2290

HOSPITAL ACQUIRED AKI – ELECTRONIC ALERTS ALONE MAY NOT BE ENOUGH

WhiteleyS¹, Ghalli F²,Meran S², Roberts G¹ ²

¹Royal Gwent Hospital, Newport, ²University Hospital of Wales, Cardiff

BACKGROUND: Acute Kidney Injury (AKI) is a well recognised cause of avoidable patient harm. In an attempt to improve AKI management, many hospitals across the UK have introduced automated electronic AKI alerts (e-alerts). It is hoped that the introduction of e-alerts will lead to more timely intervention and improved outcomes. At present there is a paucity of data on the impact of e-alerts, in particular in those hospitals without on site nephrology services. Here we report on the impact of an e-alert system, on the outcomes of patients developing hospital acquired AKI in two large district general hospitals serving a population of 600,000.

Methods: In April 2013, a computer program was developed that detected and highlighted instances of AKI occurring in hospital inpatients.As per the Acute Kidney Injury Network (AKIN) criteria, a 50% increase in creatininewithin a 48 hour period was classed as an AKI event. All hospital acquired AKI events occurring between April and October 2013 were analysed. To assess any potential impact of the e-alert, data were compared with a previous AKI outcomes study undertaken in the same region over a similar time period.

RESULTS: 380 AKI events were recorded over the 6 month observation period. The median patient age was 75, with 53% of patients being male. The majority of patients with AKI were under a medical speciality (66%). The commonest primary diagnosis during admission was infection(27%), followed by cardiovascular disease (13%) and cancer (9%).With regard to peak stage of AKI, 28% and 32% progressed to stage 2 and 3 respectively. Inpatient mortality increased according to peak stage of AKI, with respective inpatient mortality of 24%(stage 1), 32% (stage 2) and 48 %( stage 3). At time of discharge, 68% of patients had creatininevalues above their previous baseline values, with a median increase of 67μmol/L.

Compared to the findings of a previous study from our region (pre-dating the e-alert), there was an increase in the number of patients having a timely acid-base status check (45% vs. 14%). TheAKI guidelines linked to the e-alert were accessed (on average) 110 times per month in the 6 months following the e-alert introduction. In spite of this apparent increased recognition of AKI, the need for renal replacement therapy during admission (2% current study vs. 3% previous study), the overall mortality (40% current vs. 42% previous ), and median length of stay (16 dayscurrent vs. 17days previous)were similar.

CONCLUSION: In this study we have looked at the impact of an e-alert system on the outcome of hospital acquired AKI in two district general hospitals served by a nephrology outreach service. Though the e-alert appears to have improved recognition of AKI, it does not appear to have had a significant impact on length of stay, inpatient mortality nor the need for RRT during admission. Our study is limited by its retrospective nature, the lack of case note review, and the use of an alert based on AKIN criteria (which has now been superseded by an e-alert system based on KDIGO criteria, which also captures community acquired AKI). We speculate that in order to maximise its potential impact, an e-alert system in a district general hospital needs to be supported by additional innovations. An example of this may be that stage 2 and 3 alerts trigger review by a rapid access outreach team.