Enhancing Chronic Kidney Disease Diagnosis Through Deep Learning-Based Predictive Analytics

Abstract

Kidney disease stands as a pressing contemporary health issue, impacting millions worldwide. Glomerular filtration rate (GFR), a pivotal indicator of kidney function, exhibits a significant positive correlation with blood metabolite creatinine levels. Given the challenges in directly measuring GFR, the presence of Chronic Kidney Disease (CKD) is initially gauged through creatinine levels. Despite its diagnostic utility, creatinine testing remains absent from routine health check-ups in many countries due to cost constraints associated with comprehensive examinations. In response to this gap, this study proposes the inclusion of creatinine testing in routine fitness examinations to facilitate early CKD detection. Leveraging classifier models, our suggested approach demonstrates superior performance compared to alternative techniques, achieving an impressive accuracy rate of 98.5%. By integrating creatinine testing into routine check-ups, practitioners gain access to definitive and clear data, leading to improved diagnostic outcomes and enhanced interpretative abilities. Furthermore, this project employs the Flask framework to develop a predictive web application, ensuring accessibility and scalability of the proposed CKD detection method. Through the amalgamation of advanced analytical techniques and user-friendly technology, this initiative endeavors to streamline CKD diagnosis, ultimately contributing to improved public health outcomes.