A study by the Scripps Research Translational Institute published in The Lancet Digital Health in recent days illustrates how Fitbit users’ data can be used to track diseases.
The researchers obtained data, anonymously, of 200,000 people who used a Fitbit wearable device from March 1, 2016 to March 1, 2018 in the United States.
The researchers included users who worn a Fitbit for at least 60 days and used it throughout the trial period, residing in five states: California, Texas, New York, Illinois and Pennsylvania.
Inclusion criteria included having a self-reported birth year between 1930 and 2004, height greater than 1 m, and weight greater than 20 kg. They excluded daily measurements with missing RHR, missing wear time, and wear time less than 1000 min per day.
The researchers were able to count on a data set of 47,249 patients with over 13 million measurements of resting heart rate (RHR) and sleep duration.
They then compared sensor data with weekly estimates of flu-like illness rates (ILI) at the state level provided by the U.S. Centers for Disease Control and Prevention (CDC), identifying weeks when Fitbit users showed high RHR and increased sleep levels.
For each state, the researchers modeled the ILI case counts with a negative binomial model that included the 3-week CDC ILI delay data (null model) and the proportion of weekly Fitbit users with high RHR and longer sleep duration above a certain threshold (full model). The researchers also evaluated the weekly change in the ILI rate by linear regression using the change in the proportion of high Fitbit data. The Pearson correlation was used to compare the expected ILI rates against those reported by CDC.
The researchers found that Fitbit data significantly improved ILI predictions in all five states, with an average increase in Pearson correlation of 0-12 (SD 0.07) compared to baseline models, corresponding to an improvement of 6.3 – 32.9%. The correlations of the final models with ILI CDC rates ranged from 0.84 to 0.97.
Wearable devices are increasingly being used in the United States and around the world to monitor individual health. By accessing this data, according to researchers, it may be possible to improve flu surveillance in real-time with great geographical accuracy. This information could be vital to implement timely epidemic response measures to prevent further transmission of influenza cases during outbreaks.