Antibiotic Alerts: Building Better Processes to Encourage Stewardship

In the battle against resistant infections, response efforts have been focused on developing and deploying new tools to help reduce antimicrobial use. It is estimated that roughly 50% of antibiotic prescriptions in hospital and outpatient settings in the United States are unnecessary or inappropriate. Therefore, any tool that can enhance antimicrobial stewardship is a welcome addition to the toolkit.

Given these startling numbers, it’s not surprising that many hospitals are looking to more automatic hard-stops to prevent the misuse of antibiotics. Mercy Hospital in St. Louis, Missouri, sought to make this a reality by developing and implementing an automatic antibiotic time-out alert that would de-escalate broad-spectrum antibiotics. A new study published in Infection Control & Hospital Epidemiology details the program.

The 1252-bed community hospital worked to develop this automatic approach because, like so many of us working in infection prevention, they saw that despite education, efforts to de-escalate broad-spectrum antibiotics were rolled out inconsistently. The research team defined the outcome as the proportion of patients who had their broad-spectrum antibiotics de-escalated at 72 hours in the year prior to the initiation of the antibiotic time-out alert that was developed in 2016. Furthermore, they assessed the total antibiotic days, cost per day, hospital length of stay, antibiotic-related adverse events, and in-hospital mortality of patients whose antibiotics were de-escalated versus those who continued treatment with broad-spectrum antibiotics.

The goal was simply to measure the impact of the de-escalation alert and determine whether patient outcomes were also affected. By reviewing medical records and excluding patients who had antibiotic use prior to hospitalization, the researchers were able to include hundreds of patients in their assessment.

Of the patients assessed, 113 fell into the pre-alert group and 107 in the post-alert group. A total of 91 of the 107 post-alert group patients were eligible for de-escalation. Ultimately, chart-review found that among the patients eligible for de-escalation, 35.1% of the pre-alert group and 55% of the post-alert group actually had their broad-spectrum agents de-escalated. The investigators noted that “optimal antibiotic de-escalation in the pre-alert group was 13.4% compared to the 31.9% in the post-alert group.”

In terms of secondary outcomes, investigators found that those patients whose antibiotics were de-escalated not only had fewer antibiotic days, but shorter lengths of stay, and also less antibiotic-related adverse events such as allergic reactions or incidences of Clostridioides difficile. The authors did note that there was no significant change in the antibiotic cost-per-day in those 2 patient groups.

This study is particularly helpful in assessing the efficacy of building alerts into electronic medical records systems to help guide best practices through “best practice alerts” or BPAs. In this case, the alerts helped guide the de-escalation of broad-spectrum antibiotics and encouraged better antibiotic stewardship. As health care and medicine become increasingly complex, built-in logic alerts can help guide physicians and reduce errors and/or unnecessary antibiotics.

The challenge, though, comes in the debate of moving so much critical thinking to more of a task-driven approach. Are these built-in alerts removing the critical thinking from medicine? Or is this a burden we can help take off the shoulders of medical providers? For many, the potential reduction of antibiotic usage and subsequent antibiotic resistance is well worth it.