Introduction
Nearly 32 million adults in the U.S. report an allergy to penicillin and formal penicillin allergy testing reveals only 5% of patients who report a penicillin allergy have a true allergy to penicillin. Reference Castells, Khan and Phillips1,Reference Shenoy, Macy, Rowe and Blumenthal2 Patients reporting a penicillin allergy receive less effective and more costly antibiotic treatment, and are at increased risk for the development of Clostridioides difficile infection and infections with antibiotic-resistant bacteria. Reference Blumenthal, Peter, Trubiano and Phillips3,Reference Norton, Konvinse, Phillips and Broyles4 Penicillin allergies are particularly consequential among patients undergoing surgery. Surgical patients with a documented penicillin allergy versus those without a documented penicillin allergy are more likely to receive less-effective, second-line, prophylactic antibiotic treatments that increase their risk for mortality, surgical site infections, and longer hospital stays. Reference Blumenthal, Ryan, Li, Lee, Kuhlen and Shenoy5,Reference Wilhelm, Bonsall and Miller6 Comprehensive penicillin allergy histories enable the risk-stratification and appropriate management of reported penicillin allergies. Reference Shenoy, Macy, Rowe and Blumenthal2 Yet, up to 40% of documented penicillin allergy histories lack descriptions of the index reaction, Reference Jones, Morris, Santos, Nasser and Gouliouris7 which impedes risk stratification and perpetuates the unnecessary avoidance of penicillins and cephalosporins.
In 2019, the U.S. Centers for Disease Control and Prevention (CDC) Core Elements for Hospital Antimicrobial Stewardship Programs specified that nurses may play an important role in stewardship by improving the evaluation of reported penicillin allergies. 8 We developed a multifaceted implementation strategy (actions to facilitate the adoption of evidence-based interventions) to amplify facilitators and minimize barriers to nurses’ evaluation of reported penicillin allergies. Reference Leeman, Birken, Powell, Rohweder and Shea9–Reference Greendyke, Carter and Salsgiver12 Our development of the implementation strategy was guided by the Capability, Opportunity, Motivation Model for Behavior Change (COM-B) model of behavior change that specifies behavior results from three interacting components, Reference Michie, van Stralen and West13 ie, Capability (nurses’ knowledge and skill to evaluate penicillin allergies), Opportunity (external factors that influence nurses’ ability to evaluate penicillin allergies), and Motivation (nurses’ emotional responses and analytic decisions to evaluating penicillin allergies). We defined nurses’ evaluation of reported penicillin allergies as nurses’ implementation of two interventions: 1) documenting penicillin allergies using the STORY mnemonic (Symptoms of allergy, Timing of allergy, Onset of symptoms, Resolution, and Yet again use of penicillin) Reference Carter, Schramm, Baron, Zolla, Zavez and Banach14 and 2) notifying prescribers of patients with low-risk symptoms of reported penicillin allergy. Reference Shenoy, Macy, Rowe and Blumenthal2 We provide a rationale for and description for interventions in Supplemental Materials, Appendix A.
The implementation strategy included five components and targeted nurses’ capability, opportunity, and motivation to comprehensively evaluate penicillin allergies. We categorize implementation strategy components according to the Expert Recommendations for Implementing Change (ERIC) project, Reference Powell, Waltz and Chinman15 map implementation strategy components to COM-B, and describe implementation strategies in Table 1. First, we built an interdisciplinary coalition to purposely develop, examine, reexamine, and adjust implementation strategy components. Second, we hosted educational meetings with surgeons, perioperative advanced practice nurses, perioperative physician assistants, anesthesiology physicians, nurse anesthetists, clinical pharmacists, and perioperative nurses. Educational meetings with surgical prescribers oriented them to and ensured their support of nursing interventions. Educational meetings with perioperative nurses covered the following topics: prevalence and harms of misclassified penicillin allergies, recommendations that nurses improve the evaluation of reported penicillin allergies, the STORY mnemonic to improve penicillin allergy histories, and an evidence-based toolkit that specifies low-risk symptoms of penicillin allergy, See Supplemental Materials, Appendix B. Reference Shenoy, Macy, Rowe and Blumenthal2,8,Reference Carter, Schramm, Baron, Zolla, Zavez and Banach14 Third, we developed dot phrases in the electronic medical record to minimize documentation burden associated with the interventions. Fourth, we made and distributed educational pocket-cards for nurses that outlined the STORY mnemonic, symptoms of low-risk penicillin allergies, and associated dot phrases. Lastly, we provided monthly feedback to nurses to further encourage intervention uptake. We provided a rationale for implementation strategy components in Supplemental Materials, Appendix C. In this study, we aimed to evaluate the feasibility of the implementation strategy as defined by perioperative nurses’ uptake of interventions and the acceptability of interventions as experienced by key stakeholders.
