CMU to research sensors to track elderly
Anind Dey, Associate Professor, and PhD student Matthew Lee, both of the Human Computer Interaction department at Carnegie Mellon, with their smart pill box.
Share with others:
Somewhere in the city today, researchers from the Quality of Life Technology Center will go to the high-rise residence of an older adult living alone and equip such everyday items as a pillbox, telephone, or sofa with wireless, computerized sensors. Their hope is to use those small sensors to track any decline in the resident's cognitive or physical abilities long before either normally would be noticed.
"We'll deploy the prototypes for the first time," said Matthew Lee, a doctoral student in the Human-Computer Interaction Institute at Carnegie Mellon University and a research team member.
Mr. Lee made it sound like the center, a joint program of CMU and the University of Pittsburgh, is going to war. And, in a way, the researchers are doing just that, for their ultimate goal is to identify these declines so early that they might be halted, or at least delayed, by medical or occupational interventions.
"If we can introduce these interventions earlier, they might have more effect on the individual," said Anind Dey, CMU associate professor in the institute and another team member.
Eventually, the team will install sensors in the residences of between 30 and 50 people. The participants must be at least 65, capable of giving their consent, and be at risk for the cognitive and physical declines as determined by medical and behavioral assessments and screenings.
In the first stage of the research, Dr. Dey will work with Linda Kent, an occupational therapist at Presbyterian SeniorCare, an industrial partner of the Quality of Life Technology Center, to identify participants. Most, but not all, of the participants will live in SeniorCare residences. The other team member is Diane Collins, an assistant professor of rehabilitation science and technology at Pitt, who will work with Ms. Kent to provide clinical evaluations of the participants.
The researchers will do their outlined work with the help of a $480,000 grant, announced Wednesday, from the Robert Wood Johnson Foundation's Project HealthDesign. The local team is among five nationally to be funded for the two-year project.
Dr. Collins has a different take on the use of the technology. Instead of measuring decline, she sees it as a way for families to keep watch over loved ones.
"As jobs become more scarce and people have to move away, they have this worry of caring for elderly family," she said. "This remote sensing is much less threatening than if you have a video camera there. So, for instance [the data] might tell me some information. Say it's my mom ... and I know she's not doing well and she just had surgery, I can tell if she got out of bed by a sensor in the mat by her bed."
The information collected can be used to make helpful changes in an elderly person's environment, Dr. Collins said, with tools such as commode chairs, or to discover the living arrangement no longer suits their needs.
Dr. Dey noted that the judges probably liked that the project "could have so much impact on a big part of our population."
"A big win for our technology is the speed with which it can measure subtle changes," Dr. Dey said. "People take a long time to go for assessments ... they go later than it needs to be.
"We see the changes more quickly."
The team is looking at four different activities: taking medications, preparing meals, phone use, and restlessness in bed or on a chair.
Sensors attached to the everyday items used in those activities transmit data to the team's main computer. "We analyze the data and produce visualizations for the parties," Dr. Dey said, referring to the interested doctors, occupational therapists, families and even the residents themselves, who will have monitors in their homes.
What Mr. Lee called a "smart pillbox" can, for example, provide a lot of information in the team's study of taking medications. The model the team displayed Wednesday had 28 enclosed boxes to accommodate medications taken at four different times of the day, seven days a week. Dr. Dey estimated it would be equipped with about 35 sensors, 28 in the little boxes and the rest along the side of the main box.
"If you open six different ones at once, that might indicate a little confusion," Dr. Dey said. "If one is open for a long time, that might be something of interest." There also will be accelerometers that can measure a hand shaking, which might be the sign of palsy.
Sensors placed on the water faucet and the glass or cup also will come into play in the medicine-taking observations.
Only one sensor each is needed in the sofa or couch, where increased moving around at night could indicate discomfort from osteoarthritis.
But multiple sensors will be needed to research meal preparations. For example, to cover a breakfast of cereal, Mr. Lee said, sensors might be installed on the cereal box or the cabinet it's stored in, the cabinet the bowl is in, the drawer the silverware is in, and on the refrigerator. Items equipped with sensors, however, would be tailored to the individual's daily habits.
Making coffee also would warrant multiple sensors to tell the researchers if the participant is putting the filter in place, filling the carafe with water, putting in the coffee, and putting the carafe in place.
"If it works, it's cost-effective," Dr. Dey said. "I can see the day when Medicare would pay for it ... and when it would be in a lot of different homes.
"It could reduce overall the cost of care and the overall cost of assessment, which is very expensive and time-consuming."
The other grant-winning teams are RTI International and Virginia Commonwealth University; San Francisco State University; University of California, Berkeley; and University of California, Irvine; and Charles Drew University.
First Published March 4, 2010 12:00 am