The proposed research will determine the feasibility of using non-invasive, unobtrusive continuous monitoring of vital signs and gross movement in bed as input to an early warning system to forecast future bed-exit, or attempted bed-exit, by elderly persons at risk for falling. The intended system is an application of automated pattern-recognition algorithms to an electronic medical record (EMR) in order to provide predictive capability, thus prompting timely interventions to avert impending acute health events. The research objectives support the goals of the NIH Institute on Aging (NIA) by identifying a means of extending the healthy, active years of life. The consequences of falls from attempted or successful bed exits among the elderly include hip fractures, soft tissue or head injuries, fear of falling, anxiety, depression and death. The proposed research seeks to identify a reliable, non- invasive method of predicting attempted bed exits with the goal of preventing subsequent falls and their medical consequences. The proposed research is completely aligned with the NIA's mission to "support and conduct genetic, biological, clinical, behavioral, social, and economic research related to the aging process, diseases and conditions associated with aging, and other special problems and needs of older Americans." The research uses a non-invasive, non-contact sensor array to collect heartbeat rate, respiratory rate, gross motion in bed, and bed-exit data. The specific technical question being studied is whether identifiable patterns of variations in three monitored measurements (heart rate, respiratory rate, and physical movement) can be used to reliably predict bed-exit or attempted bed-exit on a time scale of one to three minutes in the future. The broader aim of the research is to introduce a cost-effective, non-invasive method of reducing fall related injuries among the elderly, thus extending health in later years.
The specific aim i s to demonstrate how technology can be used to reliably predict bed-exit or attempted bed-exit with sufficient lead time to intervene prior to occurrence. The emphasis in Phase I will be testing whether it is possible to achieve statistical significance in predictive accuracy on a holdout sample. If statistical significance is demonstrated, Phase II will concentrate on achieving practical and commercial significance. A successful outcome on the proposed research suggests the potential for similar technological innovation in other electronic medical report (EMR) systems. These technical innovations, built on the successful outcome of the proposed research project, may be software applications, or in- house or web-based services.
The proposed research seeks to apply innovative technology as a non-invasive method of reliably forecasting bed exit attempts among the elderly to allow caregivers lead time to intervene promptly and appropriately to prevent falls. Fall and fall-relate injuries are the leading cause of injury-deaths among older adults. Therefore, with a rapidly aging population, ever-increasing healthcare costs, and limited healthcare resources, there is significant and critical relevance to reducing the number of falls, and fall-related injuries among the elderly.