Implementation strategy components, mapping to COM-B, and implementation strategy description
* COM-B refers to the Capability, Opportunity, and Motivation Model of Behavior.
Methods
This was a single-site, pre-post, feasibility implementation study conducted in the outpatient surgical areas of an academic medical center the Northeast United States. At the time of the study period, penicillin allergies were documented in the allergy module of the electronic medical record (EPIC) and there was a lack of formal partnership between the antimicrobial stewardship program and the department of nursing in addressing antibiotic allergy assessment and documentation. We used the Standards for Reporting Implementation Studies (StaRI) to guide the reporting of our implementation study. See Supplemental Materials, Appendix D.
The pre-implementation period was from October 28, 2022, to January 24, 2023, and the post-implementation period was from January 25, 2023, to April 23, 2023. We retrospectively obtained the structured and unstructured penicillin allergy documentation of eligible outpatient surgical patients at the study site during the study period. Patients were eligible for study inclusion if they had a documented allergy to penicillin, amoxicillin, ampicillin, nafcillin, oxacillin, or piperacillin in the electronic medical record, presented from an outpatient setting, and underwent a surgical procedure during the study period. We excluded duplicate records, patients with cancelled outpatient procedures, and patients captured in the pre-and post-implementation periods (as details of penicillin allergy record updates were unspecified in our data pull).
Two researchers (EC and KZ) independently characterized the information contained in each penicillin allergy record according to the STORY mnemonic. Reference Carter, Schramm, Baron, Zolla, Zavez and Banach14 We then used percentage agreement to determine the level of agreement between researchers by dividing the number of records in agreement by the total number of records and multiplying by 100. Using published toolkits, Reference Shenoy, Macy, Rowe and Blumenthal2 we identified patients that met low-risk penicillin allergy criteria in the post-implementation period. Low-risk penicillin allergies included the following symptoms and/or descriptions: nausea, vomiting, diarrhea, headache, fatigue, itchiness, patient has no recollection of allergy, family history of allergy, and yeast infections. Reference Shenoy, Macy, Rowe and Blumenthal2 Records containing no symptoms (ie, field was blank or patient was unable to specify symptoms) were classified as having uninterpretable risk. Among patients with low-risk penicillin allergies, we evaluated nurses’ use of a structured note, which documented their notification of prescribers concerning the low-risk allergy. Our primary outcome was nurses’ comprehensive documentation of penicillin allergy histories, in which we compared the information contained in penicillin allergy documentation pre- and post-implementation period. We used a 3-month pre, post-test design to achieve the recommended sample size for feasibility studies to estimate group differences and allow key stakeholders sufficient exposure to implementation strategies. Reference Teresi, Yu, Stewart and Hays16 We also quantified nurses’ notification of prescribers concerning low-risk penicillin allergies by determining the proportion of patients meeting low-risk penicillin allergy criteria for whom nurses notified prescribers.
We conducted informal formative evaluations with key stakeholders as recommended in implementation studies to learn the experiences of those directly impacted by the implementation effort and to judge the need for refinements to the implementation strategy. Reference Stetler, Legro and Wallace17 During the implementation period, we attended a regularly scheduled perioperative nurse meeting and a regularly scheduled antimicrobial stewardship committee meeting, in which team members provided a brief overview of the status of the initiative and asked attendees to share their experiences and feedback concerning the initiative. Formative evaluations were attended by a minimum of two study team members to ensure a shared understanding of conversation and to reexamine the implementation effort with the larger team.
