Bayesian spatial Durbin modelling of stunting prevalence across Indonesian districts.

Abstract

Stunting remains a significant public health concern in Indonesia, characterized by wide regional disparities and persistent prevalence in rural and underserved communities. This study applies the Bayesian Spatial Durbin Model (BSDM) to analyze the spatial distribution and interregional dynamics of childhood stunting across 31 provinces in Indonesia. District level data on stunting prevalence were obtained from the community-based health survey RISKESDAS survey of 2023, focusing on three programmatically salient covariates: the proportion of households with adequate housing, the proportion of children under five who received complete basic immunization and the proportion of infants aged 0-5 months who were exclusively breastfed. The BSDM quantifies direct and spatial spill-over effects while accounting for spatial autocorrelation and parameter uncertainty. Results indicate that adequate housing and complete immunization are associated with lower stunting prevalence and that exclusive breastfeeding are directionally protective. The study finds that spatially coordinated investments in housing quality, immunization outreach and infant feeding support accelerated stunting reduction.