The implementation of antimicrobial stewardship programs (ASPs) is paramount to the optimization of antibiotic use in hospitalized children.Reference Gerber, Newland and Coffin1, Reference Levy, Swami, Dubois, Wendt and Banerjee2 Children’s hospitals in the United States are increasingly implementing stewardship activities. In a 2011 national survey of 38 children’s hospitals, 16 (38%) had a formal ASP and 15 (36%) were in the process of implementation.Reference Newland, Gerber and Weissman3 When this survey was repeated in 2016 with 52 children’s hospitals, 49 (94%) reported having a formal ASP.Reference McPherson, Lee and Terrill4
The urgency surrounding the need for ASPs and the rapid increase in their number has stimulated research into the barriers to implementing these programs.Reference Pakyz, Moczygemba, VanderWielen, Edmond, Stevens and Kuzel5–Reference James, Luu, Avent, Marshall, Thursky and Buising8 Implementation is stymied by prescriber resistance due to a fear of threatened autonomy, Reference Perozziello, Routelous and Charani9–Reference Broom, Broom, Plage, Adams and Post11 a hierarchical hospital culture,Reference James, Luu, Avent, Marshall, Thursky and Buising8, Reference Broom, Broom, Kirby, Gibson and Post12 inadequate information technology resources,Reference Pakyz, Moczygemba, VanderWielen, Edmond, Stevens and Kuzel5 lack of dedicated personnel,Reference Perozziello, Routelous and Charani9 and lack of leadership support.Reference Jeffs, Thampi, Maione, Steinberg, Morris and Bell13 Most research on implementation in antimicrobial stewardship focuses on the experiences of new programs. We know less about prescriber perceptions of established ASPs. This information is critical to optimizing the ongoing impact of ASPs.
To investigate this issue, we conducted a cross-sectional survey of prescribers at a large freestanding children’s hospital with one of the oldest pediatric ASPs in the United States. The objective of the study was to examine prescriber perceptions of the utility, efficiency, and value of the ASP.
Methods
Study design, sample, and recruitment
We conducted a cross-sectional survey of physicians and advanced practice providers (APPs) working at the Children’s Hospital of Philadelphia (CHOP). Respondents eligible for inclusion included all physicians (attending physicians, fellows, and resident physicians) and APPs (ie, certified registered nurse practitioners [NPs] and physician assistants [PAs]) working at the hospital.
CHOP has had an ASP for 15 years.Reference Metjian, Prasad, Kogon, Coffin and Zaoutis14 It is primarily a prior authorization program with >50 targeted formulary antimicrobials that require ASP approval prior to use, in addition to all nonformulary and inhaled antimicrobials. Prescribers are required to contact the ASP for antimicrobial approval from 7:00 am to 10:00 pm daily. They are allowed to order all antimicrobials overnight without approval but must contact the ASP the following morning for approval of subsequent doses. Approvals are conducted by 2 full-time ASP pharmacists Monday through Friday, 9:00 am to 5:00 pm. Infectious diseases (ID) fellows perform approvals from 7:00 am to 9:00 am and from 5:00 pm to 10:00 pm, Monday through Friday and on weekends and holidays. Antibiotic approvals are communicated to the pharmacy through a note placed in the patient’s electronic medical record. Since 2012, the ASP pharmacists have performed daily audits with feedback for antimicrobials administered >72 hours (ie, “time out” recommendation). Infectious diseases attending physicians provide second review on antimicrobials that are not approved.
The survey was administered via research electronic data capture from February through June 2017.Reference Harris, Taylor, Thielke, Payne, Gonzalez and Conde15 Respondents were recruited via e-mail. The survey was voluntary and no incentive was offered for participation. Reminder e-mails were sent 3 times during the study period. Respondents were made aware that the survey was being administered by the CHOP ASP team to inform improvement work. Because the study was undertaken as a quality improvement effort, the study was deemed exempt from institutional review board approval.
Survey instrument
Our survey instrument was designed in 2 stages. First, we conducted formative interviews with 15 prescribers. Second, we reviewed previous survey research on prescriber perceptions of the ASP.Reference Seemungal and Bruno16–Reference Steinberg, Dresser and Daneman19 We identified thematic domains around which we designed our fixed-response survey questions, including knowledge of the ASP, perception of antibiotic approval mechanisms, frequency and reasons for working around ASP approvals, perception of value of the ASP, perception of communication with ASP staff, and level of trust in ASP recommendations. Once drafted, we circulated the instrument to a convenience sample of 7 physicians for comment on question comprehensibility and length. Modifications were made based on this feedback. The final instrument included 43 closed-ended questions, 18 open-ended prompts for respondents to further elaborate their answer to a closed-ended question, and 4 stand-alone open-ended questions. Closed-ended questions utilized 4- and 5-point Likert-type scales as well as true/false answer choices (see supplemental material online for survey instrument). All questions were optional.
