Development of a Longitudinal Diagnosis and Prognosis in Patients with Chronic Kidney Disease: Intelligent Clinical Decision-Making Scheme.

Chin-Chuan Shih,Gin-Den Chen,Chi-Chang Chang,Yu-Lin Shih,Ssu-Han Chen

doi:10.3390/ijerph182312807

Abstract

Previous studies on CKD patients have mostly been retrospective, cross-sectional studies. Few studies have assessed the longitudinal assessment of patients over an extended period. In consideration of the heterogeneity of CKD progression. It’s critical to develop a longitudinal diagnosis and prognosis for CKD patients. We proposed an auto Machine Learning (ML) scheme in this study. It consists of four main parts: classification pipeline, cross-validation (CV), Taguchi method and improve strategies. This study includes datasets from 50,174 patients, data were collected from 32 chain clinics and three special physical examination centers, between 2015 and 2019. The proposed auto-ML scheme can auto-select the level of each strategy to associate with a classifier which finally shows an acceptable testing accuracy of 86.17%, balanced accuracy of 84.08%, sensitivity of 90.90% and specificity of 77.26%, precision of 88.27%, and F1 score of 89.57%. In addition, the experimental results showed that age, creatinine, high blood pressure, smoking are important risk factors, and has been proven in previous studies. Our auto-ML scheme light on the possibility of evaluation for the effectiveness of one or a combination of those risk factors. This methodology may provide essential information and longitudinal change for personalized treatment in the future.

Highlights

The progression of chronic kidney disease (CKD) is multifactorial and complex, proper management of CKD to slow the progression of this condition is of considerable significance.According to the Global Burden of Disease (GBD) study 2017, CKD resulted in 1.2 million deaths and was the 12th leading cause of death worldwide [1]
The classification and regression tree (CART) is chosen as the classifier due to the flooded categorical and ordered variables in our dataset, unnecessary prior data distribution assumption, and the ability of the tree-based method to deal with missing data and perform a little bit well on imbalanced datasets compared to other methods
Data were collected from individual CKD case administration and care systems of 32 chain clinics and three special physical examination centers

Summary

Introduction

The progression of chronic kidney disease (CKD) is multifactorial and complex, proper management of CKD to slow the progression of this condition is of considerable significance. According to the Global Burden of Disease (GBD) study 2017, CKD resulted in 1.2 million deaths and was the 12th leading cause of death worldwide [1]. Ministry of Health and Welfare’s annual report, CKD accounts for the largest number of health insurance claims in 2018 [2]. In the 2019 annual report of the US Renal Registry. System (USRDS) [3], Taiwan has the highest prevalence and incidence of end-stage renal disease in the world [4]. In consideration of patterns of CKD progression, it is critical to conduct risk diagnosis and prognosis for CKD patients. CKD risk factors, such as hypertension, 4.0/)

Methods

Results

Conclusion