BACKGROUND: Indonesia, home to the world's third-largest smoking population as reported by the WHO, faces a significant public health challenge requiring robust tobacco control strategies. Despite extensive research on smoking behavior, the spatial analysis of smoking prevalence across Indonesia's 514 districts/cities remains underexplored. This study aimed to examine spatial distribution of smoking prevalence and its relationship with education, marital status, smoke-free area policy, and elevation at 514 district/city in Indonesia.
METHODS: This ecological study analyzed aggregate data from the 2018 National Basic Health Research and Information Report provided by Statistics Indonesia, encompassing 514 districts/cities across 34 provinces on seven major islands in Indonesia. Spatial weights were calculated using the distance method. Spatial autocorrelation was assessed with Moran's Index, and the determinants were examined using the Spatial Error Model (SEM).
RESULTS: This study indicated 231 districts/cities with smoking prevalence exceeding the national average. This study identified a positive spatial autocorrelation of smoking prevalence across five major islands in Indonesia, Sumatra (Global Moran’s I = 0.548, p = 0.001), Java (Global Moran’s I = 0.628, p = 0.001), Kalimantan (Global Moran’s I = 0.3525, p = 0.001), Bali and Nusa Tenggara (Global Moran’s I = 0.5939, p = 0.001), and Sulawesi (Global Moran’s I = 0.647, p = 0.001). Statistically significant high-high clusters (p < 0.05) identified across the five islands encompassed a total of 55 districts/cities in Indonesia. The result of the SEM analysis identified the percentage of marital status and elevation of regencies/cities with SEM modelling estimation results (R² = 9%)
CONCLUSIONS: This study identified significant positive spatial autocorrelation of smoking prevalence across five major islands in Indonesia, with 231 districts/cities exceeding the national average and 55 identified as hotspots. Spatial Error Model (SEM) analysis revealed marital status and regional elevation as significant factors, explaining 9% of the variance.