Risk Assessment of Acute Kidney Disease and Chronic Kidney Disease for In-Hospital Patients with Acute Kidney Injury

Abstract

Acute kidney injury (AKI) is a common and important complication in hospitalized patients. They may progress into acute kidney disease (AKD) or chronic kidney disease (CKD), and predispose to dialysis or death. These complications have become great burden on the society and National Health Insurance, Taiwan. Therefore, how to predict the risk for the kidney disease has been a hot topic over the past few decades. To tackle this issue, in this paper, we propose risk assessment methods that integrate techniques of data engineering and data mining to achieve high-fidelity prediction for AKD and CKD. Based on the real data from Kaohsiung Chang Gung Memorial Hospital, the factors are retrieved first. Next, the mining techniques of K-Nearest-Neighbors (KNN) and Support-Vector-Machine (SVM) are performed to classify the potential kidney disease. The experimental results reveal that, the precisions can reach around 75%. In the future, more precise models will be developed. Accordingly, the prediction models will be established and integrate into clinical practice to facilitate decision-making process for medical professionals.