The decision-making process in general practice of when to use antibiotics to treat acute rhinosinusitis

Abstract

Background

The Danish primary care sector is responsible for the majority of antibiotic prescriptions, with upper respiratory tract infections, particularly acute rhinosinusitis (ARS), being a key contributor. This study explored the decision-making process regarding antibiotic use for ARS among general practitioners (GPs) and doctors in training working in general practice.

Methods

Nine semi-structured interviews were conducted with 10 doctors. Interviews were audio-recorded, transcribed verbatim and analyzed using systematic text condensation.

Results

Four key themes were developed: (1) ‘It all starts before the consultation’: staff-led triage shaping ARS care, (2) two diagnostic worlds: experienced intuition vs. rule-bound caution, (3) C-reactive protein (CRP) in the crossfire: guide, reassurance or irrelevant in antibiotic decision-making and (4) ‘Sometimes it’s a negotiation’: managing patient expectations without losing stewardship. Although diagnosing ARS was often straightforward, distinguishing bacterial from viral infections proved difficult. C-reactive protein testing was common, but its reliability was questioned. Antibiotic prescribing decisions were influenced by patient history, comorbidities and risk factors. While antibiotics were rarely considered necessary initially, patient expectations and doctor–patient relationships sometimes influenced prescribing.

Conclusions

This study illustrates the complexity of diagnosing and managing ARS, shaped by clinical uncertainty, competing priorities and non-clinical pressures. Variation in doctor involvement, diagnostic practices and CRP thresholds reveals the challenges of standardization. Training, guidelines and patient education are valued but may not ensure optimal antibiotic use. Diagnostic decisions are relational, negotiated and context-specific, influenced by public health concerns, patient needs and workflow demands. Without acknowledging this complexity, antibiotic stewardship efforts could potentially fall short.

Introduction

Antimicrobial resistance (AMR) is an escalating global health crisis, posing a critical threat to public health worldwide [1]. The overuse of antibiotics remains the main driver of AMR, accelerating the emergence of resistant pathogens. This not only complicates infection management but also increases hospital stays, healthcare costs and mortality rates [2]. In Europe alone, multidrug-resistant infections are responsible for an estimated 35,000 deaths annually [3]. Reducing unnecessary antibiotic consumption is therefore imperative to mitigate the progression of AMR. In Denmark, general practice is responsible for about 75% of all antibiotic prescriptions, with acute respiratory tract infections accounting for a large proportion of the scripts [4,5]. Importantly, studies indicate that antibiotics are often used inappropriately for treatment of acute respiratory tract infections [6–8], and it has been suggested that the management of patients with symptoms of acute rhinosinusitis (ARS) could be optimized, both in relation to the diagnostic process and antibiotic treatment [9].

ARS is a common reason for encounters in general practice and is characterized by inflammation of the nasal and paranasal sinus mucosa. It is typically classified into acute viral rhinosinusitis, acute post-viral rhinosinusitis and acute bacterial rhinosinusitis [10,11]. The diagnosis is primarily clinical, based on symptoms such as nasal congestion, purulent nasal discharge, pain in the face and teeth, headache, reduced sense of smell, possibly fever and general feeling of illness [10]. Only a small subset of patients (0.5–2.0%) develops a bacterial infection [10,11]. However, no single symptom can reliably confirm or exclude the diagnosis of ARS, which complicates the diagnostic process [10]. Despite ARS being generally self-limiting, distinguishing between a viral and a bacterial infection poses a persistent clinical challenge, often resulting in antibiotic overuse [12,13]. In Denmark, national guidelines emphasize that diagnosis should be based on clinical evaluation, with bacterial ARS suspected when at least three of the following are present: discolored nasal discharge (often unilateral), severe facial pain (often unilateral), fever (≥38.0 °C), elevated C-reactive protein (CRP) (≥50 mg/L) or a so-called ‘double sickening’ pattern, in which symptoms worsen after initial improvement [10]. Antibiotics are recommended only for such cases, with phenoxymethylpenicillin as the initial choice of treatment [10]. CRP testing is not mandatory but may be used in cases of diagnostic uncertainty and results should always be interpreted alongside clinical findings [10]. Even so, studies indicate that over 60% of Danish patients with ARS receive antibiotics [9,14], reflecting a pattern of overtreatment similar to that observed in the UK and US, where prescribing rates range from 86% to 91% [15–17]. Although antibiotics can provide modest early symptom relief in ARS, solid research shows that most cases resolve spontaneously [18]. Given the risks of adverse effects and the growing threat of AMR, prescribing should be restricted to well-justified, clinically selected cases [18]. A 2018 Cochrane systematic review further demonstrated that approximately 85% of patients with suspected ARS recover within 7–15 days without antibiotics [19].

