Background

SAP in the context of global resistance

Surgical antibiotic prophylaxis (SAP) is an extremely high-volume area of hospital antimicrobial use, accounting for 15.5% of all hospital antimicrobial prescriptions in Australia, and significant antimicrobial use internationally [1,2,3]. In the context of a global environment of escalating antimicrobial resistance, SAP is therefore a critical area in which to optimise antibiotic use, to reduce the pressure for antimicrobial resistance (AMR) development, and therefore to preserve antimicrobial options for the future, and to reduce toxicities and risk of Clostridium difficile that is associated with inappropriate antimicrobial use [4, 5]. Long-term optimisation of antimicrobial use, of which SAP appropriateness is an important part, will be a crucial strategic to avoid the impending antimicrobial crisis.

Surgical antibiotic prophylaxis: an evidence-based intervention

The evidence for the administration of SAP is well established and is therefore a critical area for optimising quality of care in hospitals to reduce hospital acquired infection rates and improve morbidity and mortality associated with operative interventions [6,7,8,9]. Interventions to improve compliance with SAP guidelines have been published, with the successful use of collaborative change processes that include audit and feedback [10, 11]. However, there is little known about the effect of such interventions over prolonged time periods, and limited information about the factors that cause such significant overprescribing in this area.

Non-concordance with therapeutic guidelines in SAP

Guidelines are used to streamline the application of a shared body of theoretical knowledge to individual patient clinical scenarios. However, when guidelines are applied to clinical practice, the anticipated uniformity of practice often does not occur. Within SAP, suboptimal practice despite established clinical guidelines occurs both in Australia and internationally [1, 12]. SAP is often inappropriately prolonged [13], re-dosing intra-operatively in prolonged operations is commonly forgotten [14] and prophylaxis is frequently given for operations in which it is not indicated [15]. Non-compliance in this area of antibiotic use and its importance for the antimicrobial optimisation agenda are recognised in international strategies for AMR management, including the EU Guidelines for the Prudent Use of Antimicrobials in Human Health, which lists “audit of perioperative antimicrobial prophylaxis indication, choice, timing and duration” as an essential component of hospital-based AMS programmes [16].

Understanding non-concordance: the social influences on antibiotic prescribing, and SAP

There is a growing body of research exploring the social and behavioural influences on antibiotic prescribing decisions more broadly [17,18,19,20]. The design and implementation of behavioural interventions to optimise prescribing is a strategy in international AMR action plans. A previous qualitative study of anaesthetists, surgeons and nurses, described disagreement about basic aspects of SAP such as the value of prolonged prophylaxis, antibiotic choice, and the definition of SAP [21]. Country-level analysis associates cultural characteristics such as uncertainty avoidance with the use of prolonged SAP [22]. Yet, little detail is known about what underpins enduring practices in SAP (e.g. clinician emotions, institutional cultures, professional norms, and site idiosyncrasies, etc.). In this study, we aimed to explore through semi-structured interviews the experiences and perceptions of surgeons and anaesthetists around SAP prescription and administration, to provide insight into social factors which may be barriers to implementation of evidence-based practice in this area of antibiotic use.

Methods

This study was conducted at a teaching hospital in New South Wales, Australia, in 2017. Ethical approval was granted by South Eastern Sydney Local Health District Human Research Ethics Committee HREC/15/POWH/246.

Data collection

A convenience sampling strategy was used to choose the hospital setting and potential participants. Convenience sampling is a recognised sampling technique whereby participants are recruited and selected based on their proximity and accessibility to the researchers (in this case in a nearby collaborating hospital). A formal invitation letter and participant information and consent form (via email) was sent to 71 doctors which included all of the surgical doctors within the participating hospital. Participating doctors voluntarily responded to the invitation, and completed an individual face-to-face interview semi-structured interview. Of these, 20 doctors (17 surgeons and 3 anaesthetists) volunteered to participate in semi-structured qualitative interviews, during 2016 and 2017. Of the 17 surgeons, 10 were senior and 7 were junior. The 3 anaesthetists included 2 senior and 1 junior participant. 14 participants were male, and 6 were female. The interviews were based on a preformulated guide, initially informed by existing literature, and continually adapted to incorporate emerging issues raised by participants [23]. Interviews were focused on the following domains: experiences of antibiotic use and AMR more broadly; experiences and perspectives on the use of surgical antibiotic prophylaxis; experiences of interprofessional work within the operating theatre; and perspectives on antimicrobial stewardship and the place of AMS within surgery.

