Determining the Reliability of Several Consumer-Based Physical Activity Monitors

Abstract

Limited research exists on the reliability of consumer-based physical activity monitors (CPAMs) despite numerous studies on their validity. Consumers often purchase CPAMs to assess their physical activity (PA) habits over time, emphasizing CPAM reliability more so than their validity; therefore, the purpose of this study was to investigate the reliability of several CPAMs. In this study, 30 participants wore a pair of four CPAM models (Fitbit One, Zip, Flex, and Jawbone Up24) for a total of eight monitors, while completing seven activities in the laboratory. Activities were completed in two consecutive five-minute bouts. Participants then wore either all wrist- or hip-mounted CPAMs in a free-living setting for the remainder of the day. Intra-monitor reliability for steps (0.88–0.99) was higher than kcals (0.77–0.94), and was higher for hip-worn CPAMs than for wrist-worn CPAMs (p < 0.001 for both). Inter-monitor reliability in the laboratory for steps (0.81–0.99) was higher than kcals (0.64–0.91) and higher for hip-worn CPAMs than for wrist-worn CPAMs (p < 0.001 for both). Free-living correlations were 0.61–0.98, 0.35–0.96, and 0.97–0.98 for steps, kcals, and active minutes, respectively. These findings illustrate that all CPAMs assessed yield reliable estimations of PA. Additionally, all CPAMs tested can provide reliable estimations of physical activity within the laboratory but appear less reliable in a free-living setting.