New framework offers closer look at patterns in hospital antibiotic use
Hospitals across the country have programs dedicated to reducing unnecessary antibiotic use. However, determining how well those programs are working is not always straightforward.
Researchers from the University of Iowa and Iowa City Veterans Affairs Health Care System have developed a new way to look more closely at how hospitals prescribe antibiotics. And, importantly, what may be driving differences in antibiotic use between hospitals.
The study, funded by the VA Office of Research & Development, was led by Michihiko Goto, MD, MS, associate professor in Infectious Diseases, alongside Daniel Livorsi, MD, associate professor in the division, and Shungo Imai, MD, PhD, who contributed to the work during his time as a postdoctoral research scholar at Iowa. Imai, an associate professor at the Keio University Faculty of Pharmacy in Japan, has joined the Division of Infectious Diseases as an adjunct assistant professor.
Published in Clinical Infectious Diseases, the study introduces a framework to evaluate three decisions that shape antibiotic prescribing: whether treatment is initiated, how long it continues, and whether the selected antibiotic has a broad or narrow spectrum. The researchers call these components Start, Stop, and Select, or S3. The team applied the framework to data from all 118 acute-care hospitals in the Veterans Health Administration, covering more than 700,000 patients, to demonstrate how this novel framework can be applied in the real world.
“Every hospital is required to have a program focused on antimicrobial stewardship, or reducing unnecessary antimicrobial use,” Livorsi said. “However, it’s very difficult to assess how well a hospital is doing with its stewardship.”
This matters because antibiotic use comes with tradeoffs. Antibiotics can be lifesaving, but unnecessary or excessive use contributes to antimicrobial resistance. Antimicrobial stewardship programs work to preserve the effectiveness of these drugs by helping clinicians use them only when needed, for an appropriate length of time, and with an appropriate spectrum.
Pinpointing patterns in antibiotic use
Existing measures can tell a hospital how much antibiotics it is using, but that number does not necessarily explain what drives the use.
“An existing metric may simply show that a hospital is prescribing antibiotics in large volumes,” Livorsi said. “The S3 metric tries to pinpoint the areas of overuse by assessing the likelihood to initiate antibiotics, the duration of antibiotics, and the spectrum of antibiotics.”
For example, a hospital might start antibiotics more frequently than expected but generally prescribe them for an appropriate duration. Another could start treatment less often but favor broader-spectrum antibiotics. Those patterns could call for very different responses from their stewardship teams.
The framework also accounts for differences in the patients treated at each hospital. A facility caring for a larger proportion of seriously ill patients with more complicated infections may have legitimate reasons for using more antibiotics. Livorsi said that to facilitate fair inter-hospital comparisons, it is important to adjust for differences in patient populations across facilities.
Expanding the framework internationally
Imai came to Iowa to learn how antimicrobial stewardship research could be conducted using the VA’s large, integrated health care data. Working with Goto, Livorsi, and other Iowa investigators also changed how he thought about developing new measures for clinical research.
“Through my work with Dr. Goto and Dr. Livorsi, I learned that developing a new metric is only the first step,” Imai said. “It must also be carefully tested to determine whether it truly reflects prescribing behavior and whether it can support meaningful clinical and stewardship decisions.”
That distinction remains important for S3, because the framework cannot determine on its own that a hospital is prescribing antibiotics inappropriately. Instead, it can point stewardship teams toward an area worth examining.
“The framework has the potential to connect measurement directly to practice,” Imai added. “Rather than simply telling a hospital that its antimicrobial use is high or low, it can help indicate which type of prescribing decision may warrant attention and which type of intervention may be most appropriate.”
Imai returned to Japan after his postdoctoral training but has maintained his connection to Iowa. In August, he began an appointment as an adjunct assistant professor in the Division of Infectious Diseases.
The collaboration has also pointed to the next step for this research. Imai and the Iowa team hope to evaluate and adapt the framework using Japanese health care data, testing whether an approach developed with VA data can work in a substantially different health care system.
“The work began while I was in Iowa, but it is continuing after my return to Japan,” Imai said. “By combining the University of Iowa team’s expertise in large-scale Veterans Affairs data and antimicrobial stewardship research with experience from the Japanese healthcare system, I hope we can further develop and validate this framework in Japan as an important first step toward broader international application.”