Reliability and validity of bilateral thigh and foot accelerometry measures of walking in healthy and hemiparetic subjects.

OBJECTIVE Measures of walking ability in large clinical trials are usually limited to a timed short-distance walk and the distance walked in a fixed time. A new integrated system of 5 accelerometers was tested for reliability and compared to a footswitch system to determine if the accelerometers offered a practical option for the acquisition of spatiotemporal gait parameters. METHODS Leg accelerations and decelerations were defined in relation to simultaneous kinematic and electromyographic data acquired from a healthy subject. Eight healthy adults and 6 independent ambulators with hemiparetic stroke walked 15 m at 2 different speeds wearing both the accelerometers and footswitches. Twelve healthy subjects walked at 5 different speeds repeated 3 times on each of 2 days wearing the accelerometers. Walking speed, cadence, stride length, and single- and double-limb support, swing, and stance times were calculated. RESULTS No differences (t test, P > 0.2) were found between footswitch and accelerometer variables when comparing all left or right legs in healthy subjects and all paretic or unaffected legs in stroke subjects. A 2-way nested ANOVA model (speed, left and right legs, trial, and session) with the accelerometers at walking speeds from 0.5 to 1.8 m/s revealed high reproducibility of all measures. CONCLUSIONS The accelerometry system provided reliable and valid spatiotemporal measures of gait for the upper range of speeds likely to be targeted for rehabilitation interventions in ambulatory subjects.