Comparability of Daily-Life Walking Speed Measured by Smartphone GPS and Ankle-Band Accelerometer: Cross-Sectional...

Daily-life walking speed (DWS), which is a critical health indicator in older adults, can be measured using smartphone GPS technology. Although this method is becoming more widely accessible, it is restricted to outdoor walking. In contrast, accelerometers can capture walking speed indoors; however, there is limited information on the comparability of DWS data between these two techniques.

In this study, we aimed to investigate the agreement and systematic error between DWS measured using the built-in GPS of a smartphone and an ankle-band accelerometer.

Participants were recruited from a previously selected cohort of community-dwelling older adults. Their DWS was assessed using both a smartphone app and ankle band accelerometer. Data from the two devices were matched based on simultaneous recordings, and agreement between the walking speeds was evaluated using the intraclass correlation coefficient (ICC) and a Bland-Altman plot.

A total of 99 participants (38 men, 61 women; mean [SD] age 71.5 [4.9] y) were included, yielding 3652 paired data points. The mean (SD) DWS as measured by GPS was 1.30 (0.19) m/s, and this was significantly higher than the value of 1.12 (0.23) m/s measured by the accelerometer (P<.001). The ICC(2, 1) (95% CI) value was 0.523 (-0.022 to 0.765), and the ICC(2, k) value was 0.687 (-0.045 to 0.867). The Bland-Altman plot revealed a fixed error of 0.18 m/s with 95% limits of agreement ranging from -0.49 m/s to 0.13 m/s in the GPS-measured walking speed compared to that measured by the ankle-band accelerometer.

The GPS app consistently recorded a faster walking speed than the ankle-band accelerometer. The agreement between the measurements of the two devices was poor. The data suggest that a correction is necessary when comparing the DWS between these two devices.