Abstract
The proliferation of electronic health records has led to an increasing volume of routine data about patients, their conditions and clinical interventions. While these data are widely used by clinicians to inform treatment, there is little standardised aggregation for clinical support or operational insights. This research explores a possible approach to redressing this. We describe a scalable 'clinical intelligence tool' allowing researchers and managers to aggregate and analyse routine data in a generalised form, generating on-demand analyses for clinical planning and hospital operations. We demonstrate a proof-of-concept implementation, called PICTURE. In addition to detailing the architecture of this tool, we show how it can address analytics questions, such as providing an informatics consult to address clinicians' questions. We show that aggregating health record data in a clinical intelligence tool has the potential to improve clinical practice, healthcare provider operations and more. The key innovations are the generalisation and standardisation of analytical methods for reuse across a range of EHR data; the use of generic data models to support the reusability and composability of analyses; and the approach of developing analytic functionality with open-source languages and modern software development practices.