Hodgson L.E.Selby N.TAO-MIN HUANGForni L.G.2021-08-242021-08-2420190270-9295https://www.scopus.com/inward/record.uri?eid=2-s2.0-85071992746&doi=10.1016%2fj.semnephrol.2019.06.002&partnerID=40&md5=25d310407a6a955e35fb6361c3255749https://scholars.lib.ntu.edu.tw/handle/123456789/579325Summary: Acute kidney injury is a major health care problem. Improving recognition of those at risk and highlighting those who have developed AKI at an earlier stage remains a priority for research and clinical practice. Prediction models to risk-stratify patients and electronic alerts for AKI are two approaches that could address previously highlighted shortcomings in management and facilitate timely intervention. We describe and critique available prediction models and the effects of the use of AKI alerts on patient outcomes are reviewed. Finally, the potential for prediction models to enrich population subsets for other diagnostic approaches and potential research, including biomarkers of AKI, are discussed. ? 2019Acute kidney injury; electronic alerts; prediction models[SDGs]SDG3biological marker; acute kidney failure; attitude to health; care bundle; clinical decision support system; clinical feature; clinical outcome; health care delivery; human; medical education; mortality rate; online system; priority journal; prognostic assessment; protocol compliance; Review; risk assessment; acute kidney failure; forecasting; procedures; prognosis; risk assessment; statistical model; Acute Kidney Injury; Decision Support Systems, Clinical; Forecasting; Humans; Models, Statistical; Prognosis; Risk Assessmentreview10.1016/j.semnephrol.2019.06.002315149062-s2.0-85071992746