Category: Measurement; Technology (e.g. robotics, assistive technology, mHealth); Clinical Practice (assessment, diagnosis, treatment, knowledge translation/EBP, implementation science, program development)
Objective : to verify 1) adaptive gait recognition algorithm is available to estimate basic gait parameters by comparing those results to GAITRite system and 2) evaluate the validity of spatiotemporal variables
Design : A cross-sectional design
Setting : The study was performed in a neurological physical therapy laboratory
Participants (or Animals, Specimens, Cadavers) : 14 healthy subjects were included in this study(age: 28.1± 7.2 years, men/mowen: 5/9). They were asked to complete normal walking trials along a 5-m walkway while being detected by both Kinect and the standard instrument (GAITRite).
Interventions : Not applicable
Main Outcome Measure(s) : 10 common gait outcomes of cadence, speed, stride length, step length, gait cycle, step time, stance phase time, swing phase time, single support times, and double support time were derived using a developing algorithm. The correlation of these measurements between kinect based system programed by favorable algorithm and GAITRite was obtained.
Results : Significantly high correlation(r>0.75, p<0.001) was observed in most variables(6/10) with average stride length reached largest point among all(r=0.855, p<0.001). However, cadence, speed, step time, double support time only showed modest correlation level(r= 0.57-0.744, p<0.05). Generally, the finding of validity was typically good to excellent.
Conclusions : We here presented an adaptive algorithm could reasonably support kinect based system and determine the general gait measurements that were acceptable and accurate enough for assessing spatiotemporal aspects of gait. Furthermore, our results revealed the overall good validity in most gait parameters which indicate the kinect based system are well-suited, proper, and useful device likely introduced to clinical context. Future works are needed to consider analysis of more populations, especially for the people with certain disease or abnormal gait pattern and include other implementation of algorithm for data comparisons.
MengChe Shih– product manager, Longgood meditech, Taipei, Taipei
JunHong Zhou– Student, Department of Physical Therapy and Assistive Technology, National Yang-Ming University, Taipei, 11221, Taiwan, Taipei, Taipei
Yea-Ru Yang– Professor, National Yang-Ming University, Taipei, Taipei
ChihJui Chen– CEO, Longgood meditech, Taipei, Taipei