5Q with Eric Rose: Statistics-Enabled AI to Support Mental Health Care

A man wearing a light blue button-down shirt smiles for a portrait against a gray backdrop.

Eric Rose, assistant professor of epidemiology and biostatistics. (Photo by Patrick Dodson)

ALBANY, N.Y. (Sept. 25, 2025) — Assistant Professor Eric Rose joined the Department of Epidemiology and Biostatistics at UAlbany’s College of Integrated Health Sciences in Fall 2022. A statistician by training, he discovered his passion for the field while studying statistics at North Carolina State University.

During graduate school, Rose became fascinated by precision medicine’s potential as a way to apply statistical theory to improve people’s lives. He focused on mental health care and developed statistical methods for studies exploring ways to use artificial intelligence (AI) to inform treatment selection. Today, his research centers on using data-driven methods to help develop tailored mental health treatment plans according to individual patient needs.

Rose is an active collaborator with the Global Center for AI in Mental Health, a research hub founded by UAlbany, SUNY Downstate Health Sciences University and the United Nations Health Innovation Exchange.

We caught up with Rose to learn about how he applies statistics to mental health care, the growing potential for AI to improve these approaches, and what excites him most about the future of this field.

How did you become interested in using data to help people?

I have always loved math and knew that I wanted my career to be related to math in some way. When I was in college, I became interested in statistics because it allowed me to use math to address real-world problems, which led me to graduate school to study statistics at N.C. State. It was during graduate school that I was exposed to the field of precision medicine and was drawn to it immediately. It combined interesting statistical theory with the potential for clear, significant impact that could improve people’s lives. I started studying precision medicine in the context of mental health care because it stood out as a way I could make a significant impact.

How can precision medicine, together with AI, improve mental health care?

Precision medicine, also called personalized medicine, means tailoring treatment decisions to individual patient characteristics to improve patient outcomes. This does not mean creating new unique interventions for individual patients, but instead, making smarter decisions using all the available information on a patient.

AI can be very effective for synthesizing complex information from multiple data sources, which can be used to help make the best decisions for patients. Though, the goal is not to create AI models that dictate treatment, but instead to create models that help inform treatment decisions. Creating statistical models that can be interpreted by clinicians — and applied to their patients — is key to this approach.

How could your work improve someone’s experience with mental health treatment?

There are many examples of precision medicine showing potential for improving care for many different conditions such as HIV, cancer and Alzheimer's disease. In the mental health context, predicting patient response to treatments is a key area that we are exploring.

Individual responses to antidepressants are notoriously variable, due to a combination of factors. Identifying the right treatment for an individual is often a process of trial and error. Precision medicine can be used to predict individual responses to different treatment options. By identifying how multiple personal factors influence treatment efficacy, and how they interact, it becomes possible for clinicians to prescribe the best-fit treatment faster. Much of my work is focused on creating statistical methods to properly design and analyze studies to examine these types of questions.

One specific project in this vein, led by Julia Hastings, grew out of collaborations among faculty at UAlbany and SUNY Downstate through the Global Center for AI in Mental Health. Our goal is to simulate responses to different treatments at the individual level, using data-driven statistical methods that account for a mix of personal traits. Someday, this sort of method could help guide treatment selection for patients diagnosed with major depressive disorder. Instead of undergoing a lengthy and expensive search for the right antidepressant, a patient could land on the one that helps them sooner.

How can we ensure that AI health care tools are trustworthy and equitable?

Carefully considering the data that is used to train the algorithm in a given tool is extremely important. If there is existing bias in the data, AI tools built using this data will only perpetuate and potentially exacerbate that same bias. Similarly, we also need to consider the data that we use to evaluate the algorithm. Once we have trained and evaluated an AI tool, it is also very important to assess how it is being used and who has access to it to ensure that it is equitable.

What excites you most about future possibilities for mental health care enabled by statistical analysis and AI?

There are many exciting research opportunities for understanding how to best design studies and analyze data to inform treatment selection as well as many different areas across health care in which clinical decision-making could be improved by these types of studies. What excites me most, though, is the enthusiasm from students. Seeing students motivated by the chance to apply statistics and AI to improve lives is very rewarding, and it makes me optimistic about where the field is headed.