The persistent overuse of antibiotics for treatment of patients with ARS raises concerns about the factors driving clinical decision-making. Several studies suggest that entrenched prescribing habits and personal clinical experience often override adherence to guidelines [20,21]. One study demonstrated that individual general practitioners (GPs) exhibit remarkably consistent antibiotic prescribing patterns [22], while a recent Norwegian study revealed that many GPs lack confidence in national guidelines, leading them to rely on their judgment rather than evidence-based recommendations [23]. Although antibiotic use for respiratory tract infections has been widely studied, most research has so far paid limited attention to ARS as a distinct condition. ARS presents particular challenges due to diagnostic uncertainty, symptom overlap with non-viral conditions, and large variations in guideline adherence. Despite updated national guidelines and ongoing efforts to reduce antibiotic prescribing, the high rates of antibiotic use for ARS persist in Danish general practice, suggesting that further insight into the underlying decision-making process remains timely. To our knowledge, there is only one previous Scandinavian qualitative study [23], and none from Denmark, that explores in depth how GPs approach prescribing decisions specifically for ARS. Qualitative studies using individual interviews to examine this issue are also limited, despite their potential to provide nuanced accounts of clinical reasoning. This present study contributes to addressing this gap by examining how experienced GPs, as well as doctors in training, approach clinical work with patients who present with symptoms of ARS. Specifically, we investigate how doctors manage patients presenting with ARS symptoms and explore their considerations, attitudes and practices that shape their decisions about antibiotic prescribing. By offering condition-specific insights from a Danish general practice context, the study could potentially add to the understanding of national prescribing patterns and may help inform future efforts to promote more judicious antibiotic use and contribute to broader strategies to reduce inappropriate prescribing in general practice.

Materials and methods

The study employed a qualitative interview design to explore the experiences and perspectives of Danish doctors working in general practices. It was conducted in the setting of primary care where both GPs and doctors in training are responsible for prescribing antibiotics to patients. The use of interviews was chosen to capture nuanced insights into clinical practices and decision-making processes. The study was conducted within an interpretivist paradigm, aiming to explore and understand participants’ perspectives and reasoning in their clinical context. Hence, the approach was inductive, allowing patterns and themes to be developed from the data, while drawing on relevant concepts such as diagnostic uncertainty and antibiotic stewardship to inform interpretation. Methodological rigor was ensured by adhering to the Standards for Reporting Qualitative Research (SRQR), which offer comprehensive guidelines for the design, conduct and transparent reporting of qualitative studies [24].

Recruitment

The recruitment process was designed to ensure participant diversity in clinical experience, age, gender, education level and practice type (solo or partnership). Recruitment focused on densely populated areas of Copenhagen to maximize the participant pool. A convenience sampling approach was employed for its practicality and ability to include a broad range of doctors within the study timeframe [25]. General practices were identified using Sundhed.dk, the official Danish health portal, and practice websites were assessed to confirm the presence of both GPs and doctors in training. After identifying relevant practices (n = 73), all were visited in person. Secretaries were briefed on the study’s purpose and procedures, and if the practice agreed, an informational invitation was left in a visible area for doctors. Interested doctors could sign up by providing their name and email. One week later, practices were revisited. Doctors who had signed up were contacted the same day via email to arrange interviews. If no interest had been expressed, secretaries were asked to inform the first author if any doctor later wished to participate. This two-step approach offered flexibility and minimized pressure on participants, respecting their workload. The recruitment process is outlined in Figure 1. General practices that declined the invitation cited excessive workload or staff shortages as the primary reasons. Practices that were either closed or lacked an affiliated secretary to handle correspondence never received an invitation to participate.

Figure 1.

Illustration of the recruitment process.

Participants

Ten doctors were recruited for the study, four men and six women. Their ages ranged from 28 to 75 years, with professional experience varying from 11 months to 40 years. Educational backgrounds included four doctors with doctors in training qualifications and six with GP credentials. Eight participants were affiliated with partnership practices, while two operated independent solo practices. Details are shown in Table 1.

Table 1.

Characteristics of doctors and general practices.

