Factors Influencing Uptake of Evidence-Based Antimicrobial Prophylaxis Guidelines for Electrophysiology Procedures

Abstract

Background:

Clinical guidelines support early discontinuation of antimicrobials after cardiac device procedures, however, prolonged courses of antimicrobials are common.

Methods:

We conducted semi-structured interviews with 13 electrophysiologists (EPs) representing diverse geographic and clinical settings of care to identify perceived barriers and facilitators to discontinuing post-procedure antimicrobial prophylaxis as part of a formative evaluation prior to implementing a program to improve uptake of guideline recommendations. A directed content analysis approach was used to map responses to the Implementation Outcomes Framework.

Results:

Data indicated that EPs were not willing to stop post-procedural antimicrobials, indicating a lack of acceptability of clinical guidelines. Feasibility, fidelity, cost and appropriateness were also frequently cited. Factors associated with prolonged antimicrobial prescribing included beliefs about lack of harm and possible benefit. There was a strong “cultural inertia” to conform to institutional normative practices. Reasons for conforming ranged from streamlining processes for clinical staff and concerns about being perceived as an “outlier.”

Conclusions:

Institutional culture and beliefs about consequences of cardiac device infections versus antimicrobial use appeared to be major drivers of current practice. The desire promote institutional standardization suggests that strategies to enhance implementation of prophylaxis guidelines must include facility-level changes, rather than individual-provider level interventions.

BACKGROUND:

More than one million cardiovascular implantable electronic devices (CIEDs), such as pacemakers and defibrillators, are inserted annually. Infections complicate 20,000 cardiac device procedures every year, directly resulting in over 5,000 deaths and $1 billion in excess medical costs.^1–3 Evidence-based guidelines for optimizing outcomes following CIED procedures include appropriate administration of pre-incision antimicrobials with discontinuation within 24 hours after skin closure.^4,5,6 These guidelines are based on high-quality evidence demonstrating that continuing antibiotics after skin closure does not reduce infections but does increase the incidence of other adverse events, such as Clostridiodes difficile infections.^7,8,7,9 However, despite the strength of the evidence,¹⁰ implementation of evidence-based antimicrobial use practices in the electrophysiology (EP) laboratory is limited.¹¹

Key questions about the strategies for promoting the uptake of evidence-based practices—and de-implementing low-value interventions, such an inappropriately prolonged antimicrobial use--remain. Little is known about and what factors influence and drive provider antimicrobial decision making and behavior in the EP laboratory. Thus, we sought to identify factors that are driving current practice patterns and may have an impact on implementation of an evidence-based infection prevention program. We also sought to assess factors associated with adoption and lack of adoption of guideline-concordant practices.

METHODS:

Semi-structured interviews with EP providers across the country were conducted as part of a formative evaluation prior to developing a multi-faceted implementation strategy to promote unlearning of ineffective antimicrobial use.^12,13 A major goal of these interviews was to identify barriers and facilitators to adoption of EBP and de-adoption of guideline-discordant practices.

Study participants:

Study participants were recruited through multiple avenues, including referral from the American Heart Association network of clinicians and researchers, clinician networks with requests and referrals from colleagues, and a snowballing approach in which participants were asked for other providers who might be willing to participate.

Semi-structured interviews:

After development of a preliminary interview guide, clinician input from specialists in infectious diseases and cardiology was requested and the questions were refined. A final interview guide was then reviewed and approved by authors prior to data collection. Interview questions focused on processes of care, patient flow, and infection prevention and antimicrobial stewardship strategies in the EP laboratory. A major focus was current antimicrobial prescribing practice and how current practice was developed and changed over time. Semi-structured 30 minutes interviews were conducted in person and over the phone and directly transcribed to Microsoft Word during the interview process. All Interviews were conducted by a single person (WBE). As new, information emerged, specific data exploration was actively pursued through additional questions and adjustment of the interview guide.¹⁴ Interviews were continued until data saturation was reached and few new concepts were emerging.

