Antibiotic allergy is a growing public health concern in the pediatric population. Up to 10% of children are labeled betalactam allergic [1], yet the majority can tolerate beta-lactams after being evaluated by an allergist, with more than 90% able to pass a challenge to amoxicillin with no reaction [2]. Penicillin and its derivatives remain the preferred treatment for many childhood infections. Prior studies in children and adults have demonstrated adverse outcomes with a penicillin allergy label including increased hospital length-of-stay and use of broadspectrum antibiotics [3,4]. Penicillin allergy management algorithms for adults have been used to improve antibiotic stewardship [4]. At the time of inception of this project, there was a paucity of data on the effect of applying a penicillin allergy de-labeling algorithm to pediatric inpatients and there was no protocol in place at our institution to address penicillin allergies in pediatric inpatients. We implemented an algorithm to optimize the approach to penicillin allergic patients as part of a longterm quality improvement (QI) project to improve antibiotic stewardship.

A retrospective chart review of pediatric inpatients with a penicillin allergy admitted to a tertiary care center in New York City 4/2018-5/2019 was conducted (approved by the institution QI committee and safety team). Patients with a listed allergy to penicillin or one of its derivatives were identified using a reporting  tool in the electronic health record (EHR) system (Epic Slicer Dicer). An inpatient penicillin allergy algorithm (Figure 1), was designed using information published by Shenoy and colleagues [5].This algorithm was disseminated over a 3-month-period via paper copies displayed in workrooms, e-mail, and educational sessions to residents and faculty emphasizing the importance of addressing penicillin allergy in children. The primary outcome measure was efficacy of the algorithm, which was measured as percentage of inpatients with penicillin allergy labels that were de-labeled during their admission. Secondary outcome measures included the percent change in type of inpatient antibiotic used and the number of allergy referrals upon discharge. Data from the pre-/postintervention periods were summarized as proportions and percentages. Chi-square test or Fisher’s exact test was used to compare the frequency of selected antibiotics prescribed during admission for patients in which beta-lactam antibiotics were preferred to treat their infection. Post-intervention data collection was performed 8/2019 to 8/2020 to assess the algorithm’s impact.

Ninety-four of 2810 (3.3%) carried a penicillin allergy label in the pre-intervention group and 99 of 2227 (4.4%) patients in the post-intervention group. Table 1 lists additional information gathered for each patient who had been labelled as allergic to penicillin. In the post-intervention group, there were no patients that had an intolerance history. As such, none were de-labeled inpatient.

Based on the type of infection, β-lactam antibiotics were the preferred treatment in 72.2% of patients in the pre-intervention group and 81.3% in the post-intervention group. Due to the listed penicillin allergy, some of these patients received alternative antibiotics such as clindamycin and fluoroquinolones. Table 2 presents the frequency of selected antibiotics prescribed during admission for patients in which beta-lactam antibiotics would be the antibiotic of choice to treat their infection.

Three of 94 children (3.2%) were referred for outpatient penicillin allergy evaluation in the pre-intervention group and 19/99 (19.2%) in the post-intervention group. Five of the 19 referred presented to the allergy clinic. Testing was not indicated in two of those five patients due to history of anaphylaxis to beta-lactam antibiotics, and instead complete avoidance was recommended.

Broad-spectrum antibiotic use contributes to antibiotic resistance, increased hospital costs, and overall undermines antibiotic stewardship initiatives. Our intervention showed a decrease in clindamycin and fluoroquinolone use, each by approximately 10% and an increase in cephalosporin use in those that were labeled penicillin-allergic. There was a statistically significant increase in use of cephalosporins between the pre- and post-intervention groups, 38% vs. 71%, respectively (p=.008). After being introduced to the algorithm and to the other educational materials emphasizing the importance of addressing penicillin allergy, providers appear to have chosen cephalosporins more often. It should be noted, however, that use of penicillins between the pre-/post-intervention groups was unchanged and our primary outcome of de-labeling patients of their penicillin allergy was not achieved since none were delabeled post-intervention.