Doctor	Gender	Age	Education level	Experience	Type	Nurse
1	M	28	DIT	11 M	Partnership practice	$-$
2a	F	43	GP	16 Y	Solo practice	$+$
3a	F	34	DIT	4 Y	Solo practice	$+$
4	M	75	GP	37 Y	Partnership practice	$-$
5	F	29	DIT	2 Y	Partnership practice	$+$
6	F	61	GP	30 Y	Partnership practice	$+$
7	M	69	GP	40 Y	Partnership practice	$-$
8	M	62	GP	35 Y	Partnership practice	$+$
9	F	38	GP	7 Y	Partnership practice	$+$
10	F	34	DIT	5 Y	Partnership practice	$-$

M: month; Y: years; (+) nurse may see the patients with ARS; (−) only doctors see the patients.

^a Part of the group interview.

Interviews

Between September and October 2023, nine semi-structured interviews were conducted in the Capital Region of Denmark. Eight of the interviews were conducted as individual interviews and one was conducted as a group session with two doctors to promote dynamic discussion, allowing participants to exchange insights, reflect on shared practices, and explore collective decision-making processes. Semi-structured interviews were chosen as the data collection method due to their flexibility and ability to capture in-depth insights. Malterud et al.’s concept of information power guided the assessment of sample adequacy, focusing on the richness and relevance of the data rather than a fixed number of interviews and ensuring internal validity [26]. As data collection progressed, factors such as the narrow study aim, the relatively specific group of GPs and doctors in training from Danish primary care, and the emergence of the concepts diagnostic uncertainty and antibiotic stewardship to guide the study and inform the analysis, contributed to building information power. The interviews provided detailed accounts for the research question, and the analytical approach identified patterns and variations in the data. After nine interviews our data collection concluded, as no new codes, themes, or topics were identified or developed, and the ongoing analysis indicated that additional participants were unlikely to contribute substantially new insights.

The interview guide was developed by the first author in collaboration with three experienced researchers with different professional backgrounds (see details in the researcher characteristics and reflexivity section below), integrating theoretical and empirical perspectives. Before each interview, participants were briefed on the study’s aims, methods and ethical considerations. All interviews were conducted at the participants’ respective general practices, ensuring a familiar and comfortable environment that fostered open, reflective dialogue. The first author conducted all interviews, using open-ended questions to elicit reflexive and descriptive responses. The interviews explored three key themes: (1) management of patients with ARS symptoms, (2) considerations and decision-making around antibiotic use for ARS and (3) general attitudes toward antibiotic prescribing and AMR. Interviews ranged from 15 to 55 min, with an average duration of 30 min, and were audio-recorded and subsequently transcribed verbatim.

Data analysis

The transcribed data were analyzed using systematic text condensation (STC), a method that facilitates the identification of patterns and themes across cases while preserving the complexity and richness of individual experiences [27]. The analysis was conducted in four structured steps.

The first author read all nine transcriptions to develop a coherent understanding of the material. Initially, four preliminary themes were identified, after which a systematic line-by-line review identified meaning units, which are text fragments that contain information relevant to the research question. The meaning units, ranging from initial patient management to education and stewardship, were then grouped into 12 subgroups capturing specific aspects of the broader themes. Content was condensed to synthesize four core meanings, which comprised our four key themes. Condensates were created by integrating representative meaning units, while unmatched units were reassigned or excluded if irrelevant. In the final stage, the condensates were synthesized into coherent narratives and conceptual frameworks supported by illustrative quotations. Transcripts were revisited to validate the findings, ensuring accuracy and consistency. Contradictions and outliers were incorporated to capture data complexity. Table 2 provides an overview of the analytical process. Below we present the four key themes, ‘It all starts before the consultation’: staff-led triage shaping ARS care, two diagnostic worlds: experienced intuition vs. rule-bound caution, CRP in the crossfire: guide, reassurance or irrelevant in antibiotic decision-making and ‘Sometimes it’s a negotiation’: managing patient expectations without losing stewardship, that influenced the decision-making of GPs and doctors in training regarding antibiotic prescribing for ARS.

Table 2.

Examples of the analytical steps.

Meaning unit	Condensation	Code	Theme
‘When patients report being ill for several days… the staff usually advise over-the-counter medication and rest.’	Frontline staff manage initial contact and triage patients before doctor involvement	Importance of gatekeeping in patient flow	‘It all starts before the consultation’: staff-led triage shaping ARS care
‘A diagnosis cannot rely on a single finding. It requires a careful synthesis of all clinical information.’	Doctors synthesize multiple symptoms due to variability	Complexity and uncertainty in symptom-based diagnosis	Two diagnostic worlds: experienced intuition vs. rule-bound caution
‘There is no simple test… to provide a definitive answer.’	Lack of diagnostic tools limits certainty	Challenges of treating ARS without clear diagnostic support	CRP in the crossfire: guide, reassurance or irrelevant in antibiotic decision-making
‘It feels like a negotiation… without compromising clinical judgment.’	Balancing patient expectations with responsible prescribing	Shared decision-making to manage expectations	‘Sometimes it’s a negotiation’: managing patient expectations without losing stewardship

All steps of the analysis were carried out by the first author, with regular dialogue and discussion with the rest of the research group to ensure rigor and consistency. The interviews, transcriptions and analyses were conducted in Danish. Only the selected illustrative quotations used in the results section were subsequently translated into English, maintaining the authenticity of participants’ statements while ensuring accessibility for an international audience.

