Reliability of Smartphone Accelerometers for Measuring Gait During Data Collection Over Zoom

This study examined whether gait data could be reliably collected by homebound participants using iPhones under online supervision. Eighteen healthy young adults met with investigators through Zoom and installed an app to record acceleration from their iPhones' accelerometers. Half of the subjects walked normally; the other half walked while spelling words backward. During the gait tasks subjects recorded their anterior-posterior (AP), medial-lateral (ML), and vertical (V) accelerations. Data collection was repeated the following week. Seven maximum and minimum peak accelerations in the AP, ML, and vertical directions associated with events in gait were determined. Significant main effects of week and direction were observed for the first and second vertical acceleration measures. Cronbach alpha values were >0.60 for all acceleration measures, but the maximum and minimum AP accelerations that showed fair to good levels of consistency. The findings suggest gait data collected inside the home setting may be of clinical use.