Observational study of antibiotic prescribing patterns by age and sex in primary care in England: why we need to take this variation into account to evaluate antibiotic stewardship and predict AMR variation.

Background: The drivers of antimicrobial resistance (AMR) likely vary substantially by different demographics. However, few complete, open, national detailed data exist on how antibiotic use (ABU) varies by both age and sex. Here, we aimed to describe the variation in ABU and consider how these age- and sex-specific patterns influence targets and methods for AMR control. Methods: Prescriptions of all antibiotics from general practices in England for 2015-23 disaggregated by 5 year age bands and sex were analysed at the national and Integrated Care Board (ICB) level. A descriptive analysis of the relative rates of ABU by age and sex was conducted, followed by an evaluation of comparison metrics of prescription levels between regions. ABU by age and sex were compared with global AWaRe targets, and ABU by age-group was correlated against influenza vaccination introduction, to evaluate the impact of such vaccines on ABU. Results: From a total of 249 578 795 prescriptions (across 9 years), 63% were given to women and the most prescribed were amoxicillin, nitrofurantoin and flucloxacillin sodium. Prescriptions per 100 000 population varied substantially across sex, age, geographical region, season, year, COVID-19 pandemic period and drug. Most antibiotics were prescribed more to women across most age bands (84% of antibiotics had more prescriptions to females across 50% of age bands). We show how this variation requires a more nuanced approach to comparing ABU across geographies and highlight that AWaRe targets are not met uniformly [prescribing in men aged 11-20 does not fulfil the Access, Watch and Reserve (AWaRe) 80% Access target]. We also show the impact on ABU of time-sensitive interruptions (38% of antibiotics showed a seasonal pattern in the absence of disruptions), including differential age-targeted influenza vaccination, COVID-19 restrictions and a shortage of amoxicillin due to a group A Conclusions: These detailed differences in ABU across England suggest that there should be large variation in AMR burden by age and sex. Linkage of this ABU data with similarly detailed open-access AMR data is now needed for better intervention design.