Guidance document development

Interested parties with relevant expertise across the VA were identified and recruited for participation (by D.J.L.). The 15 respondents, as coauthors, used an iterative process to identify research targets to include in this report. First, we presented the goals of this draft research agenda to the VA’s Antimicrobial Stewardship Taskforce’s monthly educational webinar audience (298 attendees) and the VA’s antimicrobial stewardship listserv (552 members) in May 2023. Using an electronic questionnaire, we sought feedback from these audiences about (1) which healthcare settings should be the focus of stewardship research and (2) which specific knowledge gaps are most important to address. In total, 40 attendees responded to this questionnaire. Second, we held virtual meetings of the coauthor group (May 1, 2023, May 18, 2023, July 7, 2023, and August 31, 2023) to discuss key knowledge gaps that should be highlighted.

To create an initial draft of the 20 most important research targets, 2 authors (D.J.L. and R.L.J.) combined feedback from these 2 sources. The coauthors achieved consensus through multiple rounds of written review and discussion, which included transforming the knowledge gaps into questions, ranking them from most to least important, and then considering those questions in the context of different healthcare settings: acute-care hospitals, outpatient care, long-term care (including spinal cord injury and disorders), periprocedural and surgical settings, dental clinics, and veterans receiving care in the community.

Research agenda

Four research domains emerged during the synthesis and subsequent ranking of the knowledge gaps: stewardship strategies, clinical evidence, dissemination and implementation, and measurement (Table 1). Some questions applied to >1 domain and some were interconnected. The 3 highest-ranked research questions were prioritized for discussion below. We also addressed research questions specific to the different healthcare settings detailed above.

Table 1. Twenty Important Knowledge Gaps Related to Antibiotic Stewardship

Acute-care settings

The acute-care setting was an early target of VA antibiotic stewardship activities, and studies of several inpatient stewardship interventions have demonstrated tangible improvements in the selection, spectrum, duration, and proper dosing of antibiotic therapy. ^{Reference Kelly, Jones and Echevarria7,Reference Graber, Jones and Goetz10,Reference Suda, Clore and Evans28–Reference Grigoryan, Naik and Lichtenberger30} Meta-analyses and systemic reviews, including one from the VA Evidence Synthesis Program (ESP), have demonstrated that inpatient stewardship activities can safely improve antibiotic-prescribing and reduce antibiotic resistance. ^{Reference Wagner, Filice and Drekonja31–Reference Baur, Gladstone and Burkert33} However, limited evidence is available to inform the staffing and information technology resources essential for the success of hospital-based antibiotic stewardship programs, including how these resources should vary by facility size and complexity. Contributing to this knowledge gap is a lack of detail on resources needed to implement successful interventions that have been described in the literature. Future implementation studies should quantify not only outcomes but also resources needed for implementation in an objective manner, including time, expertise and informatic resources [Question 10].

In hospitals that lack on-site infectious disease (ID) physicians and ID pharmacists, telehealth is a promising tool for stewardship, and the VA has specific expertise in how to use telehealth for this purpose. ^{Reference Livorsi, Abdel-Massih and Crnich34} Strategies to advance the implementation and sustainment of telehealth-supported stewardship activities in resource-limited, acute-care settings are needed [Question 9]. ^{Reference Livorsi, Abdel-Massih and Crnich34}

Future studies, preferably randomized-controlled clinical trials, could further inform several targets for acute-care stewardship activities. First, while oral antibiotic therapy is gaining greater acceptance for infections traditionally treated with parenteral therapy (eg, osteomyelitis), there is room to better define the optimal selection and dosing of oral agents [Question 3]. Second, more data on the comparative safety profiles of commonly prescribed antibiotic regimens would be valuable. A recent pragmatic trial comparing cefepime and piperacillin-tazobactam may serve as a viable model for future investigations. ^{Reference Qian, Casey and Wright35} The VA Cooperative Studies Program research infrastructure and clinical trials networks have demonstrated success in the conduct of point-of-care and pragmatic clinical trials. ^{Reference Ishani, Cushman and Leatherman36–Reference Branch-Elliman, Ferguson and Doros38} Thus, the VA is an ideal setting for pursuing similar pragmatic trials designed to compare different standard-of-care antibiotic options. Third, large clinical trials could help identify specific situations where it is safe to decrease the length of recommended antibiotic therapy for common uncomplicated infections [Question 4]. For example, in cases of pneumonia, tools such as multiple polymerase chain reaction panels could improve identification of viral pneumonia and, in turn, identify patients in whom early discontinuation or avoidance of antibiotics is feasible.

