Introduction
It has been estimated that 85%–95% of all antibiotic use in industrialized countries occurs in the community, with as much as 50% being inappropriate, making the ambulatory setting a crucial target for antimicrobial stewardship program (ASP) efforts. Reference Duffy, Ritchie, Metcalfe, Van Bakel and Thomas1 Antibiotic risks include adverse drug events, Clostridioides difficile infections, and antimicrobial resistance development. Antimicrobial resistance alone is forecasted to cause over 39 million deaths between 2025 and 2050. Reference Sanchez, Fleming-Dutra, Roberts and Hicks2,Reference Naghavi, Vollset and Ikuta3 For many common infections, shorter antibiotic courses are as effective as longer courses, and promoting use of the shortest effective antibiotic duration is an important ASP intervention. Reference Spellberg4,Reference McCormick, Cardwell, Wheelock, Wong and Vander Weide5
Decisions surrounding antibiotic durations are driven by multiple factors, including decision support systems and default settings within the electronic health record (EHR). Default medication entry selections for outpatient prescriptions orders, including duration of therapy, aim to reduce cognitive burden but may also inadvertently steer prescribers to inappropriate durations. Reference Wrenn, Slaton and Diez6–Reference Jachimowicz, Duncan, Weber and Johnson8 These default settings may be incorporated into provider-specific or department-specific EHR preference lists, intended to increase the efficiency of prescription order entry for the most commonly placed prescription orders.
During the ambulatory ASP intervention process targeting cystitis, a system-wide evaluation of prescribing revealed cystitis patient encounters in the emergency department (ED) resulted in outpatient antibiotic prescriptions for longer durations relative to other outpatient practice settings (Figure 1). The ASP team noted that a preference list routinely utilized by ED providers contained 26 antimicrobials with defaulted durations, most of which were defaulted to 10 days. We hypothesized that removal of default durations on the ED discharge medications preference list would result in an overall decrease in the average duration of antibiotic therapy for prescriptions originating in the ED.
Percentage of Mayo Clinic cystitis encounters between 1/1/21 and 4/20/23 with antibiotic prescriptions issued for durations of therapy >7 days, by department specialty.
Methods
This pre-/post-quasi-experimental study included a pre-intervention period from January 1, 2023, to March 31, 2023, and a post-intervention period from January 1, 2024, to March 31, 2024, at ED sites within the Mayo Clinic Enterprise. The Mayo Clinic Enterprise consists of three major destination medical centers in Rochester (Minnesota), Arizona, and Florida, along with the Mayo Clinic Health System, a network of hospitals and clinics throughout 4 regions in Minnesota and Wisconsin. Default durations were removed from all antibiotics on the ED discharge medications preference list on November 16, 2023. The institutional outpatient ASP structure was active before and during the study periods with multifaceted interventions targeting upper respiratory tract infections (URIs) and cystitis. Reference Ilges, Jensen and Draper9,Reference Stevens, Manz and Mathre10 These multifaceted interventions specifically targeting antibiotic duration in cystitis included prepopulated ambulatory order panels, provider education, and peer comparison data. Interventions targeted at URI focused on the elimination of unnecessary antibiotic prescriptions for viral URI rather than on antibiotic duration of therapy.
All ambulatory prescriptions for an oral antibiotic included on the ED discharge medications preference list were included. Prescriptions for antivirals, antifungals, and dose packs (pre-packaged antimicrobial courses) were excluded. Liquid formulations were also excluded given inability to accurately determine the duration of therapy based on package size. Data were extracted from the EHR using Epic SlicerDicer (Epic, Verona, QI) to determine the mean duration of therapy and percentage of all antimicrobial prescriptions issued from the ED with durations of therapy of ≤ 3 days, 4–5 days, 6–7 days, 8–10 days, 11–14 days, or 14–29 days. Prescriptions issued for >29 days were excluded to eliminate capture of antimicrobials intended for prophylactic or suppressive use. Outcomes were analyzed using an independent t test and Pearson’s X 2 test with p-values of <0.05 considered statistically significant.
