The Log-Normal zero-inflated cure regression model for labor time in an African obstetric population

Area of Science:

• Obstetrics and Gynecology
• Biostatistics
• Survival Analysis

Background:

• Understanding factors influencing childbirth duration is crucial for clinical guidelines and hospital stay management.
• Survival models are essential for analyzing time-dependent variables in obstetrics.
• Standard models face challenges with zero-duration labor (fetal death) and censored data (interventions).

Purpose of the Study:

• To present the Log-Normal zero-inflated cure regression model for analyzing childbirth duration.
• To evaluate the performance of likelihood-based parameter estimation for this model using a simulation study.

Main Methods:

• Development and application of the Log-Normal zero-inflated cure regression model.
• Simulation study to assess parameter estimation accuracy.
• Analysis of the Better Outcomes in Labor Difficulty project dataset.

Main Results:

• The proposed model effectively handles zero-inflated and censored data in childbirth duration analysis.
• Inference procedures performed better with larger sample sizes and lower proportions of zero inflation and cure.
• Parity and educational level were identified as significant factors associated with childbirth outcomes in the dataset.

Conclusions:

• The Log-Normal zero-inflated cure regression model is a valuable tool for studying childbirth processes.
• The model provides insights into factors affecting labor duration and outcomes.
• Statistical advancements can improve the understanding of obstetric dynamics.