Just be aware. Article came out yesterday on the glories of Fitbit for tracking flu-like symptoms for our public health agencies. No surprise.
Fitbit Data Significantly Improves Prediction of Influenza-like Illnesses
Dibash Das, PhD
Fitbit data significantly improved detection of influenza-like illness (ILI) in US patients.
Fitbit data significantly improved detection of influenza-like illness (ILI) in US patients, according to a study published in
The Lancet.
Acute infections can cause individuals to have higher-than-normal resting heart rates (RHR), increased sleep duration and frequency, and a decline in activity levels, all of which can be measured using wearable technology. Because an increasing number of people are using personal activity monitors, researchers assessed whether population trends data from one
wearable device, Fitbit, may help improve real-time influenza surveillance and limit the potential impact of the disease.
The investigators obtained de-identified heart rate and sleep activity data from 200,000 individuals who used a Fitbit from March 1, 2016 to March 1, 2018 in the United States. They used the data from the 5 states with the most device wearers in order to determine whether changes in an individual’s weekly RHR and sleep activity as determined by the Fitbit correlated with the influenza-like illness (ILI) rates reported by the Centers for Disease Control and Prevention at a statewide level. Data from the first year (March 2016 to March 2017) was used to build an analytic model and data from the second year (March 2017 to March 2018) was used for validation. Inclusion criteria included having a self-reported birth year between 1930 and 2004, weight greater than 20 kg, and height greater than 1 m. Overall, adequate data came from 47,249 users who wore a Fitbit consistently during the study period and included 13,342,651 total RHR and sleep measurements.
Fitbit data significantly improved ILI predictions in all 5 states, with a mean increase in Pearson correlation of 0.12±0.07 over baseline models, corresponding to improvements ranging from 6.3–32.9%. The correlation between the model predictions and CDC-reported ILI rates varied from 0.84 to 0.97 for the individual states. Week-to-week changes in the percentage of Fitbit users with abnormal data were associated with week-to-week changes in ILI rates in most cases. Fitbit users were classified as having a week with abnormal data if their weekly average for RHR or sleep measurements exceeded a given threshold, such as overall elevated RHR or longer-than-average sleep duration. Additionally, it was discovered that the percentage of participants with Fitbit data above the threshold was higher during the 2017-2018 influenza season compared with the 2016-2017 season.
The researchers noted several study limitations. First, external factors other than illness can influence an individual’s RHR and sleep. Secondly, owners of wearable devices are typically wealthier than the general population, potentially making them less likely to have comorbidities that could make them susceptible to more severe infections.
The researchers concluded that “the large amount of real-time data generated by Fitbit and other personal devices will continue to prove useful for public health and augment traditional surveillance systems. The ever-expanding big data revolution offers unique opportunities to mine new data streams, identify epidemiologically relevant patterns, and enrich infectious disease forecasts.”