The effect of a GP's perception of a patient request for antibiotics on antibiotic prescribing for respiratory tract infections: secondary analysis of a point-prevalence audit survey in 18 European countries.
Domen J., Aabenhus R., Balan A., Bongard E., Böhmer F., Bralić Lang V., Bruno P., Chlabicz S., Colliers A., García-Sangenís A., Ghazaryan H., Kowalczyk A., Jensen S., Lionis C., van der Linde TM., Malania L., Pauer J., Tomacinschii A., Vellinga A., Zastavnyy I., Goossens H., Butler CC., van der Velden AW., Coenen S.
BACKGROUND: Illness severity, comorbidity, fever, age, and symptom duration influence antibiotic prescribing for respiratory tract infections (RTI). Non-medical determinants, such as patient expectations, also impact prescribing. AIM: To quantify the effect of a GP's perception of a patient request for antibiotics on antibiotic prescribing for RTI and investigate effect modification by medical determinants and country. DESIGN & SETTING: Prospective audit of general practices in 18 European countries. METHOD: Consultation data were registered of 4982 patients presenting with acute cough and/or sore throat. A mixed-effect logistic regression model analysed the effect of GPs' perceptions of a patient request for antibiotics. Two-way interaction terms assessed effect modification. Relevant clinical findings were added to subgroups of lower RTI (LRTI), throat infection, and influenza-like-illness (ILI). RESULTS: A GP's perception of a request for antibiotics meant they were four times more likely to prescribe antibiotics (odds ratio [OR] 4.4, 95% confidence interval [CI] = 3.4 to 5.5). This effect varied by country: lower in Spain (OR 0.06), Ukraine (OR 0.15), and Greece (OR 0.22) compared with the lowest prescribing country. The effect was higher for ILI (OR 13.86, 95% CI = 5.5 to 35) and throat infection (OR 5.1, 95% CI = 3.1 to 8.4) than for LRTI (OR 2.9, 95% CI = 1.9 to 4.3). For ILI and LRTI, GPs were more likely to prescribe antibiotics with abnormal lung auscultation and/or increased or purulent sputum and for throat infection, with tonsillar exudate and/or swollen tonsils. CONCLUSION: GPs' perceptions of an antibiotic request and specific clinical findings influence antibiotic prescribing. Incorporating exploration of patient expectations, point-of-care testing, and discussing watchful waiting into the decision-making process will benefit appropriate prescribing of antibiotics.