At the end of the study period, we conducted one focus group with perioperative nurses (N = 7) to understand the acceptability of interventions. An experienced qualitative researcher (EC) facilitated focus group discussion and a second researcher (KF) maintained field notes of contextual information and insights gained during focus group discussion. Reference Phillippi and Lauderdale18 Focus group questions were informed by the Theoretical Framework of Acceptability Reference Sekhon, Cartwright and Francis19 and addressed constructs of intervention acceptability, ie, affective attitude, burden, ethicality, intervention coherence, perceived effectiveness, and self-efficacy. See Focus Group Guide, Figure 1. To enhance the credibility of findings, we conducted member checking during the focus group, Reference Birt, Scott, Cavers, Campbell and Walter20 in which EC summarized in her own words participants’ descriptions of intervention acceptability and asked participants to comment on the accuracy of the summary. Focus groups were audio recorded, transcribed by a professional transcription service and analyzed using thematic analysis according to the Theoretical Framework of Acceptability. Reference Nowell, Norris, White and Moules21 The academic medical center’s institutional review board approved this study (IRB # 23-027S-1).
Focus group guide.
Results
The implementation strategy reached a total of 171 nurses and prescribers from the operating room and anesthesia (n = 49), obstetrics (n = 30), pharmacy (n = 23), antimicrobial stewardship (n = 22), orthopedics (n = 20), pre-op (n = 8), otolaryngology (n = 7), general surgery (n = 6), and vascular surgery (n = 6). A total of 426 patients with 487 penicillin allergy records (216 records pre-implementation period, 271 records post-implementation period) met eligibility criteria and were analyzed. The drug implicated in allergy records were listed as a specific penicillin drug (eg, penicillin V, penicillin G, etc.,) or “penicillins” (n = 355, 73%), amoxicillin or an amoxicillin combination drug (n = 126, 26%), and ampicillin (n = 6, 1%). Among records, 424 (87%) contained a description of the penicillin allergy reaction. The most common symptoms documented were: rash (n = 150, 31%), hives (n = 131, 27%), gastrointestinal symptoms (n = 71, 15%), anaphylaxis (n = 46, 9%), swelling (n = 37, 8%), and itching (n = 34, 7%).
Penicillin allergy records contained the following information in the pre- versus post-implementation period: symptoms of the reaction (87.0% vs 87.1%), timing/years since reaction (7.9% vs 25.8%), onset of reaction in relation to taking penicillin (0% vs 20.7%), how symptoms resolved (0% vs 20.7%), and penicillin re-exposure (2.8% vs 21.0%), Table 2.
Content of penicillin allergy records pre versus post implementation strategy
* STORY mnemonic used to characterize the information contained in penicillin allergy records.
Percentage agreement in evaluating the content of penicillin allergy documentation ranged from 97% to 100% across STORY fields.
In the post-implementation period, 53 patients (20%) met low-risk penicillin allergy criteria. Reasons for low-risk criteria included, gastrointestinal symptoms (n = 37, 70%), itchiness (n = 9, 17%), yeast infections (n = 5, 9%), and patient denying allergy (n = 2, 4%). Nurses documented their notification of prescribers for 14 (26%) of patients meeting low-risk penicillin allergy criteria. Symptoms of allergy were entered as structured data in the pre- and post-implementation period (n = 174, 80.6% records pre-implementation; n = 216, 79.7% records post-implementation). Free-text information was provided in 69 (31.9%) of records in the pre-implementation period and 121 records (44.7%) in the post-implementation period, of which 56 records (46.2%) included the use of the.STORY phrase.
Formative evaluations with members of the hospital antimicrobial stewardship committee and perioperative nurses revealed their positive experiences with the initiative. One pharmacist noted caring for a patient with STORY information and described it as “So helpful.” Nurses described using STORY in practice, pointed to the STORY pocket cards that appeared on their ID badge clips, and reflected that the STORY pocketcard was taped to nurse workstations. Perioperative nurses also recommended that a listing of penicillin-type antibiotics be posted to workstations to remind nurses of the specific antibiotic allergies that are targeted for STORY. This was the only modification made to the implementation strategy.
