Prinicpal Investigator: SkubicCo-Investigators: Keller, Koopman, Lane, Popescu, Rantz, Robinson
Sensing technologies hold enormous potential for early detection of health changes that can dramatically affect the aging experience. Embedded health assessment can enable functional independence, improve self-management of chronic or acute conditions, and thus, improve quality of life. Problems in chronic disease management are often the cause of losing independence for aging Americans. In 2012, 1 in 2 American adults (117 million) had at least one chronic condition, and 1 in 4 had multiple chronic conditions, accounting for 84% of US health care costs. Chronic diseases especially affect older adults in whom exacerbations result in dramatic changes in health status, hospitalization, complex treatments, and high cost. Sensing technologies address this challenge by monitoring for health changes so that early interventions can be offered when treatment is most effective, prevention of decline is still possible, and costs can be controlled. Early illness recognition and early treatment improves health status with rapid recovery after an exacerbation of a chronic illness or acute illness and also reduces morbidity and mortality in older adults.
In previous work, we developed a health alert system that captures and analyzes data from sensors embedded in the home, and flags possible health changes. In a pilot study (NINR R21, Rantz, PI), we showed significant differences in health outcomes with health alerts from motion and bed sensor data (restlessness; low, normal, and high pulse/respiration rates). In the proposed project, we will extend this work by incorporating additional, more finely grained in-home sensing (gait and quantitative pulse and respiration), exploring the use of wearable health and fitness health sensors, and studying new algorithms that integrate changes in health status and medication use for customized alerts that recognize health change very early. We will also redesign our existing clinician-focused interface for displaying sensor data to empower older adults and family members to better self-manage chronic health conditions while addressing their privacy concerns.
The purpose of the proposed work is to refine a health alert system by developing new algorithms using data from sensors and an electronic health record (EHR), that provide alerts of very early changes in health status and that are customized to the individual consumer. A consumer-appropriate interface will also be developed to help consumers better manage their own health, and we will explore their opinions on how the system could be used. We will incorporate a recently developed bed sensor that passively captures quantitative pulse, respiration, and restlessness; we have validated the importance of in-home bed sensing. Gait parameters (in-home walking speed, stride time and stride length) will be captured using silhouettes from a depth sensor. Walking gait has also been shown to correlate to fall risk and health status In addition, commercially available, wrist-worn, health and fitness sensors will be investigated, to explore their potential for health alerts as well as seniors’ preferences and usage patterns. We propose to integrate new features from these sensors with new algorithms that track health status, medications, and trajectories of health status changes. Longitudinal health trajectories and ratings on clinical relevance, provided by clinicians, will be used to evaluate the health alerts. The new integrated health alert system will offer more customized and more sensitive alerts to better meet the needs of the older adults and care givers, and the new consumer interface will display sensor data and alerts in a format better suited for older adults and their family.
The project leverages a current study of 70 deployed health alert systems (with bed and gait sensing) in 12 senior housing sites (NINR R01, Rantz, PI). These sites provide a pool of diverse subjects for conducting a retrospective study on customized health change alerts with in-home sensing. Additional subjects will be recruited to explore the consumer view on customization, interfaces, and wrist-worn sensors. Using and evaluating this sensor system is important for the rapidly expanding elderly population. If we can help older adults remain healthier, active, and control their chronic illnesses with early detection of health changes and early intervention, seniors will have improved quality of life and independence as they age, avoiding or reducing debilitating and costly hospital stays, and for many, avoiding or delaying the move to a nursing home.
- Develop new health alert algorithms to improve the clinical relevance of the health alerts by incorporating new sensor features, multi-dimensional feature combinations, and algorithms that track trajectories in health status.
- Customize health alerts for each older adult’s needs using on-line machine learning and by integrating user feedback, health status and medications from an Electronic Health Record (EHR).
- Capture consumer input on customized health alerts and user interfaces. Design a consumer interface to display information in a format that older adults and family members find easy to use and interpret, to empower them to better self-manage chronic health conditions, while addressing their privacy concerns.
- Explore wearable sensors for older adults, including preferences, adherence, use patterns, usefulness, and the potential for automated health alerts to indicate changes in health.