F5‐03‐05: A ubiquitous sensing environment to detect functional changes in assisted living apartments: The Tiger Place experience
Marjorie Skubic · Alzheimer s & Dementia · 2010
To address an aging population, we have been investigating the use of sensor networks for monitoring older adults in their homes. In addition to recognizing urgent conditions such as falls, we are interested in capturing patterns of activity from the sensor data that correlate to functional ability and health conditions. Our premise is that the continuous assessment of physical and cognitive function can provide an early indication of decline in health and functional ability. Identifying and assessing problems while they are still small can provide a window of opportunity for early treatment that may prevent catastrophic problems and allow the older adult to retain a higher quality of life. Sensor networks with passive motion, bed, and stove sensors have been installed in 26 apartments in TigerPlace, an eldercare facility in Columbia, MO. The work has been ongoing for over four years with an average installation time of 23 months 10 networks have been operational for over 2 years. This longevity in data collection facilitates the investigation of data trends and algorithms that track the daily patterns for an individual over time and recognize when activity patterns begin to deviate from the norm. Analysis comparing the sensor data to health records has been done in a retrospective study. A web interface provides a means of examining the data in multiple forms including visualizations such as the motion density map. We are also investigating algorithms for automatic alerts that provide early indications of health problems. Case study analysis has shown that the sensor data can track activity changes resulting from a variety of conditions, including depression, knee replacement, cardiac events, and cognitive impairment. The data particularly useful in tracking health changes have been the bed sensor which captures pulse, respiration and restlessness and the density of the motion sensor events which capture lifestyle differences such as sedentary vs. active and regular vs. irregular routines. The sensor data have been found to correlate to changes in health and function in a retrospective study. We are preparing to conduct a prospective study that will begin this year.