Related Experiment Video
Updated: Oct 2, 2025
Design and Analysis for Fall Detection System Simplification
Published on: April 6, 2020
Evaluation of unsupervised 30-second chair stand test performance assessed by wearable sensors to predict fall status
Lindsey J Tulipani1, Brett Meyer1, Dakota Allen1
1M-Sense Research Group, Department of Electrical and Biomedical Engineering, University of Vermont, Burlington, VT, United States.
Background:
One in two people with multiple sclerosis (PwMS) will fall in a three-month period. Predicting which patients will fall remains a challenge for clinicians. Standardized functional assessments provide insight into balance deficits and fall risk but their use has been limited to supervised visits.
Research Question:
The study aim was to characterize unsupervised 30-second chair stand test (30CST) performance using accelerometer-derived metrics and assess its ability to classify fall status in PwMS compared to supervised 30CST.
Methods:
Thirty-seven PwMS (21 fallers) performed instrumented supervised and unsupervised 30CSTs with a single wearable sensor on the thigh. In unsupervised conditions, participants performed bi-hourly 30CSTs and rated their balance confidence and fatigue over 48-hours. ROC analysis was used to classify fall status for 30CST performance.
Results:
Non-fallers (p = 0.02) but not fallers (p = 0.23) differed in their average unsupervised 30CST performance (repetitions) compared to their supervised performance. The unsupervised maximum number of 30CST repetitions performed optimized ROC classification AUC (0.79), accuracy (78.4%) and specificity (90.0%) for fall status with an optimal cutoff of 17 repetitions.
Significance:
Brief durations of instrumented unsupervised monitoring as an adjunct to routine clinical assessments could improve the ability for predicting fall risk and fluctuations in functional mobility in PwMS.
