Labor estimation by informational objective assessment (LEIOA) for preterm delivery prediction

Area of Science:

• Obstetrics and Gynecology
• Computational Medicine
• Biomedical Engineering

Background:

• Premature delivery poses significant risks to newborns and healthcare systems.
• Accurate prediction of imminent preterm birth is crucial for timely intervention and improved outcomes.
• Current clinical tools have limitations in precisely forecasting delivery timelines.

Purpose of the Study:

• Introduce LEIOA, a novel screening method for predicting delivery within 7 days for patients with suspected preterm labor.
• Enhance the prognostic and therapeutic capabilities for threatened premature delivery.
• Provide a tool to assist clinicians in managing high-risk pregnancies.

Main Methods:

• Developed the LEIOA model using tocography data from 286 women (24-37 weeks gestation).
• LEIOA combines Generalized Hurst Exponent and Approximate Entropy of uterine contractions via logistic regression.
• Compared LEIOA against a purely obstetric model and a combined model, followed by cross-validation.

Main Results:

• The combined LEIOA and medical model improved predictive values by an average of 12% over the medical model alone.
• Achieved high performance metrics: sensitivity 0.937, specificity 0.747, positive predictive value 0.907, negative predictive value 0.819.
• Incorporating LEIOA reduced misclassified cases by nearly 50% compared to the medical model alone.

Conclusions:

• LEIOA significantly enhances predictive accuracy for imminent preterm delivery when integrated with clinical variables.
• The LEIOA model offers a valuable tool for improving the estimation of delivery immediacy in threatened preterm labor.
• This method holds potential for optimizing management strategies and improving neonatal outcomes.