Test-retest reliability of daily life gait speed as measured by smartphone global positioning system

Background: Gait speed is useful in predicting adverse health outcomes among older adults. In previous studies, gait speed has typically been measured when subjects walk in laboratory settings, where they are able to intentionally change their gait speed. Thus, it is unclear whether the gait speed captured in a laboratory setting is representative of the subjects' actual walking pace in daily life.

Research question: This study proposes using the more accurate "daily life gait speed" (DGS), measured as the subject's average gait speed over a week-long period using the global positioning system (GPS) in their smartphone. We examined the test-retest reliability of the DGS measure in the present study.

Methods: Three daily life gait parameters with 186 volunteers (57 men and 129 women), aged 19 to 84 years, were measured using a smartphone application: DGS, average of daily gait cycle during a week (DCY), and average of daily cadence during a week (DCA). Test-retest reliability of the daily gait parameters between test week (T1) and retest week (T2) was assessed with the intraclass correlation coefficient, ICC (2,1), and systematic biases were observed via Bland-Altman plots.

Results: The ICCs between the daily gait parameters at T1 and T2 were 0.902 for DGS, 0.916 for DCY, and 0.917 for DCA. The Bland-Altman plots showed no significant fixed or proportional bias between the measurements at T1 and T2.

Significance: These results verify that the test-retest reliability of the daily gait parameters in the present study was adequate.