We interviewed 20 clinicians. Participants included 3 hospitalists, 5 inpatient pharmacists, 1 infectious disease physician, 2 anti-microbial stewardship pharmacists, 4 primary care providers, 2 outpatient pharmacists, 2 resident physicians, and a nurse case manager for the allergy service. We should note that the outpatient clinicians we interviewed for this study did not participate in the previous quality improvement based initiatives at our facility. Therefore, they had not received any penicillin allergy education and were not provided access to the CDST prior to our outpatient interviews. However, given the need for future involvement of our outpatient healthcare team, we recruited them to gather preliminary data surrounding possible and perceived barriers to expanding penicillin allergy de-labeling interventions to outpatient settings.

The factors that contributed to barriers to penicillin allergy evaluation and de-labeling were classified under six TDF domains spanning both individual and system-level determinants. In our study, we found that the factors related to knowledge, skills, beliefs about capabilities, beliefs about consequences, environmental context and resources, and professional role and identity were the most prominent barriers to penicillin allergy evaluation (Fig. 2).

Fig. 2

Key barriers to penicillin allergy de-labeling process for inpatients

Inpatient settingKnowledge, skills, beliefs about capabilities and consequences

All inpatient clinician groups were aware of the scientific evidence supporting the penicillin allergy evaluation. Specifically, participants were familiar with data showing increased risk for comorbidities in patients labeled as penicillin allergic. However, they reported a need for further education on de-labeling benefits and reassurance for its safety. Health care professionals in all groups cited apprehension about inducing an allergic reaction with test dose challenges and having inadequate skills and resources to treat a possible allergic reaction. Despite this, both inpatient pharmacists and hospitalist physicians were amenable to gaining the skills to identify low-risk patients and perform test dose challenges in low-risk patients in the future. In addition, pharmacists and hospitalists felt that they needed frequent practice to maintain familiarity and comfort with the process.

When asked questions regarding using the CDST available to aid history taking, risk stratification and de-labeling of penicillin allergic patients, clinicians in all groups reported a lack of knowledge on where the CDST could be found within the EMR or how to apply it. Upon reviewing the penicillin allergy algorithm (Additional file 2: Appendix S1b), residents and pharmacists noted that the tool was straightforward. “I think [the] algorithm is really helpful. I think the biggest barrier is if [patients] don’t remember the reaction, or can’t get enough information to feel confident, but the algorithm itself is very straightforward” (Resident 1).

However, despite this positive feedback about the CDST, infrequent engagement and lack of practice with the task diminished clinicians’ beliefs about their self-efficacy to effectively participate in the penicillin de-labeling process. Infrequent engagement with de-labeling also influenced a perception that identifying suitable patients took a lot of time. Participants stated that they would need to hone their skills to navigate the EMR system and access the patient history and lists of previous and current medication lists. A lack of self-confidence in their skills with risk stratification and treating possible adverse reactions were noted as specific barriers. For example, a resident said they are not very comfortable with treating patients who may have adverse reactions during the drug challenges: “I know drips and epinephrine things. I just haven't had the opportunity to treat many patients with an acute reaction” (Resident 2).

Even though most participants acknowledged the positive consequences of removing allergy labels from patient records, the fear of erroneous de-labeling and patients having a serious allergic reaction as a consequence prevented clinicians from engaging in the de-labeling process. For example, hospitalists were worried that adverse reactions would add complexity to patients’ clinical care or extend their hospital stay. Pharmacists raised concerns regarding disciplinary action following an error. Several pharmacists noted that if there was evidence in the EMR that a patient was prescribed penicillin in the past, they felt confident to update EMR with a note that the patient tolerated penicillin in the past. However, they were still reluctant to de-label without consulting the patient’s hospitalist, primary care doctor or an allergy specialist, suggesting a lack of confidence or trust in interpreting allergy data in the EMR. “I think it is really tough to take that allergy off the chart unless the patient gets that specific antibiotic while they're here, we have the discussion with the providers that the patient tolerated it just fine and are comfortable that documenting that we're pulling it off the chart completely” (Pharmacist 5).

Professional role and identity

As demonstrated in Fig. 1, penicillin evaluation and de-labeling is a multidisciplinary process requiring a collaborative approach. The interdependent nature of the process requires high levels of coordination and communication within and among teams. When a patient is admitted to the hospital, the pharmacy team (either inpatient pharmacist or antimicrobial stewardship pharmacy) usually initiates the de-labeling process by identifying patients with penicillin allergy and conducting a medication reconciliation based on patient’s medical records. In the next step, the patient is risk stratified and direct drug challenge is recommended if the patient is low risk. In the current workflow, the Allergy Consult service evaluates and de-labels the patient, although ideal future workflows would empower inpatient pharmacists and hospitalists to fulfill this role for low-risk patients. Currently, pharmacists and hospitalists expressed that they needed the Allergy service’s approval to de-label a patient especially if there is disagreement among teams.