Data analysis:

Extensive interview notes were taken, and details from the interviews were analyzed using a directed content analysis approach,¹⁵ using an a priori coding frame developed from the Implementation Outcomes Framework, which was selected based on study goals. The Implementation Outcomes Framework focuses on eight key concepts that have been shown to impact implementation; these include acceptability, adoption, appropriateness, cost, feasibility, fidelity, sustainability, and penetration.¹⁶ In addition to these eight domains, coding through a directed content analysis approach is flexible and allows for the emergence of new themes to develop from the data, outside of the a priori identified coding framework.¹⁵ Each line of interview data was read and early ideas about which implementation outcome, or a potential new code, was described in the margins of the interview notes. New codes, for example, those that did not map to one of the eight implementation outcomes, was initially defined in a codebook for discussion and classification at a later point. Two analysts (WBE, RE) independently coded one interview, and then met to discuss codes and assess reliability and validity of this early coding. Discordant coding was resolved through discussion. Following this discussion and consensus session, a further four interviews were coded independently by two analysts, repeating the discussion and consensus process, until all coding disagreements were identified, and codes were well defined. Per the a priori analysis plan, at least 25% of interviews were also coded by a second analyst (RE) to ensure accuracy and consistency of coding results.

Microsoft Excel was used to organize the qualitative data and coding.

RESULTS:

19 electrophysiologists were contacted through various approaches; 13 participated in the semi-structured interviews (response rate, 68.4%); 11/13 over the phone and 2 in-person. 5/13 interviews (38.4%) were double-coded for reliability. Providers represented diverse geography (Northeast, Northwest, South, Southwest, Midwest) and work in different types of healthcare settings (private, academic, VA hospital).

General overview:

Responses to questions about antimicrobial use patterns and barriers and facilitators to adopting evidence-based practices most commonly mapped to the acceptability outcome; feasibility, fidelity, cost and appropriateness were also frequently identified; sustainability and penetration were infrequently cited. See Table 1 for interview guide question, coding, and exemplar quotes. Within the implementation outcomes, several subthemes emerged. These included beliefs and knowledge, which impacted acceptability, appropriateness, and adoption; local culture and normative behaviors, which impacted feasibility, acceptability, and adoption; organizational structure, including the hierarchy of providers, environment, cost and resources, which impacted feasibility, adoption, and acceptability. People—particularly local content experts—emerged as key facilitators of practice by influencing the acceptability (or lack thereof) of guideline-based practices.

Table 1.

