U. researchers unveil AI-powered tool for disease prediction with ‘unprecedented accuracy’

SALT LAKE CITY — It’s long been a goal in medicine to better understand the long trajectories of diseases in hopes of engaging in better prevention and early intervention.

“Collectively, they’re (chronic and progressive diseases) responsible for about 90% of the health care costs in this country and the vast majority of morbidity and mortality,” said Nina de Lacy, a professor of psychiatry and member of the One-U Responsible AI Initiative’s executive committee.

Now, University of Utah researchers have taken a crucial step in doing so, unveiling a new, open-source software tool kit that uses artificial intelligence to predict whether individuals will develop progressive and chronic diseases years before symptoms appear.

Enter RiskPath, a new technology that analyzes patterns in health data collected over multiple years to identify at-risk individuals with “unprecedented accuracy” of 85% to 99%, according to National Institute of Mental Health-sponsored research published last week by the U.’s Department of Psychiatry and Huntsman Mental Health Institute.

The program harnesses explainable AI, which is designed to explain complex decisions in ways humans can understand.

“Explainability means, can I explain enough about how AI accomplished this prediction such that it becomes understandable to humans?” de Lacy said. “That would be things like what RiskPath does.”

De Lacy explained something that has always been a challenge in biomedicine is building models and analyzing longitudinal data, meaning it’s collected over many time periods.

“One of the major use cases in using longitudinal data is course development, understanding how people grow up and develop over time,” de Lacy said. “And one of the other ones is what RiskPath is aimed at, which is understanding progressive or chronic disease. There are many progressive and chronic diseases out there, and some of the big ones are things that are the major diseases that affect humans.”

The research shows current medical prediction systems for longitudinal data often miss the mark, correctly identifying at-risk patients only about half to three-quarters of the time. Unlike existing prediction systems for longitudinal data, RiskPath uses advanced time-series AI algorithms that deliver crucial insights into how risk factors interact and change in importance throughout the disease process.

“By identifying high-risk individuals before symptoms appear or early in the disease course and pinpointing which risk factors matter most at different life stages, we can develop more targeted and effective preventive strategies. Preventative health care is perhaps the most important aspect of health care right now, rather than only treating issues after they materialize,” de Lacy said.

De Lacy and the rest of the research team validated RiskPath across three major long-term patient cohorts involving thousands of participants to successfully predict eight different conditions, including depression, anxiety, ADHD, hypertension and metabolic syndrome.

The technology offers several key advantages:

Enhanced understanding of disease progression: RiskPath can map how different risk factors change in importance over time, revealing critical windows for intervention. For example, the study showed how screen time and executive function become increasingly important risk contributors for ADHD as children approach adolescence.Streamlined risk assessment: Though RiskPath can analyze hundreds of health variables, researchers found that most conditions can be predicted with similar accuracy using just 10 key factors, making implementation more feasible in clinical settings.Practical risk visualization: The system provides intuitive visualizations showing which time periods in a person’s life contribute most to disease risk, helping researchers identify optimal times for preventive interventions.

While RiskPath is primarily a research tool to help researchers build better risk stratification models, de Lacy hopes it will eventually be used in a health care setting to improve disease management.

“Some may be using that to build models that can be implemented in health care, and we kind of hope that they do that. But … a big part of what my lab is interested in doing is building tools that do a better job of risk stratification. We’re very interested in prevention,” de Lacy said. “The ultimate aim of RiskPath and tools like RiskPath is to help people build better risk stratification tools and decision support tools.

“And what those do is help clinicians, and maybe one day patients, be able to understand their risk for a chronic or progressive disease better and earlier,” she said.