Development of a risk prediction model for dialysis disequilibrium syndrome in hemodialysis patients based on decision tree algorithm
- VernacularTitle:基于决策树的血液透析患者透析失衡综合征风险预测模型的构建
- Author:
Shanshan YAN
1
;
Ya WANG
;
Dongfang YAO
Author Information
- Publication Type:Research Article
- Keywords: hemodialysis; dialysis disequilibrium syndrome; risk factors; decision tree; chi-squared automatic interaction detection; prediction model
- From: Journal of Clinical Medicine in Practice 2024;28(18):51-55
- CountryChina
- Language:Chinese
-
Abstract:
Objective To establish a risk prediction model for dialysis disequilibrium syndrome in hemodialysis patients using the Chi-squared Automatic Interaction Detection (CHAID) decision tree algorithm. Methods A total of 200 hemodialysis patients were enrolled as study subjects. Patients who developed dialysis disequilibrium syndrome after hemodialysis were included in occurrence group, while those who did not develop dialysis disequilibrium syndrome were included in non-occurrence group. Clinical data were collected and analyzed, and univariate and multivariate Logistic regression analyses were performed to screen independent influencing factors of dialysis disequilibrium syndrome in hemodialysis patients. Based on these independent factors and the decision tree model, a risk prediction model for dialysis disequilibrium syndrome in hemodialysis patients was constructed. Results Among 200 hemodialysis patients, 40 developed dialysis disequilibrium syndrome, while 160 did not. Multivariate Logistic regression analysis revealed that age and urea nitrogen were independent risk factors for dialysis disequilibrium syndrome in hemodialysis patients (OR > 1,
P < 0.05), while albumin was an independent protective factor against DDS (OR < 1,P < 0.05). The decision tree model showed that age, located at the first level, was a most important influencing factor for dialysis disequilibrium syndrome (the sample was divided into three subgroups, with an incidence rate of 95.0% in patients older than 56.5 years, which was significantly higher than that in patients aged ≤56.5 years). Albumin and urea nitrogen were influencing factors for patients aged > 55.5 to 56.5 and > 56.5 years, respectively, located at the second level. Conclusion Analyzing the independent influencing factors of dialysis disequilibrium syndrome and risk prediction model constructed based on the decision tree algorithm can predict the probability of dialysis disequilibrium syndrome in hemodialysis patients.
