Accuracy of consumer-based activity trackers as measuring tool and coaching device in patients with COPD and healthy...

Consumer-based activity trackers are used to measure and improve physical activity. However, the accuracy of these devices as clinical endpoint or coaching tool is unclear. We investigated the use of two activity trackers as measuring and coaching tool in patients with Chronic Obstructive Pulmonary Disease (COPD) and healthy age-matched controls.

Daily steps were measured by two consumer-based activity trackers (Fitbit Zip, worn at the hip and Fitbit Alta, worn at the wrist) and a validated activity monitor (Dynaport Movemonitor) in 28 patients with COPD and 14 healthy age-matched controls for 14 consecutive days. To investigate the accuracy of the activity trackers as a clinical endpoint, mean step count per patient were compared with the reference activity monitor and agreement was investigated by Bland-Altman plots. To evaluate the accuracy of activity trackers as coaching tool, day-by-day differences within patients were calculated for all three devices. Additionally, consistency of ranking daily steps between the activity trackers and accelerometer was investigated by Kendall correlation coefficient.

As a measuring tool, the hip worn activity tracker significantly underestimates daily step count in patients with COPD as compared to DAM (mean±SD Δ-1112±872 steps/day; p<0.0001). This underestimation is less prominent in healthy subjects (p = 0.21). The wrist worn activity tracker showed a non-significant overestimation of step count (p = 0.13) in patients with COPD, and a significant overestimation of daily steps in healthy controls (mean±SD Δ+1907±2147 steps/day; p = 0.006). As a coaching tool, both hip and wrist worn activity tracker were able to pick up the day-by-day variability as measured by Dynaport (consistency of ranking resp. r = 0.80; r = 0.68 in COPD).

Although the accuracy of hip worn consumer-based activity trackers in patients with COPD and wrist worn activity trackers in healthy subjects as clinical endpoints is unsatisfactory, these devices are valid to use as a coaching tool.