The results from two newly published randomized clinical trials provide more evidence that a computerized stewardship alert embedded in a hospital's electronic health record can safely reduce empiric use of extended-spectrum antibiotics in patients at low risk for multidrug-resistant infections.
The two INSPIRE (Intelligent Stewardship Prompts to Improve Real-Time Empiric Antibiotic Selection) trials, led by researchers with the University of California Irvine School of Medicine and the Harvard Pilgrim Healthcare Institute, found that computerized provider order entry (CPOE) prompts based on patient-specific risk factors were associated with a 28% and 35% relative reduction in the use of empiric extended-spectrum antibiotics in non–critically ill patients with skin and abdominal infections, respectively, compared with routine antibiotic stewardship strategies. And in both studies, reduced use of broader-spectrum antibiotics did not result in longer hospital stays or different intensive care unit (ICU) transfer rates.
The idea behind the CPOE prompts, which rely on an algorithm that assesses a patient's risk for a multidrug-resistant organism (MDRO), is to reduce unnecessary broad-spectrum antibiotic use and help clinicians choose antibiotics that are as targeted toward a patient as possible.
In two INSPIRE trials published in April 2024, the same team reported similar reductions in extended-spectrum antibiotic prescribing for patients with pneumonia and urinary tract infections tied to the use CPOE prompts.
Reducing extended-spectrum antibiotic overuse
In the INSPIRE 3 trial, published yesterday in JAMA Internal Medicine, the researchers evaluated whether real-time CPOE prompts could reduce empiric extended-spectrum antibiotics for non-critically ill patients who had community-acquired skin and other soft-tissue infections (SSTIs).
Although national guidelines recommend standard-spectrum antibiotics for nonpurulent or nonsurgical SSTIs, the study authors note that concerns about possible infection with methicillin-resistant Staphylococcus aureus and Pseudomonas aeruginosa result in 30% to 50% of hospitalized SSTI patients receiving extended-spectrum antibiotics.
"However, extended-spectrum antibiotic overuse can cause harm, including Clostridioides difficile colitis, allergies, or kidney and liver adverse effects," the study authors wrote.
From January 2019 through December 2023, 92 hospitals within HCA Healthcare—the largest private community hospital system in the United States—were randomized 1:1 to routine antibiotic stewardship or the CPOE bundle intervention. Routine stewardship activities included providing hospital guidelines and protocols for antibiotic selection, requiring a documented reason for antibiotics, and prospective clinician feedback to de-escalate antibiotics after 3 days.
In the hospitals assigned to the CPOE bundle group, clinicians practiced routine stewardship and also received CPOE prompts from the electronic health record when extended-spectrum antibiotics were ordered for certain non–critically ill SSTI patients. The prompt was triggered when the CPOE algorithm determined that the patient's risk for an MDRO was 10% or less and that standard-spectrum antibiotics could be used safely. The CPOE hospitals also received clinician education and feedback reports.
Comparing a 12-month baseline period and a12-month intervention period across the two strategies, the researchers assessed empiric extended-spectrum antibiotic days of therapy (DOT) per 1,000 empiric days targeting Pseudomonas and/or MDR gram-negative pathogens. To assess the safety of the intervention, they compared length of hospital stay and days to ICU transfer in both periods.
Among the 118,562 patients (median age, 58; 56.7% male) admitted with SSTIs, 60,932 were treated at 46 hospitals in the routine-stewardship group and 57,630 at 44 hospitals in the CPOE group (2 hospitals dropped out of the study). Receipt of any empiric extended-spectrum antibiotic during the baseline and intervention periods was 57.0% and 56.0%, respectively, for the routine-stewardship group compared with 55.4% and 43.0%, respectively, for the CPOE group.
Empiric extended-spectrum DOT per 1,000 empiric days was 511.5 during the baseline period and 488.7 during the intervention period in the routine-stewardship group and 496.2 and 359.1, respectively, in the CPOE bundle group (rate ratio [RR], 0.72; 95% confidence interval [CI], 0.67 to 0.79). There was no evidence of inferiority in the CPOE bundle group for mean hospital length of stay (6.4 days vs 6.5 days in the routine-stewardship group; hazard ratio [HR], 0.99; 95% CI, 0.95 to 1.04) or days to ICU transfer (6.3 days vs 6.3 days; HR, 1.14; 95% CI, 1.00 to 1.31).
Even greater reductions with abdominal infections
The reduction in empiric extended-spectrum antibiotic prescribing was even greater in the INSPIRE 4 trial, published yesterday in JAMA Surgery. In this trial, the same research team used similar methods and outcomes to evaluate the impact of CPOE prompts in patients with intrabdominal infections (IAIs). As with SSTIs, patients with IAIs are frequently prescribed extended-spectrum antibiotics, even though most can safely receive standard-spectrum drugs.
Among 198,480 patients (mean age, 60 years; 59.8% female) at 92 hospitals from January 2019 through December 2023, 101,109 were randomized to 46 hospitals in the routine-stewardship care group and 97,371 to 44 hospitals in the CPOE bundle group. Receipt of any empiric extended-spectrum antibiotics for the routine-care group was 48.2% during baseline and 50.5% during intervention compared with 47.8% and 37.6%, respectively, for the CPOE bundle group.
Empiric extended-spectrum DOT per 1,000 empiric days fell from 519.4 in the baseline to 499.6 in the intervention period in the routine-stewardship hospitals and from 518.9 to 349.6, respectively, in the CPOE bundle hospitals (RR, 0.65; 95% CI, 0.60 to 0.71). Hospital length of stay was noninferior in the CPOE bundle group (HR, 1.02; 90% CI, 0.98 to 1.08) while mean days to ICU transfer was indeterminate (HR, 1.10; 90% CI, 0.99 to 1.23).
"Inclusion of a wide variety of abdominal infections diagnoses suggests the intervention’s broad applicability to hundreds of thousands of patients who receive extended-spectrum antibiotics for abdominal infection in US hospitals annually," the study authors wrote.
What's behind the CPOE bundle's success?
In an editorial on the two studies published in JAMA Network Open, experts from University of Iowa Health Care and the Iowa City Veterans Affairs Health Care System say the findings provide valuable insights into the effectiveness of the CPOE bundle.
"In both trials, the effect of the CPOE bundle was fairly immediate," Shinya Hasegawa, MD, and Daniel Livorsi, MD, wrote. "That is, a new baseline of improved antibiotic prescribing was achieved within only a few months at the intervention sites."
The next step, they add, is to explore what's behind the bundle's success. They note that while the CPOE prompt is a key feature of the bundle, the percentage of providers who changed from an extended-spectrum to a standard-spectrum antibiotic in response to the prompt was small—9.5% in the SSTI trial and 12.2% in the IAI trial.
"This suggests that only a small proportion of the observed reductions in extended-spectrum antibiotic use can be directly attributed to the alerts," they wrote.
Hasegawa and Livorsi say several clinician- and hospital-level factors could have played a role. For example, clinicians may have been influenced simply by learning about how MDRO risk is estimated or may have adjusted their prescribing practices after receiving the CPOE prompt once or twice. Feedback from local stewardship teams may have also played a role.
At the hospital level, the use of the CPOE bundle, which was implemented by the same hospitals for both SSTIs and IAIs during the same time period, may have shifted the norms around antibiotic prescribing.
"By next exploring the why behind the bundle's success, strategies to replicate INSPIRE can be developed and tailored to a variety of hospital settings," they concluded.
