Exploiting electronic health records to infection management and optimise antimicrobial use

Large-scale electronic health record data can potentially answer a far greater number of questions about infection management than traditional epidemiological studies using questionnaires. Their volume and scale are continuously increasing as larger amounts of healthcare data are linked and de-identified for research. Examples and challenges include
 

• Can we identify a wider range of risk factors for infection to target interventions? The sheer number of factors that could be considered, many with substantial amounts of missing data, poses challenges to traditional epidemiological approaches. A recent novel statistical analysis approach called ‘doublethink’ has been proposed which could be applied to a range of microbiologically and/or syndromically defined infections to identify novel populations to target to reduce infection risks, and compared with other methods including machine learning.
• Can we work out how best to use diagnostic tests for infection, widely considered to be a key tool to improve antimicrobial stewardship, in the real-world? In what patient populations are they and should they be used, how often, and what are their ultimate effects on both antimicrobials prescribed and patient outcomes? 
   

Projects can exploit an existing large datawarehouse of de-identified individual patient data, the Infections in Oxfordshire Database. They would suit students interested in infections/antimicrobial usage and coding, who wish to gain experience in design of studies using healthcare records to answer real-world questions. There will be opportunities to learn how to manage and use large-scale electronic health record data, and apply a range of quantitative methods including novel causal epidemiological methods.