Researcher characteristics and reflexivity

The research team brought together complementary clinical and qualitative research backgrounds. The first author is a medical doctor with recent experience in general practice. The second author is a medical doctor, PhD, with experience in general practice and clinical microbiology, and now works as a full-time researcher with extensive knowledge of antibiotic use in primary care. The third author is a practicing GP, PhD, combining clinical work with academic research. The fourth author is an anthropologist, PhD and full-time researcher whose primary research expertise is in qualitative methods. These varied perspectives supported a broad analytical approach and informed critical reflection on emerging findings. We recognize that our prior experiences with antibiotic prescribing, general practice and antimicrobial stewardship could have influenced the framing of questions and interpretation of data. To reduce potential bias, reflexive discussions were held within the research team throughout data collection and analysis to challenge assumptions and reach consensus on interpretation.

Ethics

As the study did not involve sensitive patient data or medical interventions, it did not require approval from the Regional Committee for Medical and Health Research Ethics. Nevertheless, it was conducted in accordance with recognized ethical principles for research involving human participants, consistent with good practice in national and international contexts. Interviewing healthcare professionals requires safeguarding confidentiality, minimizing risks to professional reputation, and ensuring fully informed, voluntary participation. Ethical considerations were considered during the planning, conduct, analysis and reporting of the study. Participants received written information, provided informed consent, and data were pseudonymized in accordance with GDPR regulations. Findings are presented in a way that prevents attribution of specific views to identifiable individuals.

Results

‘It all starts before the consultation’: staff-led triage shaping ARS care

Doctors described patient management as beginning well before the consultation, with initial interactions typically handled by professional healthcare staff, such as secretaries, nurses or medical students, either over the phone or at the clinic reception. These frontline personnel played a critical role in the triage process, assessing patient symptoms and determining whether an appointment with a doctor was necessary. Not every patient had a consultation offered, and many patients received initial advice and guidance from the practice staff. Doctors expressed strong confidence in their teams’ triaging abilities, emphasizing that patients were effectively assessed and directed without compromising care. This system of initial assessment was consistent across all practices.

When patients report being ill for several days without improvement, an appointment is scheduled within a few days. If the symptoms have only recently appeared, the staff usually advise over-the-counter medication and rest. (doctor 8)

However, the structure of patient consultations varied depending on practice-specific workflows, often shaped by staffing levels and time constraints. In practices facing doctor shortages or high patient volumes, more detailed protocols were developed to guide practice staff. Nurses in these settings frequently had independent consultations, with doctors stepping in only when the nurse encountered uncertainty. In situations where an antibiotic prescription was warranted, doctors often relied on the nurse’s assessment, issuing prescriptions without personally examining the patient.

The guidelines we use are created by our GPs and are designed to ensure the highest quality patient assessment. They outline the key questions to ask and proper examination techniques to use, essentially mirroring the approach we use ourselves. With these instructions, we believe the nurse can assess the patient just as effectively as a doctor in training. (doctor 6)

In contrast, other practices mandated that doctors handle all patient consultations, driven by concerns over the diagnostic limitations of practice staff. These doctors expressed reservations about the adequacy of nurses’ medical training and the lack of diagnostic tools available to them, which they felt increased the risk of misdiagnosis. For these doctors, personally conducting the consultation was seen as essential to ensuring diagnostic accuracy and maintaining a high standard of care.

While ARS can be relatively straightforward to diagnose, it can also be quite complex. Patients often present with vague or overlapping symptoms, and it’s essential to consider multiple differential diagnoses. (doctor 7)

Two diagnostic worlds: experienced intuition vs. rule-bound caution

Despite working within the same specialty, doctors approached diagnosis in strikingly different ways. For some, years of experience had shaped an almost instinctive process, where structured routines ensured efficiency and consistency.

