Unraveling spatial disparities in the multidimensional health of young Brazilian adults using the new smart K-means algorithm

The study aims to develop and validate a health quality indicator that associates social and environmental conditions with physical and mental health conditions. Together these conditions classify health into five categories: “very good” “good” “fair” “poor” and “very poor”. This study combines the Smart K-Means algorithm and spatial analysis methods to investigate health disparities across Brazil using data from the National Health Survey. The algorithm clusters 24 health and well-being sub-indicators applying Shannon entropy to retain those with high informational diversity, and identifies five groups representing different quality of life levels. The average silhouette score of 0 .63 confirms the effectiveness of clustering, facilitating the interpretation of complex, multidimensional health patterns. Spatial analysis highlights significant regional disparities: higher self-perceived health in parts of the Northeast and North regions, despite objectively poorer conditions, contrasts with lower subjective health in some wealthier southeastern areas, illustrating Brazil’s regional paradox. These findings suggest a mismatch between expected socioeconomic development and actual health outcomes, particularly in the north, where healthcare access remains limited. The results emphasize the need for targeted health interventions and improvements in healthcare infrastructure in underserved areas, especially in northern Brazil, while maintaining better outcomes in the south and southeast. Together with Smart K-Means clustering and spatial statistics, these methods provide valuable tools for understanding health inequalities and informing public policy to improve the population’s quality of life.