Using big data to solve a moody issue

Mood disorders like depression are common among U.S. adults. Still, such disorders remain challenging for clinicians to diagnose and treat effectively.

A UB public health researcher is part of a team of scientists that received a National Science Foundation (NSF) grant to use big data to develop a new approach they say will improve the classification of mood disorders and lead to more effective outcomes for psychiatric patients.

Rachael Hageman Blair, assistant professor of biostatistics in the School of Public Health and Health Professions, is one of five principal investigators on the one-year, $100,000 planning grant, funded by NSF in a joint effort with the National Institutes of Health.

Hageman Blair’s collaborators on the project include biostatistics, information science, mathematics, biomedical informatics, psychiatry and electrical and computer engineering researchers from the University of Iowa, University of North Carolina-Chapel Hill, University of Oregon and the University of Utah.

Their aim is to use big data to develop a novel methodology and visualization tools to cluster patients with mood disorders. “Existing approaches often break or are inappropriate in big data settings for several reasons,” Hageman Blair explains.

“There is not a one-size-fits-all approach, even for well-behaved data sets. Bringing together different methods under a single umbrella with strong visual interpretations holds value for a clinician,” she adds.

The collaborators point out that recent studies from the National Institute of Mental Health show that while mood disorders are prevalent, treatment is less than 25 percent effective. “An existing hypothesis is that the [Diagnostic and Statistical Manual of Mental Disorders] labels themselves are inaccurate because they do not fully integrate all available data,” says Hageman Blair, who has a PhD in mathematics.

“Our aim is to ignore the DSM label and regroup patients based on comprehensive data profiles, which include genetic, environmental, demographic and clinical data, among others. Some groups of individuals may be more responsive to treatment, which is important for precision medicine,” she adds.

The collaborators met over the summer at an innovation workshop hosted by the Statistical and Applied Mathematical Sciences Institute, a NSF-affiliated research institute located in Research Triangle Park, North Carolina.

“It was a lot like speed-dating for scientists. By the end of the week, I found six great collaborators, and then the work of developing the proposal began,” says Hageman Blair.

Over the next year, the research team will begin developing its methodology. “We’ll be focusing on applications to mood disorders, which are known to be particularly challenging to classify,” she says.