Diagnosing doesn’t take long. I’ve seen these cases so many times that I instinctively know what to do. I feel confident with my own routines. Strict adherence to guidelines isn’t necessary. (doctor 7)

Experienced GPs often relied on pattern recognition, quickly identifying familiar presentations and making confident decisions. In contrast, doctors in training approached each case with meticulous attention to formal guidelines, frequently reviewing protocols before consultations. For them, general practice was a field of hidden complexities, where even seemingly mild symptoms could mask serious conditions.

In general practice, we encounter patients with a wide range of symptoms. Those who seem straightforward can sometimes turn out to be the most complex. (doctor 1)

Beyond clinical expertise, experienced GPs highlighted the value of the doctor–patient relationship in diagnosis. Long-term familiarity with patients allowed for more intuitive and efficient assessments, while new patients required a more cautious, thorough approach.

When you see the same patients regularly, the job becomes easier. Each consultation is a continuation of the previous one. You know their medical history and you know their expectations. They trust your decisions. (doctor 4)

Yet, regardless of experience level, most doctors shared a symptom-based approach as common diagnostic foundation. Key symptoms, such as headache, purulent nasal discharge, postnasal drip, cough, nasal congestion and fever, formed the core of their assessments. Sinus pain was often evaluated through palpation and positional tests, with doctors instructing patients to bend forward to gauge discomfort.

Patients respond differently to being sick and the symptoms they present with may vary. A diagnosis cannot rely on a single finding. It requires a careful synthesis of all clinical information. (doctor 10)

However, uncertainty was unavoidable, particularly when patients presented early in the disease course with mild or nonspecific symptoms. In such cases, symptomatic treatment was the preferred course, with follow-up recommended if symptoms worsened. If no improvement was observed, referral to an otorhinolaryngologist was often the next step.

As doctors, we should never be completely certain, but it’s reasonable to commit to a diagnosis until proven otherwise. Uncertainty is an inherent and daily aspect of medical practice. (doctor 2)

CRP in the crossfire: guide, reassurance or irrelevant in antibiotic decision-making

When asked about their approach to treatment, doctors highlighted the persistent challenge of managing ARS in the absence of a definitive diagnostic test. Without a reliable method to differentiate viral from bacterial infection, the decision to prescribe antibiotics often relied on clinical judgment.

When patients present with a sore throat, we can rely on objective tools like the Strep-A test to confirm a diagnosis. But with ARS, it’s more complex. There is no simple test, no quick swab or immediate imaging like a CT scan to provide definitive answer. (doctor 7)

In an attempt to enhance diagnostic accuracy, some doctors used CRP testing. However, the use of CRP was contentious. While some practitioners found it useful in guiding treatment decisions, others questioned its reliability and regarded it more as a tool for patient reassurance than a definitive diagnostic marker. The appropriate cut-off range was another subject of considerable debate among doctors, with differing opinions on its diagnostic accuracy and clinical relevance.

I’ve seen patients with CRP levels below 5 who still had bacterial sinusitis. If the symptoms and clinical findings align with bacterial infection, CRP results become irrelevant. (doctor 2)

Despite diagnostic uncertainties, most doctors favored a conservative approach, reserving antibiotics for select cases. Instead, they relied on symptomatic management, including nasal decongestants, analgesics and saline irrigation, as the preferred strategy. Antibiotics were considered only when symptoms persisted or worsened, with doctors carefully weighing individual risk factors such as immunosuppression, comorbidities, medical history and age.

It’s important to remember that in most cases, ARS is not a life-threatening condition. You have to weigh the risks and benefits carefully. What does the patient gain from this medication? (doctor 3)

While the majority of doctors adhered to evidence-based prescribing practices, several acknowledged that external pressures such as patient insistence, time constraints and the desire to avoid conflict, occasionally influenced their clinical decisions.

It can be frustrating when patients argue. I won’t deny that I’ve prescribed antibiotics at times just to end a dispute. (doctor 6)

When asked about prescribing habits, a generational divide in prescribing attitudes was evident. Senior GPs reflected on how prescribing antibiotics for ARS used to be routine and rarely questioned. In contrast, doctors in training approached antibiotic use with greater caution, guided by current medical education that emphasizes antibiotic stewardship and the risks associated with AMR.

