Practice recommendations
- Pain is not a sufficient criterion to diagnose sinusitis clinically (B).
- Purulence makes sinusitis more likely (B).
- Antibiotics are of moderate benefit in acute sinusitis (A).
- UK family doctors predominantly use specific measures of pain in diagnosis, and rate purulence less strongly.
- Patient psychosocial factors do not appear important in the decision to prescribe.
- The decision to prescribe is influenced most strongly by the FPs belief in the effectiveness of antibiotics.
Acute sinusitis is almost always a self-limiting condition involving the maxillary sinuses. About half of cases are free of bacterial infection, and 70% to 80% of patients will be symptom-free by 2 weeks, with or without antibiotics.1,2 Only a small proportion develop chronic sinusitis (>12 weeks),3 characterized by low-grade symptoms.4
However, unlike other upper respiratory infections, acute sinusitis is still over-treated with antibiotics by primary care physicians—85% to 98% of sinusitis patients in the US receive antibiotic prescriptions,5 92% in the UK,6 80% in Norway,7 and 67% in the Netherlands.8 Why is this so, given the self-limiting nature of sinusitis, the very real threat of increasing antibiotic resistance, and the estimated annual treatment costs of £10 million in the UK and $2.4 billion in the US?4
Our study aimed to establish criteria used by family doctors to diagnose acute sinusitis, and to determine the most important factors affecting doctors’ decisions to prescribe antibiotics.
What clinical features make diagnosis more likely?
The most rigorous scientific evaluation to date, though admittedly conducted in a selected population, is that of Berg and Carenfelt,9 who identified 3 symptoms and 1 sign that are most diagnostic of bacterial infection:
- Purulent rhinorrhea with unilateral predominance
- Local pain with unilateral predominance
- Bilateral purulent rhinorrhea
- Pus in the nasal cavity.
Presence of 2 or more of these findings gave a specificity of positive bacterial culture on sinus aspiration of 77% to 78% and a clinical reliability of 86%. This is reviewed in the Agency for Health Care Policy and Research evidence-based guideline.3
Other studies have confirmed the importance of purulent secretions to a diagnosis of bacterial infection and have suggested several additional predictive features, such as lack of response to decongestants, biphasic illness after a cold, temperature >38.5°C, and tests results such as elevated erythrocyte sedimentation rate or C-reactive protein.10-13 Limited evidence suggests that nasal purulence predicts therapeutic benefit from antibiotic therapy.14
It has been unclear, however, which symptoms and signs family doctors deem important in daily practice.
What influences the treatment decision?
Howie’s classic paper on sore throats showed that the patient’s psychological and social history significantly affected the doctor’s response. Doctors’ beliefs in the effectiveness of antibiotics were not assessed in relation to these factors.15 We were unable to find any studies that have evaluated factors influencing antibiotic prescribing for acute sinusitis. More recent qualitative work on sore throat and respiratory infections has identified a number of specific and general factors that affect prescribing and are likely to hold some relevance for sinusitis.16-18
We constructed a questionnaire and case vignettes to assess physicians’ means of diagnosing and treating acute sinusitis.
Methods
Development of the questionnaire
Identifying the diagnostic criteria. First we reviewed the diagnostic and clinical literature in acute maxillary sinusitis and rhinosinusitis to identify all common symptoms and signs used to diagnose the condition. We also conducted a focus group at a Wessex Research Network (WReN) meeting, with approximately 20 family doctors and nurses contributing to the discussions on the “everyday” diagnosis of sinusitis and the development of the items to include in the questionnaire. We identified symptoms and signs using an open-ended approach, asking these professionals to reflect on aspects of their decision-making processes.19
After eliminating duplications and unclear criteria, we retained 24 symptoms and signs most often used for diagnosis. This inclusive list (TABLE 1) became part of the final questionnaire. Participating doctors were instructed to rate all 24 items for their “importance to you in diagnosis” on a 5-point Likert scale (4=very important, 3=important, 2=moderately important, 1=slightly important, 0=not important), and to further select the 5 most important symptoms and signs they actually used to diagnose acute sinusitis in everyday practice.
Construction of case vignettes. For the second part of the questionnaire we used 2 case vignettes to explore clinical judgment. The vignettes reflected the importance of psychosocial variables, as shown in the work of Howie in explaining respiratory prescribing in family practice. We also included such cognitive factors as doctors’ beliefs about therapy. Case vignettes allowed us to evaluate several real-life factors that influence judgment in prescribing decisions.
A small group of doctors and researchers (the authors) met to consider the themes in the literature and our own qualitative work databases.15-17 We discerned and agreed on 6 important features potentially relevant to clinical judgment in prescribing for acute sinusitis in every day practice. These features were included in the vignettes: