Clinical and dental predictors of preterm birth using machine learning methods: the MOHEPI study

Area of Science:

• Obstetrics and Gynecology
• Periodontology
• Data Science in Healthcare

Background:

• Preterm birth (PTB) is a leading cause of infant mortality and lifelong health issues.
• Maternal health conditions and dental status, particularly periodontitis, are implicated in PTB.
• Predictive modeling can identify key risk factors for PTB.

Purpose of the Study:

• To identify significant clinical and dental predictors of preterm birth (PTB) using machine learning.
• To analyze the directionality of associations between identified predictors and PTB/SPTB.

Main Methods:

• Prospective cohort study of 60 women delivering singleton births.
• Machine learning analysis, including Random Forest (RF) variable importance and Shapley Additive Explanation (SHAP) values.
• Inclusion of 15 independent variables (10 clinical, 5 dental).

Main Results:

• Top PTB predictors identified by RF included pre-pregnancy BMI, modified gingival index (MGI), preeclampsia, DMFT index, and maternal age.
• SHAP values indicated positive correlations between PTB/SPTB and factors like premature rupture of membranes, pre-pregnancy BMI, maternal age, and MGI.

Conclusions:

• Pre-pregnancy BMI, maternal age, and dental health (MGI) are significant predictors of PTB.
• Positive correlations highlight the importance of integrating medical and dental care during pregnancy.
• Further research is needed to validate predictors and develop interventions for PTB risk reduction.