“There's sometimes a little bit of disagreement with the history taking and the one that comes up all the time is, did the patient really have hives or true urticaria? And then almost always in that situation, we default to the most conservative or safest option, [between] skin testing, getting allergy involved, or doing an oral test dose”. (Hospitalist 1)

Clinicians noted that obscurity on which teams would take the lead on de-labeling created barriers to developing robust workflows in clinic. In addition, discomfort with the tasks that did not clearly fall under a specific specialty—such as ordering of the test dose and monitoring the patient during oral challenge—discouraged clinicians from engaging in the de-labeling process. Specifically, neither pharmacists nor hospitalists felt that removing the allergy labels from patient records fell within the inpatient teams’ current professional roles. Because there was not a point person or a group who clearly championed the initiative, the task ended up “bouncing” among teams and fell through the cracks (Table 2). This, in combination with other environmental stressors, resulted in inconsistent clinical workflows and variable application of the penicillin allergy CDST.

“I guess it's a little unclear [who takes the lead on de-labeling]. Um, I think that, you know, teams, individual medical teams do try to do something. It is certainly not very systematic amongst the teams” (Hospitalist 2)

Table 2 Representative quotes demonstrating perspectives of inpatient and outpatient cliniciansEnvironmental stressors, resources, and organizational culture

Pharmacists, hospitalists, and specialty consult services described an organizational culture where workload and competing priorities prevented implementation of penicillin allergy protocols in the inpatient setting. The teams’ abilities to focus on patients who are penicillin allergic were hindered by the need to prioritize other competing quality measures and exacerbated by the limited inpatient bed availability.

“I think from an inpatient perspective, it's probably the culture that ‘we need to address the things that need to be addressed as an inpatient, and the rest can be pushed to outpatient world.’ So that tends to be a general thought process. And it's sometimes appropriate, and sometimes it isn't, and penicillin allergy falls in that bucket. So, I think that is probably something in the organizational culture". (ID MD1)

Participants also said that using an alternative antibiotic was easier than evaluating the allergy. This perception was reflected in workflows, especially in times of stress and periods of competing priorities where individuals defaulted decision making to prioritize discharging of patients. “In terms of time to evaluation and treating the patients effectively, a lot of times using an alternative antibiotic is the path of least resistance if there is an alternative there. But if we’re kind of stuck between a rock and a hard place, and we need that one antibiotic, maybe that is the way to go then. But I feel like I've just seen so far that a lot of times a different antibiotic is picked just to steer clear of that allergy for the time being” (Pharmacist 5).

With easy access to alternative antibiotics, clinicians prioritized other competing tasks and postponed de-labeling to an unspecified time or deferred to an allergist. Although allergists assumed a leadership role by becoming the point person for patients with complex histories, insufficient resources such as staffing and clinical space prevented them from consulting with all potentially eligible patients. Overall, emphasis on rapid discharge workflows interrupted the momentum and often led to patients being discharged before evaluation.

“I think de-labeling is important but right now, the hospital is completely full every day. We are getting messages on the screen, ‘discharge your patients as fast as you can.’ So, everything becomes secondary to getting the inpatient work done and getting the patients out of the hospital as quickly as we can”. (Hospitalist 1)

Team members described how the priority to discharge patients quickly predisposed them to dismiss tasks that may delay discharge. This was exacerbated by time constraints and the precedence to make beds available in case of an urgency, especially during Covid surges. The inpatient healthcare team often deferred penicillin allergy evaluation to a later, undefined future patient encounter: “We can't be here every hour. You're kind of having to pass the buck to somebody else to take care of it” (Pharmacist 3). One exception that facilitated allergy evaluation was if the penicillin allergy specifically affected the patient’s current hospital course.

The lack of adequate staffing to complete daily tasks was also a major barrier to de-labeling. Several clinicians pointed out that shortages of critical team members such as LPNs, and variable hospitalist schedules created barriers to standardizing and implementing de-labeling processes. In addition, inpatient pharmacists were co-assigned to two teams at once, which impeded following a patient through their entire hospital stay and prevented inpatients from being identified early enough in their hospital stay to allow time for an oral challenge. If a pharmacist or ID physician sent an alert to the inpatient team toward the end of a patient’s hospital stay, the team often deferred the task to a later time to avoid discharge delay.

