UMass researchers aim to use sensors to predict unhealthy behavior

By Neil Versel
Psychologist Sherry Pagoto, UMass Medical School in Worcester

Psychologist Sherry Pagoto, UMass Medical School in Worcester

Annually since 2004, the president of the University of Massachusetts system has had discretionary funds to award to faculty research projects in science and technology. This year, President Robert L. Caret gave the largest of $750,000 in Science and Technology Initiatives Fund grants – a $185,000 award over two years – to support research into mobile health-based sensors and interventions.

"We're excited that he did put a priority on this collaboration between behavioral sciences and computer sciences," grant co-recipient Sherry Pagoto, a psychologist at UMass Medical School in Worcester, tells MobiHealthNews. Pagoto is teaming up with computer science professor Deepak Ganesan from the main UMass campus in Amherst.

"We got together and wanted to merge our expertise," Pagoto explains. "I've always been interested in using mobile phones in managing health," adds Pagoto, who focuses much of her clinical psychology practice on addressing the causes of obesity.

"I'm hoping that we can understand and predict unhealthy behaviors with technology so we can change them," she explains.

The first year of research will examine technologies such as wearable sensors – perhaps embedded in jewelry -- as well as software that can analyze data in real time and provide feedback to patients. They will be studying both physiological and environmental data for what Pagoto calls "signatures" of unhealthy behavior, for example what triggers a person to overeat or to smoke a cigarette. Environmental data might come from tapping a smartphone's GPS feature to see where patient is or linking to the phones of friends and family to see if whom the person is with has an effect on unhealthy behavior.

"Our aim is to see if we could create predictive modeling," says Pagoto, who notes that there are mathematicians on the research team. "You could not just warn them, you could insert an intervention in there," perhaps some patient education or an activity to help reduce stress."

The researchers also would like to see if they can lessen the amount of self-reporting, which tends to be highly subjective and not always accurate. "This information [about environmental factors] is helpful from a clinical perspective," Pagoto says. Rich data might be able to inform psychotherapists in ways never before possible as they develop treatment plans. Historically, Pagoto notes, "I only have access to patients when they come in once a week."

Pagoto and Ganesan are teaming up with companies that include Jawbone/Aliphcom, Rock Health, Nexercise, Runkeeper, Microsoft Research, IBM and Google, plus the Massachusetts Department of Public Health and the Massachusetts Green High Performance Computing Center in Holyoke. The latter is a collaboration among state government and major universities in the commonwealth, including UMass, Harvard, Massachusetts Institute of Technology, Boston University and Northeastern University.