Could wearable sensors provide early warning signs of respiratory infections?

30-11-2021 |   |  By Robin Mitchell

A recent paper published by Duke University showed that wristbands fitted with biosensors can be used to detect early cases of respiratory infections such as influenza and the common cold with a high degree of accuracy in asymptomatic cases. What challenges do asymptomatic individuals present, what did the research show, and how could such technology help control future pandemics and contagions?

Why asymptomatic individuals are a headache in pandemics

Someone who is asymptomatic means that they are infected with a contagion but show no sign of symptoms. For example, asymptomatic individuals carrying the flu virus will not have a bad chest, congested sinuses, or even a headache. Many would think that those who are not showing symptoms are therefore not contagious, but this is far from the truth, as the COVID pandemic has shown.

One of the biggest blunders made by researchers and governments worldwide is their action and modelling around asymptomatic individuals and those who are vaccinated. Generally speaking, the message has been that asymptomatic individuals are unlikely to transmit the disease while vaccinated people can’t spread the virus. However, the truth is that those who are asymptomatic are still able to spread the virus, as well as those who are vaccinated.

If viruses only passed amongst those showing symptoms, then COVID would have easily been controlled. This has been demonstrated with the virus’s ability to spread despite lockdowns, the large number of asymptomatic people testing positive for the virus, and the most significant contributor of virus spreaders being those who are vaccinated (most likely caused by misinformation).

The challenges of asymptomatic patients are exacerbated when it comes to identifying asymptomatic carriers. An asymptomatic individual can only be identified when being confirmed to have the virus, and such a test is hard to do when there are no signs of the virus.

This inability to identify those who are asymptomatic as well as those who are carrying the virus means that governments cannot do targeted lockdowns and isolations. If this was possible, the global economy would have unlikely to be affected by the COVID pandemic.

Researchers demonstrate wearable devices detecting asymptomatic individuals

A study conducted over the past few years by Duke University has finally been published on the use of wearable devices to detect early infection of common viruses such as influenza and the common cold. According to the study, wearable devices gathering biodata accurately determined asymptomatic individuals 92% of the time for influenza and 88% for the common cold. Furthermore, the model designed around the data collected was able to differentiate between mild and moderate infection 24 hours before any symptoms showed with an accuracy of 90% for influenza and 89% for the common cold.

The wearable sensors used by the researchers gathered data, including resting heart rate, heart rate changes, acceleration, electrodermal skin activity, and skin temperature. This data was then fed into an algorithm designed to look for changes and abnormalities that would indicate a sign of infection. While the paper does not mention the term Artificial Intelligence, it implies using Machine Learning to create an algorithm with predictive abilities. This model was created using Python Scikit.

How could such technology help control future pandemics?

Trying to lock down the entire population has proven to be problematic. The possibility of some countries (such as Austria) bringing in mandatory vaccinations raises concerns about human rights (specifically, the right to refuse medical treatment and autonomy on one’s own body). These problems are not made any better by trying to introduce track and trace systems that potentially invade individual privacy to try to only keep some individuals away from society while potentially infectious.

However, wearable devices incorporating biosensors with the sole purpose of predicting infection in those who are both symptomatic and asymptomatic could be a very good compromise. Citizens would not be required to send any personal data to the government or require tracking. Furthermore, vaccines would not need to be made mandatory as early-warning sensors could help keep potentially infected individuals away from society. The only requirement is that citizens wear a medical device that simply informs them when they are potentially infectious and take extra precautions to prevent further infection.

Such wearable devices could be based around popular smartwatches (such as the Apple Watch) or wristbands (such as Fitbit). Considering that AI silicon already exists, a small microcontroller working alongside an AI processor could arguably be used to run prediction models to detect if the wearer is potentially infectious.

But most importantly, what this research demonstrates is that AI predictive algorithms can and should be used in medical science and medical diagnosis. Making predictions about the future is often a bold move as anything can happen, but one thing that can be said for certainty about AI and medicine is the following:

AI WILL make access to medical treatment fairer, easier, and faster than has even been possible. Anyone around the world WILL be able to go online and consult with an AI doctor before being directed to appropriate treatment sources. 


By Robin Mitchell

Robin Mitchell is an electronic engineer who has been involved in electronics since the age of 13. After completing a BEng at the University of Warwick, Robin moved into the field of online content creation developing articles, news pieces, and projects aimed at professionals and makers alike. Currently, Robin runs a small electronics business, MitchElectronics, which produces educational kits and resources.

Related articles