Evaluating the Psychometrics of Accelerometer Data for Independent Monitoring of Task Repetitive Practice.

Individuals post-stroke commonly experience impairments in upper extremity function that limit participation in valued activities. Task repetitive practice is an effective intervention strategy, but accurately monitoring adherence and movement quality in home programs remains a challenge. This pilot study investigates the reliability and validity of raw accelerometer data captured by a commercially available, wrist-worn activity monitor to assess upper extremity movement in healthy adults during task repetitive practice. Measures of duration, angular velocity, and acceleration were obtained from activity monitors worn by 25 healthy adults performing four functional tasks under varying conditions. Preliminary results indicate moderate to excellent within-session reliability in these three measures when compared across repeated trials of the same task, with one exception. Across all tasks, the duration measure consistently detected differences in exercise time between sets of 5, 10, and 20 repetitions at a comfortable pace. All three measures differentiated between 10 comfortable repetitions and 10 fast repetitions on three out of four tasks. These findings provide initial psychometric properties in a healthy population and further research is required to determine whether these properties remain robust in the presence of motor impairment. This work represents an early step towards developing approaches for monitoring home exercise programs that support stroke recovery.