CASE STUDIES
Federated analytics for studying drug utilisation in primary care and hospital settings in the UK
In the realm of paediatric research, obtaining comprehensive data can be a challenge, often requiring the integration of multiple sources from both primary and secondary care. This complexity raises concerns about data governance and the risks associated with transferring sensitive information.
To tackle these issues, researchers in the UK have adopted a federated analysis (FA) approach to study the trends in Fluoroquinolone (FQ) drug utilisation, particularly in response to the 2019 risk minimisation measures (RMM) set by the Medicines and Healthcare products Regulatory Agency. Aim of the Study The primary goal of this study was to evaluate the effectiveness of a federated analysis approach for assessing drug utilisation in paediatrics. Methodology Data were locally extracted from four key UK databases: CPRD GOLD, Barts Health, Great Ormond Street Hospital, and HIC Dundee.
The study focused on subjects under 18 years old who had been registered in these databases between 1 January 2012 and 31 December 2022, ensuring they had at least 30 days of visibility before being exposed to FQ. The analysis code was developed by the UK Observational Health Data Science and Informatics team at Oxford University and tested at each site before running the final analysis.
Results from all sites were combined during a focused 5-day study-a-thon. Key Findings The federated analytics approach enabled a large-scale multi-centre study while minimising data governance risks. Over the course of the study-a-thon, results were successfully generated from six data partner sites.
An RShiny app, created by the data science team during this period, facilitated the pooling and exploration of aggregated results from all partners. The combined data allowed researchers to calculate annual incidence rates of FQ usage and analyse country-wide trends before and after the RMM. Differences in FQ utilisation between primary care (CPRD) and secondary care (Barts Hospital) were also examined.
This case study illustrates how federated analytics can effectively support large-scale characterisation and drug utilisation studies in paediatrics, an area often hindered by limited data availability.
By leveraging multiple datasets while safeguarding patient privacy, this innovative approach paves the way for more robust research in paediatric healthcare. Related How does the federated analytics approach enhance data security What specific trends in Fluoroquinolones use were identified in the study How did the study ensure data consistency across different databases What role did the UK Observational Health Data Science and Informatics team play in the study How can the findings of this study be applied to other pediatric drug studies