Risk of unintended consequences from lower antibiotic prescribing for respiratory tract infections in primary care.

Objectives: About 60% of antibiotic prescribing in primary care is for respiratory tract infections (RTIs), some of which is likely unnecessary. There is limited evidence on the association between reduced antibiotic prescribing and adverse events. We aimed to identify associations between practice-level prescribing rates for RTIs in general practice, and patient-level adverse outcomes.

Methods: We included 1471 English General Practitioner (GP) practices, linked to hospital admissions in England, from the Clinical Practice Research Datalink for 2005 to 2019. Outcomes were hospitalisations, RTI-related re-consultations and additional antibiotic prescriptions, adjusted for practice level case-mix prescribing.

Results: Prescribing rates for practices falling within the lowest and highest prescribing quintiles were 52 and 139 prescriptions per 1000 RTI-related consultations. Patients from practices in the lowest prescribing quintile did not have significantly higher risk of hospitalisation, adjusted odds ratio 0·99 (95% CI 0·96 to 1·02). Re-consultations within 30 days were significantly higher for the lowest prescribing practices, adjusted odds ratio 1·209 (1·206 to 1·212). Additional antibiotic prescriptions and subsequent prescriptions upon re-consultation were significantly lower for the lowest prescribing practices, adjusted odds ratio 0·317 (0·314 to 0·321) and 0·706 (0·699 to 0·712), respectively.

Conclusions: Our results contribute to evidence on the safety of reduced antibiotic prescribing for RTIs in primary care. Results suggest that for the majority of practices, further reductions in RTI-related antibiotic prescribing should be possible without an increase in hospitalisation for pneumonia.