The reliability and validity of the OneStep smartphone application for gait analysis among patients with gait asymmetry undergoing rehabilitation for lower limb chronic pain or functional disability

Background: An affordable, easy-to-use, reliable, and valid tool is essential for gait assessment of people with various gait pathologies. This study aimed to assess the reliability and validity of the OneStep smartphone application (app) compared to the C-Mill VR + instrumented treadmill (Motek, Nederlands), among patients with gait asymmetry undergoing rehabilitation for lower limb disability or chronic pain. Methods Seventy patients participated in the study. The data of 35 patients were used to improve the app's algorithms, which were then validated on the remaining 35. Participants were asked to walk on a treadmill for up to 15 minutes at their own comfortable pace. Spatiotemporal gait parameters were extracted from the treadmill and from two smartphones placed on each leg. Reliability was evaluated using Pearson correlation, Cohen's d, and intra-cluster correlation coefficient (ICC), comparing the readings of the app from the two phones. Validity was assessed using the same measures, comparing readings from each phone to the data obtained from the treadmill. Bland-Altman plots were used to demonstrate the level of agreement between the app and the treadmill for each leg separately. Results Twenty-eight patients completed the validation study (seven were excluded); the median age was 45.5 years, 61% males. Walking sessions lasted on average 8.4 minutes. The ICC for consistency between the phones showed high correlation (r = 0.89 to 1) and good to excellent reliability (ICC range, 0.77 to 1) for all the gait parameters examined. The correlations between the phones and the treadmill were mostly above 0.8. The ICC for consistency demonstrated excellent validity for cadence and gait speed (range, 0.92 to 1), and moderate-to-excellent validity for all the other parameters (range, 0.58–0.93). Only step length of the impaired leg parameter showed poor-to-good validity (range, 0.37–0.84). Some subtle differences, though statistically significant, were found between the readings of the treadmill and the smartphone placed on the healthy leg. Cohen's d effect size was small (d < 0.5) for all the parameters. Conclusions The OneStep smartphone application demonstrated good reliability and validity for spatiotemporal gait assessment in people with gait asymmetry undergoing rehabilitation. These findings highlight the potential for smartphone-based gait analysis in a rehabilitation setting. Trial registration: NIH identifier: NCT05009303, date of registration: August 17, 2021.