Introduction of the algorithm was associated with an increase in the percentage of patients referred to outpatient allergy (3.2% vs. 19.2%) in the pre- and post-intervention groups, respectively. Although there was a marked increase in the number of outpatient allergy referrals, few of these patients had an appointment scheduled. At the time of data analysis, 5 of 19 patients had an allergy consultation. If each of these referrals led to a penicillin allergy evaluation, this could lead to significant potential benefit by increasing de-labeling and preventing further use of alternative/broad-spectrum antibiotics. Importance of addressing the penicillin allergy label as an outpatient may not have been emphasized upon discharge, which may explain the small proportion of patients that made appointments. While a short statement of the importance of addressing a child’s penicillin allergy and clinic contact information was included in the template entered into discharge instructions as part of our intervention, families were responsible for scheduling. This may be an area of focus for the next iteration of the project.

There were limitations to our study. The algorithm was only available in paper format, which could be overlooked or forgotten when not integrated into the workflow within the EHR. Attempts to incorporate our algorithm into the EHR or to make changes to the EHR’s allergy section were unsuccessful, largely due to inflexibility of certain features of the EHR. Severity of infection may be a confounding variable that was not examined in this study, as “sicker” patients may have more use of broadspectrum antibiotics before an organism was identified. The postintervention timeline included some of the COVID-19 pandemic, with the pandemic’s peak during March-May 2020 in New York City. The pandemic may have affected assessment of penicillin allergies, as many specialties limited visits in the outpatient setting and inpatient consultation was reserved for urgent need.

Our study suggests that implementation of a penicillin allergy algorithm in a pediatric hospital may be associated with increases in outpatient allergy referrals, an increase in use of cephalosporins, and a decrease in use of alternative antibiotics. Future directions include working with the hospital antibiotic stewardship team to incorporate an electronic version of the algorithm into the EHR, which may increase ease of access and utilization. A recent study implemented an electronic algorithm to identify low-risk penicillin-allergic pediatric patients and showed that the electronic algorithm was superior to the paper version at identifying these patients [6]. Increasing outpatient allergy referrals has also been accomplished through targeting pediatric primary care physicians [7], which is also a next step in our study. More education should be provided to those taking care of pediatric inpatients, as front-line providers are crucial in reconciling and addressing listed allergies during the admission process. A recent study conducted among penicillin-allergic pediatric inpatients has demonstrated success in de-labeling patients using a clinical care pathway that equips these front-line providers with the protocol to proceed directly to drug challenge for low-risk patients, without the need for Allergy consultation [8,9]. In addition, increased collaboration with the infectious disease team may further promote antibiotic stewardship and lead to the goal of increased penicillin allergy de-labeling.

1. Abrams EM, Atkinson AR, Wong T, Ben-shoshan M. The Importance of Delabeling β-Lactam Allergy in Children. J Pediatr. 2019; 204: 291- 297.
2. Collins C. The Low Risks and High Rewards of Penicillin Allergy Delabeling: An Algorithm to Expedite the Evaluation. J Pediatr. 2019; 212: 216-223.
3. Lucas M, Arnold A, Sommerfield A, Trevenen M, Broconnier L, Schilling A, et al. Antibiotic Allergy Labels in Children Are Associated with Adverse Clinical Outcomes. J Allergy Clin Immunol Pract. 2019; 7: 975-982.
4. Blumenthal KG, Shenoy ES, Wolfson AR, Berkowitz DN, Carballo VA, Balekian DS, et al. Addressing Inpatient Beta-Lactam Allergies: A Multihospital Implementation. J Allergy Clin Immunol Pract. 2017; 5: 616-625.
5. Shenoy ES, Macy E, Rowe T, Blumenthal KG. Evaluation and Management of Penicillin Allergy: A Review. JAMA. 2019; 321: 188- 199.
6. Iammatteo M, Lezmi G, Confino-Cohen R, Tucker M, Ben-Shoshan M, Caubet J-C. Direct challenges for the evaluation of beta-lactam allergy: evidence and conditions for not performing skin testing. J Allergy Clin Immunol Pract. 2021; 9: 2947-2956.
7. Roberts H, Soller L, Ng K, Chan ES, Roberts A, Kang K, et al. First pediatric electronic algorithm to stratify risk of penicillin allergy. Allergy Asthma Clin Immunol. 2020; 16: 103.
8. Wang H, Kozman M, Pierce H, Ma L, Collins C. A quality improvement initiative to improve primary care referral rates for penicillin allergy delabeling. Ann Allergy Asthma Immunol. 2021: S1081-1206.
9. Bauer ME, MacBrayne C, Stein A, Searns J, Hicks A, Sarin T, et al. A multidisciplinary quality improvement initiative to facilitate penicillin allergy delabeling among hospitalized pediatric patients. Hosp Pediatr. 2021; 11: 427-434.