Proactive Diagnosis of Chronic Kidney Disease Using Machine Learning

Abstract

Chronic Kidney Disease (CKD) is a progressive disease that poses a major global health burden, requiring early detection and intervention to avoid serious complications, such as kidney failure. This study presents a machine learning-based CKD prediction model that incorporates clinical biomarkers like Specific Gravity, Hemoglobin, Albumin, Red Blood Cell Count, and Creatinine, as well as demographic variables like age, gender, and ethnicity. The Random Forest model is applied to classify the patients into CKD and non-CKD groups, delivering a stable and interpretable model that can handle intricate medical data. Moreover, the Glomerular Filtration Rate (GFR) is computed using the Modification of Diet in Renal Disease (MDRD) formula, allowing for accurate measurement of CKD severity and staging. A web application developed using Flask is employed to give real-time predictions based on which users can input health parameters and immediately receive results regarding their CKD status. Apart from this, an SQLite database is implemented to store health predictions, allowing for long-term tracking of patients and trend analysis. Integrating machine learning with clinical risk assessment serves as an effective decision-making tool for healthcare professionals. It allows for timely diagnosis of CKD so that interventions having a positive impact on patient outcomes are possible. The model is extremely effective in CKD classification and can serve as an effective tool in proactive kidney care management and personalized risk assessment.