创新链/学科链/研发链/产业链

新药研发前沿动态 / 医药领域趋势进展

数据挖掘与Logistic回归评估依诺肝素干预时机对高风险脑外伤静脉血栓栓塞的预防效果

Data Mining and Logistic Regression to Evaluate the Effect of Enoxaparin Intervention Timing on Venous Thromboembolism Prophylaxis in High-Risk Traumatic Brain Injury

  • 摘要: 目的 探讨基于数据挖掘及Logistic回归模型评估依诺肝素干预时机对高风险脑外伤患者静脉血栓栓塞(venous thromboembolism,VTE)的预防效果。方法 回顾性分析2022年1月至2024年12月宁波市第六医院收治的560例高风险脑外伤患者。所有患者均接受依诺肝素干预治疗,依据干预时间将其分为A组(脑外伤后4~5 d干预,254例)与B组(脑外伤后6~7 d干预,306例),对比2组基线资料及临床结局。再依据是否发生VTE将其分为VTE组(121例)与无VTE组(439例),对比2组基线资料及检查结果。采用Modeler软件进行数据挖掘,并通过特征选择模块筛选重要指标;采用Logistic回归模型分析高风险脑外伤患者发生VTE的影响因素,并构建综合指数,采用受试者操作特征曲线分析综合指数对高风险脑外伤患者发生VTE的预测价值。结果2组其他部位出血比较差异无统计学意义(P> 0.05),A组住院时间显著长于B组(P<0.05);A组脑出血增加、神经功能恶化、再次手术及VTE发生率显著高于B组(P<0.05)。采用Modeler软件进行数据挖掘,指标重要性评估结果显示血小板计数(1.0000)、肢体肌力(1.0000)、中心静脉置管(1.0000)是影响患者VTE发生率的最关键因素。体重指数、脱水剂使用、卧床时间、肢体肌力、中心静脉置管、血小板计数及干预时机是发生VTE的影响因素(P<0.05)。多因素Logistic回归分析结果显示,肢体肌力≤3级、有中心静脉置管及血小板计数升高是患者发生VTE的独立危险因素(P<0.05)。综合指数预测发生VTE的曲线下面积为0.865(95%CI:0.825~0.904)。结论 过早依诺肝素干预可能会增加患者不良临床结局及VTE发生风险;肢体肌力≤3级、有中心静脉置管及血小板计数升高是患者发生VTE的独立危险因素,综合指数在预测VTE发生方面具有较高价值。

     

    Abstract: Objective To explore the preventive effect of enoxaparin intervention timing on venous thromboembolism(VTE) in patients with high-risk traumatic brain injury based on data mining and a Logistic regression model. Methods A retrospective analysis was conducted on 560 high-risk traumatic brain injury patients admitted to Ningbo Sixth Hospital from January 2022 to December 2024. All patients received enoxaparin intervention. According to the intervention time, they were divided into Group A(intervention at 4-5 days after traumatic brain injury, n= 254) and Group B(intervention at 6-7 days after traumatic brain injury, n=306). The baseline data and clinical outcomes of the two groups were compared. Based on whether VTE occurred, the patients were further divided into the VTE group(n= 121) and the non-VTE group(n= 439). The baseline data and examination findings of the two groups were compared. Modeler software was used for data mining, and the feature selection module was employed to screen important indicators. A Logistic regression model was adopted to analyze the influencing factors of VTE in high-risk traumatic brain injury patients, and a composite index was constructed. The predictive value of the composite index for VTE in high-risk traumatic brain injury patients was analyzed using the receiver operating characteristic curve. Results There was no statistically significant difference in bleeding at other sites between the two groups(P> 0.05). The length of hospital stay in Group A was significantly longer than that in Group B(P< 0.05). The incidences of increased intracranial hemorrhage, neurological deterioration, reoperation, and VTE in Group A were significantly higher than those in Group B(P< 0.05). Through data mining with Modeler software, the importance assessment of indicators showed that platelet count(1.0000), muscle strength(1.0000), and central venous catheterization(1.0000) were the most critical factors affecting the incidence of VTE in patients. Body mass index, use of dehydrating agent, duration of bed rest, muscle strength, central venous catheterization, platelet count, and intervention timing were influencing factors for VTE occurrence(P< 0.05). Multivariate Logistic regression analysis revealed that muscle strength ≤ grade 3, the presence of central venous catheterization, and elevated platelet count were independent risk factors for VTE in patients(P< 0.05). The area under the curve of the composite index for predicting VTE occurrence was 0.865(95% CI: 0.825-0.904). Conclusion Premature intervention with enoxaparin may increase the risk of adverse clinical outcomes and VTE in patients. Muscle strength ≤ grade 3, the presence of central venous catheterization, and elevated platelet count are independent risk factors for VTE in patients. The composite index has a high predictive value for VTE occurrence.

     

/

返回文章
返回