AI in Infection Prevention: Promise and Pitfalls

AI-driven staffing models are being considered to optimize infection prevention workloads, and experts see both potential and significant limitations.

In this installment of their discussion with Infection Control Today^® (ICT^®), authors of the study, “Quantifying the Progressing Landscape of Infection Preventionists: A Survey-Based Analysis of Workload and Resource Needs” shared their insights on how AI could be used in infection prevention efforts.

The authors are Brenna Doran, PhD, MA, who specializes in hospital epidemiology and infection prevention at the University of California, San Francisco, and is a coach and consultant in infection prevention; Jessica Swain, MBA, MLT, director of infection prevention and control at Dartmouth Health in Lebanon, New Hampshire; and Shanina Knighton, an associate professor at Case Western Reserve University School of Nursing and senior nurse scientist at MetroHealth System in Cleveland, Ohio.

“It can help with forecasting infection trends, which will help allow for proactive staff and adjustments,” Knighton explained. “It can help efficiently distributing tasks based on real-time data. It can help hospitals be able to analyze and predict outbreaks before they happen, which could help IPs act earlier. It can help hospitals decide where to place infection preventionists to have the biggest impact.”

Complex facilities may benefit most. “Do you have a transplant center? Do you have a cancer center? Do you have a burn unit?” Swain said. “So this might be a potential way to help in those more complex facilities as well to make sure that those specific areas are getting the coverage that they need, and we're not missing things around those more complex areas.”

Yet limitations remain. “We're not at the table even thinking through how AI is fitting into our work, because we're seeing it happen, and it's here, whether we want to embrace it or not,” Knighton said. “Hospitals are utilizing it. It's all through the EHRs. It's everywhere. And so if we're not able to understand how things are working… it's going to make it very challenging for us to do a lot of tracking in our work.”

Ultimately, as Doran noted, “we have IPs that are regularly working 45, 50, 55, 60 hours a week, and that is baseline… AI may provide insights, but it cannot yet capture the unpredictability and invisible workload that defines infection prevention.”

This is the fifth installment of this interview. The first can be found here. The second can be found here. The third can be found here. The fourth can be found here.

Reference

Doran B, Swain J, Knighton S. Quantifying the progressing landscape of infection preventionists: A survey-based analysis of workload and resource needs. Am J Infect Control. 2025;53(6):669-677. doi:10.1016/j.ajic.2025.03.012