Summary of Interview Guide Questions, Codes, and Example Quotes

Interview Question	Code	Example
Tell me about the pre-procedural area	Environment of care, resource limitations	The [cefazolin] is fast and we can just infuse that in the room before the incision. The [vancomycin] is really the one that gives us trouble. We often don’t’ have space. Lots of space constraints. There just isn’t room in the recovery room. We really hope they are MRSA negative.
What are some common barriers that you encounter?	Feasibility, environment	[Vancomycin] we have had problems with. With weight-based dosing, [vancomycin] takes 4–5 hours, on average, to infuse. The [operating room] just runs it in quickly, but our nurses have to follow pharmacy instructions. In the [operating room], the anesthesiologists can do what they want and infuse it themselves…. But we have problems. For example, some guy shows up at 8AM for the procedure. They can be stuck waiting 4–5 hours to complete the infusion and then not make it home and then they are very angry and frustrated. This was a big point of contention. The OR can operate what it wants due to its staffing, but the EP lab is not the OR. We don’t operate the same way and don’t have the same resources… There are lots of things from the operating room that don’t translate well to the [electrophysiology] lab.
What is the protocol for antimicrobial prophylaxis?	Adoption	Decision to stop post-procedure is not universally accepted… they are worried about hospital-based patients and their risk of infection.
How is antimicrobial prophylaxis administered?	Fidelity	It is all over the place. Everyone does something different.
Tell me about your infection prevention protocols	Process	Most of our patients, especially if outpatients, they go home 1 hour after the procedure is complete. We don’t send them home with any antibiotics or give them any more intravenous antibiotics before they go. It just doesn’t make sense. For inpatients, there is an order set. In the order set, there is the option for ordering post-op antibiotics. If you do, then it is 2 more doses of cefazolin, or one more dose of vancomycin. But most of us don’t do that, we just give the one dose before the procedure.
Tell me about the pre-procedure infection prevention process	Process, Standardization	There is not a whole lot that patients do in terms of pre-procedure… We have not been really protocol-driven.
Tell me about antibiotic use.	Anchoring, tradition	I only give a single dose because it is what is encouraged by the hospital, by infectious diseases’s view of this. But I think that old habits die hard.
Tell me about post-procedure antibiotic use	Fidelity, knowledge and beliefs, anchoring/medical training and education	It kind of varies by each person’s level of superstition and how much you believe the devil is in there… I think that we often hold so fastly to what we learned in fellowship. When I was in fellowship, I learned that it is a black and white world, you do what you are told and everything works out. Then you get out into the real world and nothing is clear…we have a natural tendency to harken back to a time when things were determined for you and you didn’t have to make any choices.
Tell me about what lead you to change practice.	Acceptability	When I saw that no one was really giving them here, I checked the literature and there really is no evidence for it… There really is no data for post-operative antibiotics.
What do you think lead you to change your practice?	Acceptability, normative behaviors and local culture, knowledge, adoption	I think everyone sort of adopted [stopping antibiotics] on their own. It was pretty apparent was that most people were not giving antibiotics. As we all transitioned from keeping people over night to going home the say day, we all decided what is the point? If they are safe to go home, then why would they need antibiotics? This became the consensus… we all decided on our own and we all decided this is what we have to do, it’s what makes sense to do.
How did the change occur in your institution?	Knowledge, normative behaviors/institutional culture, influencers/local experts	We talk about it at faculty meetings and have group didactics. It has also helped that, within the last year or so, a large Canadian trial (the PADIT trial)4, came out. It was a weird study design, weird and it took forever and was way too complicated, but the upshot was that infections were very unusual and extended antibiotics did not seem to make a difference, either clinically or statistically… I think that trial went a long way toward reassuring my colleagues. One of the rhetorical ploys I used was to point out that electrophysiology in our hospital was a relative outlier, so compared to cardiothoracic surgery, orthopedics, neurosurgery, where they also put in devices that are hard to take out, they don’t do this. A major part of the argument [for continuing antibiotics] is that, if a device has been in there for decades and it is scarred down, it is really hard to take out, so what’s the big deal to give a couple of days to prevent an infection when it is really hard and complicated to take out the device? So just told them, cardiothoracic surgery, orthopedics, neurosurgery, they all do procedures with devices and implants that are also really hard to take out, it is hard to take out a deep brain stimulator, it is hard to take out a total hip replacement, yet despite that, orthopedics and neurosurgery don’t do this, they don’t give post-op antibiotics. They are also dealing with immunoprivledged spaces, so the ability to clear even a small smoldering infection, is basically non-existent, but they still don’t do it. Yet we are still giving way more antibiotics than these people seem to. For some, that was a powerful argument. The other thing was the antimicrobial stewardship support team, [they were] helpful.
Tell me about how infections and other adverse events are tracked.	Process	We have a homemade system for tracking EP lab complications, but I am not sure how comprehensive that is. It kind of depends on self-reporting, and I am not sure how accurate that is or how much gets captured
Do you ever use antimicrobial coated pockets?	Cost, resources	We have them. We really don’t use them. They are pretty expensive.
Tell me about antibiotic pockets	Adoption, cost, resources, knowledge	It is hard to know what the infection rate would be if you don’t use it because now I have been using it for a while, but I am convinced that if there is an infection risk of 3%, then it does reduce it to 1%. On 200 procedures, it is hard to know if you are preventing 1–2 infections per year. It does not make sense for the hospital budget for everybody, but for those patients with a 3–4% risk of infection, it is the reasonable thing to do.
How are infections defined?	Knowledge, standardization	It is really the physical exam. Is the pocket hot, red, fluctuant, is the patient febrile?

