SO041
INCIDENCE AND MORTALITY OF PATIENTS WITH ACUTE KIDNEY INJURY IN HOSPITALISED PATIENTS USING THE NHS ENGLAND AKI ALGORITHM

Introduction and Aims: Acute kidney injury (AKI) is associated with high mortality, health care cost and is common in hospitalised patients. A national AKI alert algorithm has standardised the definition of AKI and has recently been implemented by NHS England. The idea is to ensure a timely and constant approach to the detection and diagnosis of patients with AKI across the NHS. In our hospital, we have developed a fully automated AKI detection system that runs in our Laboratory Information Management System (LIMS) since 2012 and has undergone extensive internal validation. We have also done validation work on NHS England AKI algorithm and then the system was upgraded to NHS England AKI Algorithm in March 2015. This audit looked into the incidence and outcome of AKI cases using NHS England AKI alert algorithm.

Methods: We looked in all AKI alerts generated over a month period (May 2015). We also looked into the demography, in- patient mortality, length of stay, recovery of renal function and mortality in 3 months post AKI.

Results: A total of 327 alerts were generated (stage 1 = 68.2 %, stage 2 = 17.7%, stage 3 = 14.1%; table 1). The audit data identified that 89.6% of AKI alerts were true AKI after reviewing the results and trend electronically (10.4% false alerts and mostly stage 1 AKI false alerts). There was high in-hospital mortality in all stages as outlined in Table 1. However, there was a reduction in terms of stage 3 AKI mortality comparing to a previous audit in 2013¹ (stage 3 AKI mortality 32.6% vs.43.7% in 2013) with the implementation of a dedicated AKI team and AKI care bundle. Total mortality within three months also improved with stage 3 AKI1¹ (41.3% compared to 50%). Although these figures have improved since our last audit, mortality from AKI still remains high for all stages of AKI following discharge at 3 months (table 1).

We also looked into the recovery and mortality at 3 months post AKI. Our data shows that full recovery from AKI for stage 1 is 71.8%, for stage 2 is 71.0% and for stage 3 is 37.5% (table 2). However, there was significant missing data in the electronic lab informatics system at 3 months results (41-47 % depending on stages) which could be due to various reasons including not having repeat blood tests in the community or, having blood tests at local centers and results not being available in our laboratory reporting system.

Conclusions: AKI alert system is useful to identify the cases but the alert system needs to be integrated into wider clinical care. Over a period of 18 months our inpatient hospital mortality has improved for all stages of AKI particularly stage 3. Implementation of our dedicated AKI team including our AKI care bundle for all wards has helped improve our outcomes. Our data shows that 3 month mortality rates remain high for all stages of AKI and more needs to be done for these patients with robust post AKI discharge plan and close follow up in the community. Ref : ¹ The usefulness of an electronic AKI alert system for early diagnosis and intervention in hospitalised patients with AKI, Abstract and Poster, ERA-EDTA , London 2015.