The Differences in Antibiotic Decision-making Between Acute Surgical and Acute Medical Teams: An Ethnographic Study of Culture and Team Dynamics

1(See the Editorial Commentary by Szymczak on pages 21–3.)

2Across hospitals, healthcare professionals from a range of specialties diagnose and treat community- and hospital-acquired infections. Antimicrobial stewardship programs (ASPs) are implemented in hospitals to optimize antibiotic use. These programs are not, however, contextually designed or implemented with end users of different specialties in mind. In healthcare, culture influences the shape and outcome of interventions and impacts on patient outcomes. Culture has been defined as “the shared knowledge people use to interpret, experience, and generate behavior” as members of a group. The key components of culture are norms, values, tools, and rituals that people of a specific group adopt, that identify them as belonging to that group. To optimize patient care, hospital-based specialties constantly diversify and evolve in response to scientific and technological breakthroughs and developing technical expertise. This diversification creates microcultures within the macro healthcare culture. The microcultures within specialties have the potential to shape intervention outcomes.

3ASP studies emphasize quantitative methodologies, to answer “what works,” and “by how much” with an emergent body of research investigating “why” interventions do or do not work. Studies have explored the contextual determinants of antibiotic prescribing decision-making, identifying the influence of hierarchies and etiquette, and the complexities of rising to expectations set in policies and guidelines. Such insights are critical for the development of targeted and sustainable ASP. Qualitative research methods provide in-depth understanding of the subject being studied leading to development of new theories. Ethnographic studies in hospitals provide a rich understanding of the contextual complexities that influence outcomes. Qualitative research methods have investigated education and training of doctors, and effective team communication on ward rounds (WRs). These studies report on the complexity of WRs. In surgery, studies have used WR simulation to test interventions including efforts to reduce disruptions, and checklists to improve patient outcomes. We conducted a mixed-methods study of antibiotic decision-making with a focus on WRs in medicine and surgery. The study included ethnographic research into how culture and team dynamics within specialties affect antibiotic decision-making. We drew on Spradley’s definition of culture to define and describe the cultural norms and values that help generate behaviors (Figure 1). In this article, we describe the qualitative findings, focusing on the influence of culture and team dynamics on antibiotic decision-making. These insights are important in the development of contextually fit ASPs.

4Ethical approval for this study was granted by the UK National Research Ethics Committee (application number 176038).

5The study was conducted in a central London teaching hospital with 1300 beds and an existing ASP. Wards were conveniently sampled to represent a high percentage of elective admissions, and nonelective admissions; wards were identified through consultation with medical directors at the hospital. Clinical leaders on the identified medical and surgical wards identified were emailed the study information sheet and invited to participate in the study. All healthcare professionals who attended the WRs or worked on the study wards were eligible to participate in the study. The lead consultant surgeons/physicians who agreed to participate in the study were asked to provide a list of the names of the staff who participated on the WRs. An information leaflet was sent to all participants; to reduce bias and Hawthorne effect, the leaflet stated that the study investigated clinical decision-making. Informed consent was obtained before the observations and interviews. The interviews took place following the observations to ensure that participants could be questioned about their antibiotic decision-making without affecting their behaviors during the observations.

6Ethnography is the study of people within the context in which they exist, live, and work. The ethnographic study design included nonparticipant direct observations, interviews, and documentary analysis (detailed notes on the type of data collected are provided as Supplementary Data). One researcher, E. C., conducted the observations on the wards and face-to-face interviews with key informants. Detailed, descriptive notes of observations were collected. Separate reflective notes were kept detailing the observer’s perceptions and interpretations of what was recorded. Handover sheets, multidisciplinary team meeting notes, and the policy and guidelines on antibiotic prescribing were collected to provide contextual knowledge of the processes. These different methods supported cross-validation and triangulation of the findings.

7Healthcare professionals participating in the observations were invited to participate in a follow up face-to-face interview. The interviews were semistructured with an interview guide, developed through review of literature and drawing upon previous work of the research team. The interviews were recorded using an audio recorder, transcribed verbatim, and anonymized.

