Fall prediction of elderly with logistics regression model based on temporal variables by timed up and go test

The quantitative assessments of balance ability with fall risk by functional mobility testing are important. The representative test is Timed up and go (TUG) test. A previous study showed that assessing the TUG test by total duration has limitations in screening faller. Therefore, the TUG test is performed with an inertial measurement unit (IMU) sensor. In this study, the logistics regression analysis was performed to derive the characteristics of the faller through the temporal variables at the TUG test. The regression analysis results include ‘Sit-to-stand duration’ and ‘Turn-to-sit duration’ variables. The short duration of Turn-to-sit has a high risk of fall. Contrary, the risk of falls increases with Sit-to-stand duration. It is necessary to consider not only the total duration but also temporal variables in the separated phases.