Utilising Routine Population Healthcare Data for Epidemiological Research and Quality Improvement in Care: Application of a National Acute Kidney Alerting System as a Proof of Concept

  • Jennifer Holmes

    Student thesis: Doctoral Thesis


    Acute kidney injury (AKI) is a global health issue known to cause avoidable harm and death. Improvement in its prevention and management is therefore considered an important goal for the healthcare sector. While improvement science states that without measurement, improvement is not possible, at the outset of this research, the epidemiology of AKI was also not fully understood. Whereas numerous studies had described AKI that is developed during hospitalisation, less had focused on that developed in the community. This healthcare informatics thesis centres on the proposed utilisation of routine serum creatinine (SCr) data for informing both the epidemiology of AKI, and improvement in care for patients who develop and/or are at risk of AKI. It follows the recent implementation of an automated laboratory-based electronic alerting system for AKI across the entire Welsh National Health Service which, in addition to its primary purpose of aiding early intervention, also had potential to be applied as a centralised system of data collection.

    The first of three projects documented in the portfolio style thesis involves development of processes to systematically extract, cleanse and classify data on electronic (e)-alerts transmitted by the algorithm of the aforementioned system. This author’s work results in a data set which is suitable for analysis and enables reliable identification of all AKI occurring in the Welsh population, including that which is hospital and community acquired. Following this is project two and presentation of a series of peer-reviewed studies by the author which, by utilising the described novel data set, include some of the first comprehensive, population-based descriptions of AKI and its associated factors known to date. Project three follows with further practical examples of work by the author to demonstrate the potential of the same data set. It begins with studies by the author which, in helping end-users better understand the significance of AKI e-alerts, could allay speculation healthcare providers and professionals may have regarding their clinical utility. It closes by presenting a prototype of a tool which, via application of statistical process control techniques and indicators developed by the author, provides a means to robustly measure, and identify variation in, quality of healthcare provision related to AKI in Wales, a mechanism which did not exist prior.

    The original contribution to knowledge is the methodological development of a new national data source for AKI, and the demonstration and conclusion that these large validated data provide unique opportunity to reliably inform the epidemiology of, and quality improvement in healthcare provision for, AKI. Work on and of this scale and scope is yet to be achieved elsewhere. This thesis therefore significantly adds to the literature to which it relates. To validate its contribution, and support its overall conclusion, the peer-reviewed status of the author’s 14 published works appended and presented throughout are offered. Finally, while the work here concerns SCr and AKI, it is also presented as a wider proof of concept which in theory could be replicated for other routine laboratory-based healthcare data, to improve the epidemiological understanding and management of other health conditions.
    Date of Award2022
    Original languageEnglish
    SupervisorRay Higginson (Supervisor) & John Geen (Supervisor)

    Cite this