Risk factor analysis for stunting incidence using sparse categorical principal component logistic regression

Area of Science:

• Child Health and Nutrition
• Statistical Modeling
• Public Health

Background:

• Stunting in toddlers is a significant global health issue with complex risk factors.
• Existing statistical methods may struggle with the categorical and multicollinear nature of stunting data.
• Understanding these factors is vital for developing effective prevention strategies.

Purpose of the Study:

• To develop and apply a novel statistical model for identifying stunting risk factors.
• To analyze the association between various maternal and child factors and stunting incidence.
• To provide insights into the nutritional and healthcare-related determinants of stunting.

Main Methods:

• Developed a sparse logistic regression model combined with categorical principal component analysis.
• Estimated model parameters using maximum likelihood and Newton-Raphson iterative methods.
• Analyzed stunting risk factors using both child-specific and parental data (age, education, pregnancy nutrition).

Main Results:

• The model effectively handled categorical data with multicollinearity, identifying statistically significant risk factors.
• Medical history, inadequate complementary feeding, formula feeding, lack of feeding programs, and insufficient maternal iron supplementation were strongly associated with stunting.
• Parental factors like age and education also played a role in stunting risk.

Conclusions:

• The study highlights the critical importance of maternal nutrition throughout pregnancy and breastfeeding.
• Adequate complementary feeding and access to nutritional programs are essential for preventing toddler stunting.
• The developed statistical approach offers a robust method for analyzing complex categorical health data.