The association of vitamin-D level with catheter-related-thrombosis in hemodialysis patients: A data mining model.

Abstract

This study aims to investigate the association of different risk factors including vitamin-D level with catheter-related-thrombosis in hemodialysis patients by applying data mining techniques. This study used the retrospectively approach and was done based on the CRISP-DM framework. The data of 1048 hemodialysis patients of Hasheminejad Kidney Center whose first catheterization was between 2014 and 2019 was used for analysis. In this study, patients with a previous history of deep venous thrombosis, thrombophilic condition, and undergone anticoagulant therapy were excluded. The decision tree J48 in WEKA software was used for modeling. The K-fold cross-validation method was also used to evaluate the classification performance. Finally, the model was evaluated by using Confusion Matrix and F-Measure. The approved model was accurate in 70.3% of the predictions, and it demonstrated an association between patients' vitamin-D range and catheter-related-thrombosis. The results indicated that in hemodialysis patients without the history of cancer or high blood pressure, vitamin-D had a significant impact on catheter related thrombosis.