The emerging applications of machine-learning algorithms for prediction of patient outcomes using EHR data have demonstrated promise in helping physicians to anticipate future events at an expert level, and in some cases surpassing the performance of clinicians 6. Recent availability of electronic health record (EHR) data and advances in artificial intelligence (AI) have sparked growing interest in machine learning-based risk prediction models for healthcare 4, 5. However, the only FDA-approved biomarker test, Nephrocheck (Astute Medical, San Diego CA), is indicated for critically ill patients, whereas substantial proportion of patients developing AKI are cared for outside the intensive care units. Biomarkers to predict risk, diagnosis, prognosis, and therapeutic responses in patients with AKI are being identified and evaluated in clinical studies. Patients who survive an episode of AKI are also at increased risk for long-term adverse health outcomes such as the development of cardiovascular disease or progression to chronic kidney disease or end-stage renal disease 3. In-hospital mortality among patients with AKI is high, with as much as a sevenfold increased mortality risk compared with patients without AKI 1. AKI complicates the course of treatment and worsens outcomes of significant numbers of hospitalized-patients, i.e., 10–15% of all inpatients 1 and more than 50% of the critical care patients 2. Acute kidney injury (AKI) is a potentially life-threatening clinical syndrome for which there is no effective treatment other than supportive care and dialysis.

Mmd ai bases male