Thinking Beyond the “Core” Antibiotic Stewardship Interventions: Shifting the Onus for Appropriate Antibiotic Use from Stewardship Teams to Prescribing Clinicians Free

Abstract

United States guidance for hospital antibiotic stewardship has emphasized prospective audit and feedback and prior authorization of select antibiotics as core interventions. These remain the most common interventions implemented by stewardship programs. Although these approaches have been shown to reduce unnecessary antibiotic use, they incorrectly put the onus for appropriate antibiotic use on the stewardship team rather than the prescribing clinician. We propose that a primary focus of stewardship programs should be implementation of broader interventions that engage frontline clinicians and equip them with tools to integrate antibiotic stewardship into their own daily practice, thus reducing the need for day-to-day stewardship team oversite. We discuss a framework of broader interventions and policies that will facilitate this paradigm shift.

Over the last 5 years, there has been a dramatic scale-up of antibiotic stewardship activity across United States hospitals. In 2014, only 41% of hospitals reported meeting all 7 of the Centers for Disease Control and Prevention’s (CDC’s) Core Elements of Antibiotic Stewardship. By 2018, 85% of hospitals reported meeting the 7 Core Elements [1]. This remarkable increase is in large part due to regulatory requirements or strong recommendations for hospital antibiotic stewardship programs (ASPs) from the Veterans Health Administration, The Joint Commission, the Centers for Medicare & Medicaid Services, the Office of Rural Health Policy, and the President’s Council of Advisors on Science and Technology.

US antibiotic stewardship guidelines [2, 3] and the CDC’s core elements of hospital antibiotic stewardship [4] have emphasized prospective audit and feedback and prior authorization of select antibiotics as foundational or “core” antibiotic stewardship interventions. Although these 2 interventions have been shown to reduce unnecessary antibiotic use [3], they are not without their downsides. They are labor-intensive, require consistent time and effort by personnel with appropriate knowledge and expertise, and have anecdotally been associated with ASP team member burnout. They are often perceived as “top down” or “big brother” approaches that impinge on clinician autonomy. Furthermore, their impact in promoting sustainable prescribing changes among frontline clinicians is unclear. Perhaps most importantly, they incorrectly put the onus to ensure antibiotic use is appropriate on the stewardship team rather than the frontline clinician. One could reasonably argue that an ASP intervention to prevent or correct a suboptimal or inappropriate antibiotic order is not actually a success, but rather a failure in that it had to occur in the first place.

Currently, the majority of US stewardship programs focus on prospective audit and feedback and prior authorization [5]. While we believe these interventions are important, we propose that instead of focusing on how to do more of these interventions or the optimal approach to execute them, ASPs should consider their primary role to be the implementation of broader interventions that prevent inappropriate antibiotic use to begin with. Such interventions should aim to engage frontline clinicians and equip them with knowledge and tools to optimize their own antibiotic decision-making. This approach will ultimately reduce the need for audit and feedback and prior authorization. In short, our goal as antibiotic stewards should be to make our day-to-day oversight of antibiotic use unnecessary. We put forth a framework of interventions and policies to advance towards this goal, and discuss the current state and future directions of each. Some are evidence-based interventions currently ready for dissemination and implementation; others are future directions that need to be considered by the field.

UPDATING PRESCRIBERS ON EVIDENCE LIKELY TO IMPACT ANTIBIOTIC DECISION-MAKING

Despite the relatively rapid pace of clinically relevant data introduced into the scientific literature, it takes an average of 17 years for evidence to translate into clinical practice [6]. A major reason for delays in practice changes is a lack of awareness of the evidence. It is not uncommon for clinicians to be unfamiliar with the most up-to-date medical literature, including pivotal studies or guidelines. For example, there have been a number of robust trials conducted in the past few years, including ones examining the treatment of extended-spectrum beta-lactamase–producing bloodstream infections with piperacillin-tazobactam [7], the optimal duration of therapy for uncomplicated Gram-negative bloodstream infections [8], the role of oral therapy for endocarditis [9, 10] and bone and joint infections [11], the duration of therapy for intra-abdominal infections [12], and the role of antibiotic therapy for uncomplicated diverticulitis [13, 14]. However, it is unclear how many frontline clinicians are aware that these trials occurred and understand how the information gained from these studies can impact their daily antibiotic decision-making.

ASPs can play an important role in updating clinicians on relevant new evidence and society guidelines. Much of the evidence related to optimizing treatment regimens is published in infectious diseases journals, which other clinicians (eg, emergency medicine, intensivists, hospitalists) are unlikely to routinely access. Not following the evidence or guideline recommendations may be a result of not being aware of this information: this is something ASPs can help to rectify by developing summaries of important literature and effectively disseminating them.

STRUCTURED APPROACH TO EVALUATING ANTIBIOTIC DECISION-MAKING

To increase clinician engagement and consistency in the antibiotic decision-making process, ASPs should consider implementing structured approaches for care teams to critically assess antibiotic regimens. One such approach is the Four Moments of Antibiotic Decision Making framework, which prompts clinicians to consider specific questions at the various time points antibiotic decisions should be addressed [15]. Briefly, Moment 1 prompts the clinician to weigh whether it is more likely than not that the patient has a bacterial infection that requires antibiotic therapy. Moment 2 encourages the clinician to consider what diagnostic testing would be helpful and what empiric therapy should be prescribed, taking into consideration the likely source of infection, previous organisms and antibiotic resistance profiles, severity of illness, and known drug allergies. Moment 3 occurs after additional clinical and diagnostic data are available, and reminds the clinician to reconsider whether antibiotics are still indicated and, if so, whether the regimen needs to be narrowed or broadened and whether transition to oral therapy can occur. Finally, Moment 4 encourages the clinician to identify the planned duration of therapy. To be most impactful, ASPs should consider approaches to build use of the Four Moments into daily practice, such as a requirement for their standardized documentation within daily notes, with audit and feedback to promote adherence. This may foster a more consistent, logical approach to antibiotic decision-making and reduce the need for ASP prompting.

SYNDROME-BASED LOCAL GUIDELINES

For clinicians to be sufficiently equipped to incorporate the Four Moments into their daily practice, local, syndrome-specific guidelines must be available at the point of care. ASP should therefore prioritize the development and effective dissemination of such guidelines. Although the list of potential infectious syndromes in hospitalized patients is lengthy, at a minimum, guidelines should target the most common conditions for which antibiotics are prescribed. Such conditions are likely similar whether the institution is a critical access hospital or large academic medical center, and include community-acquired pneumonia, urinary tract infections, skin and soft tissue infections, hospital-acquired pneumonia, and perioperative antibiotic prophylaxis [16]. To be most effective, the ASP should draft guidelines using a structured approach (such as the Four Moments) that specifies (1) criteria that make the infectious process likely; (2) appropriate diagnostic testing; (3) empiric treatment choices; (4) de-escalation considerations; and (5) duration of therapy. They should be developed after a thorough literature review—including publications in the non–infectious diseases literature—and incorporate input from a multidisciplinary group of stakeholders. Final guidelines should be shared with relevant clinician groups, pharmacists, and nurses to promote knowledge of and uptake of the guidance. It is essential that local guidelines are easily accessible to clinicians at the point of care, such as through storage on a uniform platform (eg, handbook, Intranet, mobile application). Ideally, local guidelines should be accompanied by order sets to reinforce guideline-based prescribing decisions.

INDIVIDUALIZED FEEDBACK WITH PEER COMPARISON

In the primary care setting, providing clinicians with data on their own prescribing patterns with a comparison to peers improves antibiotic use. For example, among pediatric providers, regular, individualized feedback on the use of broad-spectrum antibiotics led to substantial reductions in use of these agents [17]. Among adult primary care providers, the comparison of individual providers’ prescribing rate for antibiotic-inappropriate respiratory infections with peers led to a notable reduction in antibiotic prescriptions [18]. In both cases, withdrawal of the regular feedback led to reversion of prescribing patterns toward preintervention practices, highlighting the need for ongoing feedback [19, 20]. As a field, we must explore the longer-term sustainability of prescribing changes associated with these interventions, determine efficient approaches to scale up implementation, and, importantly, evaluate their effectiveness in non–primary care environments.

The hospital setting is an important area for potential application of such behavioral interventions. Most hospitals do not report hospital antibiotic use at the provider level, so inpatient providers are likely to know little about their own prescribing patterns in relation to their peers. As one example, an intervention providing feedback to hospitalists on their use of antibiotics with activity against methicillin-resistant Staphylococcus aureus and antibiotics with a broad spectrum of Gram-negative activity, with a comparison to their peer hospitalists, could increase attention to appropriate use of these agents, prompt more frequent critical assessments of their need, and promote outliers to examine their own prescribing patterns, thereby driving down unnecessary or inappropriate use. Published studies of this type of intervention in the hospital setting are lacking but represent an attractive opportunity for future work.

Although standardized antibiotic utilization metrics have been developed (eg, days of therapy per 1000 patient-days), there are currently no standardized approaches to quantify and report the appropriateness of antibiotic use or antibiotic-associated adverse events to frontline clinicians. Because measures of antibiotic utilization do not convey benefits or avoidance of harm to patients, feedback of appropriateness or adverse events to clinicians may be more effective at fostering behavioral modifications in prescribing; novel metrics in these domains are needed.

