Adapting Antibiotic Stewardship to the Community Hospital

Antibiotic stewardship programs (ASPs) are an important tool for slowing the emergence and spread of antibiotic resistance. Although commonly implemented and studied in large hospitals, ASPs are less likely to be implemented and evaluated in smaller community hospitals even though their antibiotic prescribing rates are similar to those of larger hospitals.¹ In JAMA Network Open, Anderson and colleagues² provide a welcome addition to the literature by evaluating 2 approaches to antibiotic stewardship in the community hospital setting.

While ASPs can take on several different forms, a recent set of guidelines highlighted 2 core components of any ASP: prior authorization (PA) and postprescription audit and review (PPR).³ Prior authorization is a restrictive strategy that requires the antibiotic prescriber to justify the rationale for the desired antibiotic before the agent can be dispensed. In contrast, PPR is a persuasive strategy that involves the stewardship team auditing antibiotic use 24 to 72 hours after initiation and providing immediate feedback to prescribers, with the goal of optimizing the current antibiotic treatment. These core strategies have recently been studied in a large academic medical center using infectious diseases (ID)–trained specialists, and the results favored PPR.⁴

However, community hospitals frequently lack dedicated ID resources to manage ASPs. Surveys of US community hospitals have found that 41% to 50% had no ID physician and 93% had no ID-trained pharmacist.¹ While these community hospitals are increasingly expected to have ASPs, the optimal approach for implementation without ID expertise is unclear.

Anderson and colleagues² describe one approach to implementing PA and PPR in community hospitals, namely, training local pharmacists to be antibiotic stewards. This was a multicenter nonrandomized clinical trial with crossover design study and with historical controls that compared PPR with a modified form of PA across 4 community hospitals in North Carolina. Each intervention lasted 6 months. The antibiotics targeted by these interventions were vancomycin hydrochloride, piperacillin-tazobactam, and antipseudomonal carbapenems (eg, meropenem). At least 1 designated pharmacist at each site underwent a 3-hour training in conflict management and antibiotic stewardship. All hospitals were also provided clinical pathways and suggested criteria for appropriate use of the targeted antibiotics but were not provided remote clinical ID assistance. Only 1 of the 4 hospitals had daily access to a local ID clinician.

The primary outcome of the study by Anderson and colleagues² was the feasibility of implementing PPR and modified PA. Four additional implementation outcomes⁵ were not specifically cited but were measured, including (1) feasibility (all 4 sites agreed to implement the interventions, with the caveat that PA was modified so that the first antibiotic dose was allowed before stewardship review), (2) fidelity (the trained pharmacists performed a PA or PPR review on 2692 patient admissions who received a targeted antibiotic, or roughly 31% of all eligible admissions), (3) adoption (most pharmacist recommendations were either fully or partially accepted by clinicians [79.2% during PA and 69.0% during PPR]), and (4) acceptance and appropriateness (more than half of clinicians reported that the interventions improved patient care and were not burdensome to their workflow, although, more than half of trained pharmacists found PPR to be either burdensome or were equivocal on this point).²

Also, an opportunity cost clearly existed with these interventions, as pharmacists spent 5 to 19 hours per week conducting either PA or PPR.² Importantly, hospital pharmacy departments were compensated for their time spent on study-related efforts.

Significant differences among secondary outcomes were observed between the PA and PPR intervention periods.² During the PPR period, the trained pharmacists were twice as likely to determine that antibiotics were inappropriate, more likely to contact the treating clinician with recommendations, and more likely to recommend de-escalation. These factors likely contributed to why PPR was found to reduce antibiotic use compared with historical controls (mean difference in days of therapy per 1000 patient-days, −40.1; 95% CI, −71.7 to −8.6), with no change observed during the PA intervention (mean difference, 4.4; 95% CI, −55.8 to 64.7).

These findings may represent a starting point for stewardship at these hospitals. Across the PA and PPR periods, recommendations were communicated to clinicians in only 8.4% of all eligible admissions.² The basis for such a low percentage could be multifactorial. For one, PPR was protocolized to occur at 72 hours (±24 hours) after order entry, but the average length of stay was 48 to 72 hours for the study hospitals. Earlier PPR would be more responsive to microbiological cultures and rapid diagnostics, while increasing the chance of a stewardship review before discharge.

Anderson and colleagues² have shed some much-needed light on the practical aspects of implementing core stewardship strategies in community hospitals. Their findings are likely generalizable because the 4 participating hospitals were similar to community hospitals across the United States: they lacked ID resources to support the ASP and were not trained in antibiotic stewardship. Some key differences would be that the hospitals were relatively large (median bed size, 305) and were part of the Duke Antimicrobial Stewardship Outreach Network, which probably made them more aware of stewardship principles than the traditional community hospital.

To our knowledge, there is only 1 other multicenter study evaluating stewardship implementation in community hospitals. Stenehjem and colleagues⁶ assessed a multifaceted stewardship intervention across 15 small community hospitals within a health care network. Using a cluster randomized design, the study allocated hospitals to 1 of 3 antibiotic stewardship implementation strategies. One cluster received minimal stewardship training; another was trained and implemented locally controlled PA and PPR, in line with the approach in the present study; while the third group had remote ID experts prospectively reviewing positive microbiological results, remote ID experts actively managing the PA, and trained pharmacists performing PPR on most antibiotics used. All groups had access to remote ID clinicians via a telephone hotline. A reduction in total and broad-spectrum antibiotic use was only demonstrated in hospitals with the most intensive stewardship intervention with ID physician and pharmacist support and PPR of most antibiotics.

While Anderson and colleagues² demonstrated that targeted PPR can result in a reduction of antibiotic use without ID specialist involvement, Stenehjem and colleagues⁶ showed that the involvement of remote ID specialists and extensive PPR was critical in reducing antibiotic use. Regardless, and most importantly, both studies convincingly demonstrate that implementation of stewardship practices in community hospitals can improve antibiotic prescribing.

Using innovative study designs and tracking implementation outcomes can help us learn important lessons about the spread of ASPs in community settings. One key implementation outcome that was not measured by either of the above-mentioned studies is sustainability.⁵ For example, Anderson and colleagues² found that PPR and modified PA were feasible over the 12-month study period. However, were these interventions challenging to sustain once the project ended and grant funds expired? A concerning trend, as shown in their supplemental eFigure 3B, was that the use of all study antibiotics increased over both strategies during period 2. This may indicate a “voltage drop,” with the stewardship pharmacists or the clinicians exhibiting intervention fatigue.⁷

Arthur Ashe once said, “Start where you are. Use what you have. Do what you can.” Ashe was not referring to ASPs, but he might as well have been. Both Anderson and colleagues² and Stenehjem and colleagues⁶ have demonstrated that PA and PPR can be modified based on community hospitals’ local resources and local personnel. The need to improve antibiotic use is universal, but the approach to stewardship need not be the same. Adaptation, not imitation, is the key to success.

Published: August 16, 2019. doi:10.1001/jamanetworkopen.2019.9356

Open Access: This is an open access article distributed under the terms of the CC-BY License. © 2019 Livorsi DJ et al. JAMA Network Open.

Corresponding Author: Daniel J. Livorsi, MD, MSc, Iowa City VA Health Care System, 601 Hwy 6 W, Iowa City, IA 52246 (daniel-livorsi@uiowa.edu).

Conflict of Interest Disclosures: Dr Livorsi reported receiving grant funding from Merck and the US Department of Veterans Affairs Health Services Research & Development Service. No other disclosures were reported.