Results
A total of 21,678 encounters with antimicrobial prescriptions issued were evaluated, with 10,351 prescriptions (83,843 days of therapy) and 11,327 prescriptions (82,687 days of therapy) in the pre-intervention and post-intervention periods, respectively. In total, 18,686 (86.2%) of antimicrobial prescriptions originated from the targeted preference list. A significant reduction in the mean duration of therapy for prescriptions was observed following removal of default durations from the ED preference list (8.1 vs 7.3 days; P < 0.001). A subgroup analysis of orders only originated from the preference list showed similar results (8.1 vs 7.3 days, P < 0.001). This decrease in mean duration of therapy corresponds to an elimination of approximately 9,061 days of antibiotic therapy over the 3-month period.
There were significant increases in the percentage of prescriptions written for a duration of 4–5 days and 6–7 days, with a corresponding, and significant, decrease in the percentage of prescriptions written for a duration of 8–10 days (Figure 2). An analysis of prescriptions written for exactly 10 days, the previous default, revealed 43.6% pre-intervention and 24.8% post-intervention (P < 0.001). The percentage of prescriptions written for a duration of ≤ 3 days, 11–14 days, and 14–29 days were not significantly different pre- and post-intervention (Figure 2).
Frequency of prescription durations.
Discussion
This study adds to existing literature that default settings in the EHR influence prescribing behaviors. Reference Jia and Linder7 In one health system, modification of defaulted durations of therapy for antibiotics used for urinary tract infections led to an overall increase in guideline-based treatment adherence. Reference McCormick, Cardwell, Wheelock, Wong and Vander Weide5 In another study, removal of default durations from prescriptions for fluoroquinolones failed to result in an overall decrease in mean days of therapy but was associated with a 1-day reduction in median antimicrobial durations during hospital discharge transition of care. Reference Wrenn, Slaton and Diez6
In the year preceding this study, our ED prescribers chose the default of 10 days for antibiotic prescriptions nearly half of the time, despite guideline recommendations for shorter durations. Although providers were able to modify the default durations within the preference list, there was no prompt promoting active decision-making. Despite implementation of an ASP intervention bundle, removal of default durations from the routinely used preference list was the single most effective intervention for reducing antimicrobial durations of therapy in the ED. Our team carefully considered implications of reducing the default durations of therapy (ie changing to 3, 5 or 7 days) versus removing the defaults all together. However, appropriate durations are influenced by several factors including syndrome, severity, and individual patient characteristics; therefore, it was felt that any widely applied, non-syndrome-specific standard default duration may continue to result in guideline discordant durations.
There are several study limitations, most notably retrospective design and lack of randomization. We did not control for confounding variables or compare patient characteristics in the pre- and post-intervention groups. Additionally, the ASP bundle implemented prior to and concurrently with this intervention likely also had an impact on prescriber decisions regarding antimicrobial durations, specifically in cystitis. An interrupted time series analysis may have been more effective in assessing secular trends in prescribing over time, perhaps accounting for the influence of other ASP bundle interventions. Safety variables such as repeat ED encounters or other repeat healthcare contact and the concordance of prescribed durations with available guideline recommendations were also not examined. Nevertheless, this study highlights the impact of defaulted durations of therapy and the importance of carefully considering default settings within the EHR prior to implementation. This relatively easy intervention had a large impact and could be considered for implementation by other health system ASPs.
Supplementary material
The supplementary material for this article can be found at https://doi.org/10.1017/ice.2025.31
Acknowledgments
We thank all members of the Mayo Clinic Enterprise Antimicrobial Stewardship Program for their contributions to the ambulatory ASP programmatic development for the Mayo Clinic Enterprise, making the detailing of this study possible.
Financial support
No financial support was provided relevant to this study.
Competing interests
All authors report no conflicts of interest relevant to this study.
