中国血液净化

中国血液净化 ›› 2026, Vol. 25 ›› Issue (03): 202-206.doi: 10.3969/j.issn.1671-4091.2026.03.006

• 临床研究 • 上一篇    下一篇

基于机器学习的维持性血液透析患者抑郁症状及影响因素研究

李芳河    叶淑施    周祖木   

  1. 325000 温州，温州医科大学附属第一医院1神经内科 2产科
    325007 温州，3温州医科大学附属康宁医院 浙江省精神心理疾病临床医学研究中心
  • 收稿日期:2025-08-05 修回日期:2025-11-18 出版日期:2026-03-12 发布日期:2026-03-12
  • 通讯作者: 周祖木 E-mail:zhouzumu@126.com
  • 基金资助:
    温州市基础性科研项目课题(Y20240642)

A machine learning-based study on depressive symptoms and influencing factors in patients undergoing maintenance hemodialysis

LI Fang-he, YE Shu-shi, ZHOU Zu-mu   

  1. Department of Neurology, 2Department of Obstetrics, The First Affiliated Hospital of Wenzhou Medical University, Wenzhou 325000, China; 3The Clinical Medical Research Center for Mental and Psychological Diseases of Zhejiang Province, The Affiliated Kangning Hospital of Wenzhou Medical University, Wenzhou 325007, China
  • Received:2025-08-05 Revised:2025-11-18 Online:2026-03-12 Published:2026-03-12
  • Contact: 325007 温州，3温州医科大学附属康宁医院 浙江省精神心理疾病临床医学研究中心 E-mail:zhouzumu@126.com

摘要： 目的 本研究旨在探讨维持性血液透析(maintenance hemodialysis,MHD)患者并发抑郁症状及其影响因素。 方法 研究采用多中心设计，共纳入397例MHD患者，收集患者社会人口学资料、实验室检测指标及患者健康问卷-9(patient health questionnaire-9,PHQ-9)评分等，并采用单因素分析、Logistic回归和卡方自动交互检测器(Chi-square automatic interaction detector，CHAID)决策树分析方法确定影响因素。 结果 MHD患者抑郁症状的罹患率为21.91%。Logistic回归分析显示:失眠(OR=2.583，95%CI:1.472～4.533，P＜0.001)、糖尿病(OR=1.796，95%CI:1.026～3.144，P=0.040)和乙型肝炎(OR=3.996，95%CI:1.786～8.942,P＜0.001)是抑郁症状的独立影响因素。CHAID决策树分析显示失眠(χ2=20.046,P＜0.001)、糖尿病(χ2=5.543,P=0.019)、低蛋白血症(χ2 =19.198,P＜0.001)和β2-微球蛋白(β2-microglobulin‌，β2-MG)(χ2=10.969,P=0.008)是影响抑郁症状的主要因素。 结论 失眠、糖尿病、乙型肝炎、低蛋白血症和β2-MG与MHD患者的抑郁症状密切相关。本研究可为早期识别和干预MHD患者的抑郁症状提供科学依据。

关键词: 机器学习, 血液透析, 抑郁, 危险因素, Logistic回归, CHAID决策树

Abstract: Objective This study aims to investigate the prevalence of depressive symptoms and their influencing factors in patients undergoing maintenance hemodialysis (MHD). Methods A multicenter design was adopted, and a total of 397 MHD patients were included. Data on sociodemographic characteristics, laboratory test indicators, and scores from the Patient Health Questionnaire-9 (PHQ-9) were collected. Univariate analysis, logistic regression, and Chi-squared Automatic Interaction Detector (CHAID) decision tree analysis were used to determine the influencing factors.  Results The prevalence of depressive symptoms in MHD patients was 21.91%. Logistic regression analysis indicated that insomnia (OR=2.583，95%CI: 1.472～4.533，P＜0.001), diabetes (OR=1.796，95%CI: 1.026～3.144，P = 0.040), and hepatitis B (OR=3.996，95%CI: 1.786～8.942, P＜0.001) were independent influencing factors of depressive symptoms in MHD patients. CHAID decision tree analysis showed that insomnia (χ2=20.046, P＜0.001), diabetes (χ2=5.543, P=0.019), hypoproteinemia (χ2=19.198, P＜0.001), and β2-microglobulin (β2-MG) (χ2 =10.969, P=0.008) were the main factors affecting depressive symptoms.  Conclusion  Insomnia, diabetes, hepatitis B, hypoalbuminemia, and β2-MG are closely associated with depressive symptoms in MHD patients. This study provides a scientific basis for the early identification and intervention of depressive symptoms in MHD patients.

Key words: Machine learning, Hemodialysis, Depression, Risk factors, Logistic regression, CHAID decision tree

中图分类号: