BS READS: Can fitness trackers detect Covid-19?

As the world faces the coronavirus pandemic, researchers and technologists are figuring out if fitness trackers and smartwatches could be used to detect early symptoms. But is it possible?

After his father suffered a heart attack, Aditya Awasthi got him an Apple watch. The wrist-strapped wearable would monitor his heart rate and alert if the readings breached permissible limits. This intervention not only saved Awasthi senior many visits to the doctor, as he could monitor his own stats real-time, but also helped him get a better grip on his lifestyle.

Now, as the world faces the coronavirus pandemic – over seven million people infected globally, and counting – researchers and technologists are figuring out if smart wearables can also be used as a weapon in fight against the virus.

Modern-day trackers, also called wearables or fitness bands, sport an assortment of sensors. They can track movement, sleep, heart rate, breathing, pulse rate, and body temperate. The trove of health data that they generate — and given that millions of such devices are already in use — makes a strong case for trackers to be repurposed as devices to detect Covid-19. But is it possible? Answering this is the objective of a research at Stanford School of Medicine, the biomedical research centre at Stanford University.

The study, launched in April 2020 in partnership with Fitbit and US non-profit Scripps Research, aims to determine if data from fitness trackers can be used to determine infection in a person before they start showing symptoms.

More specifically, the research aims to ultimately devise software algorithms to determine illness, by measuring changes in body temperature and heart-rate. These typically get elevated when the body is fighting a disease.

If successful, it could be a major boon in limiting the spread of viral infections, including Covid-19. In the case of coronavirus, which creates asymptomatic carriers who might spread the virus without even knowing, this breakthrough will be a remarkable event.

“Smartwatches and other wearables make many, many measurements per day — at least 250,000. That makes them powerful monitoring devices,” Michael Snyder, head of the Stanford research, was quoted as telling the medical centre’s in-house news journal. “My lab wants to harness that data and see if we can identify who is becoming ill as early as possible — potentially before they even know they are sick.”

Snyder, the chair of Stanford School of Medicine’s genetics department, has spent a large part of his life working with biosensors. For this study, he is recruiting volunteers who will provide their health data, using five different wearable devices, including those of Fitbit, which has donated 1,000 trackers for this study, according to Stanford Medicine News Centre.

There are many facets to this puzzle, but the most important is data. Researchers will monitor health patterns of healthy and Covid-19-positive individuals over a period of time. The data, and the ensuing data analytics, will ultimately uncover health patterns that will show if a person is Covid-19-positive.

An earlier study proved this is possible. In 2017, researchers at Case Western Reserve University showed that “specific patterns of heart-rate variation could indicate illness, sometimes even when the individual was asymptomatic.” When a person catches cold or flu, there is usually a period just before symptoms set in, when they display elevated heart-beat. And this is possible to detect.

However, in the case of the coronavirus – which is still being studied – not all symptoms are currently known. A majority of patients display difficulty is breathing, fever and sore throat, but there are also many who have reported skin rashes, nausea and vomiting. Further, little is known about how a patient’s heart beats during or before the symptoms become visible.

The other challenge is determining which vitals to monitor. Does covid-19 alter sleep patterns — frequent interruptions and movement during sleep — or does it elevate or lower the heart-beat? A combination of these vital signs, or all of them, may offer some clues. When that is done, how does one manage the data coming from all sorts of wearables with varying degrees of accuracy?

Vishal Gondal, founder of Indian smart-band maker Goqii, which has launched a similar study in partnership with German firm Thryve, says this will be possible over time. “As machines get trained on more and more data, the challenge of interoperability of data will be solved,” he says.

Gondal backs body temperature as a prime indicator for Covid-19, which Goqii smartbands can monitor. Under its partnership, Goqii is offering the data of users in India for Thryve to analyse and devise algorithms. Thryve itself is a major player in biometrics research. In partnership with the German government, it has deployed an app that monitors health stats — including sleep, pulse-rate and temperature— from users’ fitness trackers for Covid-19 symptoms.

“More than 500,000 people have downloaded the application and use their wearables to provide us data, and our engine calculates on a daily basis the likelihood of symptoms in Germany,” says Thryve Chief Executive Officer Freiedrich Lammel. “This enables the German government to understand the potential onset of symptom hotspots, a lot faster and more accurately than what was possible earlier.”

Beyond projects at Stanford Medicine and Goqii-Thryve, the idea is picking up steam in many other pockets as well.

Researchers at University of California, San Francisco, are using Oura rings, known for accuracy in measuring sleep activity. About 2,000 rings have been handed out to healthcare workers in the US — “to easily track changes in their body temperature, respiratory rate, and heart rate, (through which) they may be better equipped to understand early warning signs of infection within the group.”

In terms of institutional push, the US Army has asked technology companies to develop wearable sensors to detect early symptoms. On May 6, the agency floated a $25-million contract to develop a device using existing technology.

Fitbit, which is sponsoring the Stanford study, last month announced another study. Using proprietary algorithms on heart-rate data, it plans to identify signs of atrial fibrillation or irregular heart-beat, which may cause strokes. This will ultimately power on-board electrocardiogram (ECG) feature, which Fitbit will soon bring to its devices.

“These projects are in the experiment stage,” says Manish Singhal of Pi Ventures, a venture capital firm that specialises in deep-tech companies. “My gut feeling is that none of these will be conclusive. They will only be indicative” of whether trackers can detect Covid-19 after all.

“It should be fine if they do not have very high false positive rates. Even if trackers are able to just conclusively tell me that I have nothing to worry about, it would be great,” says Singhal.

Over the years, fitness trackers have become more and more inexpensive and, therefore, more widely used. Chinese brands like Xiaomi and Huawei have done in the fitness tracker segment what they did in smartphones earlier. Now, a low-priced tracker could cost as little as Rs 3,000 in India.

Here, it is important to make a distinction between a smartwatch and fitness tracker. Smartwatches (Apple Watch for example) have high-end specifications, and LED watch faces. These cost upwards of Rs 15,000-20,000. A fitness tracker band, on the other hand, typically has a monotone screen. In terms of functionality, the higher the price, the better the functionality. For instance, ECG and heart-rate readings are possible only in high-end smartwatches. Fitness bands typically offer step-tracking only.

Among the most sophisticated features, ECG first came to the Apple Watch in December 2018. The glass dial on the back of the watch face sports electrodes. To record readings, a user has to place the finger of the other hand (not the one on which the watch is worn) on the watch to complete a circuit. The same glass dial at the back of the watch also has LED and infra-red lights. These beams penetrate the wrist to detect blood flow in capillaries to ultimately give the pulse reading. Sleep activity tracking is done by a combination of movement and heart-rate.

“In India, 1 million smartwatches and 6 million fitness tracker bands were sold last year. The total installed base in the country is 15 million at present,” says Counterpoint Research Associate Director Tarun Pathak. “However, if you compare it with the number of smartphones out there – around 500 million – the penetration still seems quite low.”