Multilevel Confirmatory Factor Analysis for Modelling the Determinants of Malnutrition Among Under-Five Children in Indonesia

Malnutrition, including stunting, wasting, and underweight, remains a serious public health challenge in Indonesia. Various factors contribute to malnutrition, ranging from individual- and household-level to environmental, socioeconomic, and health policy factors. The use of data analysis as a foundation for policy formulation has been widely adopted in many fields, including healthcare. Most studies have analyzed the determinants of nutritional status using conventional approaches, without accounting for the hierarchical structure of the data or the presence of unobserved (latent) variables. This study aims to examine the factors influencing malnutrition in Indonesia using a Multilevel Confirmatory Factor Analysis (MCFA) approach. This approach allows modeling the causal relationships among latent variables at two distinct levels of analysis, the district (level one) and the province (level two) simultaneously. Based on the Intraclass Correlation test, all malnutrition indicators have values greater than 0.05, indicating variability at the group (province) level. In conclusion, most indicators of Parenting, Health and Environmental Services, and Malnutrition are significant, with Parenting and Malnutrition showing a strong impact. Socioeconomic factors at the district level show limited influence, while provincial-level factors are more significant for wasting and underweight. Both district and pro-vincial-level factors contribute to stunting, with district-level factors having a greater impact.