This proposal will incorporate a logistic regression models developed by the investigators to predict risk for acute kidney injury in pediatric intensive care unit and pediatric ward patients into the electronic medical record system, enabling personalized decision support. Real-time surveillance using the risk prediction models will identify pediatric inpatients at increased risk for acute kidney injury. When patients exceed the threshold risk for acute kidney injury, the electronic medical record system will notify providers. Patients will be randomly assigned to be in the intervention group, for whom the notification will be implemented, or in the control group, for whom the notification will not display. The risk notification will be assessed for its impact on outcomes including rates of screening for acute kidney injury and the severity of acute kidney injury.
