Prediction and management model of preterm birth

Abstract

Background: It seems relevant to study the contribution of socio-demographic, somatic and obstetric-gynecological factors in the implementation of preterm birth.
 Aims: Assessment of the prognostic significance of socio-demographic, obstetric-gynecological and somatic factors in the prediction of preterm birth and associated adverse pregnancy outcomes with subsequent validation of the prognostic model.
 Materials and methods: Cohort study with a mixed cohort. A retrospective assessment of socio-demographic factors, harmful habits, obstetric and gynecological pathology, somatic diseases, course and outcomes of pregnancy was carried out with the assessment of the status of newborns in 1246 women with subsequent construction of a predictive model of preterm birth and adverse outcomes of pregnancy using Regression with Optimal Scaling and its prospective validation in 100 women. Results: The most significant predictors, that increase the chance of preterm birth and adverse pregnancy outcomes, were history of premature birth, irregular monitoring during pregnancy, history of pelvic inflammatory disease, smoking, obesity, the onset of sexual activity up to 16 years, cardiovascular and endocrine diseases. Intellectual job reduced the chance of preterm birth and adverse pregnancy outcomes This multivariate predictive model has a diagnostic value. The score of risk factors 25 points had a sensitivity of 73%, a specificity of 71%, the area under the ROC curve (AUC) 0.76 (good quality), p0.001. After stratification of high-risk groups by maternal and perinatal pathology the following list of diagnostic and therapeutic measures is introduced and actively implemented in antenatal clinics. To stratificate this model, we prospectively analyze the course and pregnancy outcomes of 100 women divided into 2 groups: group 1 ― 50 women with preterm delivery, group 2 ― 50 women with term delivery. A total score of 25 and above had 44% of women in group 1 and only 10% of women in group 2 (sensitivity 81.4%, specificity 61.6%, positive predictive value 44%, negative predictive value 90%, positive likelihood ratio 2.2 [1.53.0], negative likelihood ratio 0.3 [0.130.68]).
 Conclusions: We have proposed a model for predicting preterm birth and delivery and perinatal losses using the available characteristics of pregnant women from early pregnancy with moderate indicators of diagnostic value. Further validation of the model in the general population of pregnant women is required.