Acute Kidney Injury (AKI) is common in hospital admissions, effecting up to 20% of hospital admissions. It is associated with an increased risk of chronic kidney disease, prolonged length of hospital stay and a higher mortality rate. Research into this syndrome has increased rapidly in this century following the introduction of clear definitions, the most recent of which was created in 2012. This is the Kidney Disease Improving Global Outcomes (KDIGO) definition and is based on changes in serum creatinine (sCr) and urine output. In an aim to improve recognition and therefore outcomes of AKI, an electronic alert (eAlert) was rolled out across Wales in 2014-15. The eAlert is a copy of the National Health Service (NHS) England eAlert algorithm (a modification of the KDIGO criteria), which has been applied by laboratories to all sCr tests sending out alerts to the requester when AKI is detected. Based on clinical experience we suspect that there are variations in the local application of the eAlert system.
The project is being carried out using the secure anonymised information linkage (SAIL) databank at Swansea University. This is a data safe haven, where anonymised patient level data can be linked and analysed safely and securely without risk of identification of patients. Using an analyst (Gareth Davies), we will recreate the AKI eAlert algorithm applying it to the biochemical data within SAIL (Swansea, Bridgend and the Welsh Results Reporting Service (WRRS) datasets). By running this algorithm against these sCr results, we will identify those patients who trigger alerts, by using the renal dataset we can then remove dialysis patients. This will identify those with ‘true’ AKI based on the sCr criteria in all adult patients. We can then compare these alerts to the laboratory alert codes in AWLIMS, testing a hypothesis that some alerts are sent out erroneously by the laboratory (i.e. in dialysis patients). Using the primary care dataset we will be able to collect comorbidity data such as diabetes, ischaemic heart disease etc….) and we will be able to compare medication before and after the episode of AKI. Using hospital admissions data (allowing comparison with diagnosis coding and the alerts), critical care data and office of national statistics death data we will assess outcomes (length of stay, intensive care and death) of AKI.
1. Validate and assess the impact of electronic Alerts for AKI
2. Determine the temporal changes of hospital coding for AKI
3. Evaluate the impact of AKI on primary care
Dr Tim Scale
Working with Dr J Chess, Professor R Lyons, Professor S Bain and Gareth Davies