PREDICTION OF CHRONIC KIDNEY DISEASE RISK USING AI TECHNIQUES CONSIDERING VARIOUS ENVIRONMENTAL AND GENETIC FACTORS
Akshitha D. Gopikrishnan and Dr. S. Prabhakar Karthikeyan
Keywords: CKD, Machine Learning Algorithms, Risk Prediction Models, XGBoost Classifier, Feature Importance Analysis, SHAP Values, Genetic and Environmental Factors, Early Detection and Prevention, Personalized Risk Assessment, Healthcare Decision Support Systems
Abstract
Chronic Kidney Disease (CKD) is a global crisis affecting 10-15% of the world population and affects a few million people worldwide. As CKD expected to rise as a major health issue, it heavily burdens the strained healthcare systems, particularly in low- and middle-income countries. The economic impact associated with the treatments and long-term care, will further challenge global health systems. Despite the ever-increasing prevalence, a considerable proportion of affected individuals remains undiagnosed; this leads to premature mortality. Undetected CKD progresses through stages, ultimately leading to kidney failure, which requires dialysis or a transplant as an ultimate care. When this get linked to cardiovascular diseases it significantly impacts patient's quality of life, leading into long-term disability. The early detection and personalized management of CKD plays an important role in preventing disease progression and improving patient prognosis. This research aims to address a critical gap by developing a robust algorithm that calculates an individual’s risk level of CKD incorporating both genetic predispositions and environmental factors. By leveraging machine learning techniques and with extensive data analysis, the proposed tool will enable healthcare professionals worldwide to deduct the high-risk patients at an early stage. This research resulted in the development of a detailed, yet userfriendly, risk assessment tool that facilitates early detection and prevention of kidney disease. Enabling healthcare professionals by providing with a tool that can be reliable and accessible method for identifying the risk based on several influencing factors ultimately improves disease management and patient outcomes. Implementation of such tools within health systems will translate into a significant impact on the reduction of the burden of CKD worldwide.