In the early days, AMR wasn’t something we really discussed, and I’ve probably been more liberal with antibiotic prescriptions compared to younger doctors today. They’re trained to prescribe only when necessary and tend to be more deliberate in their decisions. (doctor 4)

‘Sometimes it’s a negotiation’: managing patient expectations without losing stewardship

Beyond the clinical challenge of distinguishing viral from bacterial infection, doctors encountered patient expectations as an equally complex obstacle. Cultural beliefs, past experiences with antibiotics, and the pursuit of immediate symptom relief frequently influenced patients’ demands for antibiotic treatment. Many doctors recognized that effective communication was essential in mitigating unnecessary prescriptions, highlighting the critical role of shared decision-making in antibiotic stewardship

Sometimes, it feels like a negotiation. You need to ensure that patients feel they are receiving appropriate care without compromising clinical judgment. (doctor 3)

Yet, some doctors acknowledged the influence of patient expectations on their prescribing decisions. For these doctors, maintaining the doctor–patient relationship and meeting perceived patient needs occasionally took precedence over strict clinical protocols. In such cases, antibiotics were prescribed not solely based on medical necessity but to align with patient demands and ensure satisfaction.

The first thing I do when patients arrive is gauge their expectations, whether they’re hoping for a penicillin prescription or simply seeking reassurance and an explanation. Regardless of what they want, they’re looking to me for help and guidance, and it’s my responsibility to meet that need. (doctor 9)

When reflecting on strategies to reduce unnecessary antibiotic prescriptions, doctors consistently highlighted patient education as a central tool in promoting responsible antibiotic use. They emphasized that engaging patients in meaningful discussions about AMR, the typical self-limiting course of ARS, and the critical distinction between viral and bacterial infections was essential in shifting patient expectations. Such educational efforts not only decreased the demand for antibiotics but also reinforced the rationale for symptomatic management, aligning patient understanding with evidence-based care. Moreover, doctors felt that these interactions facilitated shared decision-making, empowering patients to take an active role in their treatment while fostering trust and improving adherence to non-antibiotic management strategies.

Taking a few extra minutes to educate patients will benefit you, the patient and society. Once they understand, they rarely forget, and sometimes that is all it takes. (doctor 5)

Although most doctors leaned toward a conservative prescribing approach, their views on AMR varied. Some saw it as an urgent issue demanding stricter antibiotic stewardship, while others prioritized immediate patient care, sometimes placing public health considerations in the background. Still, despite these differing perspectives, all doctors agreed on the need for responsible prescribing, limiting antibiotics to necessary cases, performing bacterial cultures when feasible, and favoring narrow-spectrum agents over broad-spectrum ones.

Every treatment decision must be justified. While AMR is a serious concern, our primary responsibility is to treat the patient, not just the broader public health issue. (doctor 6)

Discussion

Principle findings

The flow of patient visits across general practices followed a standardized structure, but beneath this uniformity lay diverse approaches to patient assessment. In many clinics, nurses conducted initial evaluations independently, guided by protocols created by the GPs. In contrast, some practices relied on doctors to assess every patient due to concerns over diagnostic accuracy. In our study, the experienced GPs often drew on intuition and routine, while doctors in training adhered more strictly to formal guidelines for consistency. Most doctors agreed that diagnosing ARS was rarely straightforward as the distinction between viral and bacterial infections was often blurred, challenging even experienced GPs. Many turned to CRP testing to assist diagnosis, though its role remained debated. While viewed as a helpful tool, it was not considered as definitive and required nuanced interpretation. Clinical judgment typically guided antibiotic decisions, considering patient age, comorbidities, immune status and medical history. Although the doctors generally agreed that antibiotics were rarely necessary in early stages of ARS, prescribing habits varied. Patient expectations and the doctor–patient relationship influenced some GPs, while others claimed minimal impact. AMR was a shared concern, but most prioritized the immediate needs of individual patients. Patient education was widely endorsed as a means to reduce unnecessary antibiotic use, as informed patients were considered more receptive to symptomatic treatment.

Discussion of findings

Our study highlights important insights into the complexities of diagnosing and managing ARS, particularly in relation to antibiotic prescribing behavior and clinical judgment. One key finding was the variability in doctor involvement in nurse-led consultations. While a 2017 Danish study reported doctors were involved in 40% of nurse-led respiratory tract infection cases [28], our study showed that GPs often prescribed antibiotics based on nurse assessments alone. Though many viewed nurses as aligned with doctors in training, others cited differences in education and clinical background as a reason for direct doctor involvement. As our study focused solely on doctors’ experiences, the findings reflect a doctor-centered perspective on ARS management. Given that nurses also often manage patients with acute respiratory tract infections, including their perspectives could have revealed interesting additional dimensions of collaborative practices and strategies for managing diagnostic uncertainty. In addition, the extent to which nurses possess sufficient diagnostic knowledge remains unclear, with some international studies suggesting moderate understanding of diagnosis and AMR, alongside evidence from specific regions emphasizing similar challenges [29,30]. This highlights potential risks tied to antibiotic stewardship, suggesting the need for further research into the impact of reduced doctor oversight on prescribing practices.

