Goal

General practices (GP practices) across the UK prescribe antibiotics at different rates. This variation may reflect differences in the patients they see (for example, age, underlying health conditions, or infection severity) or differences in prescribing behaviour between practices. We wanted to understand which of these factors contributes most to the differences in antibiotic prescribing between GP practices.

Lead

Nam Nguyen – Nuffield Department of Primary Care Health Science, University of Oxford

What we did

Using a large UK primary care database, we examined variation in antibiotic prescribing across GP practices for 11 common infections, including acute bronchitis, cough, acute otitis media, acute rhinosinusitis, acute sore throat, asthma exacerbation, COPD exacerbation, acute gastroenteritis, impetigo, respiratory tract infections, and urinary tract infections. We investigated the factors driving these differences and explored fairer approaches for comparing antibiotic prescribing between GP practices.

Key learnings

We found that antibiotic prescribing is often higher than necessary and varies substantially between GP practices. Most of this variation was explained by differences in prescribing behaviour rather than differences in the patients seen by each practice. We also found that comparisons between practices were most accurate when they accounted for detailed patient characteristics, including infection severity. Finally, we showed that comparing practices separately for individual infections is more informative than combining all infections into a single measure. Among the infections studied, upper respiratory tract infections and acute otitis media may be particularly suitable for benchmarking antibiotic prescribing performance between GP practices.

Outputs

Explaining variation in antibiotic prescribing for common infections: a three-way variance decomposition using UK primary care data

Funder

The Wellcome Trust and NIHR