Jan. 9, 2023 — The health care field produces dozens of useful data points during each patient visit, from height, weight and blood pressure to cholesterol levels and relevant family history. However, finding the right way to synthesize these data to improve current treatments and even find new cures for diseases requires a lot of computing power.

To elevate health care systems in the digital age, where resources like artificial intelligence can be used to sift through and organize millions of data points that would otherwise be impossible to extrapolate, partnerships must be forged between front-line clinicians and data scientists behind the scenes.

At the UF College of Medicine, Jiang Bian, Ph.D., UF Health chief data scientist and a professor of biomedical informatics, is leading a project under the research pillar of the college’s strategic plan aimed at accomplishing such a task. Through the Enabling AI and Other High-Tech Research project, Bian and his College of Medicine colleagues are identifying and addressing opportunities to improve UF Health’s research information technology infrastructure to make it “AI ready.”

“Raw data, like the data collected on a smartwatch every one or two seconds, is not ready to answer questions like what the measurements detect and what a significant change in heart rate or breathing data points might indicate,” Bian said. “That data needs to be understood and converted to be an indicator of a risk factor, for example. Machine learning helps draw those connections between the data.”

Bian, who frequently collaborates with UF Health Cancer Center researchers, said one way he has applied AI to cancer research is by creating algorithms that combine patient clinical characteristics, lifestyle factors, previous image scan results, family history and social determinants of health to determine cancer screening needs at an individual level.

By improving the capability of health care systems to do these computations, patients often experience better outcomes from the combined efforts of expert physician and researcher insights and technological assistance, Bian said. This is a significant goal of learning health systems, a cyclical process that involves producing data through both clinical and research activities that provides actionable evidence of a claim or outcome. That evidence is then put into practice, and the cycle begins again to create new data from the adapted practice.

“Through that learning cycle, you continuously improve your local practice, and you also generate new insights,” Bian said. “It is also beyond just the health system — it has to go to the community level as well. Health systems need to figure out ways to work with the community, from the patient and provider to community organizations and state agencies. In the end, it’s about improving the local practice to learn and improve as a learning health community.”