Analysis of infant growth curves using multivariate adaptive splines

Area of Science:

• Perinatal health
• Biostatistics
• Infant development

Background:

• Maternal substance use, specifically cocaine, is a public health concern impacting infant outcomes.
• Infant growth trajectories following prenatal cocaine exposure require robust statistical modeling.
• Longitudinal data from irregular schedules present unique analytical challenges.

Purpose of the Study:

• To investigate the effect of maternal cocaine use on infant growth, particularly body weight.
• To apply and enhance the Multivariate Adaptive Splines for the Analysis of Longitudinal data (MASAL) method.
• To explore and estimate covariance structures in complex longitudinal datasets.

Main Methods:

• Retrospective study design utilizing existing data.
• Application and enhancement of the MASAL statistical approach.
• Analysis of infant growth curves and underlying covariance structures.

Main Results:

• Cocaine use by pregnant women was found to be a marginally significant factor affecting infant body weight post-birth.
• The MASAL method proved effective for fitting growth curves and analyzing covariance structures in longitudinal data.
• The study addressed the statistical challenge of analyzing data collected on irregular schedules.

Conclusions:

• Maternal cocaine exposure shows a marginal but statistically significant association with reduced infant body weight.
• The enhanced MASAL methodology offers a powerful tool for analyzing complex, irregularly scheduled longitudinal data in developmental studies.
• Further research is warranted to fully elucidate the long-term impacts of prenatal cocaine exposure on infant development.