One performance measure of acute-care antibiotic stewardship programs is the standardized antimicrobial administration ratio, which compares observed to expected antibiotic use at the ward and hospital-level without accounting for patient-level differences in case-mix across facilities. Although these types of gross measures may have a role in facilitating interhospital comparisons, more precise risk adjustment is necessary [Question 18]. Actionable metrics would be even more useful, such as those that quantify appropriate or optimal antibiotic use rather than overall use [Question 15]. To this end, there is a need for electronic quality indicators based on accepted clinical guidelines for common inpatient conditions (eg, pneumonia, skin and soft-tissue infection, and UTIs) [Question 20]. How these measures can be made valid, reliable, feasible, well calibrated, and interoperable across varying electronic health records, patient populations, and systems of care will be key challenges. Standardized, validated, patient-centered outcomes that measure the clinical impact of antibiotic stewardship activities would provide additional opportunities to monitor the effect of these programs, which may include reductions in length of stay and decreased hospital readmissions [Questions 14 and 16].

In summary, future antibiotic stewardship research in the acute-care setting should focus on (1) quantifying the amount of program resources necessary to achieve a measurable effect on antibiotic use; (2) building the evidence-base for using oral over parenteral therapy, comparing the safety of different antibiotic agents, and identifying opportunities to shorten duration of therapy for common infections; and (3) creating electronic quality indicators that measure guideline-concordant antibiotic use and patient-centered outcomes.

Outpatient settings

Outpatient antibiotic stewardship encompasses efforts to measure and improve antibiotic-prescribing in primary care clinics, urgent-care clinics, emergency departments, and home-based primary care settings. Outpatient settings require different solutions than those used in acute-care hospitals due to different care models and workflows. Specific challenges to stewardship in these settings include (1) the time pressure faced by clinicians in emergency departments, urgent care, and primary care clinics; (2) competing priorities of patients and clinicians in a time-limited clinical encounter [Question 11]; and (3) the lack of accepted metrics that resonate with all stakeholders [Questions 14 and 15].

Findings from several systematic reviews, including the VA ESP, suggest that outpatient stewardship activities improve antibiotic-prescribing without short-term negative outcomes. ^{Reference Drekonja, Filice and Greer17} Nevertheless, several knowledge gaps remain. First, few studies address the influence of outpatient stewardship activities on antibiotic resistance [Question 19] or on patient-specific outcomes [Question 14], such as antibiotic-related harm [Question 17] and microbiome effects [Question 2]. ^{Reference Drekonja, Filice and Greer17} Second, although studies have shown the efficacy of audit and feedback of prescription data, clinician focused education, and interventions embedded within the electronic health record (eg, order sets), which of these strategies is most effective in the outpatient setting remains unclear. Third, most outpatient stewardship studies have addressed antibiotic prescribing for acute respiratory tract infections and have not considered other common reasons for outpatient antibiotic-prescribing, such as for diarrhea, UTIs, or skin and soft-tissue infections. Fourth, both sustainability and scalability of outpatient stewardship activities have rarely been addressed. ^{Reference Drekonja, Filice and Greer17,Reference Frost, Hersh and Hyun39,Reference Sadeq, Hasan and AbouKhater40} Finally, even within the VA, a pioneer in telehealth, data describing the influence of telehealth on outpatient antibiotic prescriptions are limited. ^{Reference Jump, Mongilardi and Wilson41}

Given these knowledge gaps, we propose the following areas for future antibiotic stewardship research in outpatient settings: (1) adapting strategies that apply the successes around antibiotic-prescribing for acute respiratory illnesses to other common outpatient infections, such as UTIs and skin and soft-tissue infections; (2) defining stewardship strategies that address the unique diagnostic and communication challenges posed by the increasing use of telehealth; (3) enhancing strategies to mitigate patients’ real and perceived antibiotic-seeking behavior for unnecessary indications; and (4) developing stewardship metrics that are tailored to and will resonate with clinicians practicing in outpatient settings.

