Scientists create a smartwatch-based device to detect COVID-19 early

Scientists have published an article in the journal Nature Medicine detailing the development of a smartwatch-based alarming device that can detect the start and severity of severe acute respiratory syndrome coronavirus 2 infection (SARS-CoV-2). The system analyzes the abnormalities in activity and physiological signals that are associated with early-stage infections and can detect the signs of infection.

Study Alerting system that is real-time to monitor COVID-19 and other stress events using wearable data. Image Credit: wavebreakmedia/Shutterstock


SARS-CoV-2, the causative pathogen of coronavirus disease 2019 (COVID-19), is a highly infectious and deadly respiratory virus of the human beta-coronavirus family. To effectively control viral transmission and disease progression it is crucial to detect the virus during its early stages of infection, before the onset of symptoms.

Smartwatches and other wearable devices with real-time monitoring and alerting systems are useful interventions for the early detection of respiratory infections and other diseases. These devices act by detecting and analyzing changes in the physiological and activity parameters, such as heart rate, step count as well as sleep patterns and body temperature that are related to infection onset.

Scientists have created a wearable device that detects the early onset COVID-19. This is done by analysing the abnormal changes in the body caused by SARS-CoV-2 infection. The algorithms employed in this device can detect the disease before symptoms onset.

Smartwatch-based real-time alerting system

The real-time monitoring and alerting system developed by scientists detects aberrant physiological signals using wearable devices. The system collects physiological information (heart rate and steps) and health data (surveys for illness, medication and vaccinations) through smartwatches. It transfers the information to the cloud in real-time for analysis.

The data was analysed using three online infection detection algorithms on the cloud, specifically NightSignal, RHRAD, and CuSum. The findings obtained from NightSignal algorithm were used to alert the participants in real-time.

Validation of the system

Wearable data from 2,155 participants was used to test the system’s effectiveness. 2,117 participants received daily alerts about physiological changes. Overall, the analysis involved 278 participants who reported positive SARS-CoV-2 test results, 1,213 people who reported SARS-CoV-2-negative test results, 1,825 participants without any test results and 189 vaccine-vaccinated participants.

Participants who received an alert were prompted to provide details on their disease symptoms, diagnosis, and activities using the various COVID-19 surveys. The diagnoses included COVID-19 influenza, and adenoviral disease. The symptoms included headache, fever, cough, and adenoviral infection. The causes were a variety of lifestyle factors like intensive exercise, alcohol consumption and travel, as well as stress that can alter the physiological parameters.

Efficacy of the system

The system employed NightSignal to detect early signs of SARS-CoV-2 infection and alerted patients in pre-symptomatic stages (i.e. three to 10 days prior to symptoms onset). The algorithm was able to detect the onset of disease even in participants who were not symptomatic and had positive test results. The algorithm identified infection in both pre-symptomatic and asymptomatic cases with the sensitivity of 80%. Some missed cases identified in the study may be due to the lack of wearable data.

The other two algorithms, RHRAD and CuSum, which required high-resolution data, demonstrated 69% and 72 percent sensitivity, respectively, in detecting infection onset. The NightSignal algorithm and RHRAD algorithms showed 87 percent and 83%, respectively and 83% respectively in detecting the onset of infection among people who had negative test results. The CuSum algorithm was specific to 83.

On average, the system issued 3.4 alerts over a 21-day infection detection window for participants with SARS-CoV-2-positive test results. For those who had negative test results the number of alerts was 1.3.

Association between system-generated alerts and symptom progression

A review of symptoms reported by patients and the frequency of alerts received, revealed that fatigue and sleep disturbance were both common in COVID-19 cases. However, COVID-19 positive cases were more prone to symptoms like headaches and aches. System-generated alerts were also more prevalent in COVID-19 cases that were negative.

Concerning the physiological reactions to COVID-19 vaccinations, studies revealed that system-generated alerts could be triggered either by a single dose (one dose) or full (two doses).

In particular, the analysis of resting heart rate fluctuations prior to and after vaccination showed that vaccination with the one 1st dose of Pfizer COVID-19 vaccine triggered the maximum increase in heart rate during the night of vaccination. For the 2 2nd dose of Moderna and Pfizer vaccines the heart rate increased on the first and second night, respectively.

The analysis of the symptoms reported and received alerts indicated that the system was able to detect the symptoms triggered by vaccines as well as vaccination effects.

The significance of the study

The study found that the smartwatch-based, real-time alerting system has high effectiveness in delivering early warnings of infection with SARS/CoV-2. A large-scale application of this system could help in preventing the spread of virus and accelerate disease progression by detecting the disease early.

Journal reference:

Content Source:

Gemma Wilson

Gemma is a journalism graduate with keen interest in covering business news – specifically startups. She has as a keen eye for technologies and has predicted quite a few successful startups over the last couple of years.

Related Articles