For example, the development of a provider-specific “harm score” for feedback to clinicians may prompt clinicians to critically assess their own antibiotic use. Furthermore, since harm associated with inpatient antibiotic administration frequently occurs after a patient is discharged from the hospital [21], feedback of postdischarge serious adverse events (eg, severe Clostridioides difficile infection, Stevens-Johnson syndrome)—which clinicians currently may never know occurred—may be useful to “nudge” clinicians to be more judicious with their prescribing. Such an approach was shown to be successful with the opioid crisis. In a randomized trial including 861 clinicians who cared for patients who suffered a fatal opioid overdose, clinicians who received notification of their patients’ deaths were less likely to prescribe opioids in the subsequent 3-month follow-up period [22].

PARTNERING WITH PEER CHAMPIONS

Antibiotic stewards cannot be omnipresent in the hospital and should not strive to be so. Although many of us have likely asked ourselves what more we can be doing to improve antibiotic use in the hospital, the more appropriate question may be “who else can I engage to champion antibiotic stewardship in the hospital?” Engaging a champion clinician within a particular specialty or care location to facilitate practice change among peers is not a novel concept, and has been effectively used as a component of antibiotic stewardship interventions. In 1 study, peer champions from emergency medicine, hospital medicine, and surgery successfully improved antibiotic selection and reduced durations of therapy for patients hospitalized with skin and soft tissue infections [23]. Hospitalist-led antibiotic stewardship projects have also proven effective at identifying the need for performance improvement and driving change [24–27].

Anecdotal evidence suggests that initiatives led by peers within a specialty for which the change in prescribing is intended may be more effective than those led by individuals external to that specialty (ie, antibiotic stewards). This may be particularly important in locations with high antibiotic utilization, such as critical care, or within particular specialties with the greatest opportunity to reduce unnecessary antibiotic use. The engagement of peer champions also diffuses the responsibility for antibiotic stewardship away from the central stewardship team, and allows a greater scope of stewardship-related activities to occur within a hospital. ASPs should identify and train peer champions in relevant medical and surgical fields, and periodically meet with them to discuss successes and barriers encountered. Communication to frontline clinicians should come from the peer champion to reinforce that stewardship-related quality-improvement work is a collaborative from the ground up rather than a “top down” approach.

ELECTRONIC CLINICAL DECISION SUPPORT

As more and more facilities have moved towards electronic health records (EHRs), capitalizing on this infrastructure can guide clinicians with their antibiotic decision-making in real time. As an example, the practice of sending urine cultures in the absence of signs or symptoms of a urinary tract infection is common across health-care facilities [28, 29]. Practitioners respond to positive urine cultures by prescribing antibiotics, even in the absence of relevant clinical symptoms. One institution successfully reduced antibiotic treatment for positive urine cultures using a novel approach in which, rather than reporting urine culture results in the EHR as the default option, clinicians would find the following message: “The majority of positive urine cultures from inpatients without an indwelling urinary catheter represent asymptomatic bacteriuria. If you strongly suspect that your patient has developed a urinary tract infection, please call the microbiology laboratory.” [30] This relatively simple intervention was associated with a 36% reduction in the treatment of asymptomatic bacteriuria without unintended consequences from withholding antibiotics.

Similarly, an EHR best-practice advisory was implemented across 3 hospitals to reduce unnecessary C. difficile testing. For patients who received laxatives or underwent recent C. difficile testing, when providers attempted to order C. difficile testing, an EHR alert appeared that prompted them to call the microbiology laboratory for approval if testing was still considered necessary. This modification to the EHR led to a significant decrease in C. difficile testing at all 3 hospitals, without any related adverse events [31].

Both of these examples build off a concept in the behavioral economics literature called “asymmetric paternalism.” Asymmetric paternalism is an approach that enables individuals to determine their own path in situations where they are prone to making potentially harmful choices: not limiting their choices, but rather by biasing their decision-making in favor of more beneficial options [32]. Larger studies are warranted to validate these findings and investigate similar approaches to improving decision-making around commonly ordered laboratory tests that either positively or negatively impact antibiotic use (eg, sputum culture, procalcitonin), as they have the potential to lead to dramatic practice changes among frontline clinicians.

Although implementing changes in the EHR can require sophisticated information technology resources, they can lead to downstream benefits related to both diagnostic and antibiotic stewardship. Rather than hospitals individually navigating the process to develop supportive tools, future investigations should identify effective EHR tools that can be easily implemented across hospitals using the same EHR software (eg, Epic, Cerner).

