An Approach for Personalized Dynamic Assessment of Chronic Kidney Disease Progression Using Joint Model

Abstract

Chronic kidney disease (CKD) poses significant challenges to public health and healthcare systems, demanding a comprehensive understanding of its progressive nature. Prior methods have often fallen short in capturing the dynamic and individual variability of renal function. This study aims to address this gap by introducing a novel approach for the individualized assessment of CKD progression. A cohort of 1042 patients, comprising 700 with stage 3a and 342 with stage 3b to stage 5 CKD, treated at a veteran general hospital in Taiwan from 2006 to 2019, was included in the study. A comprehensive dataset spanning 12 years, consisting of clinical measurements, was collected and analyzed using joint models to predict the progression to hemodialysis treatment. The study reveals that the estimated glomerular filtration rate (eGFR) can be considered an endogenous factor influenced by innate biochemical markers. Serum creatinine, blood pressure, and urinary protein excretion emerged as valuable factors for predicting CKD progression. The joint model, combining longitudinal and survival analyses, demonstrated predictive versatility across various CKD severities. This innovative approach enhances conventional models by concurrently incorporating both longitudinal and survival analyses and provides a nuanced understanding of the variables influencing renal function in CKD patients. This personalized model enables a more precise assessment of renal failure risk, tailored to each patient’s unique clinical profile. The findings contribute to improving the management of CKD patients and provide a foundation for personalized healthcare interventions in the context of renal diseases.