Implementation of an automated system for the detection of acute kidney injury in a district general hospital and its impact on patient outcomes

  • Graham Andrew White

Student thesis: Doctoral Thesis


The main aim of the study was to develop real-time detection method for the notification of acute kidney injury in patients where creatinine was requested. A software application was developed that evaluated creatinine results for acute kidney injury against a variety of algorithms. In the initial stage the only algorithm reported was that developed by the Acute Kidney Injury Network but, in 2013 following an NHS England Patient Safety Notice, this was changed to the mandated algorithm and all historical results were also re-evaluated.
Between 2010 and 2014 a total of 346,780 patients aged 18 and over covering 1,800,545 individual requests were recorded in the database. Using the NHS England algorithm 25,551 (7.4%) of patients had AKI at some stage in the study period. Analysis suggested that the introduction of AKI alerts coincided with a significant reduction in the relative risk for thirty-day mortality following an AKI episode; stage 1 mortality falling from 19.5% to 14.6%, stage 2 mortality falling from 30.1% to 22.1% and stage 3 mortality falling from 35.3% to 25.9%. However, there was also a fall in mortality for stage 2 and 3 AKI at an associated Medical Unit that did not receive alerts suggesting a significant role for the educational elements and sharing of Consultant care across sites. The data also suggest that, following an episode of acute kidney injury full recovery of renal function may take in excess of 180 days.
The data support the hypothesis that the introduction electronic alerts along with education can have significant impacts on subsequent mortality. However, other improvements in healthcare such as an increased focus on sepsis, an important factor in the development of acute kidney injury, may also have contributed to the reduction in mortality.
Date of AwardDec 2015
Original languageEnglish
Awarding Institution
  • University of Portsmouth
SupervisorJohn Young (Supervisor) & Graham Mills (Supervisor)

Cite this