PUBLIC REPORTING OF ANTIBIOTIC USE

In the field of infection prevention, public reporting of health care–associated infections (HAIs) has been used as a strategy to reduce HAIs through increasing attention, effort, and resources dedicated to infection prevention. Approximately 27 states have public reporting requirements for HAIs, and a subset require reporting to the CDC’s National Healthcare Safety Network (NHSN) [33]. Several US studies have shown an association between public reporting and lower rates of central line–associated infections in intensive care units [34, 35]. A Canadian study demonstrated that public reporting of hospital C. difficile infections was associated with a 27% reduction in hospital-onset C. difficile infections [36].

It stands to reason that public reporting of antibiotic use would be associated with the same benefits that it has garnered for the field of infection prevention: attention of hospital leadership and the resources needed to improve antibiotic use. The following is necessary to make public reporting of antibiotic use a reality. First, there would need to be widespread reporting of antibiotic use by hospitals to the NHSN antibiotic utilization module [37]. Several states now mandate reporting of antibiotic utilization data to NHSN, and more than 1500 acute-care hospitals report data nationally. Second, since the appropriate or optimal amount of antibiotic use will vary among hospitals based on numerous factors (eg, number of intensive care unit beds), there would need to be a validated, risk-adjustment model to determine expected antibiotic use, to allow for comparison among dissimilar hospitals. There is ongoing research in this area [38]; however, there is still much work to be done to develop a valid method to compare antibiotic use across the spectrum of hospitals through use of a metric that clinicians and the general public can understand. Finally, a balancing measure would have to be accounted for so that poor outcomes associated with the inappropriate underuse of antibiotics would not be incentivized. We believe this is a worthy future pursuit to further incentivize responsible prescribing among frontline clinicians.

FINANCIAL INCENTIVES

Few studies involving the use of financial incentives to drive improvements in antibiotic use are available in the published literature. In 1 study from the United Kingdom, the introduction of financial incentives for local health commissioners to reduce outpatient antibiotic prescriptions were associated with an 8% reduction in all antibiotic prescriptions and a 19% reduction in broad-spectrum antibiotic prescriptions [39]. The Michigan Hospital Medicine Safety Consortium utilized data feedback with pay-for-performance to incentivize improvements in adherence to 5-day antibiotic courses for uncomplicated community-acquired pneumonia [25]. Hospitals meeting the performance target received an incentive payment from Blue Cross Blue Shield of Michigan and Blue Care Network. Those hospitals could use the incentive payments at their own discretion, and some chose to incentivize hospitalists by providing bonuses if the performance measure was met.

This unique intervention highlights 2 potential pay-for-performance models: private insurers incentivizing hospitals to improve antibiotic use and hospitals incentivizing clinicians to improve antibiotic use. Another potential model would involve financial incentives by the Centers for Medicare & Medicaid Services for hospitals to meet performance measures. The Centers for Medicare & Medicaid Services currently uses both incentives and penalties for HAIs. The Hospital-Acquired Conditions Reduction Program penalizes hospitals in the lowest-performing quartile, with a 1% payment reduction. The Hospital Value-Based Purchasing Program financially incentivizes hospitals, with a bonus of up to 3% for providing high-quality care. Similar to public reporting, such a model may increase engagement of hospital leadership and garner resources for antibiotic stewardship efforts. However, a validated method to compare antibiotic use across hospitals is needed for this to become a reality.

CONCLUSION

In summary, there has been a marked scale-up of ASPs in acute-care hospitals, and most ASPs are prioritizing traditional “core” interventions of prospective audit and feedback and prior authorization [5]. Although these are effective approaches to improve antibiotic use, they should not be viewed as the primary role of an ASP. Rather, the focus over the long term should be the development and implementation of higher-level interventions that motivate and empower frontline clinicians to integrate antibiotic stewardship into their own daily practice. This paradigm shift may be more fulfilling for ASPs and will help usher in an era in which frontline clinicians feel responsible for prescribing antibiotics appropriately and have the resources to do so, thereby reducing the need for day-to-day ASP oversight. Although some degree of audit and feedback or prior authorization may always be necessary, ASPs should not lose sight of the fact that their fundamental role is not to correct inappropriate antibiotic use, but to prevent it. The recommendations and policies we propose will need careful monitoring by the ASP to ensure that they are indeed effective at reducing inappropriate antibiotic use, are not associated with unintended negative consequences, become integrated in medical practice, and are sustainable. This will ultimately put the onus for antibiotic stewardship where it belongs: on prescribing clinicians rather than stewardship teams.

Note

Potential conflicts of interest. The authors: No reported conflicts of interest. Both authors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest.

References