Emergency Caesarean Section In Pregnant Women With Gestational Diabetes Construction And Verification Of Prediction Model

Objective:First,To investigate the influencing factors of emergency caesarean section in pregnant women with Gestational diabetes mellitus(GDM).Second,a risk prediction model for emergency cesarean section in pregnant women with GDM was constructed based on the results of multivariate analysis,and the predictive ability of the model was tested.Third,develop a dynamic line graph calculator convenient for clinical application,and evaluate the extensibility and clinical applicability of the model through external verification.Methods:Using convenient sampling method,this study retrospectively selected 1978 pregnant women with GDM who underwent vaginal trial labor in a maternal and child health hospital in Chengcheng City from May 2017 to October 2021 as research objects.General information of all pregnant women and fetuses,prenatal and perinatal related factors were collected for statistical description and construction of data set.The whole dataset was randomly divided into modeling group(n=1384)and internal validation group(n=594)at a ratio of 7∶3.In addition,295 pregnant women with GDM who underwent vaginal trial delivery in the same hospital from October 2021 to May 2022 were prospectively selected as the external validation group of the model.Through univariate,Lasso regression and multivariate logistic regression analysis,the influencing factors were screened,and P<0.05 was used to construct the best clinical prediction model of emergency caesarean section in GDM pregnant women.Receiver Operating Characteristic(ROC)curve was used to evaluate the discrimination of the model.Hosmer-Lemeshow goodness of fit test and calibration chart were used to evaluate the calibration of the model.A Clinical Decision curve(DCA)was constructed to evaluate the clinical application value of the model and quantify the net benefit within the threshold probability range.To evaluate the clinical application value of the model and quantify the net benefit within the threshold probability range.The external validation method was used to evaluate the generalization and clinical applicability of the model.To develop a dynamic nomogram for predicting the risk of emergency cesarean section in pregnant women with GDM based on the model construction and internal validation results,so as to provide a more convenient and efficient assessment tool for clinical practice.Results:First,logistic multivariate regression analysis showed that parity(OR=0.205),suspicious fetal status(OR=11.342),hypertensive disorder complicating pregnancy(OR=3.234),ICP(OR=2.598),Bishop score(4-6,OR=3.234;7-8,OR=0.357;≥9 points,OR=0.110),induction mode(oxytocin and/OR cervical balloon,OR=2.215;Xinproesen,OR=1.873;Others,OR=5.204),chorioamnionitis(OR=4.835),characteristics of amniotic fluid(OR=2.011),height(OR=0.948),pre-pregnancy BMI(OR=1.099),1h blood glucose(OR=1.128),fetal head circumference(OR=1.243),and other factors.These 12 independent variables are independent influencing factors for the occurrence of emergency cesarean section in pregnant women with GDM.Second,the AUC of the clinical prediction model in the modeling group was 0.872(95%CI:0.849,0.895),the sensitivity was 0.745,the specificity was 0.843,the positive predictive value was 0.503,the negative predictive value was 0.939,and the prediction accuracy was 0.826.In the internal validation group,the AUC was 0.868(95%CI:0.829,0.908),the sensitivity was 0.768,the specificity was 0.836,the positive predictive value was 0.471,the negative predictive value was 0.950,and the prediction accuracy was0.825.In the external validation group,the AUC was 0.884(95%CI:0.839,0.928),the sensitivity was 0.800,the specificity was 0.838,the positive predictive value was 0.530,the negative predictive value was 0.948,and the prediction accuracy was 0.831,indicating that the model had good discrimination.Thirdly,The results of the prediction model show that the actual curve and the standard curve are consistent with each other in the calibration diagrams of the modeling group and the internal and external verification group,and the degree of fit is high.At the same time,H-L goodness of fit test also verified the conclusions of the calibration chart,H-L test results of the three groups were X~2=9.762,p=0.282;X~2=5.778,p=0.672;X~2=5.145,p=0.742.In the clinical decision analysis curves of the three groups,the probability of curve risk threshold ranges from 0.05 to 0.7;0.05 to 0.82;0 to 0.79 higher than the two extremes,suggesting that the model has some clinical validity.These results suggest that intervention in this risk range may reduce the risk of emergency caesarean section in women with GDM.Conclusion:First,through case review analysis,this study found that the emergency cesarean section of GDM pregnant women is affected by many risk factors,including parity,suspicious fetal status,hypertensive disorders of pregnancy,ICP,Bishop score,mode of labor induction,chorioamnionitis,amniotic fluid characteristics,height,pre-pregnancy BMI,1h blood glucose,fetal head circumference,and maternal age.These 12 predictors are independent influencing factors of emergency cesarean section in pregnant women with GDM.Second,through the analysis of internal and external validation results,this study shows that the established prediction model has excellent performance in discrimination and calibration,and can accurately predict the risk of emergency cesarean section in pregnant women with GDM,which can provide important reference value for clinical practice.Thirdly,the dynamic nomogram calculator developed in this study may provide a more intuitive and convenient scientific tool for clinicians and pregnant women to make early delivery decisions.