On the analysis of mixed longitudinal growth data

Area of Science:

• Biostatistics
• Growth Modeling
• Child Development

Background:

• Mixed longitudinal data combines observations from multiple individuals over limited age ranges.
• Analyzing such data requires population-based approaches due to limited individual data.
• Accurate modeling of growth spurts and variability is crucial in child development studies.

Purpose of the Study:

• To propose a novel statistical method for analyzing mixed longitudinal growth data.
• To model both the mean and variance curves of observed characteristics using linear models with cubic splines.
• To apply the method to real-world data on children's height and head circumference growth.

Main Methods:

• Utilized linear models constructed from cubic splines.
• Modeled both the mean and variance curves of growth characteristics.
• Applied the method to a dataset of Victorian schoolchildren's growth.

Main Results:

• Successfully modeled the mean and variance of height and head circumference growth.
• Identified the timing of growth spurts and increases in variability in the study population.
• Demonstrated the utility of the proposed spline-based method for mixed longitudinal data.

Conclusions:

• The proposed cubic spline method provides a robust approach for analyzing mixed longitudinal growth data.
• This method effectively captures complex growth patterns, including spurts and changing variability.
• The findings offer valuable insights into child growth dynamics and statistical modeling techniques.