8The data were analyzed using classic grounded theory approach, using mainly inductive methods of inquiry. Grounded theory relies on simultaneous data collection and analysis, in an iterative manner that enables theory construction and does not rely on existing frameworks for analysis. The analysis aimed to explore categories and relationships within the data collected. During focused coding, a constant comparative method was used for the analysis of the emerging categories and themes, aided by Nvivo 11 software. The data from observations, documentary analysis, and interviews were open coded to identify key categories, which were developed into themes. The analysis was conducted using an iterative and recursive process of moving between the coded data, data collection in the field, and the higher-level themes, until the themes and the relationship between the themes reached saturation (ie, no new themes or interrelationships between them were identified).

9The results presented below are derived from the analysis of field notes, documents, and the interview data. The field notes and quotes in this article are representative of the key emerging themes. Between August 2015 and May 2017, 30 surgical and 22 medical WRs (Figure 2) were observed. More than 100 hours of observation of routine healthcare worker practices were also collected from the wards. The acute surgical team was comprised of 6 surgeons, and the acute medical team of 14 consultants. Twenty-three healthcare professionals (14 in surgery, 9 in medicine) including surgeons, medical consultants, trainee doctors, nurses, and pharmacists were interviewed (Supplementary Data).

10In surgery, individualism is clear in the influence of consultant surgeons on the team. The surgical team is vertical in structure, with the surgeons leading decision-making and with less room or opportunity for team input, particularly during WRs (Figure 3). In medicine, a more collectivist culture prevails, whereby each team member is encouraged to participate. The medical WR included a dedicated pharmacist (Table 1, quote [Q] 1). The presence of the pharmacist reinforces the necessity to review patient medications; the pharmacist will often question whether the antibiotic therapy could be de-escalated from intravenous to oral therapy or switched to a narrower-spectrum agent according to the local policy (Table 1, Q1).

11How time is viewed and managed is an important attribute of each specialty and determines the collective behavior of the teams. The sense of time in the medical team is less structured or pressurized. In contrast, the surgical team focus primarily on the “now,” requiring the team to plan and schedule, and have a sense of the time taken up by the WR, and the need to end it and move on to other tasks (Table 1, Q2 and 3). The surgical team has 3 main commitments including the operating room, clinics, and the ward. The surgeons spend much of their time in the clinics or operating room (Table 1, Q2). This leads to teams being dispersed, and working in a disjointed manner, with little opportunity for forward planning (Table 1, Q4). As a result, surgical teams rely much more on their mobile telephones for team communication, including decisions about patient care that occur via messaging apps such as WhatsApp (Figure 3). The medical teams are more likely to have face-to-face meetings, and it is rare for mobile phones to be used on the WR. One of the contributing factors to the communication challenges in the surgical team is the lack of multidisciplinary input into patient care on the WR. As in the surgical observations, information is diffuse among the team. However, the medical team are more in tune with the need to have access to multidisciplinary staff during the WR and the review of patients (Table 1, Q5–7). The physical presence of the infectious disease consultant removes a communication barrier, with the team often requesting and receiving ad hoc advice for individual patients (Table 1, Q6).

12In surgery, the diagnosis of infection is reliant on infection markers, including C-reactive protein, white cell count, and temperature. The decision-making is focused more on prevention and prophylaxis than on treatment of infections. In medicine, though infection markers are an important part of the process, the team members try to rationalize their decisions, making efforts to align them with local policy and readily involve other healthcare professionals.

13The need and expectation to intervene in surgery means that antibiotics are readily initiated for surgical patients. This process is rationalized by the surgeons as being an extension of their roles as “interventionists”. In the absence of concrete evidence of infection, what drives decision-making is not the risk of failure, but the risk of blame (Table 1, Q8 and Q9). A focus on starting, but not on reviewing or stopping, treatment can lead to unnecessary and prolonged courses of antibiotics (Table 1, Q10). The awareness of the culture of blame and responsibility that was evident in the surgical team is reflected in the medical team’s perception of how antibiotic decision-making occurs in surgery. The impetus to “do something” is greater in surgery, as “patients are not allowed to die” (Table 1, Q11 and Q12).