Within acceptability and adoption outcomes, providers reported that a major driver of current antimicrobial use and prevention practices is concern for their patient and awareness of the impact and severity of a CIED infection (Table 1). Beliefs about costs, and concerns about some clinical outcomes (e.g., CIED infections) and not others (e.g., C. difficile infection, antimicrobial-associated adverse events) impacted the acceptability (or lack thereof) of discontinuing antimicrobials after skin closure. Providers expressed a strong desire to reduce CIED infections, which they view as a devastating complication, and an important focus of prevention efforts. Electrophysiologists also indicated that while there may be no evidence to support the effectiveness of prolonged antimicrobial prophylaxis, there was a perception about a lack of harm associated with giving antimicrobials and a lack of evidence that stopping the antimicrobials immediately after skin closure is safe. For example, providers reported that they were unaware of any of their patients ever developing C. difficile infection after receiving prophylactic antimicrobials (Table 2).

Table 2.

Existing Surveillance Infrastructure, Variability in Reporting, and Definitions of Infections

Process, adoption, standardization	I see all of the patients and I keep my own list of infections. I see them within 1–2 weeks and then again at 2–3 months, then as needed. If I saw an infection at one of these visits, I would note it.
Adoption, resources	It is a weakness here because we don’t have much infrastructure.
Processes of care	We track hospitalizations and ED visits. The coordinator comes to us with visits or procedures, then the physical adjudicates whether it was related or unrelated to the procedure.
Adoption, fidelity, process of care, organizational structure, local experts/influencers	The hospital sends out surveys to all providers to follow up on their patients to see if they have had any infections. Then if you have had an infection, you have to check the box. It requires honesty on your part. We also have a quality analyst, who tracks for us, tracks all hospitalized patients with infections for 90 days after the procedure. Then the [infectious diseases] service has its own method for tracking infections… Other than infections, the other thing we track is mortality. We had been using the ICD registry to track other infections, but as of last year, Medicare made it no longer a requirement to do the registry so most of us stopped using it.
Adoption, fidelity process, resource limitations, organizational structure	It is easy. There are only two electrophysiologists and we are responsible for reporting our infections. There might be underreporting of a stitch abscess, or cellulitis or something, but I don’t even think that would even count in our complication list… If the pocket is involved, then the device comes out and that’s a big deal and we would report it and we would know about it. Since I started keeping track of every complication 12 years ago, I have a huge spreadsheet with the number of devices, so I have the denominator and, the number of complications, so we have the numerator.
Adoption, fidelity processes of care, resources	C. difficile probably wouldn’t get picked up, although [we] would probably report it. Honestly, I have not seen it. But I would probably report it if I did hear about it… I can’t think of a single case I have heard of, antibiotic-associated diarrhea, sure. If [C. difficile] was bad enough to get [the patient] admitted, we would say something about it… We don’t look for it, so we would not know about it.
Adoption, processes, resources, costs	We track anything that is a significant complication… We track them by the physician tracking their own complications… The tracking is a challenge because it is hard to do systematically… The resources we have for tracking are not extremely robust (laughs). So, it is mostly on the treating physician to recognize and then compile in a log. Then we submit to our quality and safety committee and it is reviewed quarterly by everyone… the quality people will also generate a list based on coding. Not always that accurate, but the best we have for now.
Adoption, resources, costs, local adaptation	We don’t have a tracking system, but I run the extraction program, we are working on building something through EPIC that will send alerts to me and my partner about patients with devices in place or recent device procedures who have positive blood cultures.
Adoption, resources, process, standardization	Aside from what is documented in the EMR, there is no tracking system.
Adoption, resources, process, standardization, organizational structure	There is no [surveillance and tracking system] that I am specifically aware of. If there is an infection, I report it to our nurses, and they may have a more formalized tracking system.
Resources, fidelity	Infections…we do keep track of them. As you know, device and pocket infections don’t happen right away so that makes it a lot harder.

