Criterion validity and accuracy of actigraph and Xiaomi Smart Band to count steps after stroke.

Abstract

Finding valid and accurate methods to count steps is important after a stroke. To determine the criterion validity and accuracy of the Actigraph and Xiaomi Smart Band to count steps in individuals after stroke. To explore the influence of walking speed on the accuracy of these activity monitors. Individuals after stroke performed the 2-min Walk Test while wearing the ActiGraph and Xiaomi Smart Band. The 2-min Walk Test was recorded, and the observed step count was compared with the estimated step count of both activity monitors. Fifty individuals after stroke were included. The ActiGraph had a low correlation (rho 0.41) between observed and estimated step count and a high percentage absolute error (39%). The Xiaomi Smart Band had a high correlation (rho 0.83) and a moderate percentage absolute error (12%). At high walking speeds, the Xiaomi Smart Band had a low percentage absolute error (3%) and at slow walking speeds had 18% a moderate percentage absolute error. The Xiaomi Smart Band has good criterion validity and moderate-to-high accuracy counting steps after stroke. It is a low-cost activity monitor that is a viable alternative to the more expensive activity monitors used in research.IMPLICATIONS FOR REHABILITATIONThe ActiGraph has low criterion validity and accuracy at counting steps after stroke.The Xiaomi Smart Band has good criterion validity and accuracy at counting steps after a stroke.The Xiaomi Smart Band is a low-cost activity monitor that is a viable alternative to the more expensive activity monitors used in research.