14In medicine, junior members were concerned about managing risks in the context of making decisions on behalf of senior leaders. The junior doctors are more likely to start and escalate antibiotic therapy but are reliant on consultants deciding whether therapy needs to be de-escalated (Table 1, Q13 and Q14). The reluctance of the trainee doctors to de-escalate is tied to this acute awareness that decisions will be viewed as ultimately being made on behalf of the consultant who is responsible for the care of the patient, highlighting ambiguity about ownership for antibiotic decision-making (Table 1, Q15).

15In medicine, the critical decisions about whether a patient needs to be prescribed antibiotics occur in the first 24 hours of admission. There are multiple teams involved in the care of the patient, presenting challenges for assuming ownership of antibiotic decision-making. The participants identified problems arising from pressure to act and respond rapidly in the emergency department (ED) (Table 1, Q16 and Q17). Some of this urgency to initiate antibiotics revolves around the ED clinicians having a heightened awareness of the need to rapidly diagnose and treat sepsis (Table 1, Q18). The decision-making often takes place without samples taken for microbiological investigations (Table 1, Q19 and Q20). There is also a perception that often antibiotics are initiated upon admission to buy time for the clinicians and to act as an exercise in uncertainty avoidance (Table 1, Q21). This decision to initiate therapy where there is a suspicion of infection becomes difficult to evaluate and review on subsequent WRs, particularly if patient symptoms are improving. There is a reluctance to change or override the decisions made in ED (Table 1, Q22–Q24). This reluctance to assume responsibility for the antibiotic decision-making made in the ED is not replicated for other classes of medication. Anticoagulants, analgesics, and blood pressure medications are amended during the WR and adjusted based on the patient symptoms. Cardiology issues are routinely apportioned to the cardiologist intern based on the admissions ward (field notes). It is only for the antibiotic therapy that the teams report and exhibit a hesitation to alter the prescribing decisions of their colleagues (Table 1, Q17).

16The patient care pathway in hospitals includes a multitude of medical and surgical specialties, pharmacists, nurses, and allied healthcare professionals, who may prioritize different policies, agendas, and interim goals. Different specialties have their own language, behaviors, social norms, and values. The cultural differences between specialties and healthcare professionals: (1) shape the shared knowledge within and across specialities in patient pathways, and (2) result in variation in care, thus impacting patient outcomes. Antibiotic decision-making is driven by different determinants in medicine vs surgery. There are different drivers of overuse that arise from cultural and contextual differences across these specialties. An understanding of these key differences is essential to the development of contextually fit and sustainable interventions to optimize antibiotic use.

17The interplay of collectivist vs individualistic approaches, together with fear of negative outcomes in surgery, leads to less tolerance to uncertainty. In medicine, uncertainty arises from the transition of care between ED and inpatient teams. The medical teams take a more collectivist approach than surgical teams. Though subject to communication flaws, the medical team addresses them by including other healthcare professionals, such as the pharmacist and infectious diseases consultant, in routine, daily decision-making. The presence of the pharmacist on the WR and the infectious diseases consultant on the ward means that they can have direct discussions with the consultants and decisions can be made there and then. The cultural norm of collectivism and the interdisciplinary approach to decision-making contribute to efforts to optimize antibiotic use. The gaps in care around antibiotic decision-making in medicine appear to be in the 24 hours after admission, when the patient moves from ED to medical wards. This transition of care between different teams is a challenge to antibiotic optimization. The Surviving Sepsis Campaigns have driven the culture of being “infection aware” and initiating antibiotics for patients upon admission, but can have the unintended consequence of driving inappropriate antibiotic use in the ED. The medical team in this study reported frustrations at antibiotics being initiated in the ED that are then not reviewed or de-escalated. Though it can be accepted that patients may be started on empirical antibiotics with little evidence of infection in the first hours of admission, it is not acceptable for patients to remain on antibiotics unnecessarily once the medical teams have had time to clinically review and reassess patients. The pharmacists have in this study been identified as healthcare professionals who can oversee and promote rationalized antibiotic decision-making, if they are part of the WR. If supported by the wider team, they may be able to address the gaps in ownership in antibiotic prescribing in the transition between the ED and inpatient teams for medical patients. The transition from the ED to surgical inpatients was less problematic as the surgical team assumed responsibility for patients identified as needing surgical attention in the ED. There is no transition of patients between teams, and the surgical team remains in charge of the patient for the duration of the admission episode.