Long-term care and spinal cord injury units

Both residents in long-term care (LTC) facilities and persons with spinal cord injury/disorder (SCI/D) are special populations with a high frequency of antibiotic use and an increased risk for acquiring antibiotic-resistant organisms. ^{Reference Fitzpatrick, Suda and Safdar42,Reference Daneman, Bronskill and Gruneir43} Antibiotic decision-making in these populations is often complicated by difficulty ascertaining symptoms and limited access to diagnostic tests.

Although templates exist for establishing antibiotic stewardship programs in LTC, little guidance exists for SCI/D and evidence remains limited overall on optimal structure, strategies, implementation, and metrics. ^{44–Reference Jump, Gaur and Katz46} Most published antibiotic stewardship interventions in LTC have used education and guideline dissemination. ^{Reference Crespo-Rivas, Guisado-Gil and Penalva18,Reference Grigoryan, Naik and Lichtenberger30,Reference Katz, Tamma and Cosgrove47} The benefit of other stewardship strategies remains unclear in LTC, where the feasibility and effectiveness of any strategy may be hampered by lack of on-site ID or pharmacy expertise, lack of advanced diagnostic tests, and increased pressure from patients and caregivers to prescribe antibiotics [Question 11]. Telehealth offers great potential for more rapid and sustainable stewardship implementation in LTC facilities lacking on-site ID-trained physicians and pharmacists, but only a handful of telehealth studies have included LTC settings and none evaluated SCI/D populations. ^{Reference Wilson, Banks and Crnich48–Reference Livorsi, Sherlock and Cunningham Goedken50} Only a few stewardship strategies have been evaluated specifically in SCI/D populations, with studies limited by small size and short follow-up periods. ^{Reference Alnajjar, Alrashidi and Almutairi51,Reference Patros, Sabol, Paniagua and Lans52}

Nonlocalizing signs and symptoms are common in older adults as well as in persons with SCI/D. Vague signs and symptoms are common triggers for inappropriate antibiotic prescribing, particularly for suspected UTIs. Although guidance exists on which nonlocalizing signs and symptoms should prompt evaluation for infection and empiric antibiotics in LTC residents, effective implementation of these recommendations remains an ongoing challenge. ^{Reference Loeb, Bentley and Bradley53–Reference Nicolle, Gupta and Bradley55} Differentiation of asymptomatic bacteriuria (ASB) from UTI is especially challenging in people with SCI/D, who have both a high prevalence of ASB and frequent nonlocalizing symptoms and/or urinary symptoms that are not necessarily related to infection. ^{Reference Wirth, Suda and Burns56,Reference Okamoto, Prieto and Avery57} Previous prospective research studies have demonstrated the value of standardized symptom assessments and infection definitions for ASB and UTI to support antibiotic stewardship in LTC. ^{Reference Trautner, Grigoryan and Petersen25,Reference Grigoryan, Naik and Lichtenberger30,Reference Nace, Hanlon and Crnich58} This work may serve as a foundation for similar prospective studies in people with SCI/D.

Given the dedicated and integrated care models for veterans in LTC and those with SCI/D, the VA is an ideal setting in which to study antibiotic stewardship for these populations. Future antibiotic stewardship research in LTC and SCI/D populations should (1) further evaluate the use of telehealth for implementing stewardship practices; (2) adapt traditional and newer stewardship metrics to LTC and SCI/D settings; and (3) develop validated tools to help distinguish ASB from UTI in patients with nonlocalizing symptoms.

Other settings