Technology Promises to Extend Independent Living for Elderly

This Behind the Scenes article was provided to LiveScience in partnership with the National Science Foundation.

By 2030, 20 percent of the American population will be 65 or older.

As they grow into old age, these 71.5 million Americans could one day benefit from remote-monitoring technology designed to support longer independent living.

NSF-funded researchers at the University of Missouri have 

developed a non-invasive monitoring system for older adults that can detect acute illness in early stages and help manage chronic illness.

The work of professors Marilyn Rantz, Marjorie Skubic and their research team could eventually lead to a network of remote monitoring systems for older adults living on their own.

“Our goal is to allow people to age in a healthier, socially capable and independent lifestyle,” said Skubic.

Early Stages

This interdisciplinary project began in 1996 as the Aging in Place project, a community-based study that tested the effectiveness of coordinating care among registered nurses.

“We learned that nurse care coordination improves resident outcome and lowers health-care cost,” said Rantz.

With this new knowledge, the project sought to improve residents’ health and activity levels, and to avoid functional decline — loss of certain abilities such as hearing — and promote chronic illness management.

The next step was to implement the Aging in Place intervention in a specialized housing facility.

While the facility that would be called TigerPlace was under construction, Rantz and the nursing school entered a partnership with Skubic and the engineering school to develop new technology solutions that would resolve problems associated with aging. “We started problem solving about what the technology might look like and what we might be able to do with it,” said Rantz.

Next Steps

The researchers first tackled a common stereotype: the elderly are reluctant to use technology. The team conducted focus groups with older adults, and found that elderly people were not completely reluctant to use technology, but did have specific needs that the technology would have to accommodate.

In 2004, when TigerPlace opened, the researchers installed the first sensor network, including motion sensors and the first bed sensor (developed at the University of Virginia).

The researchers first focused on analyzing physical activity; then, they started working toward various ways to interpret the data, to predict functional decline and early illness.

The team discovered that, typically, residents exhibit a certain level of functional ability until an adverse event occurs. At that point, they move to a lower functional ability level. With this finding, the team refined the sensors and algorithms to improve early detection of any events that could lessen functional ability.

Initially, there was no straightforward way to correlate health with sensor data. But throughout the years, the researchers advanced their early model and thereby improved the correlations.

Since the residents opposed having cameras monitoring them, the researchers began using a Voxel model to extract 3D silhouettes of the residents. They also used Doppler Radar technology and the Microsoft Kinect to analyze gait. To validate the information they were gathering, the researchers compared the inexpensive systems to a Vicon system — a digital gait-modeling system that records the position of reflective markers and serves as the “gold standard” in dynamic modeling of motor behavior.

The current sensor network array includes passive infrared motion detectors, bed and chair sensors that detect pulse, respiration rate and bed restlessness, and stove temperature sensors. The team also developed a hydraulic sensor that can be positioned under a bed mattress. The new bed sensor, conceptualized by graduate student David Heise, is a sensitive tool designed to capture ballistocardiogram data and respiration rates and measure subtle changes in upper torso positioning as an indicator of restlessness.

After many years of modifications to the sensor network and analysis interface, the current data and assessment system have become fairly simple to use. Nurses and care coordinator staff members receive email health alerts. Once they click on an email, they see a website display of a resident’s sensor data. Algorithms interpret the information automatically, so while webpages display all the data, meaningful changes in data are highlighted for staff to evaluate. The bed sensor data, for example, are computed every 15 seconds; the data are then averaged and displayed in temporal intervals.

Each resident at TigerPlace chose whether they wanted to have the monitoring system in their apartment. Approximately 50 residents have participated in the research, and many of them reported positive experiences. The sensors have been used to detect and/or manage congestive heart failure, urinary tract infections, strokes and other medical disorders.

“The sensors pick up patterns of activity change, bed restlessness, pulse, respiration and general activity about 10 days to two weeks before health events and before people complain about being symptomatic,” said Rantz.

With the sensor-based early detection, one resident who had been at the Intensive Care Unit twice, was able to regain control over her congestive heart failure condition — through modification of her medication — before symptomatic weight gain and other adverse events occurred.

“We helped her manage several more years before she subsequently died at a ripe old 90 something,” said Rantz.

The researchers agree that the key to managing chronic and acute illness among the elderly is to detect it early, before hospitalization and other costly methods become necessary.

Future plans

The researchers plan to continue their study to further the accuracy of the algorithms and to advance their monitoring system.

With an NSF grant, the researchers will expand their project to a facility in Cedar Falls, Iowa.

This facility will include the newer sensors used at TigerPlace, including the Kinect depth images, and a high-speed fiber optic network infrastructure. When a health alert email is sent to the Cedar Falls staff, it will also be sent to the staff at TigerPlace. Video conferencing will be used to connect TigerPlace clinicians with residents and staff members in Cedar Falls; this will test remote nurse care coordination as well as high-volume data transfer. Working with larger samples will allow the researchers to refine the algorithms and to make the system applicable to large groups.

The team has made sure that the sensors they use are affordable and that the data the system provides is intuitive, so that health providers can easily understand and utilize it. If the project in the Cedar Falls is successful, the researchers hope that the system will be commercialized.

“Part of our ultimate goal is to make the system available in private homes,” said Skubic.

Both Skubic and Rantz have the same objective.

“Eventually we hope that we will see people across the country benefitting,” said Rantz.