18The surgical team however, has its own unique challenges. Having the senior surgeons in theater and the junior doctors chasing medical jobs means that the skill mix of the team on the ward is sometimes imbalanced. Difficult decisions such as de-escalation of antibiotic therapy are often deferred. This may result in prolonged and unnecessary courses of antibiotics for surgical patients. The surgeons prioritize surgical and patient outcomes, and their fear of negative patient outcomes overrides the fear of the unintended consequences of inappropriate antibiotic use. Time is limited in the surgical team due to the demands from the clinics, operating room, and wards. The different professions do get involved in the surgical patient pathway, but they tend to work in silos, with few communication opportunities between them. Multidisciplinary teamwork is not easily practiced in such a context, and time constraints are part of the problem. The surgeons’ time on the ward is extremely limited, so they must prioritize their commitments, preferring to only communicate with their own team and expecting the junior members of the team to coordinate the multidisciplinary tasks and actions that are identified. Therefore, it may not be expedient to suggest replicating the acute medical model of including a pharmacist on surgical WRs. Equally it may be unrealistic to expect surgeons to include ASPs in their daily review of patients. We identified inconsistencies in the approach to antibiotic decision-making in surgery, often with the key identified variables not discussed as part of the WR. Some of this is due to the practicalities of the WRs in surgery and can be overcome through simple solutions (Table 2). The surgeons, though not fully engaged with the antibiotic decisions made for their patients, remain the leaders in their specialty and engaging with them in ASP will be tantamount to engaging with their entire team. They have the power to influence the behaviors of their entire team. When constructing ASPs targeting surgical teams, the language and metrics must reflect the priorities of surgical teams, who are focused on outcomes. Using length of hospital stay and surgical outcomes may be more convincing measures than antibiotic consumption rates. It is also important to recognize the different work patterns and shifts of surgeons, and to accommodate these differences in the effort to include them in ASPs. For example, the surgical handover is an opportunity for engaging with the surgical team. The first 30 minutes to 1 hour is dedicated to detailed discussion of patients, including those admitted via the ED. Pharmacist and/or infectious diseases specialists can participate in the handover to discuss antibiotic- and infection-related issues for the patients. The surgical team uses messaging groups and texting as a preferred means of communication. Developing a specialty-level ASP-specific group on a messaging platform may get the attention of the team more effectively than the pharmacist trying to page the junior team, who must ratify any decisions verbally or via messages with the senior team. Becoming more engaged and involved with the surgical teams in the management and prevention of infection will also provide opportunities for the ASP teams to understand the priorities and difficulties in the surgical pathway. This will help them develop more context-driven interventions that are more likely to be adopted by the surgeons. Furthermore, it may result in a more interdisciplinary ASP in surgery.

19This is a single-center study limited to acute medical and surgical teams in one hospital in England. While the findings may be broadly applicable to other settings, we are replicating the study in centers in India and South Africa investigating antibiotic decision-making across surgical pathways. The study methodology has also been replicated in a study in Australia.

20The observed variation in the social norms, values, and behaviors in medicine and surgery defines the approach to antibiotic decision-making. The medical team adopts a more policy-driven, interdisciplinary approach that includes pharmacist and infectious diseases input. The transition of care between the ED and inpatient care is a critical point for optimizing antibiotic prescription decisions in medicine. The surgical teams perceive themselves to be highly skilled interventionists, and see ASPs as having a low priority. Medical and surgical specialties represent different tribes, with different cultures and challenges. To optimize antibiotic use across these specialties, it is imperative that interventions are designed with these unique challenges in mind.

21Supplementary materials are available at Clinical Infectious Diseases online. Consisting of data provided by the authors to benefit the reader, the posted materials are not copyedited and are the sole responsibility of the authors, so questions or comments should be addressed to the corresponding author.