Numerous research organizations set up studies asking users to submit data from their wearable fitness trackers as soon as the coronavirus outbreak began. Most gadgets only needed a few clicks to sign up, and users did it with enthusiasm. More than 500,000 people participated in the biggest study, the Corona Data Donation initiative run by the Robert Koch Institute in Germany. The California-based Scripps Research Institute’s study, detect, drew more than 30,000 participants.
The most helpful biomarker for disease surveillance is fever, which is an instantaneous indicator of infection. But since exact readings are challenging to obtain, the majority of wearables do not measure temperature. So it was necessary to develop a proxy by using the common metrics they do use, such as heart rate, sleep, and activity level. Any heart rate between 50 to 100 beats per minute is considered normal when people are sitting static and their resting heart rate is measured. However, each person’s rate is typically steady. But the rate increases, often dramatically, as the body battles an infection. In the case of COVID-19, data from wearable technology revealed that this increase occurred four days before participants began to experience symptoms. According to an estimate, 63% of coronavirus cases could be identified from variations in resting heart rate before the onset of symptoms.
Prior to the coronavirus outbreak, a team from Scripps led by Jennifer Radin had demonstrated that, in the US, weekly changes in the ratio of people with abnormal heart rate, sleep, and activity results—all measured from wearables—align well with the prevalence of flu-like symptoms as measured by specified surveillance systems. These monitor flu epidemics by calling multiple medical offices to see whether more patients are beginning to show these symptoms. By the time these statistics are compiled, an epidemic is typically at a different stage and may require new public-health actions because people typically seek medical attention 3–8 days after symptoms start to show up. As a result of which more pertinent insights are required.
However, data from wearables has its own peculiarities. The Koch Institute team was developing a measurement based on step count and heart rate as a precursor for fever when one day they saw a sudden peak in the measurement. It was found that Apple had modified the formula used to determine resting heart rate on their devices. The team struggled with these software updates because their data originates from about a dozen different devices. They must also fill in the various gaps. As the Apple Watches are mostly charged at night, the sleep data won’t be provided. But after reaching beyond its early hardships, the initiative turned out to be a success.
Dirk Brockmann, who leads the team, said, “It is not 100% accurate but it does a pretty good job.”
With wearable technology, some research teams have opted for an approach toward population-based surveillance in a different way. Based on the data their specific device gathers, they have created algorithms that look for variances in each person’s parameters. They establish the subject’s initial levels of several biomarkers and then look for any changes that might point to a physiologic anomaly. It is logical to assume that many individuals are becoming unwell, most likely for the same reason, when a lot of these changes appear all of a sudden, regardless of how diverse they may be from one individual to another.
Leo Wolansky from the Rockefeller Foundation’s Pandemic Prevention Institute, during an interaction with the Economist, stated that researchers must now determine if the disease-surveillance algorithms based on wearable technology might routinely overlook what is transpiring with certain types of people.
For instance, algorithms might unintentionally be optimized to detect outbreaks in affluent locations where people are more likely to have been using premium wearables for a long time duration. The algorithm for wearables may be far more likely to miss an outbreak in poorer areas where residents may have a variety of underlying health issues.
Wolansky said, “As they often say in this field, ‘Garbage in, garbage out, and we still have to better understand whether the data we’ve captured has some garbage in it.”
Similar things happened during the widespread scanning of human bodies made possible by the data derived from wearables. The German team determined that resting heart rate was lower in areas that had been in West Germany than those in East Germany.
“We still don’t know why this is. Is it because women work more in East Germany? Or is it because people eat differently?” Brockmann questioned.
Germans in all regions of the country are sleeping less in 2022 than they did in 2020, and the resting heart rate of the country has risen. Nobody really knows for sure why this happened, but one theory is that this may be related to the extra pounds that people gained during the multiple lockdowns. According to Mr. Brockmann, the information from wearables has served as “a question generator,” prompting inquiries about health that otherwise would not have been asked.
The way clinical investigations of new medicines are conducted is also evolving as a result of the ability to observe numerous human bodies as they go about their regular lives. A research company called iqvia estimates that in 2020, 10 percent of late-stage clinical studies will monitor participants with linked devices, up from 3 percent in 2016. More than 300 examples of digital biomarkers used in trials are listed in a catalogue by the US organisation the Digital Medicine Society.
