Adekanmbi Victor, Jones Hywel, Farewell Daniel, Francis Nick A
Division of Population Medicine, School of Medicine, Cardiff University, Cardiff CF14 4YS, UK.
Primary Care, Population Sciences and Medical Education, University of Southampton, Southampton SO17 1BJ, UK.
J Antimicrob Chemother. 2020 Aug 1;75(8):2363-2371. doi: 10.1093/jac/dkaa168.
To examine the association between socioeconomic status (SES) and antibiotic prescribing, controlling for the presence of common chronic conditions and other potential confounders and variation amongst GP practices and clusters.
This was an electronic cohort study using linked GP and Welsh Index of Multiple Deprivation (WIMD) data. The setting was GP practices contributing to the Secure Anonymised Information Linkage (SAIL) Databank 2013-17. The study involved 2.9 million patients nested within 339 GP practices, nested within 67 GP clusters.
Approximately 9 million oral antibiotics were prescribed between 2013 and 2017. Antibiotic prescribing rates were associated with WIMD quintile, with more deprived populations receiving more antibiotics. This association persisted after controlling for patient demographics, smoking, chronic conditions and clustering by GP practice and cluster, with those in the most deprived quintile receiving 18% more antibiotic prescriptions than those in the least deprived quintile (incidence rate ratio = 1.18; 95% CI = 1.181-1.187). We found substantial unexplained variation in antibiotic prescribing rates between GP practices [intra-cluster correlation (ICC) = 47.31%] and GP clusters (ICC = 12.88%) in the null model, which reduced to ICCs of 3.50% and 0.85% for GP practices and GP clusters, respectively, in the final adjusted model.
Antibiotic prescribing in primary care is increased in areas of greater SES deprivation and this is not explained by differences in the presence of common chronic conditions or smoking status. Substantial unexplained variation in prescribing supports the need for ongoing antimicrobial stewardship initiatives.
研究社会经济地位（SES）与抗生素处方之间的关联，同时控制常见慢性病的存在以及其他潜在混杂因素，并考虑全科医生诊所和医疗群组之间的差异。
这是一项利用全科医生数据与威尔士多重贫困指数（WIMD）相链接的电子队列研究。研究背景是参与2013 - 2017年安全匿名信息链接（SAIL）数据库的全科医生诊所。该研究纳入了339个全科医生诊所中的290万患者，这些诊所又嵌套在67个全科医疗群组中。
2013年至2017年期间共开出约900万张口服抗生素处方。抗生素处方率与WIMD五分位数相关，贫困程度越高的人群接受的抗生素越多。在控制了患者人口统计学特征、吸烟情况、慢性病以及按全科医生诊所和医疗群组进行聚类分析后，这种关联仍然存在，最贫困五分位数人群的抗生素处方比最不贫困五分位数人群多18%（发病率比 = 1.18；95%可信区间 = 1.181 - 1.187）。在零模型中，我们发现全科医生诊所之间[组内相关系数（ICC）= 47.31%]以及全科医疗群组之间（ICC = 12.88%）抗生素处方率存在大量无法解释的差异，在最终调整模型中，全科医生诊所和全科医疗群组的ICC分别降至3.50%和0.85%。
在社会经济地位较低的地区，初级医疗中的抗生素处方量增加，这不能用常见慢性病的存在或吸烟状况的差异来解释。处方中存在大量无法解释的差异，这表明持续开展抗菌药物管理举措的必要性。
