目的 构建子痫前期（preeclampsia，PE）患者阴道试产失败风险预测模型，以期为护理人员提供PE患者阴道试产失败风险筛查工具。 方法 采用便利抽样法，选取2018年2月至2023年4月在我院阴道试产的269例PE患者为研究对象。对纳入患者随机抽取70%作为建模集（188例），剩余30%作为验证集（81例）。在建模集中采用单因素分析和多因素Logistic回归分析PE患者阴道试产失败的影响因素并建立预测模型。模型效能采用受试者工作特征曲线（receiver operating characteristic，ROC）和拟合优度、决策曲线评估。 结果 年龄、试产前体质量指数、产前胎儿脐动脉血流收缩期/舒张期峰值流速的比值、医院焦虑抑郁评分是PE患者阴道试产失败的危险因素（均P＜0.05），Bishop评分为保护因素（P＜0.05）。构建预测模型的ROC曲线下面积为0.908，拟合优度检验的χ²=2.743，P=0.318。模型在验证集测试的ROC曲线下面积0.872，拟合优度仍良好（χ²=2.194，P=0.335）。决策曲线分析表明风险阈值＞0.38时，此预测模型可提供显著额外的临床净收益。 结论 构建的预测模型可作为预防PE患者阴道试产发生失败措施的参考依据，值得推广。

Objective To construct a risk prediction model for vaginal trial failure in patients with preeclampsia(PE),in order to provide nursing staff with a risk screening tool for vaginal trial failure in PE patients. Methods Adopting convenience sampling method,269 PE patients who underwent vaginal trials in our hospital from February 2018 to April 2023 were selected as the research subjects.70% of the included patients were randomly selected as the modeling set(188 cases),and the remaining 30% as the validation set(81 cases).Single factor analysis and multivariate logistic regression were used in the modeling set to analyze the influencing factors of vaginal trial failure in PE patients and establish a predictive model.The model effectiveness was evaluated using receiver operating characteristic(ROC) curves,goodness of fit,and decision curves. Results Age,preoperative body mass index,ratio of peak systolic/diastolic flow velocity of fetal umbilical artery blood flow,and hospital anxiety and depression score are risk factors for vaginal trial failure in PE patients(all P＜0.05).The Bishop score is a protective factor(P＜0.05).The area under the receiver operating characteristic of the prediction model is 0.908,goodness of fit test χ²=2.743,P=0.318.The area of the model under the receiver operating characteristic of the validation set test is 0.872,goodness of fit test χ²=2.743,P=0.318.When the decision curve analysis shows that the risk threshold is greater than 0.38,this predictive model can provide significant additional clinical net benefits. Conclusions The prediction model constructed can serve as a reference for effective measures to prevent vaginal trial failure in PE patients,and is worth promoting.

R473.71

南通市科技项目合同（MS22021004）