Journal: Frontiers in Digital Health

Article Title: Wearable Sensor Systems for Fall Risk Assessment: A Review

doi: 10.3389/fdgth.2022.921506

Figure Lengend Snippet: Studies using inertial sensor-based fall-risk assessment and fall detection systems.

Article Snippet: Qiu et al. ( ) , 100 community-dwelling Korean older women (≥65 years) • 114 non-fallers • 82 fallers , • 5 Xsens inertial sensors−3D Accelerometer + 3D Gyroscope (pelvis, thighs, and shanks) , • Time and frequency domain features from sensory integration test (SIT) • Limits of Stability (LOS) • Sit-to-Stand Five Times (STS5) • Gait parameter from timed up and go (TUG) test: gait velocity, step time and length, tuning time • Range of motion (ROM): knee flexion; knee extension • Choice reaction test (CRT) • Computerized falls efficacy scale (FES) , • Logistic regression • Naïve Bayes (NB) • Decision tree (DT) • Random forest (RF) • Boosted tree (BT) • Support vector machine (SVM) , • Support vector machine for faller classification achieved the highest overall accuracy of 89.4% with 92.7% sensitivity and 84.9% specificity • fallers exhibited worse performances of visual and vestibular systems, a smaller knee ROM, a slower information processing speed, a higher fear of falling, and experienced more difficulties during the complex tasks such as TUG, LOS, and STS5.

Techniques: Functional Assay, Plasmid Preparation, Medications, Standard Deviation