Multivariate Logistic Regression Analysis Predicted Preterm Premature Rupture Of Membrane

Objective: To discuss the feasibility and clinical value of multivariate logistic regression analysis predicting premature rupture of membranes(PPROM).Methods:Collect 237 patients with premature rupture of membranes(observation group) and 237 normal full-term pregnancy patients(control group),between January 2016 and February 2016 and in the second affiliated hospital of chongqing medical university obstetric inpatient,and their age,number of pregnancies,multiple pregnancies,late pregnancy blood routine examination(Hb,LMR,NLR,PLR),gestational diabetes mellitus,abnormal fetal position,abnormal placental position,excessive amniotic fluid.Univariate analysis was performed on these clinical data.Multivariate Logistic regression analysis was conducted for risk factors with statistical differences.Results the independent risk factors and ratio ratio of premature rupture of membranes were obtained.Evaluate the effectiveness of Logistic regression model.Results: Single factor analysis results showed that age,the times of pregnancy and bearing birth,gestational diabetes mellitus,fetal number,abnormal fetal position,hemoglobin,platelet count and lymphocyte count ratio(PLR),lymphocyte count and mononuclear cell count ratio(LMR),neutrophil count and lymphocyte count ratio(NLR) between two groups are different,the difference is statistically significant(P<0.05).Multivariate analysis showed that age,the times of pregnancy,hemoglobin,PLR and double fetus were independent risk factors for PPROM(P<0.05).Accuracy,sensitivity,specificity,positive predictive value and negative predictive value of PPROM multifactor regression model was 80.38%(381/474),78.90%(187/237),81.86%(194/237),81.30%(187/230),79.51%(194/244).Conclusions: This model has certain clinical value in predicting PPROM.