In the focus group, nurses described their comprehensive assessment of penicillin allergy histories as highly acceptable. We provide the drivers of intervention acceptability and representative quotes in Table 3. Major drivers of intervention acceptability included the appropriate ethicality of the intervention, high self-efficacy to perform the intervention, and the perceived effectiveness of the intervention. Nurses reported that taking a detailed penicillin allergy history fit well with the nurses’ role of communicating important patient information to prescribers to guide patient care. Nurses also expressed confidence in their ability to use STORY to gather a detailed penicillin allergy history, although noted that time constraints and poor patient recall presented barriers to STORY detail and completion. In describing poor patient recall as a barrier, one nurse said, “There’s a lot of [patients] who had it when they were little and they can’t even remember it.” Despite these barriers, nurses believed their thorough documentation of penicillin allergies would improve patient care in the future as prescribers could use the information for antibiotic selection. Nurses cited the STORY mnemonic, pocket card, and EPIC dot phrase as helpful in recording a detailed penicillin allergy history.
Drivers of intervention acceptability and representative quotes
Nurses’ notification of prescribers regarding low-risk penicillin allergies had poor acceptability. Drivers of poor intervention acceptability included low self-efficacy to perform the intervention, the perception that the intervention did not fit with nursing responsibilities and roles (poor ethicality), and the perception that the intervention was ineffective. Nurses expressed a lack of confidence in identifying patients with low-risk penicillin allergies and perceived the risk-stratification of penicillin allergies to be a prescribing clinician—not nursing responsibility. Similarly, nurses perceived their notification of prescribers to have no impact on patient care as antibiotic prophylaxis orders remained unchanged and as penicillin allergy records persisted without update. To improve the management of low-risk penicillin allergies, nurses highlighted the need to better engage prescribers and patients in recategorizing allergies as appropriate (side effect vs allergy). Drivers of intervention acceptability and representative quotes are provided in Table 3. We are aware of no harms or unintended effects resulting from the study.
Discussion
To the best of our knowledge, this is the first study to evaluate the feasibility of an implementation strategy to support recommendations posed by the CDC encouraging nurses to comprehensively evaluate penicillin allergies. 8 We found perioperative nurses’ documentation of penicillin allergy histories using the STORY mnemonic was acceptable and resulted in improvements in penicillin allergy documentation, whereas perioperative nurses’ notification of prescribers regarding patients with low-risk penicillin allergy symptoms had poor acceptability, despite nurses’ notifying prescribers of patients with low-risk allergies. Defining intervention elements and evaluating their acceptability and tolerability are foundational to progressing from proof-of-concept studies to effectiveness trials. Reference Czajkowski, Powell and Adler22 To date, efforts to improve nurses’ evaluation of penicillin allergies have examined nurses’ use of a penicillin allergy risk-stratification algorithm, nurses’ initiation of a penicillin allergy delabeling questionnaire, and nurses’ monitoring of patients undergoing penicillin allergy testing, with mixed results. Reference Bediako, Dutcher and Rao23–Reference Berger, Singh and Loo25 To better understand nursing practice(s) that have the greatest promise for antimicrobial stewardship, we analyzed the acceptability and implementation of two nursing interventions to improve the evaluation of penicillin allergies. We found nurses’ evaluation of penicillin allergies using the STORY mnemonic was highly acceptable and our implementation strategy provides the nuts and bolts for how to engage nurses in this practice, thus addressing calls to improve the evaluation and documentation of penicillin allergies. 8,Reference Guyer, Macy and White26–Reference Manning, Pfeiffer and Larson29
Drivers of intervention acceptability among nurses included self-efficacy and the perceived effectiveness and ethicality of interventions. Nurses believed their thorough documentation of penicillin allergies would improve patient care and cited the STORY mnemonic and associated dot phrase as “straightforward” and “easy to follow”. Recent studies show the importance of STORY fields in predicting penicillin allergy status. A machine learning model using retrospective data from the U.S. found its ability to predict positive penicillin allergy skin testing was strongest when model variables included the symptoms of reaction (particularly hives/urticaria), sex (female), time since reaction, and treatment received for reaction. Reference Gonzalez-Estrada, Park and Accarino30 Similarly, a model developed and validated using retrospective and prospective data from Australia and the U.S. found time since reaction, symptoms of reaction (anaphylaxis, angioedema, severe cutaneous adverse reaction), and treatment received for reaction had a negative predictive value of 96.3. Reference Trubiano, Vogrin and Chua31 Lastly, a multivariate logistic regression model using retrospective data from the UK found the absence of anaphylaxis, time since reaction, and unknown name of the index drug had a negative predictive value of 98.4%. Reference Siew, Li and Watts32 Future research may use natural language processing and STORY information to further improve the accuracy of these models.