We also revealed significant variability in diagnostic approaches among doctors in our study, despite standardized patient flow protocols. Experienced GPs often favored individualized routines, while doctors in training adhered more closely to guidelines. In an editorial paper on diagnostic practices in general practice, Malterud et al. argue that uncertainty is an inherent component of general practice, managed in diverse ways by doctors [31]. They posit that uncertainty should not be perceived as a weakness or failure, but rather as an essential facet of clinical practice, and embracing and acknowledging this uncertainty, might ultimately enhance diagnostic accuracy and patient care. In this context, uncertainty is not a problem to be eliminated, but a reality that demands thoughtful engagement [31]. In our study, GPs managed uncertainty differently, some relied on pattern recognition and clinical experience, while others mitigated ambiguity through strict adherence to protocols. This variability should not be seen as indicative of suboptimal practice, but rather as a reflection of the flexible, context-sensitive nature of primary care decision-making. GPs adjusted their diagnostic strategies based on the clinical context, their expertise and their relationships with patients, supporting the idea that uncertainty, when engaged thoughtfully, can enhance diagnostic accuracy and care quality.

Our findings confirm that doctors rely on key diagnostic criteria such as sinus pain, headache, purulent nasal discharge and fever, which align with established guidelines [14,32,33]. However, distinguishing viral from bacterial infection remained difficult. Many used CRP testing as a support tool, though interpretation varied. This is in accordance with findings from Norway emphasizing the CRP’s nuanced role [23]. In our study, we observed a wide range of CRP thresholds for antibiotic initiation, ranging from 11 to 60 mg/L, with an average of 50 mg/L, highlighting inconsistency. While another Danish study found that most patients presenting with respiratory tract infections in general practice were prescribed antibiotics when CRP levels exceeded 40 mg/L [34], other studies suggest that CRP levels just above 10 mg/L are used as an indicator for a bacterial infection [32,35,36]. Notably, current Danish guidelines recommend considering antibiotics for ARS only when CRP levels are above 50 mg/L [10].

Although some doctors used CRP as a guide, many prioritized clinical judgments, emphasizing experience and relational knowledge over numerical thresholds. This tension reflects the broader challenge of defining standardized diagnostic parameters in a field shaped by uncertainty. The popularity of CRP testing likely reflects a desire to manage diagnostic ambiguity. Though, it has been suggested that no test can completely resolve the interpretive challenge of diagnosis in general practice, where clinical judgment, rooted in experience, intuition and relational knowledge, remains central to navigating these diagnostic gray zones [31]. Danish data show that more than half of the patients diagnosed with a respiratory tract infection have a CRP test performed [14,34]. However, our findings suggest that its role in ARS management is not clearly defined. In our study, some doctors described CRP as useful for guiding treatment decisions, while others questioned its reliability and viewed it primarily as a means of reassuring patients. This raises the possibility that CRP testing might also have potential as a communication aid during consultations. Presenting results to patients could help make the diagnostic reasoning more transparent, provide a concrete basis for discussing why antibiotics may or may not be indicated, and encourage shared decision-making. Previous studies suggest that combining CRP testing with effective communication strategies might lower antibiotic prescribing without affecting patient satisfaction [37,38]. However, its specific relevance and impact in ARS remain uncertain, and there is a risk that CRP results could be overemphasized or misunderstood if not carefully contextualized within the broader clinical picture. The limited size of our sample could also help explain why the communicative dimension of CRP use did not feature more prominently in our findings, underlining the need for further research on how CRP testing can shape communication and prescribing practices in ARS.

Our findings also suggest that in some cases, antibiotic prescribing may be prioritized over diagnostic accuracy. While this may be driven by attempts to meet patient expectations or manage time pressures, such a shift away from a diagnosis-first approach could present significant risks to public health. Given that 46% of ARS cases resolve within a week without antibiotics and 64% within two weeks [19], premature prescribing could potentially risk both overtreatment and missed alternative diagnoses. The shift from a symptom–diagnosis–treatment model to a symptom–treatment paradigm has been a concern since the 1970s [39], and underscores the risks of bypassing uncertainty rather than managing it. As Malterud et al. note, simplifying diagnostic complexity may compromise both patient care and stewardship efforts [31].