Data analysis
Quantitative data were analyzed using Stata statistical software.20 Descriptive statistics (frequencies, percentages, means, and SDs) were used to summarize the fixed-response questions. For survey items that were categorical in nature, clinician responses are presented as frequencies, and comparisons were made using the χ2 test of significance at the level of P < .05. We collapsed ratings of “occasionally” with “rarely” and “strongly agree” with “agree” and “strongly disagree” with “disagree” for clarity of presentation. Free-text responses were entered into NVivo 12 software for analysis.21 Two coders identified recurrent patterns in the data, developed codes, and applied them across the dataset in a process of line-by-line document review. Intercoder reliability was periodically assessed, and disagreements were resolved by consensus.
Results
Respondents
Of 394 recipients, 160 (41%) completed the survey. Of all respondents, 46 (28.8%) were attending physicians, 43 (26.8%) were general pediatrics residents, 25 (15.6%) were fellow physicians, and 46 (28.8%) were advanced practice providers. Both general and subspecialist pediatric providers were well represented, with 37 (21.9%) working in general pediatrics, 27 (23.3%) in the neonatal intensive care unit (ICU), 18 (15.5%) in the pediatric ICU, 10 (8.6%) in oncology, 9 (7.8%) in the cardiac ICU, and 8 (6.9%) in surgery. Most respondents had been working at CHOP between 2 and 5 years (n = 64, 40%) and between 6 and 10 years (n = 30, 18.8%) years. In addition, 14 respondents (8.8%) had worked at the hospital for >30 years.
Closed-ended questions
Familiarity and interaction with ASP
Almost all of the respondents 140 (89%) reported being familiar or very familiar with the ASP. According to role, NPs were more likely to be unfamiliar with the program than other providers (25% vs 8% P < .05). Residents and NPs were most likely to confuse the role of ASP with ID consultation service (29%). Half of the respondents (51.7%) stated that they use the ASP to determine whether an ID consultation is necessary.
Perceptions of the value and trustworthiness of ASP
Overall, respondents had a favorable opinion about the value of the ASP, believing that it provides education to improve antibiotic use, improves the quality of patient care, improves clinical decision making, respects their clinical judgment, and facilitates appropriate antimicrobial use (Table 1). Most respondents reported trusting ASP personnel while finding them professional, credible, and fair. Most respondents (n = 92, 63%) did not feel that the ASP interferes with their clinical decision making. The most common criticism of the ASP was that it threatened efficiency (26.0% agreed).
Table 1. Perceptions of Antimicrobial Stewardship Program
a Numbers do not sum to 160 because data were missing for some questions.
Working around ASP approval
When asked how frequently they engage in workarounds to get antibiotics, even if they are against ASP recommendations (eg, by waiting to order restricted antibiotics until the ASP went home for the evening), 13 respondents (8.8%) selected “frequently,” 101 (68.7%) selected “occasionally,” and 33 (22.5%) reported that they “never” engaged in workarounds.
Open-ended questions
In total, 133 free-text responses were given by 66 respondents. Comments ranged in length from a sentence to multiple paragraphs. Systematic analysis revealed insights to further contextualize the responses to the closed-ended questions in 2 salient domains: (1) perceptions of the causes of inefficiency (mentioned by 44 respondents) and (2) reasons why prescribers engage in workarounds to ASP approval (mentioned by 34 respondents). Exemplar quotations for each domain are included in Table 2.
Table 2. Themes in Free-Text Responses and Sample Quotes
Inefficiency
When describing their perceptions of inefficiency due to ASP, the most common criticism was a sense that the approval process required too many phone calls. Respondents described how the ASP approval process felt cumbersome at times and generated excessive phone calls. Needing to make multiple phone calls grew frustrating for many respondents, especially when they were caring for a particularly sick patient whom they were anxious about. In those cases, the wait to receive a phone call back was especially uncomfortable.
One of the major causes of inefficiency identified by respondents was communication breakdowns between the ASP team and the dispensing pharmacy. Numerous respondents described situations in which the ASP team or ID consult service had approved an antimicrobial but there was a delay in the note reaching the electronic health record. The dispensing pharmacy received the order but did not see the approval note, so they called the prescriber to verify. This additional phone call was seen as unnecessary and annoying.
