AI-Enhanced Anesthesiology Part One: The Future of Patient Care

Artificial intelligence (AI) is emerging as a transformative force in medicine, promising to revolutionize health care. The potential of AI to enhance the field of anesthesiology extends beyond data research and analysis to clinical technologies and decision support. A number of initiatives within the UF College of Medicine’s Department of Anesthesiology are working to integrate AI in beneficial ways.

François Modave

“Both through the department and through the Quality and Patient Safety Initiative (QPSi), we’re using AI across UF Health: in research—because it’s still really important to raise the body of clinical knowledge in AI—but also in education, quality improvement, and elsewhere across the entire infrastructure,” says François Modave, Ph.D., associate chair of research for the department and assistant dean of the QPSi Academy and Training Programs. “When we’re able to do this, we can change how we practice and deliver care.”

Data-Driven Insights and Applications

Doctor Tighe explaining the A-I and Q-I graph

Artificial intelligence is, at its core, technology that thrives on information—the more, the better. “Anesthesiologists process a massive amount of data with every single patient,” points out Patrick Tighe, M.D., M.S., associate dean for AI Applications & Innovation, executive director for QPSI, and professor of anesthesiology and orthopaedic surgery. “Artificial intelligence can help us look at data in new ways, and hopefully we can use it to help us make better decisions for our patients.”

Chris and Heidi Goldstein

Chris Goldstein, M.D., and Heidi Goldstein, M.D., affiliated anesthesiology faculty based at the Malcolm Randall VA Medical Center, noted that modern machine learning algorithms have a significant advantage over traditional statistical methods when it comes to data analysis. “These algorithms have the ability to process the vast and diverse data sets found in electronic records, giving us the opportunity to gain new insights.” These insights have led to significant progress in implementing AI-powered technologies in anesthesiology, as the Goldsteins point out: “Some clinical examples include smart devices, such as vital sign monitors with automated EKG rhythm detection that can alert us to abnormal and dangerous rhythms, ventilators that can automatically adjust certain parameters, and more recently an AI-based device predicting whether a patient will suffer from low blood pressure minutes before the situation arises, allowing us to take proactive treatment steps.”

Meghan Brennan

With regards to these insights, a common refrain emerges again and again: the hope of improving quality and safety for patients. As Meghan Brennan, M.D., M.S., assistant professor of anesthesiology, states, “I always think about applications of research, like using natural language processing to identify rare events in medical record data. We could find out a lot of information from the data that we collect every day, and I think it could really help us identify at-risk patients and intervene earlier.”

AI-Assisted Quality Patient Care

The integration of AI to positively affect patient care is central to the College of Medicine’s AI-QI Program and the Quality and Patient Safety Initiative (QPSi). The QPSi Academy will offer online courses across a broad range of areas teaching AI concepts and use in medicine to health care professionals. To support the AI innovations in the College, grant funding for projects under the Rapid AI Prototyping and Development for patient Safety (RAPiDS) framework has been implemented.

red blood cells
Image created with AI software DALL-E 3

In the Department of Anesthesiology, this support has led to a Patient Blood Management (PBM) RAPiDS Cycle 2 initiative focused on patient-specific maximum surgical blood ordering to guide lab ordering and blood product availability for elective surgical cases, with Keith Howell, M.D., associate professor of anesthesiology serving as principal investigator. According to Imke Casey, DNP, RN-BC, quality officer for PBM, “This initiative is directly linked to improving patient quality and safety by predicting surgical blood loss. The goal is to advance from an estimation of blood loss to precision medicine so we can reduce blood waste and blood draws for diagnostic testing.”

Keith Howell, MD

Howell detailed the benefits of developing an AI-assisted model for patient- and surgeon-specific factors that influence clinical decision-making around blood product utilization, including a patient-specific risk of bleeding, transfusion, and patient blood management. “I would like to have enough information on every patient to come up with a bleeding risk score,” he said. “Ultimately the goal is to manage patients in a better way and decrease the risk of some of the complications they may have.”

Imke Casey and Thorsten Haas

Another initiative in development by the PBM team is a preoperative anemia screening program, led by Thorsten Haas, M.D., director of PBM, who received the W. Martin Smith Interdisciplinary Patient Safety Award for the project. “One of the goals of PBM is to identify and correct preoperative anemia in patients who are scheduled for elective surgery,” says Casey. “The preoperative anemia screening program will help us to understand how many UF Health patients have preoperative anemia and how much blood they need. This information will help us to improve patient care. The data that we will collect will give us a comprehensive understanding of the UF Health patient population with preoperative anemia and consequent blood product utilization.”

According to Tighe, “We are looking at how we can use artificial intelligence to predict patients who might be anemic before surgery, who might need a blood transfusion unexpectedly, or maybe are at extremely low risk of ever needing a blood transfusion for the given surgery so we can make better decisions about the labs we might need to order or decisions we might need to make during surgery with regards to conserving blood, administering IV fluid, and making the decision to transfuse blood products when necessary.”

Clinical Decision Support

Making effective decisions in clinical settings is critical, and AI technologies hold immense promise for supporting health care teams in their clinical decision-making. Tighe notes, “We’re advancing the application and science of artificial intelligence with health care through multiple prongs of the academic mission. We’re not just researching it; we’re training up in it, we’re applying it to quality improvement activities, and we’re working towards applying it to clinical practice as well.”

Chris Giordano, MD

“I am most excited about the clinical decision support that anesthesiologists will be able to tap into as AI becomes more refined and robust,” says Chris Giordano, M.D. “I haven’t met an anesthesiologist who wouldn’t like to have more information at their fingertips to help better understand situations and likelihoods. Furthermore, being able to have information quickly harmonized, curated, and called upon will substantially reduce our screen time so that we can spend more time with our patients, and working through possibilities of precision-based health care: what’s best for this patient, at this time, under these circumstances.”

Educational Initiatives

Educating health care professionals on emerging AI technologies is critical to providing a foundation for the successes ahead. Dedicated teamwork within the Department of Anesthesiology is helping pave the way. We’ll take a look at these areas in part two of “AI-Powered Anesthesiology: The Future of Patient Care.”

Check out part two – AI-Enhanced Anesthesiology: Educating Future Health Care Providers

Check out opportunities to join our faculty.