Remote Gait Analysis as a Proxy for Traditional Gait Laboratories: Utilizing Smartphones for Subject-Driven Gait Assessment across Differing Terrains

Abstract

Gait analysis has applications in medical diagnosis, biometrics, and development of therapeutic rehabilitation interventions (such as orthotics, prosthetics, and exoskeletons). While offering accurate measurements, gait laboratories are expensive, not scalable, and not easily accessible. In a pandemic-afflicted world, where telemedicine is crucial, there is need for subject-driven data remote collection. This study proposed a remote and purely subject-driven procedure for reproducible and scalable collection of real-life gait data. To evaluate the feasibility of our proposed procedure, the spatiotemporal parameters of gait were compared across two real-life terrains using a smartphone application on a focus population of healthy middle-aged individuals. Previous research validated smartphone motion sensors as accurate instruments for gait analysis, but required highly supervised, controlled environments on smaller sample sizes, thereby limiting application in real-life gait analysis. To this end, a custom-designed mobile application was developed to record lower extremity angular velocities on 69 healthy middle-aged adults; factoring in a subject-driven data submission error rate (DSER) of 17.4%, there were 57 usable data sets for analysis. Comparisons of spatiotemporal gait parameters across primary outcome measures on grass versus asphalt revealed significant measurable increases in gait duration (stride time), valley depth (max swing phase), and peak-to-valley (max stance phase to max swing phase). These results demonstrated the feasibility of using smartphones for a remote and fully subject-driven gait data collection. Additionally, our data analysis showed that even in short trials, a physical environmental load has a substantial and measurable effect on the gait of the understudied middle-aged population.