Infrastructure Barriers:

Within the feasibility, adoption, and acceptability outcomes, environmental factors and resources were major themes that impacted guideline uptake. Environmental factors, such as the size and capability of the pre-procedural area lead to substantial implementation barriers. Several providers reported shared pre-procedural areas with other specialties (e.g., interventional cardiology, interventional radiology) and that, in many cases, the pre- and post-procedural areas were the same. This translated into a limited ability to implement certain types of infection prevention interventions that are commonly used in traditional surgical settings, such as pre-operative showering protocols. Federal Drug Administration (FDA) regulations dictate that the commonly used antibiotic vancomycin be infused over at least 90 minutes, and often longer, however, following these regulations requires clinical space that may not be available.

In the OR, the anesthesiologists … infuse it very quickly and take on that risk. But we [electrophysiologists] have problems… We don’t operate the same way and don’t have the same resources. There are lots of things from the operating room that don’t translate well to the EP lab. Its’ just not the same (Table 1).

The feasibility and acceptability challenges of administering vancomycin are so acute that several providers reported developing their own work arounds to avoid using this medication, some of which lead to unintended practice improvements. One provider reported,

In my recent practice, I have encountered a lot of patients with penicillin allergies… I started sending to the allergist and most of them have fake penicillin allergies. If they do have a fake allergy, I give [cefazolin].

Beyond limitations of the physical facility, providers also reported limited IT and other infrastructure to support surveillance services, including measurement and tracking of infections and other adverse events. Lack of knowledge about what happened to patients contributed to lack of acceptability about adopting evidence-based antimicrobial use practices as electrophysiologists were not convinced that if they made a change that resulted in an increase in the rate of CIED infections it would be detected. Further, non-cardiac complications are not routinely tracked. Among providers who did report having a surveillance system, it was often home-grown with individual work-arounds in place. A consistent finding was that surveillance systems relied on self-reporting, focused on severe CIED infections requiring re-operation or readmission, and that limited systematic resources are available to support measurement. Hematomas and re-admissions were commonly measured; other adverse events, such as C. difficile infections and acute kidney injuries resulting from antimicrobial use were not measured, with some providers acknowledging that they would not learn about these types of complications (Table 3).

Table 3.

Implementation Outcomes and Emerging Codes Associated with Change and Adoption of Evidence-Based Practices

Acceptability, adoption, resistance to change, local culture	There are people who buy into (stopping antibiotics) and there are people who don’t buy into it… They don’t believe the data for stopping antibiotics is robust.
Acceptability, fidelity, knowledge, medical training, anchoring, normative behaviors and local culture, processes	For antibiotics, typically we have not been giving post-procedure antibiotics to the bulk of the patients. Some patients who are deemed to be high risk by the treating physician will be given post procedure antibiotics, but that is relatively rare. In fellowship, we were treating everybody with a standard course of antibiotics, usually for about 3 days for a de novo implant and then 5–7 days for a generator change. That was standard. But, when I came to practice, they [the other providers in my new practice] had moved away from that [giving antibiotics post-procedure]. In fellowship, there was talk about moving away from it, but there was not a lot of consensus. When I came to the new practice, there was sort of a protocol in place… most people had not been doing it. That made stopping easy. It was definitely a change. It has been something that has been on our minds for a while from a stewardship and patient safety perspective… and I was interested in probably giving it up before I left fellowship. It was nice to come to a place where people were doing it routinely so I could make that shift. Definitely facilitated the change to come to a place where people were already doing it.
Acceptability, fidelity, standardization, normative behavior, knowledge	Where I trained, everyone got 3 days for a new implant and 7 days for generator changes. When I joined my practice, I gave no orals and I have not noticed any changes in my infection rates. Also, there is really no evidence to support it in EP. You will find wide variation in that, how people deal with antibiotics after implants. I just followed practice patterns at my new place, to keep things simple and to make it easy for the staff. I also made the change based on their low infection rates even though they were not using post-op antibiotics. There’s really no evidence for it in EP and looking at the data from orthopedics and also surgical data, that really shows that for surgical site infections, it’s preoperative antibiotics and not necessarily post-procedure antibiotics, but I admit that I am not familiar with that data. I would say that overall, half give, half don’t. Where I trained, it is still pretty common.

