Estimating variations in the use of antibiotics in primary care: Insights from the Tuscany region, Italy

Background Practice variation is a well-known phenomenon that affects all aspects of healthcare delivery and leads to suboptimal health outcomes as well as poor resource allocation. Given the global rise of antimicrobial resistance, practice variation is of particular concern when it comes to the prescription of antibiotics. A growing number of healthcare systems are tackling this issue at all levels of healthcare governance. Aims and objectives This study sought to estimate the variation in antibiotic use across different levels of Tuscany's primary care, and assess the extent to which the organization of primary care delivery is responsible for this variation. Methods We analysed the performance and variation for seven indicators related to the use of antibiotics at three levels of healthcare governance: (i) the clinician level (2619 general practitioners [GPs]); (ii) the peer-group level (all 116 GP group practices) and (iii) the institutional level (all 26 health districts). For the statistical analysis, we built three-level mixed effects models that were fitted with 2619 GPs, 116 GP group practices and 26 health districts. Results The multi-level models suggested that the grand majority of the variation in antibiotic use was located at the GP level (75% to 97%). However, the percentage of variation associated with GP group practices and health districts ranged from 3% to 25%, depending on the type of indicator analysed. Conclusion While the variation was found to be in large part due to differences between GPs themselves, the influence exerted by peer groups and institutional mechanisms does have a significant impact as well. Further research needs to be conducted regarding the institutional and contextual factors that prompt GPs to harmonize their prescribing behaviour in line with best practices and lead to not only improved patient outcomes but also large cost-savings.