Predicting preeclampsia and related risk factors using data mining approaches: A cross-sectional study

Area of Science:

• Medical Informatics
• Data Science in Healthcare
• Obstetrics and Gynecology

Background:

• Preeclampsia is a serious pregnancy hypertension disorder impacting maternal and fetal health.
• Current diagnostic methods for preeclampsia lack adequacy, necessitating improved screening tools.

Purpose of the Study:

• To develop a data mining-based model for preeclampsia screening.
• To identify key risk factors associated with preeclampsia development.

Main Methods:

• A cross-sectional study analyzed clinical records of 1452 mothers (726 with preeclampsia, 726 without).
• Six data mining techniques were evaluated: logistic regression, k-nearest neighborhood, C5.0 decision tree, discriminant analysis, random forest, and support vector machine.
• Model performance was assessed using accuracy, sensitivity, and specificity.

Main Results:

• Key risk factors for preeclampsia included underlying conditions, maternal age, pregnancy season, and parity.
• Support vector machine achieved the highest accuracy (0.791), followed by C5.0 decision tree (0.788) and random forest (0.758).

Conclusions:

• The support vector machine model demonstrated superior accuracy in predicting preeclampsia.
• This data mining approach offers a promising screening tool for diagnosing preeclampsia in clinical settings.