Modeling low birth weights using threshold regression: results for U.S. birth data

Area of Science:

• Biostatistics
• Perinatal Epidemiology
• Developmental Biology

Background:

• Low birth weight and preterm birth are associated with significant infant health risks, including mortality and developmental impairments.
• Existing models may not fully capture the complex interplay of factors influencing fetal development and birth outcomes.
• Understanding these factors is crucial for developing targeted interventions.

Purpose of the Study:

• To develop and present novel statistical models for birth weight prediction.
• To conceptualize birth weight as a first hitting time (FHT) in a Wiener process model of fetal development.
• To explore the association between maternal characteristics, birthing environment, and birth weight using threshold regression.

Main Methods:

• Utilized a Wiener process framework to model fetal development as a stochastic process.
• Applied threshold regression to link process parameters and boundary conditions to covariates.
• Developed two specific FHT models: a mixture model and a competing risks model.
• Tested models using a large-scale dataset of US live births from 2002.

Main Results:

• The study introduces a novel conceptual framework and two distinct FHT models for analyzing birth weight.
• Methodology demonstrated using a significant sample of US live births, highlighting potential for empirical application.
• Focus remains on the conceptual and methodological advancements, with full empirical results deferred.

Conclusions:

• The proposed Wiener process and threshold regression approach offers a new perspective on modeling birth weight.
• These models provide a framework for investigating the impact of maternal and environmental factors on fetal development and birth outcomes.
• Further empirical research is warranted to fully validate and apply these innovative methodologies.