Walking Quality During Inpatient Stroke Rehabilitation Assessed by Wireless Sensing (P5.176)

OBJECTIVE: Demonstrate that the quality of gait, in addition to walking speed and distance, can be derived from inertial sensors worn at both ankles. BACKGROUND: The international Stroke Inpatient Rehabilitation Reinforcement of ACTivity (SIRRACT) trial deployed wireless sensors and activity-recognition algorithms to monitor and provide feedback about the quantity of stroke patients’ daily walking activity (epub, Neurorehabil Neural Repair, 2014). This follow-up study aimed to characterize the quality of walking by calculating spatiotemporal gait metrics. DESIGN/METHODS: In SIRRACT, 135 trial participants wore sensors throughout the average 3-week inpatient rehabilitation admission. Stopwatch-timed 10-meter walks (performed weekly for system calibration) were selected for the current analysis. Individual steps were delimited by identifying heel strike and toe-off times. Metrics including gait cycle duration, symmetry of stance and swing times, and double-limb support time were calculated for both the paretic and non-paretic limbs. The magnitude of peak lower leg acceleration during the swing phase of gait was also calculated. RESULTS: Data from five trial participants are presented as exemplars. Walking speeds averaged 0.24 ± 0.14m/s on study entry and 0.81 ± 0.35m/s at the time of discharge. Gait cycle duration decreased during rehabilitation (median:334ms [211,853]), as did double-limb support time (median:188ms [72,681]). Stance and swing time asymmetries were present throughout rehabilitation. Peak swing-time acceleration of the paretic leg was greater at discharge (median:0.20g-units [0.15,0.31]). CONCLUSIONS: The laboratory-quality gait metrics calculated by our wireless sensor system were sensitive to functional improvements during a period of known clinical recovery. Commercial sensor systems, for which steps counts are inaccurate at the slow speeds typical of hemi-paretic walking, have difficulty producing outcomes related to motor control in persons disabled by neurologic disease. Measurement of the quantity and quality of movements performed during daily activities enables clinicians and researchers to supervise gait training and skills practice during rehabilitation. Disclosure: Dr. Dorsch has nothing to disclose. Dr. Thomas has nothing to disclose. Dr. Dobkin has nothing to disclose.