RESUMO
OBJECTIVE: This study aimed to develop a new tool for personalised preterm birth risk evaluation in high-risk population. STUDY DESIGN: 813 high-risk asymptomatic pregnant women included in a French multicentric prospective study were analysed. Clinical and paraclinical variables, including screening for bacterial vaginosis with molecular biology, cervical length, have been used to create the nomogram, based on the logistic regression model. The validity was checked by bootstrap. A downloadable calculator was build. RESULTS: Nine risk factors were included in this model: history of late miscarriage and/or preterm delivery, active smoking, ultrasound cervical length, term of pregnancy at screening, bacterial vaginosis, premature rupture of membranes, daily travel more than 30min. Discrimination and calibration of the nomogram revealed good predictive abilities. The area under the receiver operating characteristic curve was 0.77 (95% CI; 0.72-0.81). The mean absolute error was 0.018, which showed proper calibration. The optimal risk threshold was 23.2% with a sensitivity of 74%, a specificity of 72.7% and a predictive negative value of 90.6%. CONCLUSION: The nomogram can help to better define individual preterm birth risk in high-risk pregnancies.
