Accuracy of wearable physical activity trackers in people with Parkinson's disease

The purpose of this study was to determine the accuracy of the Fitbit Charge HR™ and Garmin vívosmart^® HR in measuring steps and reflecting intensity of activity in people with Parkinson's disease (PD).

Thirty-three people with mild-moderate PD performed six, two-minute indoor walks at their self-selected walking pace, and at target cadences of 60, 80, 100, 120 and 140 beats/min. A 500 m outdoor walk with terrain challenges was also performed. Step count was recorded by the two wrist-worn activity trackers (Fitbit Charge HR™ and Garmin vívosmart^® HR) and compared to an accelerometer (ActivPAL³™). Intensity was recorded by a portable breath-by-breath gas analyser (VO₂), heart rate and Borg scale.

Both commercial activity trackers had low error (<3%) and moderate to high consistency at self-selected pace both indoors and outdoors (ICC 0.88-0.97; p < 0.05) compared to the ActivPAL³™. The Garmin recorded low error (<5%) and high agreement (ICCs > 0.68; p < 0.001) for all target cadences ≥80steps/min. The Fitbit had higher error was less consistent for all target cadences ≥80steps/min. Cadence measured by the Fitbit and Garmin weakly reflected increases in heart rate (ICCs 0.27-0.28; p < 0.05), and did not reflect VO₂ or Borg (ICCs 0.08-0.15, p > 0.05).

The Garmin device was more accurate at reflecting step count across a broader range of walking cadences than the Fitbit, but neither strongly reflected intensity of activity. While not intended to replace research grade devices, these wrist-worn devices may be a clinically useful adjunct to exercise therapy to increase physical activity in people with PD.