Math Being Used To Predict Suicide Patterns Amid Rising Rates

From SIAM: In the wake of sobering statistics indicating someone in the U.S. dies by suicide every 12 minutes and that suicide is now the second leading cause of death among Americans aged 10 to 34, a leading team of U.S. researchers is working to save lives using an unlikely method — math. The goal is to gain insight into patterns associated with high-risk populations — such as war veterans — and develop an artificial intelligence tool that can predict when a suicide attempt is looming, enabling health practitioners to intervene before it’s too late.

“Suicide is an urgent public health crisis and yet preventing it remains a largely unsolvable problem,” said Xinlian Liu, an associate professor in the Computer Science and Information Technology department at Frederick, MD-based Hood College, and key member of the research team. “Our research approach is challenging because we’re analyzing many risk factors at once, beyond just mental health, and identifying complicated patterns of human behavior that can then be used like an early warning system. Math and computer science are essential for this analysis.”

Liu presented his research team’s project at a Society for Industrial and Applied Mathematics (SIAM) conference last month. The multi-faceted effort spans eight U.S. laboratories and multiple universities and colleges, and is funded by the U.S. Departments of Energy and Veterans Affairs. Liu is part of the Lawrence Berkeley National Lab (LBNL)-based team, which is led by principal investigator Silvia Crivelli, a project scientist at LBNL.

Previous suicide research has focused on biology, attempting to identify genetic and environmental factors that contribute to suicide. However, the recent availability of large electronic health records (EHR) datasets combined with deep learning techniques have fueled a new wave of research that looks at tackling the problem using healthcare information. “Whereas biologists are more interested in genomes and pinpointing specific physiological factors that put people at a higher risk of suicide, we are focused on analyzing thousands of complex factors in order to establish and recognize patterns of behavior,” he explained.

For example, the most common means of attempted suicide is drug overdose, which often takes multiple unsuccessful tries that end in a hospital visit. Assuming that an unplanned hospitalization therefore correlates strongly with a suicide attempt, the researchers are training their model to recognize the pattern. The work requires developing deep learning algorithms capable of analyzing a vast amount of both unstructured data, such as clinical observations of social, environmental, mental and family interactions, as well as structured data such as prescriptions, examinations, vitals, lab results and other medical information.

“We’re trying to predict when a patient may have an unplanned readmission to hospital a month before it happens,” explained Liu, noting that if clinicians can be provided with a warning, they will have time to intervene. “Even if they simply call and ask if the patient is doing okay, it will make a difference since we already know that reaching out to patients is critical for suicide prevention.”

This summer, Liu and his students are heading to LBNL where they will be working on data from the Million Veterans Program (MVP), one of the world’s largest medical databases containing health information from about 700,000 veteran volunteers. It is estimated that roughly 22 veterans die by suicide per day and the researchers are aiming to curb that. “If we can stop even one or two suicides, we’ll consider our work successful,” Liu said.

A recent article published in the SIAM News blog outlines the initial work undertaken by the team using a smaller, more widely available database of de-identified clinical health data called MIMIC-III. Now that they have access to live, veteran data in the MVP, Liu is excited by the possibilities to refine the model.

“We’re working to provide a predictive tool for doctors and clinics so they can quickly identify a high risk group and prioritize resources to reach out to them,” he said, adding that preliminary results are expected to be available this fall. “We’re using high-end math to solve a complicated problem that will ultimately save lives.”