Accuracy of Heart Rate Monitoring by Some Wrist-Worn Activity Trackers

TO THE EDITOR: Cadmus-Bertram and colleagues' study (1) suggests that some wrist-worn activity trackers perform similarly to electrocardiography at rest but that none was as accurate as electrocardiography during moderate exercise (1). We agree that additional research is needed before these monitoring devices can be deemed useful for clinical advice or trials. In fact, most consumers are probably unaware of potential inaccuracies and may experience health risks related to under- or overexercise. On the basis of our recent experiment for an Advanced Placement high school research project (Singh J. Unpublished data.), we would like to highlight the additional need for policy changes to raise consumer awareness related to health applications on the market. To evaluate the accuracy of exercise-related applications, we compared data on measurements of steps and heart rate collected from a Fitbit with those manually collected by a practicing physician. Participants included 8 persons of varying heights, weights, ages, and physical fitness levels. To determine the accuracy of the devices, we compared the number of steps and heart rate measured by the 2 methods. We measured the heart rate at baseline (resting), after briskly walking on a 300-m track (postwalk), and after running on a 300-m track (postrun). We also documented steps taken during the walk and run. We found more than a 5% difference between data recorded by the Fitbit and those manually recorded for most variables, namely, baseline heart rate in 5 participants, postwalk and postrun heart rate in all 8 participants, steps while walking in 5 participants, and steps while running in 7 participants. Two situations in which participants' postrun heart rate was greater than 140 beats/min but the Fitbit failed to provide a reading were perhaps most worrisome. The heart rate data appeared only after this value decreased to less than 140 beats/min. Given these inconsistencies, especially during intense exercise when data simply disappeared, we are concerned about reassuring messages to consumers. One example is the following CNN headline from 20 April 2017 that quotes Cadmus-Bertram and colleagues' study: Fitness Trackers' Heart Rate Monitoring Accurate Enough for Most, Study Says. We believe that, beyond clinicians and researchers, consumers also need better awareness and transparency about these inaccuracies. Although we were fortunate to obtain such meaningful data on a high school research project, our society needs additional channels for ensuring rigorous device oversighta need that is not currently addressed (2). Analysis of such postmarketing surveillance data could be useful for device companies to help them recalibrate these applications to make them more accurate and reliable.