By Anne Riker Garlington
Members of the University of Florida College of Public Health and Health Professions came together recently to discuss the power and potential of artificial intelligence for improving health care delivery as part of the college’s AI Seminar Series, sponsored by the PHHP Artificial Intelligence Work Group.
“With recent advancements in computing capabilities and the capacity to gather and store large datasets, the application of AI in health care offers numerous potential benefits, particularly in moving toward personalized medicine. It’s like creating a digital health fingerprint for everyone to maximize treatment outcomes,” said Aprinda Indahlastari, Ph.D., an assistant professor in the PHHP department of clinical and health psychology and co-chair of the PHHP Artificial Intelligence Work Group.
Jim Hoover, D.B.A., presented the seminar titled “What really is AI and how can it be implemented to improve health care.” He is a clinical professor of marketing in UF’s Warrington College of Business, and director of the Business Analytics and Artificial Intelligence Center at UF.
Hoover is also the Jack Faricy Professor within the College of Business and serves as the 2023-2024 Faculty Fellow for the university’s Artificial Intelligence Academic Initiative Center, known as AI2.
“The synergy between generative AI and humans has been shown to outperform the capabilities of either AI or humans alone,” Hoover said. “Therefore, it’s important for medical experts to examine AI solutions results to ensure their accuracy in diagnosing and treating patients.”
AI can be used for pattern recognition by health care organizations, Hoover said, and in several cases, AI has been shown to be a more effective predictor of patient outcomes than individual assessments alone.
Hoover shared some specific examples of how AI technology can improve health care and prevention:
The discovery of certain hard-to-detect lumps in a breast. When AI technology was used as a third reader, early detection rates improved, particularly in cases where the cancer was very difficult to detect and ultimately more invasive.
Predicting estimated blood loss during surgery. Previously, blood loss estimates were calculated by rounding to the nearest 50 milliliters, which was not precise. AI can provide more accurate estimates, enhancing patient care during operations.
Predicting suicidal ideation. Current prediction accuracy is around 50%. AI tools could potentially increase prediction accuracy to 80%, which could save lives through early intervention. Current practices rely on questionnaires, but AI technology could help reduce bias inherent in historical data.
Responsible AI has the potential to transform patient care for the better, especially for tasks where accurate prediction models can complement human decisions. However, we must emphasize its careful use to ensure safe and fair outcomes for everyone,” said Noah Hammarlund, Ph.D., an assistant professor in the PHHP department of health services research, management and policy, and co-chair of the PHHP Artificial Intelligence Work Group.