Data analysis

The thematic analysis of the data was driven by a framework approach [24], which included the following steps: (1) familiarisation—in which the researchers reviewed the interview transcripts; (2) identification of framework—key themes and issues identified around which the data were organised; (3) indexing—application of themes to text; (4) charting—use of headings and sub-headings to build up a picture of the data as a whole; and (5) mapping and interpretation—in which associations were clarified and explanations worked towards. Initially, two members of the research team (authors A and C) independently coded the data. These were then cross-checked by authors A, B and C to facilitate the development of themes, moving towards an overall interpretation of the data. Analytic rigour was enhanced by searching for negative, atypical and conflicting or contradicting cases in coding and theme development [25,26,27]. Inter-rater reliability was ensured by integrating a number of research team members in the final analysis [24, 27]. All audio recordings, transcripts, coding reports and notes were retained and added to documentation of research aims, design and sampling and recruitment processes and practices to form an audit trail. The COREQ qualitative research reporting checklist was used to ensure comprehensive reporting [28].

Results

Participants

Twenty doctors (17 surgeons and 3 anaesthetists) volunteered to participate during 2016 and 2017. Interviews lasted between 20 and 60 min, and participant recruitment continued until research team members agreed that data saturation was reached. Of the 17 surgeons, 10 were senior and 7 were junior. The 3 anaesthetists included 2 senior and 1 junior participant. Fourteen participants were male, and 6 were female. The surgical specialties represented included; general surgery, neurosurgery, orthopaedic surgery, colorectal surgery, urology, transplant surgery, cardiothoracic surgery, vascular surgery and renal surgery.

Antimicrobial prophylaxis as a low priority in the operating theatre

Multiple participants reflected on the complexity of processes occurring in theatre, particularly in emergency situations or in prolonged operations, which resulted in antimicrobial prophylaxis being perceived as a low priority. Indicative quotations are shown in Table 1. The inconsistency of re-dosing of antibiotics in prolonged operations was discussed by a number of participants. The timing of re-dosing was described as reliant on the memory of the anaesthetist and the surgeon (when and if they remembered, relative to other pressures and priorities). The likelihood of re-dosing occurring was influenced by a number of issues including the emergency of the operation, the clinical concern for infection of the surgeon for that specific operation, and the complexity of the operation. An interesting contrast reported by two participants was the situation of high risk, complex organ transplant operations where participants described that procedures were highly protocolised and therefore guideline-based SAP was strictly followed.

Table 1 Antibiotic prophylaxis is a low priority in the operating theatre: indicative quotations
Full size table

Guideline relevance and lack of confidence in their ability to protect against adverse consequences

There were diverse opinions among participants around the application and relevance of therapeutic guidelines in SAP prescribing (see Table 2 for indicative quotations). Participants perceived that there was an increased awareness of guideline-based SAP in recent years and indeed a number of participants described an increasing trend towards compliance with SAP guidelines, but some described mistrust in the evidence around guidelines. One participant described mistrust in the ability of the guidelines to protect the surgeon from fault if an infectious complication occurred. However, this participant reflected that if AMS advice had been sought, they perceived there was an additional layer of protection against litigation to the surgeon. Multiple participants discussed the use of prophylaxis for operations in which SAP is not indicated by guidelines, or the addition of antibiotics over and above usual guideline recommendations. This was described by several participants as driven by fear of infectious complications. Junior doctors were reported as more likely to request inappropriate prophylaxis.

