reimaging or adapting an existing device. That’s what researchers at Chicago’s Northwestern University and the Shirley Ryan AbilityLab have done. The Northwestern medical team has developed a wearable device, called ADAM, that can detect the early onset of COVID-19 symptoms and can monitor these symptoms as the disease progresses. Originally designed as a device to help develop speech and swallowing protocols for stroke patients in rehabilitation, the researchers adapted it to detect COVID-19 symptoms.

About the size of a postage stamp, the flexible, wireless monitor adheres like a bandage to the skin just below the suprasternal notch—the divot at the base of the neck between the collarbones. It tracks the wearer’s heart rate, temperature, and respiratory and coughing patterns, which are stored in the device’s memory module. 

That data is sent to the cloud and compared with data from people who have COVID-19, who have recovered from COVID-19, sick people who don’t have COVID-19, and healthy people. When the algorithm indicates a person is at risk for having COVID-19, he or she is sent for a nasal swab test.

From Detection to Prevention

ADAM, short for Advanced Acoustomechanic system, acts like a digital wireless stethoscope. Using a suite of clinical-grade sensors, it measures vibrations on the skin. According to John Rogers, the Northwestern professor who led the research team, ADAM monitors the frequency and intensity of coughs and can sense labored breathing. Because it is located on the throat near the carotid artery, it can also measure the “mechanical signatures of blood flow” and, thus, a person’s heart rate.

In addition to monitoring heart rate and respiratory indicators, ADAM tracks temperature and oxygen levels. A flexible pulse oximeter monitors for silent hypoxia, typically an asymptomatic indicator of COVID-19. Checking blood oxygenation levels is critical because of the respiratory nature of the disease. Future versions might be able to monitor other biometrics, such as brain signals.

Continuous Digital Testing 

One of the benefits is that digital testing is continuous, so preventative measures can be taken immediately after variations are identified, further reducing the spread of the disease. Digital testing also enables physicians to see how the disease responds to treatment. Molecular or nasal swab testing requires more material resources, and results are based on health at the time the test occurred. Wearables can provide early detection, giving patients and healthcare providers actionable insights. 

Other device manufacturers, including Fitbit, have reported that wearables that continually monitor data can help with COVID-19 detection. A report from Conor Heneghan, PhD, the director of research and algorithms at Fitbit, says that based on the findings of a 2020 study, Fitbit devices can detect nearly 50 percent of COVID-19 cases one day before participants reported the onset of symptoms with 70 percent specificity.

“This is important because people can transmit the virus before they realize they have symptoms. If a wearable can let people know they should get tested a day before symptoms begin, they can isolate and seek care sooner, which can help reduce the spread of COVID-19,” Heneghan states. 

Adding Data to Boost Algorithm Accuracy

The ADAM COVID-19 monitor has been tested on more than 50 patients and medical staff who contracted the disease. That translates to more than 3,000 hours of collected data, which is fed into the device’s machine-learning algorithms. The goal is to use that data to enable the algorithms to distinguish between COVID-19 coughs and less threatening coughs from colds or allergies.

white monitoring tools on black background

Image credit: Northwestern University 

It is also being used to monitor pulmonary patients. The device transmits data to the cloud, where algorithms create and graphically display a patient’s respiratory status. Doctors can follow patients’ progress even after they are released from the hospital. The remote monitoring capabilities are a bonus during the pandemic as the use of telehealth continues to rise.

The ADAM device is being manufactured by Sonica Health, a startup company led by John Roberts. Sonica Health is an offshoot of Northwestern University and is partially funded by the Biomedical Advanced Research and Development Authority (BARDA), part of the U.S. Department of Health and Human Services.