首页|基于Logistic和决策树模型探讨老年肿瘤化疗患者居家期间PICC相关性血栓形成的影响因素

基于Logistic和决策树模型探讨老年肿瘤化疗患者居家期间PICC相关性血栓形成的影响因素

扫码查看
目的:基于Logistic和决策树模型探讨老年肿瘤化疗患者居家期间经外周静脉置入中心静脉导管(PICC)相关性血栓形成的影响因素。方法:根据置管1个月内是否发生PICC相关性血栓将533例老年肿瘤化疗患者分为发生组(n=53)和未发生组(n=480)。采用多因素Logistic和决策树模型分析影响因素。受试者工作特征(ROC)曲线分析预测价值。结果:高BMI、有高血压、有既往血栓史、每日活动时间<1 h、未掌握PICC置管维护知识为PICC相关性血栓形成的独立危险因素,使用抗凝药物则为独立保护因素(P<0。05)。有既往血栓史、每日活动<1 h的PICC相关性血栓形成的比例高于无既往血栓史、每日活动≥1 h的比例。在无既往血栓史节点中,有高血压的PICC相关性血栓形成的比例高于无高血压者;在有高血压的患者中,BMI>26。5 kg/m2的PICC相关性血栓形成的比例高于23。2 kg/m2<BMI≤26。5 kg/m2及BMI≤23。2 kg/m2者。ROC曲线分析结果显示,两种模型的AUC比较无明显差异。结论:高BMI、有既往血栓史、每日活动时间不足1 h是Logistic回归与决策树模型统计得出的非住院老年肿瘤化疗患者居家期间PICC相关性血栓形成的影响因素。
Explore the Influencing Factors of PICC-Related Thrombosis in Elderly Cancer Patients Undergoing Chemotherapy at Home Based on Logistic and Decision Tree Models
Objective:To explore the influencing factors of peripherally inserted central catheter(PICC)-related thrombosis in elderly cancer patients undergoing chemotherapy at home based on Logistic and decision tree models.Methods:533 elderly cancer patients were divided into the occurrence group(n=53)and the non-occurrence group(n=480)according to whether PICC-related thrombosis occurred 1 month within catheter.The influencing factors was analyzed by multivariate Logistic and decision tree models.The predictive value was analyzed by receiver operating characteristic(ROC)curve.Results:High BMI,hypertension,previous history of thrombosis,daily activity time<1 h,and failure to master PICC catheter maintenance knowledge were independent risk factors for PICC-related thrombosis,and the use of anticoagulant drugs was an independent protective factor(P<0.05).The proportion of PICC-related thrombosis with previous history of thrombosis and daily activity time<1 h were higher than those without previous history of thrombosis and daily activity time≥1 h.In the nodes without previous history of thrombosis,the proportion of PICC-related thrombosis in hypertension patients was higher than that in without hypertension patients.In hypertension patients,the proportion of PICC-related thrombosis in patients with BMI>26.5 kg/m2 was higher than that in patients with 23.2 kg/m2<BMI≤26.5 kg/m2 and BMI≤23.2 kg/m2.ROC curve analysis showed that,there was no significant difference in AUC between the two models.Conclusion:High BMI,previous history of thrombosis and daily activity time of less than 1 h are the influencing factors of PICC-related thrombosis in non-hospitalized elderly patients with tumor chemotherapy during home care obtained from Logistic regression and decision tree model statistics.Logistic regression and decision tree model can complement each other,and can replace the risk factors of PICC-related thrombosis,which is conducive to the accurate formulation of clinical prevention strategies.

ElderlyTumorChemotherapyAt homePeripherally inserted central catheterThrombosisDecision tree model

刘洋、李响、王冰洁、梁琪、张春梅

展开 >

天津中医药大学研究生院 天津 301617

天津市肿瘤医院放疗科 天津 301617

老年 肿瘤 化疗 居家 经外周静脉置入中心静脉导管 血栓 决策树模型

2024

现代生物医学进展
黑龙江省森工总医院 哈尔滨医科大学附属第四医院

现代生物医学进展

CSTPCD
影响因子:0.755
ISSN:1673-6273
年,卷(期):2024.24(21)