Prediction of all-cause mortality for chronic kidney disease patients using four models of machine learning

Abstract

Background The prediction tools developed from general population data to predict all-cause mortality are not adapted to chronic kidney disease (CKD) patients, because this population displays a higher mortality risk. This study aimed to create a clinical prediction tool with good predictive performance to predict the 2-year all-cause mortality of stage-4 or stage-5 CKD patients. Methods The performance of four different models (deep learning, random forest, Bayesian network, logistic regression) to create four prediction tools was compared using a 10 fold cross validation. The model that offered the best performance for predicting mortality in the Photo-Graphe 3 cohort was selected and then optimized using synthetic data and a selected number of explanatory variables. The performance of the optimized prediction tool to correctly predict the 2-year mortality of the patients included in the Photo-Graphe 3 database were then assessed. Results Prediction tools developed using the Bayesian network and logistic regression tended to have the best performances. Although not significantly different from logistic regression, the prediction tool developed using the Bayesian network was chosen because of its advantages and then optimized. The optimized prediction tool that was developed using synthetic data and the 7 variables with the best predictive value (age, erythropoietin-stimulating agent, cardiovascular history, smoking status, 25-OH vitamin D, parathyroid hormone, and ferritin levels) had satisfactory internal performance. Conclusions A Bayesian network was used to create a 7-variable prediction tool to predict the 2-year all-cause mortality in patients with stage 4–5 CKD. Prior to external validation, the proposed prediction tool can be used at: https://dev.hed.cc/?a=jpfauvel&n=2022-05%20Modele%20Bayesien%2020000%20Mortalite%207%20variables%20Naif%20Zou%20online(1).neta for research purposes.

Frailty, but not cognitive impairment, improves mortality risk prediction among those with chronic kidney disease—a nationally representative study

“…Nonetheless, there is face validity in this study's findings; it was expected that those with CKD would have a greater burden of frailty, cognitive impairment, and co-occurrence of the two conditions; given their higher prevalence of comorbidities, particularly in vascular disease. Further studies should aim to replicate these findings among those with and without CKD in other, diverse population studies with validated measures of frailty and cognitive function to better compare the magnitude of burden.This diverse, nationally-representative study further corroborates the potential vascular underpinnings of the frailty-cognition link demonstrated in prior studies [27][28][29][30]35]. Our findings support those prior studies in that the strongest associations between frailty and cognitive function were found for executive function and processing speed (Cohen's d = -0.44, 95%CI: -0.60, -0.28) compared to other memory-related domains.…”

“…Prediction models with traditional risk factors poorly predict mortality for patients with CKD. Models that only use factors such as demographics, lifestyle, and kidney function to predict mortality in patients with CKD have C-statistics ranging from 0.71 to 0.81 [28][29][30]. Therefore, efforts to assess the predictive ability of novel risk factors are needed to improve the accuracy of prediction survival models in CKD [9].…”

A Random Forest Algorithm for Assessing Risk Factors Associated With Chronic Kidney Disease: Observational Study

“…A previous study reported the performance of 4 prediction tools, namely deep learning, plain Bayesian, RF, and LR, for predicting all-cause mortality in patients with CKD. The study showed that Bayesian networks and LR showed superior prediction abilities [ 21 ]. However, another study reported that plain Bayesian, RF, and LR performed adequately well and showed high sensitivity for screening end-stage renal disease in patients with CKD, which is inconsistent with previous reports [ 22 ].…”

Vitamin D and Chronic Kidney Disease Association with Mineral and Bone Disorder: An Appraisal of Tangled Guidelines

“…Although controversies still exist [ 30 , 35 , 36 , 37 , 38 ], diagnosis and treatment of VD deficiency [with either nutritional (native) or active forms of VD] have not yet been excluded from nephrology practice and guidelines [ 22 , 31 , 39 , 40 ], and despite the poor quality of available evidence, much data may still support VD use in populations with VD deficiency or certain special characteristics, such as CKD patients [ 36 , 39 , 40 ]. Furthermore, when an optimized prediction tool was recently developed using machine-learning techniques, both parathyroid hormone (PTH) and calcidiol were found to be among the seven variables identified as having the best predictive value for 2-year all-cause mortality in patients with CKD G4-G5 [ 41 ].…”