Establishment and Validation of a Logistic Regression Model for Premature Delivery in Twin Preg-nancies
[Objective]By analyzing the relevant data of twin pregnancy delivery patients,a logistic regression model is constructed to screen and predict high-risk factors for premature birth in twin pregnancy,and to reduce the risk of prema-ture birth in twin pregnancy.[Methods]A retrospective analysis was conducted on the clinical data of 215 twin pregnant women delivered in our hospital.They were divided into premature delivery group(102 cases)and full-term group(113 ca-ses)according to different delivery times.A single factor analysis was used to identify factors related to preterm birth,and factors with statistical differences were included in the logistic regression analysis.A logistic regression model for twin preg-nancy preterm birth was established,and its predictive performance was verified.[Results]Univariate analysis showed that there were significant differences between the two groups in body mass index(BMI),education level,pregnancy times,the incidence of pregnancy hypertension,pregnancy diabetes,the length of cervix in the second trimester and the incidence of premature rupture of membranes(P＜0.05).Logistic regression showed that high BMI,low pregnancy times,low educa-tion level,premature rupture of membranes,pregnancy hypertension and pregnancy diabetes,and low cervical length in the second trimester were independent risk factors for preterm twin pregnancy.The receiver operating characteristic(ROC)curve analysis showed that the sensitivity of the predictive model was 80.4％,the specificity was 92.0％,the area under the curve was 0.940,and the 95％CI was 0.910-0.970.[Conclusion]This study established a logistic regression model that can effectively predict the risk of premature birth in twin pregnancies.It can evaluate the pregnancy and perinatal conditions of twin pregnant women,identify high-risk factors for premature birth,and is of great significance for the prevention of premature birth risk in twin pregnancies.
万方数据
维普