Nurses’ notification of prescribers concerning patients with low-risk penicillin allergies had modest uptake (nurses notified prescribers of 14/53 patients identified as low-risk in post-implementation period), which likely reflects nurses’ descriptions of low self-efficacy to perform the intervention, poor ethicality of the intervention, and the perception that the intervention was ineffective. While the implementation strategy included education on low-risk penicillin allergy criteria that were agreed upon by boards of the American Academy of Allergy, Asthma, and Immunology (AAAAI), the Infectious Diseases Society of America (IDSA), and the Society for Healthcare Epidemiology of America (SHEA), Reference Shenoy, Macy, Rowe and Blumenthal2,Reference Carter, Schramm, Baron, Zolla, Zavez and Banach14 nurses described hesitancy in classifying penicillin allergies as low-risk. Nurses also perceived their implementation of the intervention to have no effect on surgical prescriber behavior, which reflects limitations of the implementation strategy. Surgical prescribers were made aware of nursing interventions but were not given specific guidance on the evidence-based management of low-risk penicillin allergies. It is possible that the acceptability of this intervention would have been greater had we more purposefully engaged prescribers.
As a feasibly study, we identified aspects of the implementation strategy that require modification, namely more potent interventions that target nurse self-efficacy and the perceived effectiveness of nurses’ notification of prescribers regarding low-risk penicillin allergies. More frequent and structured formative evaluations may foster an earlier awareness of aspects of the implementation strategy that are working and those in need of refinement. Reference Stetler, Legro and Wallace17 Future studies are needed to determine the effectiveness of nurses’ evaluation of penicillin allergies on clinical outcomes among patients, while identifying optimal implementation strategy approaches. Such effectiveness-implementation hybrid designs allow for the dual evaluation of intervention effectiveness and implementation outcomes Reference Curran, Bauer, Mittman, Pyne and Stetler33 and have been used in studies to implement prevention bundles for non-ventilator-associated hospital-acquired pneumonia and screening strategies for sexually transmitted infections. Reference Wolfensberger, Clack and von Felten34,Reference Medina-Marino, Cleary and Muzny35
Study strengths and limitations
This study has several strengths. Our implementation strategy was guided by the COM-B model of behavior change and our evaluation of intervention acceptability was guided by the Theoretical Framework of Acceptability, Reference Michie, van Stralen and West13,Reference Sekhon, Cartwright and Francis19 which provided a comprehensive structure to systematize research procedures and guide further inquiry. Similarly, we characterized implementation strategies using definitions posed by the ERIC project to facilitate the replication and comparison of these strategies in further study. While we did not conduct a formal economical evaluation of the implementation strategy, costs were limited to the printing of educational pocket cards and resources associated with educational training. The minimal costs associated with the initiative support future scalability. Because this study was limited to one setting, the external generalizability of results is unknown. Similarly, because of practical considerations, we conducted informal formative evaluations with key stakeholders and only one formal summative evaluation (focus group) with perioperative nurses. While our process evaluation did not lend itself to data saturation, Reference Saunders, Sim and Kingstone36 similar perspectives and experiences were conveyed from participants and member checking supported the credibility of study findings. It is also possible our data pull of patients with a penicillin allergy missed relevant records. Although we used the same EPIC data query pre- and post-implementation, thereby reducing systematic error.
Conclusion
In this feasibility study, we found nurses’ thorough documentation of penicillin allergies was highly acceptable and improved following a theory-informed implementation strategy. We offer an innovative, theory-informed approach to improve penicillin allergy documentation and include implementation strategy components to foster adoption among those interested in engaging nurses to improve the evaluation of penicillin allergies.
Supplementary material
The supplementary material for this article can be found at https://doi.org/10.1017/ice.2024.119.
Acknowledgements
We thank the nurses and prescribers who participated in this study. We also thank Katharine Falotico, BSN, RN, and CIC for taking detailed field notes during focus group discussion.
Financial support
This study was supported by the Research Foundation of the Competency and Credentialing Institute (CCI). E.J.C. received grant support from the Patterson Research Mentored Award.
Competing interests
None.
Open access
