Repeatability of the accelerometric-based method to detect step events for hemiparetic stroke patients

This study is to evaluate the repeatability of the accelerometric-method to detect step events for hemiparetic stroke patients. To evaluate this method, four adults with chronic hemiparetic stroke were participated. The repeatability of this method using a single three-axis accelerometer was evaluated with a six optical camera motion capture system. The correlation statistics and Bland-Altman plot were then used to evaluate the agreement between the step-time differences from the accelerometer data and the reflective markers data. The correlation coefficient of each two data was 0.99 (p <; 0.001) and retest result was 0.99 (p <; 0.001). The mean ± standard deviation (SD) between each two data along with the 95% limits of agreement (LOA = ±1.96 SD) was 2.58±2.37 ms (LOA = -2.07 ms and 7.23 ms), and retest result was 3.73±2.02 ms (LOA = -0.22 ms and 7.68 ms). These results show that the suggested method is useful to detect step events for hemiparetic stroke patients.