Longitudinal data analysis by nonparametric bi-response spline truncated regression model

Abstract

This study aims to construct a growth model for toddlers based on longitudinal data using a nonparametric bivariate response truncated spline regression approach. The data used consisted of routine toddler weighing records from the Integrated Health Post in Bontosunggu Village, South Bontonompo Sub-District, Gowa Regency, encompassing age, body weight, and height of toddlers from January to December 2021. The analysis was conducted by developing two regression models for each response variable, namely body weight and height, using four knot points (12, 24, 36, and 48 months), and parameter estimation was carried out using the Ordinary Least Squares (OLS) method. The analysis results indicate that the growth of toddlers’ weight and height does not follow a linear pattern with age but rather displays a change in growth rate at specific age points. This model provides a more realistic and flexible depiction of growth and can serve as a useful tool for monitoring the nutritional and health status of early childhood populations.