Table 2 Contesting guideline credibility: indicative quotations
Full size table

Benevolence and non-concordant prolonged prophylaxis

Prolonged antibiotic prophylaxis, beyond antibiotic guideline duration recommendations, was discussed by multiple participants. As shown in Table 3, the influences resulting in prolonged prophylaxis were reflected upon by participants, and included; providing an extra perceived layer of safety for the surgeon (safety from both litigation and also from personal responsibility for a complication), conferring a sense of having done everything possible for the patient to prevent an infectious complication, and conforming to perceived peer practice (including the significant influence from specialist training years) for a particular operation. Participants discussing these factors all reflected on the recognised discord between evidence and these practices.

Table 3 Benevolence and non-concordant prolonged prophylaxis: indicative quotations
Full size table

An organisational culture of improvisation as the norm

Multiple participants from different surgical specialties reported improvised antibiotic prophylaxis strategies such as irrigation of wounds with antibiotic solutions (such as gentamicin) and soaking grafts or prostheses in antibiotic solutions prior to implantation. Indicative quotations are shown in Table 4. Participants reflected on the lack of evidence for such procedures, but perceived that there was limited harm from such techniques and reported increased surgical “comfort” with a procedure through using such idiosyncratic techniques.

Table 4 An organisational culture of improvisation as the norm: indicative quotations
Full size table

Discussion

The persistent mis-use of SAP necessitates an in-depth understanding of what drives enduring suboptimal practices in Australia and beyond, and indeed, what limits change through AMS. In this study, we sought to provide novel insight into some of the perceived factors which mediate current practices, in the Australia surgical context. The insights emergent in this study should be viewed in relation to broader work revealing the disjunctions between attempts at regulation to support evidence-based practice, and the potential conflict of regulatory practices with traditional medical values [17]. There has been increasingly sophisticated and widespread dissemination of clinical guidelines around SAP at a national level (both in Australia and internationally), and regulatory frameworks are placing considerable emphasis on improving practices across the health sector. However, this study demonstrates that where there is a low priority/attention on the antibiotic decision, high prioritisation of a position of benevolence and risk reduction, and high regard for the preservation of clinical autonomy (including a right to improvisation), the influence of clinical guidelines and regulatory practices may be significantly limited.

There are two concurrent dynamics evident here. The presence of norms and the importance of improvisation—both of which may work against guideline concordance. Norms may not be guideline-defined norms, but may represent unwritten rules of practice determined by social, professional and institutionally specific influences. Norms may also represent what is considered a priority in a given situation—for example, as described in this data, it may be the norm for SAP decision-making to be a peripheral issue for the surgical team. Improvisation, and the appropriateness of utilising improvisation in any given situation, is key to high-quality clinical practice, but identifying when improvisation is both appropriate and safe is critical.

The perception of antibiotics and antibiotic decision-making being a peripheral issue is not isolated to SAP—previous work by the authors has demonstrated that antibiotic decision-making is sidelined both in ward rounds and in discussions with consultants, with prioritisation of other clinical issues that are considered more important [17]. The acute nature of the operating theatre environment, particularly in emergent situations, is described in this study to reduce attention on antibiotic re-dosing intra-operatively. Antibiotic decision-making in theatre involves communication between different team members—primarily the surgeon and anaesthetist. Non-technical skills (NTS) such as communication have previously been found to be suboptimal in the majority of operating theatre adverse events, and also are demonstrated to be reduced in crisis situations [29, 30]. The results of this study indicate SAP decision-making as an inherently peripheral issue, and secondly, the sidelining of SAP decision-making in operative emergencies. Errors in communication around SAP and SAP prescribing are likely to increase under such circumstances.

