Date Published: July 5, 2017
Publisher: Public Library of Science
Author(s): Omar Aziz, Jochen Klenk, Lars Schwickert, Lorenzo Chiari, Clemens Becker, Edward J. Park, Greg Mori, Stephen N. Robinovitch, Yih-Kuen Jan.
Falls are a major cause of injuries and deaths in older adults. Even when no injury occurs, about half of all older adults who fall are unable to get up without assistance. The extended period of lying on the floor often leads to medical complications, including muscle damage, dehydration, anxiety and fear of falling. Wearable sensor systems incorporating accelerometers and/or gyroscopes are designed to prevent long lies by automatically detecting and alerting care providers to the occurrence of a fall. Research groups have reported up to 100% accuracy in detecting falls in experimental settings. However, there is a lack of studies examining accuracy in the real-world setting. In this study, we examined the accuracy of a fall detection system based on real-world fall and non-fall data sets. Five young adults and 19 older adults went about their daily activities while wearing tri-axial accelerometers. Older adults experienced 10 unanticipated falls during the data collection. Approximately 400 hours of activities of daily living were recorded. We employed a machine learning algorithm, Support Vector Machine (SVM) classifier, to identify falls and non-fall events. We found that our system was able to detect 8 out of the 10 falls in older adults using signals from a single accelerometer (waist or sternum). Furthermore, our system did not report any false alarm during approximately 28.5 hours of recorded data from young adults. However, with older adults, the false positive rate among individuals ranged from 0 to 0.3 false alarms per hour. While our system showed higher fall detection and substantially lower false positive rate than the existing fall detection systems, there is a need for continuous efforts to collect real-world data within the target population to perform fall validation studies for fall detection systems on bigger real-world fall and non-fall datasets.
Falls are the number one cause of injuries and injury-related deaths in older adults [1, 2]. About half of older adults who fall are unable to get up without assistance, even when no injury occurs [3, 4]. The ensuing “long lie” on the floor often leads to dehydration, muscle damage, and fear of future falls [5, 6].
Data analysis focused on determining the accuracy of our SVM classifier, trained on laboratory data, in distinguishing falls from normal ADLs based on real-world fall and non-fall data.
Our fall detection system showed perfect accuracy with all five young participants (Table 1). No false alarms were generated in the 28 hours and 38 minutes of real-world recorded data. All three sensor-locations (head, sternum and waist) provided 100% specificity.
In the current study, we examined the accuracy of an accelerometer-based automatic fall detection algorithm in distinguishing falls in older adults. The algorithm incorporated an SVM machine learning algorithm that was trained using laboratory-based falls and non-fall data, and tested with sensor data acquired from real-world falls and during daily activities by older adults. We found that, with 3D acceleration data from a single sensor, our algorithm showed 80% sensitivity (8 out of 10 real-world falls were successfully detected) and 99.9% specificity (false positive rates from 0.05 to 0.15 false alarms per hour depending on the older adult dataset of approximately 214 hours and 172 hours of ADLs). This is comparable to, if not better than the best performing algorithm reported by Bagalà et al.  which showed 83% sensitivity (23 out of 29 falls were successfully detected) and 97% specificity (false positive rate of 0.21 false alarms per hour on the older adult dataset of approximately 24 hours of ADLs).