Workarounds
Respondents identified a number of specific workarounds to ASP approval that they had engaged in, including repeatedly reordering 48-hour “rule-out” antimicrobials until a patient receives a full course, “stretching” a patient’s diagnosis to fit one of the ASP pre-approved indications, waiting until the evening to obtain the “overnight approval” indication (eg, stealth dosing), ordering a blood culture to facilitate approval and, for patients who were soon-to-be discharged, ordering restricted antibiotics to be filled as an outpatient to avoid the ASP approval process.
In elaborating the reasons why they engage in workarounds, respondents reported that they did not want to stop antimicrobials that appeared to be working on critically ill patients despite being culture negative and accommodating parent request for specific antimicrobials. Residents and NPs reported that they engaged in workarounds to satisfy the demands of their attending or specialty consult service (especially surgery), who may have disagreed with ASP approval. While many described attempts to negotiate, sometimes they “gave in” and worked around ASP approval to minimize conflict.
Discussion
We conducted a cross-sectional survey of prescribers at a large freestanding children’s hospital with an established pediatric ASP that primarily utilizes prior authorization. Prescribers held a generally favorable impression of the hospital’s ASP, believing that it facilitated appropriate antimicrobial use. Respondents did not perceive that the ASP threatened their autonomy or disrespected their clinical judgment. They reported finding ASP personnel professional, credible, and trustworthy. Nevertheless, most respondents reported occasionally engaging in workarounds to ASP approval. The major critique of the ASP program was inefficiency. Respondents reported frustration surrounding the number of phone calls required, communication gaps with the dispensing pharmacy, and perceived delays in antibiotic administration. Workarounds occurred largely for social and emotional reasons.
Consistent with previous survey research on prescriber perceptions of ASPs, most respondents felt that the program improved their use of antimicrobials, served a useful educational function, and improved the quality of patient care.Reference Seemungal and Bruno16–Reference Stach, Hedican, Herigon, Jackson and Newland18, Reference Salsgiver, Bernstein and Simon22, Reference Flannery, Swami, Chan and Eppes23 Much of the attitudinal research on restrictive ASP interventions, such as prior authorization, has found that prescribers express concern about the threat that these interventions posed to their autonomy and clinical judgement.Reference Seemungal and Bruno16–Reference Stach, Hedican, Herigon, Jackson and Newland18, Reference Salsgiver, Bernstein and Simon22, Reference Venugopalan, Trustman, Manning, Hashem, Berkowitz and Hidayat24 Contention with a hierarchical culture, facing the social norm of noninterference surrounding prescribing, and antimicrobial steward discomfort with being thought of as the “antibiotic police” are oft-cited challenges to the implementation of restrictive stewardship interventions.Reference Szymczak, Newland, Barlam, Neuhauser, Tamma and Trivedi25, Reference Sikkens, van Agtmael and Peters26 The lack of prescriber buy-in to restrictive stewardship interventions can lead to workarounds such as “stealth dosing,” in which prescribers wait until the prior approval period has ended for the day to order off-guideline or unnecessary restricted antimicrobials.Reference LaRosa, Fishman, Lautenbach, Koppel, Morales and Linkin27 Other studies have documented prescribing clinicians misrepresenting clinical information to get approval to prescribe the antimicrobials that they believe are appropriate.Reference Calfee, Brooks, Zirk, Giannetta, Scheld and Farr28, Reference Seemungal and Bruno29
Our study makes 3 key contributions. First, we found that respondents did not feel that the ASP threatened their autonomy or disrespected their clinical judgment. ASP personnel were seen as credible and fair in their approvals; there was a high level of trust in the recommendations made by the ASP. It is possible that prescriber fear of threatened autonomy diminishes with ASP age, as prescribers become more comfortable with the program over time. Because we do not have data on prescriber perceptions from the beginning of this ASP, we cannot directly attribute this attitude to duration of program existence. More experienced, long-standing ASPs have better uptake of interventions. For example, as Cosgrove et al. (2012) show in their multicenter study evaluating the impact of postprescription review and feedback, institutions with established and well-resourced ASPs decreased antimicrobial use significantly whereas institutions with less well-established ASPs did not.Reference Cosgrove, Seo and Bolon30 More research is needed to investigate changing perceptions toward ASPs over time, the factors that shape these perceptions, and their impact on program outcomes.
