Changes in Motor Behavior Predict Falls and Identify Acute Events

Abstract

This study aimed to examine how changes in motor behavior are associated with falls, delirium, and urinary tract infections (UTIs). Twenty-three (128 observations) skilled nursing residents were examined for up to 18 months. In multilevel models, motor behaviors (e.g., time and distance traveled, gait speed), measured by a real-time locating system, were used to predict falls, UTIs, and delirium. Falls were associated with decreased gait speed ( OR = 0.01; p ≤ 0.001) and path distance ( OR = 0.99; p ≤ 0.05); delirium was associated with increased distance traveled ( OR = 1.01; p ≤ 0.001), path distance ( OR = 1.02; p ≤ 0.001), and decreased time traveled ( OR = 0.99; p ≤ 0.001) and path time ( OR = 0.99; p ≤ 0.001); UTIs were associated with increased distance traveled ( OR = 1.01; p ≤ 0.001), decreased time traveled ( OR = 0.99; p ≤ 0.001), and the number of paths ( OR = 0.91; p ≤ 0.01). Subtle changes in motor behavior may be an early warning sign of falls and acute events. Continuous monitoring may enable clinical staff to prevent, identify early, and/or delay these poor health outcomes.

Real-time location systems technology in the care of older adults with cognitive impairment living in residential care: A scoping review

“…Only one study was published in 2002 ( 40 ) with remaining studies published after 2007. The included research studies used various research designs, including cross-sectional ( n = 16) ( 14 , 20 , 23 – 25 , 28 , 30 – 34 , 36 – 38 , 41 , 43 ), longitudinal ( n = 5) ( 12 , 13 , 22 , 27 , 42 ), case reports ( n = 4) ( 26 , 35 , 39 , 40 ), and one of a case series ( 21 ).…”

“…Ten studies used UWB technology ( 12 , 13 , 20 , 22 – 25 , 27 , 28 , 41 ). Other technologies used included RFID (IC tag) ( n = 8) ( 30 , 31 , 33 – 38 ), radio waves (Emerald ® ) ( n = 3) ( 21 , 26 , 29 ), Bluetooth ( n = 3) ( 32 , 39 , 43 ), and one each of GPS technology ( 42 ) passive IR ( 40 ) and an unspecified wireless mesh network (Wi-Fi) ( 14 ).…”

“…The quality of the evidence is reported in Supplementary Appendix 2 . Study quality ranged from poor ( 21 , 25 , 26 , 32 , 38 , 43 ), moderate ( 12 , 13 , 20 , 23 , 24 , 28 – 31 , 34 , 36 , 37 , 39 – 41 , 44 ), to good ( 14 , 22 , 27 , 42 ). Strengths across the studies included in-detail descriptions of the RTLS implementation in the residential setting.…”

“…Specifically, they identified residents who spend more time walking are at increased risk of falls. Bowen and Cacchione ( 22 ) used RTLS across two nursing homes and 23 residents for 18 months to examine whether mobility changes can be indicative of changes in health status. When reviewing average weekly motor behavior of the larger sample, residents with delirium walked farther (543 meters) and longer (35 min).…”

“…The duration of data collection varied widely. For example, Jansen et al ( 14 ) collected data on 69 residents for only 2 days, whereas Bowen and Cacchione ( 22 ) collected data on 23 residents for 18 months. While there is some value in building evidence through case studies and observational studies, more extensive studies using population-level data in assisted living settings and advanced data science methods such as machine learning or deep learning, are needed.…”