颅内破裂动脉瘤术前再出血风险预测模型研究

目的:探讨颅内破裂动脉瘤术前再出血的风险概率预测模型。方法:纳入颅内破裂动脉瘤患者465例,采用多因素非条件Logistic回归法来构建概率预测模型,使用ROC曲线法来评价预测模型效果;对概率预测模型进行封装,设计Excel应用小软件。结果:构建了多因素Logistic回归概率预测模型,该模型有11个独立危险因素;ROC曲线下面积为0.907,明显大于机会线下的面积0.5(P0.01)。结论:所构建的Logistic回归概率预测模型预测颅内破裂动脉瘤术前再出血较好;Excel应用软件方便医护人员使用。

Predictive Model Study on Risk Factors of Rebleeding of Intracranial Rupture Aneurysm Patients before Operation

Objective:To explore predictive model study on risk probability of rebleeding of intracranial rupture aneurysm patients before operation.Methods:465 patients with intracranial rupture aneurysm were recruited.With multiple factors non-conditional Logistic Regression method building up predictive model of probability,the effect predictive model was evaluated by ROC curve method.And the model was packaged,then small software of Excel was designed according to the predictive model of probability.Results:There were eleven independent risk factors in the predictive model of probability with multiple factors Logistic Regression method.The area of ROC curve was 0.907,much higher than opportunity line 0.5 (P＜0.01).Conclusion:There is a good effect for the predictive model of probability with multiple factors Logistic Regression method.It could fairly predict the probability of rehaemorrhagia of intracranial rupture aneurysm patients before operation.And the small Excel software is convenient for medical staffto use.