康复学报 (Aug 2024)
Reliability and Validity Study of A Portable Gait Detection System Based on Wearable Sensors
Abstract
ObjectiveTo verify the reliability and validity of a portable Right Gait & Posture® gait detection system based on wearable sensors.MethodsA total of 36 healthy volunteers were recruited in the Jiangsu Zhongshan Geriatric Rehabilitation Hospital from August to November 2023. Right Gait & Posture® system was used to detect the gait of each subject for three times by two trained assessors. The assessor 1 completed the first and second tests, and the assessor 2 completed the third test. All tests were completed on the same day. Gait parameters such as gait cycle, swing phase proportion, stance phase proportion, walking speed, stride frequency, stride length, foot deviation angle, ground contact angle, ground contact inward rotation angle and propulsion angle were recorded. The intragroup correlation coefficient (ICC) was used to calculate the inter-tester reliability and the relative reliability of the repeated tests, and the standard error of measurement (SEM) and the minimum detectable change (MDC) were calculated as the absolute reliability of the repeated tests. With video gait analysis as a control, Bland-Altman method was used to calculate the 95% consistency limit, and the validity of the Right Gait & Posture® gait detection system in detecting gait cycle, swing phase ratio, stance phase proportion and walking speed was analyzed.Results(1) Inter-tester reliability: the ICC values of gait cycle, swing phase proportion, stance phase proportion, walking speed, stride frequency, stride length, foot deviation angle, ground contact angle, ground contact inward rotation angle, forward propulsion angle and foot deviation angle between two assessors were 0.901 to 0.981. (2) Test-retest reliability: the ICC values of gait cycle, swing phase proportion, stance phase proportion, walking speed, stride frequency, stride length, ground contact angle, ground contact inward rotation angle and forward propulsion angle and foot deviation angle between the two tests of the same assessor were 0.822 to 0.983. The SEM and SEM% values of gait parameters were 0.02 to 2.17 and 0.49% to 13.83%, respectively. The MDC and MDC% values were 0.06 to 6.02 and 1.37% to 38.34%, respectively. (3) Validity analysis: Bland-Altman plot showed that the mean absolute errors and 95% CI of gait cycle, swing phase proportion, stance phase proportion and walking speed were 0.003 s, (-0.059, 0.064); 0.756%, (-3.048, 4.560); -0.756%, (-4.560, 3.048); -0.001 m/s, (-0.155, 0.152), respectively.ConclusionRight Gait & Posture® gait detection system shows good reliability in spatio-temporal parameters and ankle kinematics parameters, and has high validity in gait cycle, swing phase proportion, stance phase proportion and walking speed, which is recommended for clinical application.