Prescribing behavior was also shaped by non-clinical factors. Numerous studies have indicated that doctor experience, habitual practices and the doctor–patient relationship influence prescribing [21,22,40]. Our findings support this. Experienced GPs acknowledged that patient satisfaction, habits and relationship dynamics affected their decisions. This highlights a potential tension between clinical evidence and interpersonal pressures. In our study, doctors in training were generally more protocol-driven and cautious about unnecessary prescriptions. While addressing patient expectations is understandable, it raises questions about how best to balance individual patient satisfaction with the broader need for antibiotic stewardship. A 2017 Danish study found that 32% of patients expected antibiotics, making them eight times more likely to receive them [41]. Similar patterns have been observed in Norway and Australia, where patient insistence, time constraints and doctor stress contributed to higher prescribing, especially among parents, migrants and individuals from diverse cultural backgrounds [23,42]. Other studies suggest such factors can increase prescribing by up to 10-fold [43,44].

A persistent misconception is that antibiotics are effective against viral infections, a belief held by a large proportion of the public and primary care patients. Several studies have highlighted this widespread misunderstanding and its influence on antibiotic prescribing practices [45–47]. Many GPs in our study emphasized the importance of patient education in countering this misconception. Educated patients were more likely to accept symptomatic care and less likely to demand antibiotics. This aligns with research showing that face-to-face educational interventions can improve antibiotic-related knowledge and attitudes, potentially reducing inappropriate expectations [48]. Additional methods, such as posters and written educational materials have also demonstrated potential as effective supplementary interventions [49–52]. Combining these educational strategies with diagnostic training might strengthen stewardship and improve outcomes. Still, our findings suggest that prescribing is rarely a purely biomedical decision, it involves navigating expectations, habits and interpersonal dynamics that resist simple standardization.

Strengths and limitations

Using a qualitative design with semi-structured interviews allowed us to explore these perspectives in-depth, with the possibility to clarify and follow up. However, limitations must be acknowledged. First, one interview was conducted as a group interview, which although chosen in order to elicit group dynamics and divergent perspectives, may also have inhibited individual expression, restricted deeper reflections, and may have led to more superficial responses. Although 10 doctors participated, and after nine interviews, the ongoing analysis indicated additional participants were unlikely to contribute substantially new insights, a larger and more diverse sample might have provided a broader range of perspectives and potentially revealed themes not captured in our study. The relatively small number of participants, combined with the geographical focus in the Capital Region of Denmark, may have limited the diversity of clinical approaches, patient populations and practice settings represented in our findings. However, as is often the case in qualitative studies, contextual factors influenced how certain themes were developed, the emphasis they received, and whether other themes appeared at all. Consequently, these factors may have limited the comprehensiveness and nuance of the analysis, and the generalizability of our results should be interpreted with caution. Conversely, our geographically focused sample provides an in-depth understanding of a regionally defined group of doctors, which could serve as a basis for a comparative study. A further limitation of this study is leaving out interviews with nurses. Importantly, several of the interviewed GPs emphasized that nurses routinely conduct initial assessments and play an essential role in diagnostic and treatment decisions for ARS. Their omission may have limited the scope and depth of insights into clinical decision-making. Also worth mentioning is the fact that the recruitment was mediated by practice secretaries and may have introduced bias but was nevertheless considered the most effective strategy for reaching participants outside existing networks. In addition, it is possible that the doctors who chose to participate were more engaged with issues of antimicrobial stewardship, which could potentially limit the extent to which the findings reflect the diversity of perspectives within the wider population of doctors working in general practices in The Capital Region of Denmark. Finally, the translation of quotes from Danish to English introduced the potential for misinterpretation, as nuances in meaning may have been lost or altered during the process. Despite these limitations, this study contributes with in-depth knowledge about how doctors approach clinical decision-making and the factors influencing their choices.

Implications

Our study highlights the complexity of diagnosing and managing ARS, revealing the many factors that shape antibiotic prescribing. Variation in doctor involvement, diagnostic practices and CRP thresholds stresses the broader challenges in establishing standardized clinical protocols. While experienced doctors often rely on routine and non-clinical pressures, such habits may risk misdiagnosis and inappropriate antibiotic use, undermining efforts against AMR. Strengthening diagnostic training, patient education and promoting evidence-based guidelines could potentially help, but assuming these alone ensure optimal prescribing overlooks the nuanced reality of clinical work. Our findings demonstrate that the process of diagnosis is mediated by uncertainty and shaped by a variety of competing priorities. Some doctors are guided by public health goals, others by patient demands or workflow efficiency. In sum, improving antibiotic stewardship may not only require enhanced diagnostic training and guideline adherence but also an appreciation for the complexity and uncertainty inherent in clinical decision-making.

Funding Statement

No funding applicable.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Data availability statement

No additional data are available.