Commercial self-monitoring devices are becoming increasingly popular, and over the last decade, the use of self-monitoring technology has spread widely in both consumer and medical markets. The purpose of this study was to evaluate five commercially available self-monitoring devices for further testing in clinical applications. Four activity trackers and one sleep tracker were evaluated based on step count validity and heart rate validity.
Methods: The study enrolled 22 healthy volunteers in a walking test. Volunteers walked a 100 m track at 2 km/h and 3.5 km/h. Steps were measured by four activity trackers and compared to gyroscope readings. Two trackers were also tested on nine subjects by comparing pulse readings to Holter monitoring.
Results: The lowest average systematic error in the walking tests was -0.2%, recorded on the Garmin Vivofit 2 at 3.5 km/h; the highest error was the Fitbit Charge HR at 2 km/h with an error margin of 26.8%. Comparisons of pulse measurements from the Fitbit Charge HR revealed a margin error of -3.42% ± 7.99% compared to the electrocardiogram. The Beddit sleep tracker measured a systematic error of -3.27% ± 4.60%.
Conclusion: The measured results revealed the current functionality and limitations of the five self-tracking devices, and point towards a need for future research in this area.
Keywords: activity tracker; gait; heart rate; physical activity; pulse; slow walking; step detection.