Contactless fall detection with radio frequency sensors for the aging population

Abstract

AbstractBackgroundFalls remain a primary concern for People with Dementia (PwD) living in long‐term care facilities (LTCFs) despite many prevention and intervention methods that have tried. Current fall detection monitoring systems use wearable or vision‐based sensors. The problem is research has shown that many seniors will not consistently wear a device on their body or allow cameras in their living space because of privacy concerns that severely limit the technology user acceptance. The current standard of care is self‐report or eyewitness and therefore fall events are often undetected and underreported due to human error. Hence, we are proposing a solution to address this unmet need with a novel fall management system engineered with radio frequency (RF) sensors that do not require an individual to wear or interact with a device.MethodSix healthy participants simulated three types of falls, medial and lateral, anterior and posterior, and collapse in the Dementia and Elderly Care Robotics and Sensing (DECRS) Lab, University of Minnesota Duluth. The sensor was installed in the lab as it would be in the resident room of an LCTF. The simulated room included a bed, chair, and mattress on the floor to protect the participants.ResultOur study found that the contactless fall detection and activity system identified falls at overall sensitivity rate of 89% with a specificity of 100% in 748 tests and activity was identified at 90%. Our system identified “medial and lateral” falls with a 92% sensitivity, “anterior and posterior” falls with a 91% sensitivity, and “collapse” falls with a 75% sensitivity.ConclusionContactless sensing holds tremendous potential for monitoring individuals living with AD/ADRD and the elderly population for aging in place. LTCFs that are short staffed could benefit by utilizing tools that augment the capabilities of nurses with a continuous collection of patient‐generated health data (PGHD) and a notification system that is integrated with existing workflows and only send alerts for potential events that would require immediate attention.