Second, we found that despite not feeling that their autonomy was threatened by the ASP, prescribers still engaged in workarounds to evade the restrictions of prior authorization. Working around the ASP guidelines is largely conceptualized as a problem of prescriber resistance that emerges from a feeling that their autonomy is being threatened.Reference Szymczak, Newland, Barlam, Neuhauser, Tamma and Trivedi25 Our study illustrates that workarounds can persist even when prescribers are supportive of an ASP, an important issue to consider when evaluating a program. Free-text analysis revealed that the reasons given by prescribers for workarounds are more nuanced than an outright rejection of being told what to do by an outsider. Our respondents described several social and emotional influences that shape their decision to work around prior authorization: their sense of obligation to patients, managing their own fear related to clinical uncertainty, and being “stuck in the middle” between a superior and the ASP guidelines. Our respondents described making decisions in a system characterized by competing priorities, the navigation of which required attention to goals beyond the optimization of antimicrobial use.Reference Dixon-Woods, Suokas, Pitchforth and Tarrant31–Reference Szymczak33
The third contribution of our study is that it highlights the importance of efficiency as it relates to the implementation of ASP interventions. Two previous studies have found that prescribers perceive ASP interventions to be too time-consuming.Reference Cantey, Vora and Sunkara10, Reference Seemungal and Bruno16 Our study provides more detail on prescriber experiences of inefficiency. The repeated phone calls required to obtain approval were seen as burdensome, especially in the context of a busy day in which preapproval becomes one of many logistic hurdles clinicians face in providing care to their patients. Our respondents also identified occasional problems in communicating about prior authorization between the dispensing pharmacy and the ASP team. Both excessive phone calls and communication breakdowns are logistical challenges that can be addressed through thoughtful modifications to ASP procedures.
Although our study has limited generalizability because it was conducted at a single institution, our findings have relevance for ASPs in other settings. CHOP has a particularly restrictive ASP, with prior authorization as the main intervention. Most hospitals use a combination of strategies, with postprescription review and feedback as the primary approach.Reference Newland, Gerber and Weissman3, Reference Kronman, Banerjee and Duchon34, Reference Livorsi, Heintz, Jacob, Krein, Morgan and Perencevich35 Both of these interventions communicate recommendations about antimicrobials in either a restrictive or persuasive manner. Our findings suggest that the way this communication is incorporated into the everyday workflow of frontline prescribers is critical to their perception of the value of the program and, ultimately, to their level of engagement with ASP recommendations. This factor is likely to be important in hospitals that use both restrictive and persuasive interventions. Although we conducted this study in a pediatric hospital, many of the social dynamics we identified, including deference to senior colleagues, prescribing for emotional reasons and frustration with inefficient communication mechanisms, have been reported across many types of healthcare settings.Reference Szymczak, Newland, Barlam, Neuhauser, Tamma and Trivedi25
Our study demonstrates the value of investigating prescriber perceptions of established ASPs both in generating knowledge of how clinicians think about stewardship and in identifying process defects that can be addressed through systems redesign. Established ASPs should incorporate periodic “end user” assessments of program activities to determine areas in which improvement is needed. In general, there is a dearth of research on factors that influence sustainability in antimicrobial stewardship, despite this being identified as a critical challenge to the field of healthcare improvement.Reference Lennox, Maher and Reed36
Our study has several limitations. We were unable to assess response bias, and we did not have information on the characteristics of those who chose not to take the survey. Our survey instrument was not validated. Also, the number of respondents working in different clinical settings within the hospital were too small to permit meaningful comparisons, an analysis that could provide valuable insight. Despite these limitations, our relatively robust response rate and the iterative process we utilized to design our survey indicate that we have captured a meaningful range of responses.
This survey investigated the perceptions that prescribers held toward an established pediatric ASP that utilizes prior authorization. Respondents held the ASP in high regard and believed that it improved the quality of patient care. The primary critique of the prior authorization process was that it is inefficient. Engagement with prescribers via surveys that assess perceptions can identify areas for improvement to ensure that long-standing ASPs have maximum impact.
Supplementary material
To view supplementary material for this article, please visit https://doi.org/10.1017/ice.2019.47
Author ORCIDs
Julia E. Szymczak 0000-0002-3230-8670
Acknowledgments
We thank the anonymous respondents who took the time to fill out our survey. Drs. Szymczak, Coffin and Gerber receive support from the CDC Cooperative Agreement FOA#CK16-004-Epicenters for the Prevention of Healthcare Associated Infections.
Financial support
No financial support was provided relevant to this article.
Conflicts of interest
All authors report no conflicts of interest relevant to this article.
