A Healthy Future
AI is reshaping the healthcare industry to improve patient outcomes.
AI is driving groundbreaking advancements across medical research at the University of Minnesota, revolutionizing how diseases are diagnosed, treatments are developed, and patient care is optimized.
From predicting critical conditions like sepsis in emergency settings to accelerating cancer treatment processes, AI-powered innovations are reshaping the landscape of healthcare. As researchers harness the potential of artificial intelligence, their efforts promise to redefine standards of care and improve outcomes for patients worldwide.
As of today, AI is primarily used to increase speed and accuracy in the healthcare industry. Some of the most common uses of AI in this field include:
Diagnosing Patients
AI algorithms analyze medical data, such as CT scans, MRIs, and X-rays to assist doctors in accurate diagnoses.
Transcribing Medical Documents
AI models can turn spoken language into written text, offering doctors a more efficient and precise means of documenting medical information.
Drug Discovery
AI speeds up drug discovery by analyzing extensive datasets to pinpoint potential drug candidates and forecast how effective those drugs may be.
Administrative Tasks
AI optimizes administrative processes, including billing and scheduling, making healthcare institutions more efficient.
Research across the University is looking to expand how AI is used and the impact it has on treating and helping patients.
For instance, the Medical School’s Program for Clinical AI develops, validates, and implements artificial intelligence (AI) tools to improve healthcare delivery. A key objective of the Program for Clinical AI is to investigate AI-enabled tools in real-world settings including monitoring AI model performance for drift, equity, and fitness for answering questions across settings and subpopulations. The program’s team includes members with a wide range of expertise across the University of Minnesota including the Medical School, School of Public Health, Institute for Health Informatics, and the College of Science and Engineering.
Recent work by the program sought to predict sepsis in patients and identify which patients would benefit from early antibiotics. The research team created a predictive model that triggers a sepsis score for patients in the M Health Fairview emergency department (ED) one hour and six hours after admission. The model uses data such as vitals, labs, medications prescribed, and the patient’s chief complaint in the ED.
The model results showed that if a patient received antibiotics within one hour of passing the 1-hour sepsis score threshold, their mortality and length of stay was significantly reduced. Mortality and length of stay was reduced even more when patients were administered antibiotics within one hour of passing their six-hour sepsis score.
Overall the model performed better than similar sepsis prediction models and will improve patient outcomes.
That is one of countless ways U of M research is exploring AI and its impact on the future of the healthcare industry.
Below are four more ongoing projects at the University:
Improve Outcomes of Breast Cancer Survivors
A collaborative project between the Medical School and the College of Science Engineering is utilizing artificial intelligence (AI) to address a common issue in healthcare research: imbalanced data.
In medical research, datasets often have unequal numbers of patients with different health outcomes. This creates a challenge because traditional AI methods may produce inaccurate or biased results when faced with such data imbalances. To tackle this problem, the researchers, led by Rui Zhang, a Masonic Cancer Center researcher, and Ju Sun, a computer science assistant professor, are developing innovative AI techniques specifically designed to learn from imbalanced datasets.
Their primary objective is to improve the accuracy and fairness of disease diagnoses and prognoses, with a particular emphasis on predicting cardiotoxicity in breast cancer patients undergoing treatment. Cardiotoxicity refers to heart damage resulting from cancer treatments, and it's a significant concern in breast cancer care.
By successfully developing and implementing these AI methods, doctors worldwide will have access to more precise and unbiased predictions regarding heart issues in breast cancer patients. This advancement has the potential to significantly enhance post-treatment care and improve outcomes for breast cancer survivors.
“The technology at our disposal today is unprecedented. We’re excited and proud that the University of Minnesota is leading the way in using AI to anticipate post-treatment health challenges for breast cancer survivors,” says Ju Sun, computer science assistant professor in the College of Science and Engineering.
Their research is funded by a $1.2 million grant over the next four years from the National Cancer Institute.
This is adapted from a story by the Masonic Cancer Center.
The Cost of Healthcare
Could the cost of healthcare go down thanks to AI?
The introduction of AI technology into healthcare holds promise for improving clinical and diagnostic processes, potentially leading to cost savings. However, the extent of these savings and their impact on overall healthcare spending are not well understood. While there are anecdotal examples suggesting that AI can reduce costs by streamlining tasks such as test result analysis and potentially replacing expensive diagnostic procedures, empirical evidence is lacking.
To address this gap, researchers from the University of Minnesota School of Public Health are embarking on a study to investigate the impact of AI-enabled Software as a Service (SaaS) applications on healthcare costs and other aspects of healthcare delivery. Using Medicare reimbursement data spanning from 2016 to 2024, the researchers will compare the healthcare spending, testing rates, diagnostic practices, and overall health outcomes between clinicians who have adopted AI-enabled SaaS applications and those who have not.
By analyzing this data, the researchers aim to provide the first causal evidence on how the adoption of AI in healthcare affects spending and health outcomes.
Hannah Neprash, an assistant professor at the U of M School of Public Health and the lead researcher on the study, emphasizes that if their research demonstrates financial savings and improvements in health outcomes associated with AI-enabled tools, it could accelerate their adoption by healthcare delivery organizations. This evidence could potentially inform policy decisions and encourage broader integration of AI technology into healthcare systems.
This is adapted from a story by the School of Public Health.
Ethically Using AI
While doctors harness the power of AI for diagnosing and treating severe illnesses, nurses are faced with ethical considerations regarding its implementation in their practice. Although nurses have long utilized AI tools for various tasks like clinical decision support and patient monitoring, the increasing accessibility of generative AI and the heightened attention surrounding AI has prompted concerns about its ethical implications for patient care.
To address these concerns and equip nurses with the knowledge to navigate AI integration responsibly, the School of Nursing has initiated an examination into the ethical dimensions of AI in nursing. This initiative aims to develop a framework or strategy for ethical AI usage, with plans to implement changes soon.
“Nurses are very ethical creatures and know that any technology can be highjacked for sinister use,” says Pamela Cipriano, president of the International Council of Nurses. “Nurses are eager to further understand AI’s positive applications as well as limitations. Ensuring there are guardrails that lead to virtuous use of this powerful technology is a necessary first step to embracing the benefits of AI for nursing.”
This is adapted from a story by the School of Nursing.
Research Leverages AI to Identify Dogs at Higher Risk for Cancer