Implementation Outcomes Associated with Adoption of Evidence-Based Practices:

Acceptability and fidelity emerged as significant challenges to de-implementation of inappropriate antimicrobial use practices in the EP laboratory. Although several providers were concrete about typical clinical process questions, later responses from the same provider suggested significant lack of fidelity in the form of variation both within a single provider and between different providers. This emerged as providers discussing lack of standardization in their practices and wide variation within a facility and between facilities (Table 1).

Adoption of evidence-based practice was heavily influenced by a strong pressure to conform to normative practices within an institution, which impacted acceptability and feasibility of change; thus, providers who moved to an institution where guideline-concordant practices had already been implemented reported frequently changing their behavior and adopting the practices standard in their new facility. Reasons for this pressure ranged from reports of streamlining processes for non-clinical staff and concerns about being perceived as an “outlier” engaging in non-standardized care (Figure 1).

Figure 1.

Factors associated with adoption and lack thereof of guideline-based practice

To consolidate care, I followed what the majority of my partners were doing in my new group, to make things easier on the nurses and the other staff who are involved in taking care of the patients. I wanted it to be simple and straightforward for them and to limit confusion.

Anchoring, tradition, and resistance to change were frequently cited as reasons for maintenance of guideline-discordant practice and lack of acceptability of change. Beliefs about the quality of the available evidence were also cited as reasons for lack of acceptability. Early physician education was another factor driving long-term practices and acceptability of change, regardless of more recent evidence refuting the intervention. Beliefs about the lack of harms associated with antimicrobial use were cited as reasons to continue the practice, rather than to de-implement it.

A major part of the argument is …what’s the big deal to give a couple of days to prevent an infection when it is really hard and complicated to take out the device?

Among providers who did make practice changes and adopted de-implementation of post-procedural antimicrobial use, adherence to local culture and practices were again cited as major factors that facilitated the transition. Local experts and clinical trials evidence from EP and traditional surgeries with implantations, such as orthopedic total joint replacement procedures and neurosurgical procedures, were also reported as important drivers of adopting and sustaining practice change (Table 3). Providers cited other subspecialists generally accepted as experts in the areas having significant influence into their clinical care decisions (Figure 1).

Discussion:

Surgical site infection prevention guidelines recommend discontinuation of antimicrobials within 24 hours of skin closure for cardiac device implantation procedures.^5,6 These guidelines are based on strong clinical trials evidence that continuation of antimicrobials after skin closure does not provide additional risk reduction but does increase adverse events.^8,7 Implementation outcomes associated with adoption of guideline recommendations frequently identified during these formative interviews included acceptability, feasibility, fidelity, cost, and appropriateness. Sustainability and penetration were rarely identified, which is not surprising given the formative nature of this work, since these outcomes occur in the later stages of implementation.

Implementation outcomes associated with adoption of clinical antimicrobial prophylaxis guidelines included acceptability, appropriateness and fidelity. Themes associated with de-implementation of prolonged antimicrobial use included knowledge about clinical evidence from EP studies and other areas of clinical medicine, normative behaviors, and local culture. Providers who adopted guideline-based practices reported that they were primed to adopt these practice changes prior to switching healthcare systems, but that a major driver of uptake was moving to a medical system where guidelines were already implemented.

A key factor driving inappropriate antimicrobial prescribing is the belief that antimicrobials may be helpful for reducing infections and that they do not carry risk. Providers are aware of and concerned about the potentially highly morbid consequences of CIED infections but reported that they did not remember seeing any patients who developed negative effects from prolonged antimicrobial use, such as C. difficile infections. Electrophysiologists generally do not receive information about non-cardiac adverse events, because they are not involved in their diagnosis, treatment, or management. This imbalance in information – substantial feedback about cardiac device infections but no feedback about harms of antibiotics – may contribute to biased decision-making. This elucidates an opportunity to improve acceptability and ultimately adoption of early discontinuation of antimicrobials following skin closure (Figure 2). Putting these pieces together, development and dissemination of a centralized surveillance system – with tracking and audit and feedback of infections and other non-cardiac adverse events to provide a complete picture of the patient’s health– may be an effective de-implementation strategy to promote unlearning of ineffective antimicrobial use practices.¹⁷ Such a centralized system for feedback has the potential to address two barriers to practice change identified in these formative evaluations. First, a centralized surveillance system could provide information about benefits (e.g., reduction in CIED infections when pre-procedural antimicrobials are administered) and harms (e.g. C. difficile infections in patients with inappropriate post-procedural antimicrobials) of the prevention strategies. Second, if benchmarking and reporting of other providers is included, cultural and normative influences may be impacted, thus changing the apparent acceptability of early discontinuation of antimicrobials. Several providers reported peer and culture pressure to conform to local practices. Demonstrating a larger cultural shift therefore might be helpful for supporting uptake of guideline-concordant care (Figure 3). Further, this model of a feedback system that includes diverse, rather than limited clinical outcomes, could be applied in other areas of medicine, as siloing and lack of information about the complete picture of clinical outcomes are common barriers, particularly in cases with multiple sub-specialists each managing only a small aspect of the patient’s overall care.

Figure 2.

Model for information flow and impact on clinical decision making in the electrophysiology laboratory

Figure 3.

Model to promote de-implementation of guideline-discordant antimicrobial prophylaxis in the electrophysiology laboratory

This study was limited by the small number of study participants, who may not be representative of all electrophysiologists. However, reports of clinical practices are consistent with what has been reported through surveys and electronic data extraction, somewhat mitigating that concern.^7,18,19 The methodology in this study goes further than survey and structured data to elucidate provider impression of what drives their current practice. It is possible that providers may have tailored their responses due to understanding the nature of the study and the background of the interviewer. This would tend to bias the results toward finding higher rates of uptake of clinical guidelines and would further support the need for de-implementation research. Lack of recordings may have reduced the quality of the data collected, however, direct transcribing of interviews is an accepted practice for collecting qualitative data. Use of one interviewer to conduct all interviews may have impacted the nature of the interviews and the data that was collected. Finally, it is possible that coding of qualitative data would have been different had different investigators completed the coding; the impact of this limitation was minimized by the double-coding check to ensure agreement.

CONCLUSIONS:

Electrophysiologists expressed a strong desire to optimize clinical outcomes for their patients but had limited understanding about the harms associated with guideline-discordant antimicrobial prescribing. Based on current understanding of unlearning, a critical part of promoting guideline uptake may include audit and feedback about non-cardiac outcomes, such as C. difficile infections, to promote understanding of risks and benefits of prolonged prophylaxis regimens.¹⁷

HIGHLIGHTS:

• Prolonged antimicrobial is driven by a desire to optimize outcomes.

• Providers know about harms related to their specialty.

• Providers have limited knowledge about other adverse events outside of their field.

• A critical factor for promoting de-adoption may include expanding feedback to include all types of adverse events, not just specialty-specific ones.