For example, step count is a defined outcome in treatment trials for cystic fibrosis, Parkinson’s disease, asthma, arthritis, and heart failure. Instead of asking individuals to rate a medicine on a scale, measuring how much a person walks can offer a more accurate, or at least complementary, picture of the drug’s influence on pain or mood.
First and foremost, medical experts have been able to see patients’ reactions to a particular condition and treatment for the first time thanks to tools that quietly watch over patients as they go about their daily lives. In the sleep lab of a pharmaceutical business, nobody gets adequate sleep. The “six-minute walk test,” which measures how far a person can walk in six minutes, is the most popular test of cardiovascular and physical fitness. It involves a patient walking up and down a hospital hallway as a nurse notes down the outcome on a clipboard.
Fitness trackers, some of which have included the six-minute test in their library of movement measurements, have made this easier to understand. For instance, an Apple Watch calculates its estimates based on a variety of metrics from its sensors that are passively observed over time during a user’s typical behavior. This algorithmic assessment is highly accurate, according to validation tests in individuals over 65.
The individuals may be able to select treatments that best fit their priorities if the patient’s quality of life is included in drug trials. New cancer drugs are currently deemed successful even if they only add a few more months to patients’ lives. Though many cancer patients would prefer to live a little longer, they are much more concerned with what they can do in the months that they do survive cancer.
They would select a treatment that may offer fewer additional days but a higher likelihood that they would be according to Andy Coravos of HumanFirst, a company that assists pharmaceutical companies in deploying connected devices for monitoring trial participants at home, wearable sensors have also allowed patients who would not otherwise be eligible for clinical trials to participate in them. She uses the muscle-wasting disease Duchenne muscular dystrophy as an example. A six-minute walk test and a four-stair climb test are the primary outcomes for medications created to treat the disease. But because 60 per cent of the patients are wheelchair-bound, they are unable to participate. What the treatments can accomplish for them is thus unknown. The ability to include them in trials is made possible via an armband that tracks upper-body mobility.
More data from fitness trackers are being used in academic research on non-drug therapies, such as behavioral cues to enhance physical activity, rather than participants being asked to keep diaries or answer questionnaires. According to one survey of clinical trials that were registered in the US, the number of trials that used linked devices increased from 88 in 2007 to over 1,100 in 2017. The majority of those trials have not been conducted by pharmaceutical firms, but rather by research organizations, such as the precision medicine-focused team at Stanford University led by Euan Ashley.
In 2019, Dr. Ashley’s team was one of the first to conduct a fully digital trial in which participants never really interacted with a researcher in person. He claims that not very long ago, finding trial volunteers meant hanging posters with tear-off pieces of paper that included a phone number for them to contact. After that, they would have to visit the hospital and meet with a nurse to sign 17 pages of consent documents.
He said, “If you could get 200 people in a few months, you’d be pretty happy.”
Individuals can now download the study app and enroll while standing in line for coffee. When Dr. Ashley’s team initially applied this technique to a study on physical activity, 40,000 people signed up in just two weeks, and the findings were available in a few months. Even though it was very simple to enroll in the study, it was also very simple to drop out, and by the time it was over, only two weeks in, almost 80 percent of people had done so. However, the final group was nearly ten times larger than the usual size for this line of research.
According to the findings of this study, wearable health and fitness trackers can alter how people try to prevent illness and stay healthy, how their doctors treat them, and how population-level health interventions are implemented.
The digital health care that wearables enable may improve the effectiveness, efficiency, and personalization of therapy. Many Americans who might not otherwise receive any therapy at all employ digital therapies.
An AI therapist might not always provide mental health care that is as effective as that provided by a person. But those who cannot afford the cost or time off to see a doctor, or in areas where there is a shortage of mental-health specialists, may obtain it far more easily.
Patients with chronic diseases, who make up the majority of healthcare consumers, can benefit substantially from automated, round-the-clock monitoring of their ailments and results. When done correctly, it can also enable clinicians to serve more patients without becoming overburdened. In developing nations with a shortage of specialists, this type of care can significantly improve outcomes.