Acute sinusitis: Which factors do FPs believe are most diagnostic and best predict antibiotic efficacy?
A questionnaire-and-case-vignette study reveals misjudgment of infection and antibiotic overprescribing.
None of the other psychosocial variables, nor any family doctor or health center characteristics (eg, number of partners, list size, collegiate membership, special interest in otorhinolaryngology) were significant in affecting the decision to prescribe in either case vignette, in either univariate analysis or in multivariate analysis.
TABLE 3
Vignette A (unlikely bacterial sinusitis): Belief in antibiotic effectiveness for symptoms the most important variable in decision to prescribe
| CRUDE ODDS RATIO (95% CI) | ADJUSTED ODDS RATIO (95% CI) | WALD TESTZ (P VALUE) | |
|---|---|---|---|
| Doctor's belief in effectiveness of antibiotics in treating symptoms of acute sinusitis | 2.16 (1.62–2.88) | 2.14 (1.52–3.0) | 4.41(<.001) |
| Pain (factor 4) | 1.49 (1.21–1.83) | 1.31(1.05–1.64) | 2.40 (.017) |
| Nasal (factor 1) | 1.37 (1.12–1.67) | 1.44 (1.17–1.78) | 3.37 (.001) |
| Townsend deprivation score | 1.07 (1.01–1.14) | 1.08 (1.01–1.15) | 2.12 (.03) |
| Health Centre type* | |||
| urban | 0.62 (0.42–0.93) | 1.09 (0.64–1.86) | |
| rural | 0.63 (0.38–1.07) | 1.22 (0.62–2.39) | |
| Number of years as a family doctor | 1.03 (1.00–1.05) | 1.02 (0.99–1.04) | 1.19 (.233) |
| *The comparison group were suburban practices. | |||
TABLE 4
Vignette B (likely bacterial sinusitis): Belief in antibiotic effectiveness for symptoms again the chief determinant in decision to prescribe
| CRUDE ODDS RATIO (95% CI) | ADJUSTED ODDS RATIO (95% CI) | WALD TESTZ (P VALUE) | |
|---|---|---|---|
| Doctor’s belief in effectiveness of antibiotics in treating symptoms of acute sinusitis | 2.90 (2.08–4.04) | 2.70 (1.90–3.82) | 5.58 (<.001) |
| Doctor’s belief in effectiveness of antibiotics in treating complications of acute sinusitis | 2.16 (1.52–3.06) | 1.67 (1.02–2.75) | 2.82 (.005) |
| Family history | 1.67 (1.02–2.75) | 1.53 (0.88–2.64) | 1.51(.13) |
Discussion
This study shows that family doctors in the UK diagnose and treat presumed acute sinusitis in a manner contrary to available evidence.
Family doctors mostly use measures of pain rather than presence of purulence when diagnosing acute sinusitis. The 4 highest rated symptom and sign scores were indices of pain. Unilateral pain and rhinorrhea were also important.
Doctors’ beliefs in the importance of nonpurulent nasal symptoms were significantly associated with prescribing, though sinusitis was unlikely. Purulent nasal symptoms did not reach significance for either case vignette. Doctors’ reliance on pain and nonpurulent nasal symptoms is in contrast to studies indicating that symptoms or signs of purulent nasal secretions are the most predictive of bacterial sinusitis.9,10
Doctor’s beliefs about the effectiveness of antibiotics strongly influenced their decision to prescribe, when the diagnosis of acute sinusitis was likely (vignette B) and when it was less likely (vignette A). However, the decision to prescribe when the diagnosis was less likely was also linked to the health centre material deprivation index and with doctors’ belief in the importance of non–evidence-based nasal symptoms—eg, nasal blockage with facial pain. In our study, 37% of doctors prescribed antibiotics for unlikely acute sinusitis, and 82% for likely diagnoses.
A Cochrane review concluded that antibiotics were moderately effective when acute sinusitis was confirmed by diagnostic tests. In contrast, a recent primary care trial with broader entry criteria including maxillary pain, purulent discharge, or pain on bending forward (without formal diagnosis) showed no significant improvement in recovery time following antibiotics.24 The effectiveness of antibiotics for acute sinusitis in family practice populations is yet to be clarified, and there are no clearly agreed predictors of treatment response in any healthcare setting.2,24-29
The importance of doctors’ beliefs in benefit of antibiotics for symptom relief and the high reported prescribing rate contrast with current evidence for only moderate effectiveness of antibiotics.1,3,24-29 Further evidence is needed to guide appropriate antibiotic prescribing in acute sinusitis-like illness in family practice.
Psychosocial factors seemed unimportant. The family history was associated with the decision to prescribe in probable sinusitis, but its effect was not independent when controlling for the doctors beliefs. Apart from the sociodemographic variable—the Townsend Index of the practice—none of the psychosocial factors we identified as important appeared to be significant in the decision to prescribe. This contrasts with the original work from Howie in which psychosocial factors had strong bearing on the decision to prescribe for sore throat. It may reflect a greater perceived need for antibiotics with suspected sinusitis, which may cause more discomfort and last longer, or it may reflect a change in attitude to prescribing over time. It is possible that with improved evidence about prescribing, psychosocial factors are now weighted less heavily in prescribing decisions. Qualitative work from the UK and case vignettes in Australia investigating prescribing for sore throat both lend support to this assertion.16,30
Study limitations. Our sample was taken from doctors practicing in Wessex, where there is a broad range of health care settings. The response rate of 54.4% was low.