Unwritten rules or norms of practice, including accepted (non-guideline based) prolonged SAP duration for particular operations/specialties are described in this study. The phenomenon of unwritten ‘rules’ guiding medical decision-making has previously been described in operating theatre settings where doctors have been demonstrated to disregard clinical guidelines in deference to rules determined by social groups and influences within their professional streams [31]. Similarly, in the context of SAP, this data supports the existence of a ‘back stage’ of social norms, which shape action, and which need to be acknowledged with AMS processes. Consideration of the origin, and perpetuating factors underlying such rules, will be critical in the design and implementation of sustainable strategies to optimise SAP prescribing. The ability to challenge norms must first rely on identifying that they exist, and then realising that they result in behaviour that does not conform to evidence-based practice. In this study more junior doctors were reported by participants to more commonly request inappropriate SAP than senior doctors. Education and confidence around appropriate antimicrobial prescribing is reported as variable between medical schools in a multicentre study in the United States [32]. This study would suggest that education to optimise antimicrobial use should not only incorporate appropriate prescribing choices, but the influences of hierarchy and social norms on prescribing decisions.

This data demonstrates clearly through the frequent description of unconventional SAP techniques (such as gentamicin washes), the significance and impact of improvisation in SAP decision-making. Improvisation is well recognised in healthcare (and in other industries) [33], occurring as a result of a desire to circumvent a perceived workflow block, and different professional streams may be more or less tolerant of improvisation by others within their stream. Doctors, for example, have been shown to be more tolerant of improvisation than nurses [25]. In daily clinical practice, doctors are required to judge when a patient fits within parameters that indicate a guideline-based standard of care, and to decide when the patient is significantly outside a guideline to require individualised or improvised care. It is clear from the case of SAP, and specifically detailed in the accounts in this study which document frequent occurrence of non-evidence-based strategies, that improvisations are utilised in situations where usual guideline-based practice would be expected to be safe and achieve good clinical outcomes. In addition, this data suggests that individual doctors’ decision-making and improvisation may result in care that is both non-compliant with evidence-based practice and has the potential to put the patient at risk (prolonged antibiotics, unconventional use of potentially toxic antibiotics such as gentamicin). It was evident in the accounts presented here that the improvisation behaviours described within SAP prescribing are driven by concern around adverse patient outcomes, a sense of benevolence towards the patient (held by the surgeon), and an internalised sense of what is perceived conventional practice for a particular operation. The issue raised from this data is the apparent limitation in doctors’ ability to identify when improvisation and individualisation of patient care is appropriate, and when it detracts from the quality of care they provide.

Conclusion

These results indicate the following challenges for AMS teams seeking to optimise SAP. Firstly, the importance of the SAP decision must be made more significant to critical team players in the operating theatre—that is, the surgical and anaesthetic teams. Second, identification of the existence of unwritten guidelines which significantly influence SAP choice and duration is required, and the design of interventions that address the discord between evidence-based practice and these (unwritten) norms. This will necessarily require collaborative, inter-specialty (infectious diseases, anaesthetics, surgery) based consensus building around SAP guidelines. Where group agreement (within a specialty) is present, there may be a shift in the perception of the unwritten rules within an institutional setting. In addition, consideration of the influence of professional autonomy as a barrier to guideline-based care is an important issue in streamlining practice. Consideration of these significant perceived barriers to guideline-based SAP would seem important in designing sustainable AMS interventions in this area.

This study has various limitations. First, the focus of the study was on participants’ attitudes and perspectives on SAP, thus the results cannot show actual behaviours around SAP practice. Second, although appropriate for a qualitative study, and albeit inclusive of a diverse range of surgical specialties, it only captures the experiences of participants from one hospital setting. Hence, these findings cannot be transferred to other experiences in other settings, despite providing important themes and theoretical insights likely to have resonance with many other settings. Third, participants were self-selected and thus the sample might only reflect particular views about SAP. Future research in different hospital settings and the perspectives of operating theatre nurses and antimicrobial stewardship team members would be valuable. Social influences on prescribing may be significantly different in different cultural environments and with differing resource levels. Qualitative research on surgical antibiotic prophylaxis prescribing influences in different countries would be extremely informative.