“I think there's always an inherent time limitation, the admission pharmacy med rec isn't put on the chart sometimes for, like, 24 or 48 h after admission… By the time you hit 48 h, we're already planning to get [patients] out of the hospital at that point.” (Hospitalist 1)

Within the busy inpatient workflow process, ineffective communication systems further impeded the implementation of penicillin allergy evaluations. Specifically, the inability to quickly identify eligible patients within the EMR upon admission delayed risk stratification of the patients and subsequent decision making about whether the inpatient could be challenged and de-labeled by the inpatient team. Pharmacists, residents, and hospitalist physicians cited difficulties finding the CDST within the EMR due to the unintuitive nature of the system. Small errors such as not updating the history within the allergy field and indicating the relevant clinical encounter often buried important information in the clinical record, limiting data accessibility. Residents pointed out additional challenges with accessing patient history in the system, especially if they were accessing allergy records from a different institution.

Clinicians discussed a number of factors related to the culture of the organization. The decrease in staffing due to COVID and inpatient COVID surges resulted in siloed teams and reduced opportunities for multidisciplinary discussions. For example, pharmacists noted that they no longer rounded with the teams. Multidisciplinary communications were reduced to Teams messages, which made it harder to provide the team with recommendations about de-labeling and to initiate the process. Both pharmacists and hospitalists described how increasing reliance on asynchronous messaging led to ambiguity in recommendations and created the perception that recommendations to challenge patients were less urgent than recommendations that were given in person. Similarly, suggesting penicillin challenges through CPRS notes was considered as “noncommittal,” as notes were a passive form of communication, compared to recommendations conveyed over a phone call or in person. Hospitalists acknowledged that they did not always prioritize ID recommendations documented in CPRS.

“I think it is a much more passive form of communication of just assigning people to notes. It's very noncommittal by the signature that you've received that, whereas, you know, if you had a phone call, it may convey more importance”. (Hospitalist2)

Inpatient to outpatient transitions

When the inpatient pharmacists and physicians could not de-label a patient during their hospital stay for reasons such as competing priorities, or workflow issues or pressures, they deferred the de-labeling tasks to outpatient care. However, outpatient pharmacists and primary care providers (PCP) in our study expressed several concerns with taking on penicillin de-labeling as a responsibility.

Barriers to de-labeling in primary care settings

PCPs and outpatient pharmacists echoed the barriers described by inpatient clinicians related to knowledge, skills, beliefs about capabilities, beliefs about consequences, and professional role (Table 2). Because these clinicians had not participated in previous quality improvement-based initiatives surrounding penicillin allergy de-labeling, they expressed hesitation about their level of knowledge and training surrounding risk stratification and oral challenges. They reported that they would need reassurance about the safety of the procedures through practical guidance and protocols on risk assessment, while ensuring that only low-risk patients would be de-labeled. They also expressed that even with updated training, they may still feel ill equipped to safely address patients’ potential allergic reactions during oral challenges because of infrequent practice. Several PCPs noted that assessing the accuracy of a penicillin label in patient records had not been part of their workflow in the past so they “[did] not think to assess it.” Additionally, because PCPs did not regularly assess the accuracy of penicillin allergy labels, they did not always remember to refer patients to the allergy clinic: “I think recognition is probably the biggest thing. It hasn’t been part of my workflow in the past to look for penicillin allergy and then to think to assess whether it’s real” (PCP2). While the PCPs thought they could play a role in patient identification by increasing their exploration of patients’ allergy history and referring patients to allergy for further assessment, they expressed that conducting oral challenges would fit better into a specialty role rather than primary care.

Outpatient clinicians also described barriers related to environmental stressors, organizational culture, and resources, and pointed out how those barriers would make it challenging for them to incorporate penicillin de-labeling into their workflows. In particular, they expressed that primary care already has so many other tasks they must cover in each appointment, that discussing and addressing penicillin allergy is a lower priority given their time constraints. Because “identifying low-risk patients and having them go through a 90-min test might be a tough sell to have the team available” (Outpatient Pharmacist 1), they preferred de-labeling tasks to be performed in the allergy clinic or by the inpatient team. Additionally, they felt that lack of emergency resources at community clinics to treat potential allergic reactions, lack of space to conduct challenges, and lack of support from nursing staff due to staff shortages were significant hurdles. Outpatient pharmacists also noted that the CPRS system could be “clunky,” making it difficult to find protocols and access accurate patient history.

“We're also struggling with space concerns at the facility where I work. I just don't think the building management would like to have people sitting around for 2 h when we don't have enough rooms as it is”. (PCP 1)