There is an urgent need to improve surgical antimicrobial stewardship (AMS), to enhance individual care and reduce population-level antimicrobial resistance, but it is a complex issue.

We aimed to conduct an ethnographic study asking what would work in practice to improve surgical antibiotic prescribing behaviour?

Adopting a socio-cultural-historical perspective, we undertook ethnographic observations of clinical practice (43.5 hours) and semistructured interviews (n=31) with surgical staff, AMS staff and patients at two English National Health Service hospitals. Interview transcripts and observational fieldnotes were analysed using the Framework Approach. Additionally, we integrated stakeholder engagement throughout to ensure the findings were meaningful.

Our analysis of all fieldnotes (based on 43.5 hours of observation) and interview transcripts (n=31 from interviews with 31 different participants) identified that, while surgical staff were aware of antimicrobial resistance, they seldom considered AMS urgent or important in the acute setting where lack of time and the desire to mitigate perceived risk often prevailed. Other surgical issues were perceived to dominate senior decision-makers’ focus, thus perpetuating the status quo. Furthermore, attention to AMS was not always prioritised at the organisational level or by resource-limited AMS teams. Consequently, there was an absence of relationships and tools that foreground AMS. Electronic prescribing systems frequently hindered antimicrobial review and exacerbated patterns of siloed inter-disciplinary working, and feedback on antimicrobial prescribing and patient outcomes was largely absent. To improve AMS, surgical teams wanted sustainable improvements which effectively account for the hierarchical relationships, division of labour, rapid workflow and high staff turnover. Infection experts should better integrate into surgical teams to build relationships and trust, and to proactively contribute to patient care.

We offer data-driven, theoretically informed strategies to support change. Contextually appropriate improvements that address the status and visibility of AMS in surgery will be key. Further research is needed to assess the impact and sustainability of the suggested approaches.

(1) Analyse in depth an exemplar safety-critical task required of newly qualified doctors (prescribing insulin) and (2) Provide transferable insights into how undergraduate education could better educate medical students to meet the demands of practice when they become postgraduate trainees.

Document analysis of doctors’ reported experiences of insulin prescribing, an everyday task that has an emergent logic of practice and harms not just patients but (psychologically) new doctors. Application of third-generation (social emergence) complexity theory to explore why practice can be ‘mutually unsafe’.

A system of care comprising all five Northern Irish (UK) Health and Social Care Trusts, which together provide healthcare to a population of nearly two million people.

68 postgraduate year 1 and year 2 trainees (PGY1/2s), mainly PGY1s.

Thick description of new doctors’ contexts of action, reasons for acting and specific actions. We present this as a narrative compiling all 68 stories, 13 detailed exemplar stories and a diagram summarising how multiple factors interacted to make practice complex.

Situations that required PGY1/2s to act had interacting layers of complexity: (1) disease trajectories; (2) social dynamics between stakeholders and (3) contextual influences on stakeholders’ interactions. Out-of-hours working and unsuitable wards intensified troublesome contextual influences. All three individually complex layers ‘crystallised’ briefly to create ‘moments of action’. At best, PGY1/2s responded proactively, ‘stretched time’ and checked the results of their actions. At worst, PGY1/2s ‘played safe’ in unsafe ways (eg, took no action), acted on unsafe advice or defaulted to actions protecting them from criticism. Informal, pervasive rules emerged from, and perpetuated, unsafe practice.

New doctors’ work includes acting on indeterminate, emergent situations whose complexity defies rules that are determinate enough to be taught off the job. If new doctors are to perform capably in moments of action, medical students need ample, supervised, situated experience of what it is like to take responsibility in such moments.

The UK’s medical workforce is under increasing strain, and this is compounded by increasing numbers of resident doctors diverging from specialist training pathways, instead entering non-training roles, reducing clinical hours or leaving the profession or UK workforce entirely. These decisions are shaped by both individual motivations and wider structural conditions, including unsatisfactory working conditions, limited flexibility and a perceived lack of support or autonomy. While pursuing alternative career routes offers personal and professional benefits, they can also delay progression to senior clinical roles, contributing to workforce instability. There remains limited understanding of how best to support retention, particularly given the varied contexts, settings and career trajectories of resident doctors. This realist synthesis will examine how, why and in what contexts resident doctors leave the National Health Service, and what interventions might support their retention.

This realist synthesis will follow Realist And Meta-narrative Evidence Synthesis: Evolving Standards guidance and will be conducted in five iterative steps: (1) identifying existing theories to develop an initial programme theory; (2) undertaking formal and purposive searches to identify relevant UK-based literature; (3) selecting documents based on relevance and rigour; (4) extracting and coding data to support the development of explanatory insights; and (5) synthesising findings using a realist logic of analysis to develop and refine context-mechanism-outcome configurations. An advisory group will guide the review throughout. The final programme theory will inform the development of evidence-based recommendations and design principles to support resident doctor retention.

Ethical approval is not required for this synthesis of existing literature. Findings will be disseminated through academic publications, conference presentations and accessible formats, including infographics, plain English summaries and blog posts. Target audiences include resident doctors, medical educators, workforce planners and policymakers.

PROSPERO, CRD420251004453.