Sensor-Based In-Home Frailty Assessment Based on Daily Physical Activity
Author
Pradeep Kumar, DanyaIssue Date
2021Advisor
Toosizadeh, NimaLaksari, Kaveh
Metadata
Show full item recordPublisher
The University of Arizona.Rights
Copyright © is held by the author. Digital access to this material is made possible by the University Libraries, University of Arizona. Further transmission, reproduction, presentation (such as public display or performance) of protected items is prohibited except with permission of the author.Abstract
Frailty and associated falls are among the most important reasons for hospitalization in the older adult population. Each year, three million older adults are treated in emergency rooms for fall injuries. Frailty becomes more prevalent with increasing age and confers high risk of adverse health outcomes, including falls, hospitalization, institutionalization, as well as mortality in some cases. Numerous geriatric interventions have been developed to improve clinical outcomes for frail older adults. A major obstacle to the success of such interventions has been the absence of a standardized and clinically practical method for screening of frailty and the onset of frailty to effectively target care. The development of clinically useful assessments of frailty is challenging because these assessments require adequate space, specialized equipment, trained personnel and sufficient time to administer the tests or questionnaire. Additionally, older adults with mobility difficulties may not be able to easily commute to the clinic or may require assistance to do so. In such scenarios, in-home health assessment can mitigate the mobility and communication challenges. With the advent of affordable wearable sensors, activity monitoring has become fairly simple and convenient. Daily physical activity (DPA) recording in older adults at home enables continuous activity monitoring over an extended duration in their natural surroundings with minimal constraints, and could provide an affordable and reliable frailty assessment. This work aims to assess frailty among older adults using sensor-based DPA and potentially develop a method for in-home/in-community frailty assessment.The goal of this series of studies was to evaluate the use of sensor-based DPA in a home-based environment as a measurement of frailty assessment. This was done by capturing the daily physical activities from 126 volunteers for two consecutive days using body-worn tri-axial accelerometer sensors. The data collected from the sensor was then processed to identify and categorize daily activities such as sitting, standing, walking, lying and postural transitions. Subsequently, the quality of walking or gait was used as a diagnostic element of physical function decline among the elders and its associations with frailty. The test-retest repeatability of gait parameters between day-1 and day-2 of recording in the entire sample and within each frailty group were also assessed. The variability in DPA was also examined in terms of variability in the durations of DPA and variability in the performance of DPA, and their associations within frailty groups were observed. The quality of walking provided a better frailty assessment in older adults compared to the traditionally used total number of steps and also showed a higher between-day repeatability. DPA variability also showed strong associations with frailty and improved frailty assessment when used with daily gait quality measures. Future research will build upon our findings to develop machine learning algorithms for the quality assessment of DPA and early detection of frailty by incorporating the sensor-based features characterizing gait and activity discovered within the current studies.Type
textElectronic Dissertation
Degree Name
Ph.D.Degree Level
doctoralDegree Program
Graduate CollegeBiomedical Engineering