Home // eTELEMED 2024, The Sixteenth International Conference on eHealth, Telemedicine, and Social Medicine // View article
Authors:
Kangbeen Ko
Jiwon Ryu
Sejoong Kim
Keywords: chronic kidney disease; automatic classification; machine learning; explainable artificial intelligence.
Abstract:
Chronic Kidney Disease (CKD) represents a globally prevalent condition characterized by the gradual loss of renal function over time. The covert progression of CKD accentuates the necessity for regular and continuous inspection. Conventional diagnostic methods for CKD, including blood and urine analyses to estimate the Glomerular Filtration Rate (GFR) and to measure the urine Albumin-Creatinine Ratio (uACR), while effective, are invasive and often fail to facilitate early detection due to the asymptomatic progression of CKD in its initial stages. To tackle these limitations, we propose a novel, non-invasive diagnostic technique to enhance the early detection and management of CKD. This technique utilizes the patients' voice features, caused by respiratory muscle weakness and vocal chord swelling in patients with CKD, as an auxiliary indicator, leveraging machine learning algorithms to identify subtle changes in voice patterns that may correlate with CKD progression. Our method demonstrated a diagnostic accuracy of 0.86, quantified by the F1 score, and showed promising potential as a supplementary diagnostic tool. Implementing this technique paves the way for its integration into telemedicine platforms, offering a promising avenue for remote monitoring and managing CKD patients. This breakthrough advances our understanding and capability in the early diagnosis of CKD. It expands the potential for remote healthcare delivery, ensuring timely intervention and improving patient outcomes in managing kidney conditions.
Pages: 37 to 42
Copyright: Copyright (c) IARIA, 2024
Publication date: May 26, 2024
Published in: conference
ISSN: 2308-4359
ISBN: 978-1-68558-167-1
Location: Barcelona, Spain
Dates: from May 26, 2024 to May 30, 2024