Validation of a Smartwatch in Persons With Multiple Sclerosis

Multiple sclerosis (MS) is a chronic inflammatory disease of the central nervous system. It is one of the most common neurological diseases in young adulthood. The main symptoms of MS are gait and mobility impairments, which also determine the severity of the disease. Currently, subjective clinical functional tests are commonly used to evaluate therapy outcomes. Due to advancing mechanization, body-worn accelerometers have been established as a potential instrument for the objective detection of mobility in clinical research. The accuracy of these devices for determining the mobility is therefore essential. However, current studies, especially in people with altered gait, show large deviations in step detection accuracy. In the present study we want to investigate: i) whether a commercially available smartwatch (Polar M600) counts steps in persons with multiple sclerosis (pwMS) accurately and ii) whether it counts steps as accurately as an accelerometer commonly used in clinical research (ActiGraph wGT3X-BT). In addition, the accuracy of the step detection by the algorithms of the manufacturer of the two devices (Polar M600 and ActiGraph wGT3X-BT) will be compared with an own-developed algorithm. For this purpose, 63 pwMS with low to severe functional restriction (EDSS: 0 - 6.5) will participate in this validation study. The validation protocol consists of three walking tests with different walking velocities. During these three conditions the participants will wear the Polar M600 on the wrist and the ActiGraph wGT3X-BT with a belt over the hip. In the first walking test, the subjects will complete 250 steps with comfortable walking velocity (CWV). After a sufficient break, the participants will be randomly assigned to either a fast walking velocity condition (FWV: CWV + 20%) followed by a sufficient period of rest and a slow walking velocity condition (SWV: CWV - 20%) or vice versa. Each of the walking tests with modified walking velocity also consists of 250 steps. To analyze the accuracy of step detection, the measured steps of the devices will be compared with the actually walked 250 steps.