Visual Interfaces for Practice Analytics: Transforming Routinely Collected Clinical Indicator Data into Reflective Practice Dashboards for Physician and Surgical Teams

Abstract

Hospitals generate large amounts of data from the delivery of care to patients. Clinical indicators (CIs) based on these data, such as length of stay and readmissions are extracted from patient administration systems. The datasets may be valuable in supporting professional learning and development if they could be repurposed and made available to clinicians. Teams and individual clinicians do not normally see the data, even though it has the potential to provide new insights that enable them to improve their practice. This has yet to be realised as these CIs are not frequently available in suitable forms and little is known how best to present them. First, we conducted a scoping review to map the literature on dashboards based on routinely collected CIs. Second, we conducted an international thought leader study to capture perspectives on why these CIs were not used for reflective practice. Third, we conducted a single hospital site need finding study, to identify the key design considerations when selecting and presenting these indicators to clinicians. Fourth, we drew upon the first three studies to design three new reflective practice dashboards based on real world CI data. Finally, we evaluated these dashboards by observing their use in three authentic settings of clinical team meetings: Neurology, Urology, and an Oncology multi-specialty group. The contributions of this thesis are: (1) the characterisation of state-of-the-art dashboards based on repurposed CIs; (2) the identification of the benefits, barriers, and enablers for reusing these CIs; (3) visual design guidelines for selecting and presenting the CIs for reflective practice; (4) demonstration of the potential value of reflective practice dashboards based on authentic clinical indicator data for three diverse clinical teams; (5) a dashboard system and replicable process to deploy in new specialty groups and hospitals; and (6) the underlying code base and design as a foundation for replication.