In research published January 16, 2020, in The Lancet Digital Health, scientists demonstrated that information on resting heart rate and sleep patterns captured by wearable devices such as smartwatches and fitness trackers correlated closely with flu case numbers estimated by the Centers for Disease Control and Prevention (CDC). Investigators from Scripps Research Translational Institute (SRTI) in La Jolla, California, analyzed data from 47,249 individuals in five states who wore a Fitbit consistently. The researchers’ models to predict influenza incidence not only corresponded with CDC data, but when the Fitbit information was incorporated into flu forecast calculations, it improved accuracy by 6 to 33 percent. “If this tool is able to help detect the spread of flu early, then we may be able to do more about it,” says study author Steven Steinhubl, MD, a cardiologist and director of digital medicine at Scripps Research. “From an epidemiological standpoint, having real-time data can allow you to actually minimize the spread. We can treat people, keep people home, and give them advice on how to prevent the illness from getting worse or infecting others.”
Taking Advantage of Daily Data
When someone has an infectious disease like the flu, resting heart rates tend to spike and sleep duration tends to be longer than normal. Because these physiological factors are continuously measured by Fitbit devices, lead investigator Jennifer Radin, PhD, a senior staff scientist at Scripps, hypothesized that real-time measures of heart rates and sleep might enhance influenza surveillance. Dr. Radin, along with Dr. Steinhubl and colleagues, examined data from more than 47,000 people in California, Illinois, New York, Pennsylvania, and Texas who used a Fitbit wearable device for at least 60 days during the study. The average user was age 43, and 60 percent were female. Details on user identities were all kept private, and the Fitbit Privacy Policy notified individuals that their de-identified information could potentially be used for scientific study. The information captured during the two-year period from 2016 to 2018 included more than 13.3 million total measures of resting heart rate and sleep. The investigators compared weekly averages from the sensors with weekly estimates of influenza-like illness rates at the state level, as reported by the CDC. Correlation measures were high, ranging from 0.84 to 0.97. (A correlation coefficient of 1 means that there is a strong correlation, while a value of 0 means there is none.) “This is the first hint of healthcare of the future, of truly keeping people healthy by the ability to track individual changes in vital signs,” says Steinhubl.
A Potentially Powerful Predictive Tool
Every year, 9 million to 45 million Americans get sick with the flu, according to the CDC. The illness is much more serious than the common cold, resulting in hundreds of thousands of hospitalizations and tens of thousands of deaths. The virus is highly contagious, so when public health officials see flu cases rising, they raise the alarm and do what they can to try to slow its spread. Traditional flu tracking has been problematic because surveillance has been based on reports from doctors, which can take weeks to process. The study authors see an advantage using data from wearable fitness devices because they give information immediately. The population of Fitbit fans has been rapidly expanding. The company estimates that it has about 28 million active users. The number of connected wearable devices worldwide is expected to jump from 526 million in 2016 to 1.1 billion by 2022, according to a report in Statista. “We anticipate that the large amount of real-time data generated by Fitbit and other personal devices will prove highly useful for public health and augment traditional surveillance systems,” wrote Cecile Viboud, PhD, a mathematical epidemiologist at the Fogarty International Center at the National Institutes of Health in Bethesda, Maryland, and Mauricio Santillana, PhD, director of the Machine Intelligence Lab at Harvard University in Boston, in an accompanying editorial. Michael Snyder, MD, director of genomics and personalized medicine at Stanford University in Palo Alto, California, also sees great potential for using real-time data from fitness devices to help track other illnesses. “The idea that we can follow infectious disease spread by simple smartwatches is very exciting,” says Dr. Snyder, who was not involved in the research. “The investigation follows a study we published three years ago [in PLoS Biology], where I discovered my Lyme disease and other respiratory illness from my smartwatch.” Snyder would like to see more research on how individuals might use these devices for personal diagnoses.
Prospect of Measuring Other Vital Signs
The researchers see great potential in using wearables to identify health abnormalities by including more sensors to potentially provide blood pressure, temperature, electrocardiogram, and cough analysis. “The general idea of having day-to-day measures of vital signs instead of just measures when at the doctor’s or hospital can help with infectious diseases and may help with chronic conditions like diabetes, hypertension, and even mental health issues if we get to the point where we can measure stress,” says Steinhubl. He underscores that this is an early study and more research needs to be done. “We’re guessing that heart rate changes are related to flu, but our results don’t prove that flu caused heart rates to increase in these people,” says Steinhubl. Other factors may lead to an increase in resting heart rate, including stress or other infections. Because Fitbit users were predominantly middle-aged adults and likely to have higher income than average, future research may include a broader population. Despite study limitations, the promise of wearables to change the landscape of healthcare is there. “Physiological and behavioral data from a growing number of wearable device users globally could greatly improve timeliness and precision of public health responses and even inform individual clinical care,” study authors conclude.
