足月单胎妊娠孕妇引产失败预测模型的建立

Establishment of a Predictive Model for Induced Abortion Failure in Term Single Pregnancy Pregnant Women

目的:分析足月单胎妊娠孕妇引产失败的影响因素并建立引产失败预测模型。方法:回顾性分析2019年1月1日至2019年12月31日于北部战区总医院和平院区妇产科分娩的足月单胎妊娠引产1483例孕妇的临床资料,按引产结局分为引产成功组(1108例)和失败组(375例)。采用多因素Logistic回归分析筛选出引产失败的影响因素,并构建预测模型,应用受试者工作特征(ROC)曲线、Hosmer-Lemeshow检验评价该模型的预测性能与拟合优度。结果:多因素Logistic回归分析结果示高龄初产、无阴道分娩史、受教育程度≤12年、分娩孕周<40周、孕前超重、孕前肥胖、孕期增重超标、身高<160 cm、引产前子宫颈Bishop评分<4分、新生儿出生体质量≥3750 g、联合引产、可疑胎儿窘迫、引产至临产时间≥24小时是引产失败的独立危险因素(OR>1,P<0.05),身高≥165 cm,地诺前列酮栓引产是引产失败的保护性因素(OR<1,P<0.05)。分别利用产前因素、产前结合产时因素构建引产失败预测模型,ROC曲线下面积(AUC)分别为0.914与0.940,约登指数分别为0.660与0.733,预测准确率分别为87.5%与88.9%。结论:本研究筛选出引产失败的显著影响因素,一方面为临床制定提高引产成功率的措施提供理论依据,并构建引产失败预测模型,有助于产科医生与孕妇共同决定分娩方式,保障母婴安全;另一方面旨在提升大家孕期保健意识,提升阴道分娩率。

Objective:To analyze the influencing factors of failed induction of labor(IOL)in full-term singleton pregnant women,and to establish a prediction model for failed IOL.Methods:This study retrospectively analyzed the clinical data of 1483 pregnant women with full-term singleton of IOL in the Department of Obstetrics and Gynecology,Heping Branch of General Hospital of Northern War Zone from January 1,2019 to December 31,2019.According to the outcome of IOL,the pregnant women were divided into the successful group(1108 cases)and the failed group(375 cases)of IOL.The influencing factors of failed IOL were screened to establish the prediction model through multivariate Logistic regression analysis.The receiver operating characteristic(ROC)curves and Hosmer-Lemeshow test were used to assess the predictive performance and fitting degree of the model.Results:Multivariate Logistic regression analysis showed that there were risk factors for failed IOL(OR>1,P<0.05),including elderly primiparous delivery,with no history of vaginal delivery,education level≤12 years,gestational age<40 weeks,pre-pregnancy overweight or obesity,excessive gestational weight gain,height<160 cm,cervical Bishop score before IOL<4 points,neonatal weight≥3750 g,combined IOL,suspected fetal distress,and the time from IOL to onset of labor≥24 hours,and height≥165 cm,IOL with dinoprostol suppositories were protective factors for failed IOL(OR<1,P<0.05).Antepartum factors and antepartum factors combined with intrapartum factors were separately used to establish model for predicting failed IOL.The area under the ROC curve(AUC)were 0.914 and 0.940,and the Youden index were 0.660 and 0.733,respectively.The prediction accuracy were 87.5%and 88.9%,respectively.Conclusions:This study screened the significant influencing factors of failed IOL,providing a theoretical basis for clinical measures to improve the success rate of IOL and constructing a prediction model of failed IOL,which is helpful for obstetricians and pregnant women to decide the mode of delivery together,and ensure the safety of mother and baby;on the other hand,it aims to enhance everyone′s awareness of pregnancy health care and improve the vaginal delivery rate.