mdlogix Awarded Federal Grant Expected to Improve Suicide Prevention Programs through Machine Learning

Improve Suicide Prevention

Baltimore, MD — Allen Tien, MD, MHS, founder, president, and chief science officer of mdlogix, has been awarded a one-year, $499,000 federal grant from the National Institutes of Health (NIH) to study how machine learning (ML) can improve capabilities of underserved populations, specifically Native and African American communities, to prevent suicides.

The Artificial Intelligence/Machine Learning Consortium to Advance Health Equity and Researcher Diversity (AIM-AHEAD) brings researchers and communities together to enhance capabilities for use of emerging ML. Machine learning automates developing and applying complex calculations to process data; three main types being image, unstructured text, and numerical.

There are three 12-week stages in this project which just started. First, the research team will focus on teaching high school students ML fundamental concepts, along with providing them with a user-friendly software tool for the application of those concepts, MILO (Machine Intelligence Learning Optimizer). Second, the participating students will learn about suicidality. Third, the team will teach, guide, and support the students in asking and answering questions about suicidality using MILO and rich numerical data collected with the Behavioral Health Screen (BHS), an evidence-based comprehensive assessment tool for identifying suicidality in children and youth.

“One of our top goals is to show how machine learning can be applied to solve some suicide prevention program problems, as well as to guide public health policy and program design in states across the country,” Dr. Tien explained. “We will focus on training high school students and teachers in ML basics and how to use MILO to better understand the knowledge otherwise locked away in complex data.”

MILO provides the opportunity to leverage powerful ML approaches to develop optimal predictive models to quantify individual risk for health problems, including levels of suicidality. The deidentified assessment data will come from bhworks, a HIPAA-compliant mental health software platform from mdlogix engineered for suicide prevention. The BHS within bhworks identifies mental health problems and psychosocial risk factors. It consists of psychiatric symptom scales and risk behaviors that cover the multiple areas suggested by numerous peer-review studies and best practice guidelines. Results are immediately scored, summarized, and securely sent to a designated clinician for review.

The project will recruit Native and African Americans to participate in training and education about suicidality, covering what is known and unknown within a public health framework of prevention, early intervention, and treatment; data analysis; and how to apply, sustain, and expand the research findings in a community.

The project’s starting Native American site is Nah Tah Wahsh Hannaville High School in the Michigan Upper Peninsula, with which mdlogix is already contracted to provide bhworks to support two programs, one on student safety and the other on student mental health.

The starting African American site will be one or two high schools from the existing 100,000 student contract mdlogix has with the Prince George’s Public School System in Maryland, and working with one or two of their nine behavioral health providers, including those from Sheppard and Enoch Pratt Hospital in Towson, Md.

A range of experts will be assisting Dr. Tien in carrying out this grant project. This includes Mary Ann Dutton, PhD, vice chair for Research at Georgetown University’s Department of Psychiatry, who will be a key clinical science consultant, along with Ross Jacobucci, PhD, University of Notre Dame assistant professor whose research focus is machine learning and suicidality.

mdlogix, home of bhworks, designs software that combines science and practice. We are on a mission with our clients to improve behavioral health care together. Learn more about bhworks and schedule a demo today