中华护理杂志 ›› 2026, Vol. 61 ›› Issue (16): 2209-2216.DOI: 10.3761/j.issn.0254-1769.2026.16.007

• 专科护理实践与研究 • 上一篇    下一篇

肝病患者PICC相关性血栓发生风险预测模型的构建与验证

赵亚莉1(), 李玉华2, 吕源3, 李纳新4, 孙楠楠1, 鲁师宇5, 张莉莉6,*()   

  1. 1 首都医科大学附属北京佑安医院肿瘤内科 北京市 100069
    2 首都医科大学附属北京佑安医院感染综合科 北京市 100069
    3 首都医科大学附属北京佑安医院肝病消化中心 北京市 100069
    4 首都医科大学附属北京佑安医院李纳新感染性疾病静脉输液治疗护理工作室 北京市 100069
    5 首都医科大学附属北京佑安医院重症医学科 北京市 100069
    6 首都医科大学附属北京佑安医院护理部 北京市 100069
  • 收稿日期:2025-11-13 出版日期:2026-08-20 发布日期:2026-08-14
  • *通讯作者: 张莉莉,E-mail:zhanglilivip@163.com
  • 作者简介:赵亚莉:女,本科,副主任护师,护士长,E-mail:m15011467519@163.com
    第一联系人:

    赵亚莉:研究设计、数据收集、论文撰写及修改、基金支持;李玉华、吕源、李纳新、孙楠楠:数据收集及整理;鲁师宇:资料收集和整理、数据分析、论文审校;张莉莉:研究设计、论文审校

  • 基金资助:
    首都医科大学附属北京佑安医院2023年度院内中青年人才孵育项目(BJYAYY-YN2023-28)

Construction and validation of a risk prediction model for PICC-related thrombosis in patients with severe liver diseases

ZHAO Yali1(), LI Yuhua2, LÜ Yuan3, LI Naxin4, SUN Nannan1, LU Shiyu5, ZHANG Lili6,*()   

  1. 1 Department of Medical OncologyBeijing You’an Hospital,Capital Medical UniversityBeijing 100069, China
    2 Department of Infectious DiseasesBeijing You’an Hospital,Capital Medical UniversityBeijing 100069, China
    3 Liver Disease and Gastroenterology CenterBeijing You’an Hospital,Capital Medical UniversityBeijing 100069, China
    4 PICC Nursing StudioBeijing You’an Hospital,Capital Medical UniversityBeijing 100069, China
    5 Department of Critical Care MedicineBeijing You’an Hospital,Capital Medical UniversityBeijing 100069, China
    6 Department of NursingBeijing You’an Hospital,Capital Medical UniversityBeijing 100069, China
  • Received:2025-11-13 Online:2026-08-20 Published:2026-08-14
  • * Corresponding author: ZHANG Lili,E-mail:zhanglilivip@163.com
  • Funding program:
    Young and Middle-aged Talent Incubation Program of Beijing You’an Hospital,Capital Medical University(BJYAYY-YN2023-28)

摘要:

目的 构建并验证肝病患者PICC相关性血栓发生风险预测模型,为临床护士早期识别高风险患者、制订针对性的预防措施提供参考。 方法 将2021年8月—2025年9月在北京市某三级甲等医院住院并置入PICC的701例肝病患者作为调查对象,回顾性收集其相关资料,采用1∶2倾向性评分匹配法,基于患者的年龄、性别与疾病类型平衡混杂因素。通过单因素分析、Logistic回归分析筛选肝病患者PICC相关性血栓的影响因素并构建风险预测模型,通过受试者操作特征曲线、Bootstrap法和5折交叉验证法评估模型的区分度,采用Hosmer-Lemeshow检验和Brier分数评估模型的校准度。 结果 701例肝病患者中,57例(8.13%)发生PICC相关性血栓。经倾向性评分匹配后,血栓组纳入57例、非血栓组纳入114例。Logistic回归分析结果显示,天门冬氨酸氨基转移酶和凝血酶原时间是肝病患者PICC相关性血栓的影响因素。根据以上结果构建风险预测模型,其受试者操作特征曲线下面积为0.739、基于约登指数的最佳截断值为0.362、准确度为0.661、灵敏度为0.754、特异度为0.614、阳性预测值为0.491、阴性预测值为0.833。Hosmer-Lemeshow检验结果显示,χ2=6.628(P=0.357),Brier分数为0.188。 结论 该研究构建的风险预测模型具有较好的预测效能,可作为肝病患者PICC相关性血栓发生风险的早期筛查工具。

关键词: 经外周静脉穿刺中心静脉置管, 肝病, 静脉血栓, 危险因素, 预测模型, 护理

Abstract:

Objective To develop and validate a risk prediction model for peripherally inserted central catheter-related thrombosis(PICC-CRT) in patients with liver diseases,providing a clinical reference for nurses to identify high-risk individuals and implement early preventive measures. Methods A retrospective study was conducted on 701 patients with liver diseases who underwent PICC placement at a tertiary A hospital in Beijing between August 2021 and September 2025. Propensity score matching(PSM) was employed at a 1∶2 ratio to balance confounding factors,including age,sex,and disease type. Univariate analysis and logistic regression analysis were performed to identify independent risk factors for PICC-CRT and to construct the prediction model. The model’s discriminative ability was evaluated using the receiver operating characteristic(ROC) curve,Bootstrap resampling,and 5-fold cross-validation. Calibration was assessed using the Hosmer-Lemeshow(H-L) test and the Brier score. Results Among the 701 patients,57 cases(8.13%) developed PICC-CRT. After PSM,57 patients were included in the thrombosis group and 114 in the non-thrombosis group. Logistic regression analysis identified aspartate aminotransferase concentration and prothrombin time as independent predictors of PICC-CRT in liver disease patients. The area under the ROC curve for the constructed model was 0.739,with an optimal cut-off value of 0.362 based on the Youden index. The model demonstrated an accuracy of 0.661,sensitivity of 0.754,specificity of 0.614,a positive predictive value of 0.491,and a negative predictive value of 0.833. The H-L test yielded χ2=6.628(P=0.357),and the Brier score was 0.188. Conclusion The prediction model constructed in this study demonstrated acceptable predictive performance and may serve as a screening tool for early identification of PICC-CRT in patients with severe liver disease.

Key words: Peripherally Inserted Central Catheter, Liver Diseases, Venous Thrombus, Risk Factor, Predictive Model, Nursing Care