.png){kind=icon}
.png){kind=icon}
Return
Bayesian Logistic Regression to Explore the Role of Complete Blood Count in Kidney Disease Mortality
.png){kind=icon}
PRE
AI
I
J
Z
A
A
L
DOI：.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
Kidney disease is a major global health challenge, ranking fifth in Malaysia and ninth worldwide as a leading cause of death in 2021. This growing burden highlights the need for cost-effective tools to support early identification of patients at risk of mortality. The complete blood count (CBC) is an affordable, widely used diagnostic test, while Bayesian methods offer advantages for incorporating prior knowledge and quantifying uncertainty. However, the use of CBC parameters with Bayesian approaches for mortality prediction among kidney disease patients in Malaysia remains limited. This study aimed to develop a risk stratification model for kidney disease mortality using CBC data and Bayesian logistic regression (BLR). A retrospective study was conducted using data from 5,158 patients with kidney disease treated at Queen Elizabeth I Hospital. The final multivariate BLR model identified 13 significant predictors of mortality. The strongest predictors were low haemoglobin, high mean platelet volume (MPV), and high neutrophil-to-lymphocyte ratio (NLR), followed by high white blood cells (WBC), and hospitalisation history. The model demonstrated good calibration and discrimination, with an area under the receiver operating characteristic curve (AUROC) and area under the precision-recall curve (AUPRC) greater than 0.8, supporting its reliability for mortality risk stratification. These findings suggest that combining CBC parameters with demographic information may improve early detection and clinical decision-making, particularly in resource-limited settings.
Keywords:
Bayesian logistic regression
complete blood count
kidney disease
mortality
risk stratification
Journal
IF:
0.8
Papers:
113
Citations:
2.9K
Organization
Cited Papers