Pneumonia and urinary tract infection (UTI) are the two most common infection-related conditions for which US adults are hospitalized annually. And at least 40% of the time, they receive unnecessarily broad-spectrum antibiotics on admission to treat these infections.
The reason many physicians prescribe broad- or extended-spectrum antibiotics to treat pneumonia and UTIs in the 2 to 3 days before they have bacterial culture results is because of fear their patients might be infected with a multidrug-resistant organism (MDRO) and that their condition could deteriorate without an effective antibiotic.
"It's a time of great uncertainty for physicians," Shruti Gohil, MD, MPH, an assistant professor of infectious disease and associate medical director of epidemiology and infection prevention at the University of California Irvine School of Medicine, told CIDRAP News. "We're nervous. We want the patient to get better."
If the culture results show that the pathogen that caused the infection isn't an MDRO, the physician can then de-escalate to a standard-spectrum antibiotic. Still, that brief exposure to extended-spectrum antibiotics could increase the patient's risk for a future MDRO and antibiotic-associated adverse effects like Clostridioides difficile.
But the results of two large randomized clinical trials led by Gohil, published today in JAMA, suggest that a computerized stewardship prompt in a hospital system's electronic health record (EHR) could help physicians feel more confident about prescribing a standard-spectrum antibiotic for pneumonia and UTI patients from the get-go. And she believes the impact on antibiotic stewardship and patient safety could be significant.
Assessing patients' MDRO risk
The two INSPIRE (Intelligent Stewardship Prompts to Improve Real-Time Empiric Antibiotic Selection) trials involved adult patients hospitalized with pneumonia and UTI at 59 hospitals within HCA Healthcare, the largest private community hospital system in the United States. Both trials had an 18-month baseline period and a 15-month intervention period and were conducted from April 2017 through June 2020.
In both trials, one group of hospitals was randomly assigned to routine antibiotic stewardship activities, which included providing hospital guidelines and protocols for antibiotic selection, requiring a documented reason for antibiotics, and prospectively evaluating antibiotic use with clinician feedback to de-escalate antibiotics after the return of microbiologic results. Hospital staff received educational materials and quarterly coaching calls to maintain these activities.
Physicians in the other groups of hospitals, in addition to practicing the routine stewardship activities, received computerized provider order entry (CPOE) prompts whenever extended-spectrum antibiotics were ordered in a non-intensive care unit (ICU) setting for patients with pneumonia or UTI within 72 hours of admission (the empiric prescribing period). The prompt was triggered when the CPOE algorithm determined that the patient's risk for an MDRO was 10% or less and standard-spectrum antibiotics could be used safely. The CPOE hospitals also received clinician education and feedback reports.
The CPOE algorithm, Gohil explained, uses a host of variables to assess a patient's risk for having an MDRO infection. It takes into account the individual patient's risk for a specific MDRO, the type of infectious syndrome being treated, and the antibiotic selected by the physicians.
"So what the physician will get is a risk estimate that tells them—for the specific antibiotic they have chosen—whether or not the patient has the antibiotic-resistant organism that the physician is worried about," she said. "If it's a low-risk patient, then and only then will the physician see the prompt."
Sizable reductions in extended-spectrum therapy
Gohil said the idea behind the trials was to find a way to help physicians choose antibiotics that are as targeted toward a patient as possible. The primary outcome was extended-spectrum days of therapy in a non-ICU location in the first 3 calendar days of hospitalization.
In both trials, the CPOE bundle was associated with significant reductions in extended-spectrum therapy. In the INSPIRE Pneumonia trial, which included 96,451 patients (51,671 during the baseline and 44,780 during the intervention period), the 29 hospitals assigned to the CPOE arm saw a 28.4% reduction in empiric extended-spectrum days of therapy from the baseline compared with the 30 hospitals assigned to the routine stewardship group (rate ratio [RR], 0.72; 95% confidence interval [CI], 0.66 to 0.78).
In the INSPIRE UTI trial, which included 127,403 patients (71,991 during the baseline and 55,412 during the intervention period), the 29 hospitals using the CPOE prompts saw a 17.4% reduction in empiric extended-spectrum days of therapy from baseline compared with the 30 routine stewardship hospitals (RR, 0.83; 95% CI, 0.77 to 0.89).
"This is a pretty sizable reduction," Gohil said. "It just goes to show you that if you give physicians the right information at the right time, you can really inform their prescribing habits."
In both trials, the reductions were evident within the first 3 months of the prompts going live. Similar reductions were seen for use of vancomycin and antipseudomonal antibiotics, which are frequently prescribed at admission for pneumonia and UTI. And the reduction in empiric extended-spectrum antibiotic prescribing was maintained during the initial months of the COVID-19 pandemic—a time when many hospitals were seeing an increase.
It just goes to show you that if you give physicians the right information at the right time, you can really inform their prescribing habits.
Gohil and her fellow investigators also found, in both trials, that the reduction in extended-spectrum antibiotic prescribing over time was driven in part by physicians deciding to prescribe a standard-spectrum antibiotic before they even received the prompt.
"There was a reduction in the initial [antibiotic] selection such that the prompt had to fire less," she added.
Furthermore, analysis of the primary safety outcomes in both trials indicated that reducing empiric extended-spectrum antibiotics for pneumonia and UTI patients was safe. In INSPIRE Pneumonia, mean days to ICU transfer (6.5 vs 7.1 days) and length of hospital stay (6.8 vs 7.1 days) were not significantly different between the hospitals in the CPOE arm and those in the routine stewardship arm. The results for INSPIRE UTI were similar.
Changing paradigms
Gohil said the findings are important because many antibiotic stewardship interventions in hospitals are targeted toward reducing antibiotic duration or narrowing the spectrum of antibiotics once there is more information about the causative pathogen. But the INSPIRE trials are one of the largest efforts to take on empiric prescribing.
"This work really flips the paradigm more towards early initiation of standard-spectrum antibiotics, as opposed to the current antibiotic stewardship strategy, which heavily relies on starting broad, then narrowing later," she said.
In an accompanying editorial, infectious disease experts Anurag Malani, MD, of Trinity Michigan Health, and Preeti Malani, MD, of the University Michigan, say they hope other institutions will take note and consider similar EHR-based interventions to improve empiric prescribing.
"The aptly named INSPIRE trials do just that—provide inspiration and imagination, along with a powerful paradigm to harness the EHR to optimize antibiotic prescribing and improve human health," they write. "Rigorous studies that build on the successes reported by Gohil and colleagues are urgently needed."
