Utility of electronic AKI alerts in intensive care: A national multicentre cohort study

Background: Electronic AKI alerts highlight changes in serum creatinine compared to the patient's own baseline. Our aim was to identify all AKI alerts and describe the relationship between electronic AKI alerts and outcome for AKI treated in the Intensive Care Unit (ICU) in a national multicentre cohort. Methods: A prospective cohort study was undertaken between November 2013 and April 2016, collecting data on electronic AKI alerts issued. Results: 10% of 47,090 incident AKI alerts were associated with ICU admission. 90-day mortality was 38.2%. Within the ICU cohort 48.8% alerted in ICU. 51.2% were transferred to ICU within 7 days of the alert, of which 37.8% alerted in a hospital setting (HA-AKI) and 62.2% in a community setting (CA-AKI). Mortality was higher in patients transferred to ICU following the alert compared to those who had an incident alert on the ICU (p < 0.001), and was higher in HA-AKI (45.3%) compared to CA-AKI (39.5%) (35.0%, p = 0.01). In the surviving patients, the proportion of patient recovering renal function following, was significantly higher in HA-AKI alerting (84.2%, p = 0.004) and CA-AKI alerting patients (87.6%, p < 0.001) compared to patients alerting on the ICU (78.3%). Conclusion: The study provides a nationwide characterisation of AKI in ICU highlighting the high incidence and its impact on patient outcome. The data also suggests that within the cohort of AKI patients treated in the ICU there are significant differences in the presentation and outcome between those patients that require transfer to the ICU after AKI is identified and those who develop AKI following ICU admission. Moreover, the study demonstrates that using AKI e-alerts provides a centralised resource which does not rely on clinical diagnosis of AKI or coding, resulting in a robust data set which can be used to define the incidence and outcome of AKI in the ICU setting. (C) 2017 Published by Elsevier Inc.