MENU
component-ddb-728x90-v1-01-desktop

Software program can predict treatment outcomes for patients with bipolar disorder

The same technology that can outmaneuver Air Force pilots can also predict medical outcomes. (WKRC)

CINCINNATI (WKRC) - The same technology that can outmaneuver Air Force pilots can also predict medical outcomes.

A team of University of Cincinnati researchers published the findings in the journal Bipolar Disorders. The group of engineers and medical experts used artificial intelligence to accurately predict how patients diagnosed with bipolar disorder would respond to a specific type of treatment.

The software works by exploring countless outcomes and instantaneously selecting the best one. In the study, the artificial intelligence predicted how patients diagnosed with bipolar disorder would respond to lithium treatment 100 percent of the time.

"It takes us a couple of days on a simple PC to train the system with data, but once we've trained it, operationally [it takes] less than an instant," says Kelly Cohen, professor of aerospace engineering and engineering mechanics at UC.

Cohen approached the UC College of Medicine with his idea for this research during his sabbatical in 2014. He's worked with UC doctoral graduate Nicholas Ernest to develop and utilize the AI, David Fleck, an associate professor at the college of medicine and Dr. Caleb Adler, the UC Department of Psychiatry and Behavior Neuroscience vice chairman of clinical research.

Together they're working on ways to expand the research. Now, they're using the same artificial intelligence to determine the severity of concussions. They're also going to look into how it might be used to predict PTSD.

"Predictive modeling, personalizing medicine and treatment is something that we believe constitutes a paradigm shift that . . . has a potential for a huge, huge impact on society," says Cohen.

The researchers say being able to fast forward to the patient's outcome before treatment even begins could make health care safer and more affordable.

Trending