Abstract
Health equity is a key policy priority in public health and a central component of Sustainable Development Goal 3 (SDG 3), which focuses on "Good Health and Well-Being". While SDG 3 sets global benchmarks, its local implementation-particularly in countries with decentralized healthcare systems and significant regional disparities-poses a considerable challenge. This study addresses the need for subnational analyses by moving beyond national averages and identifying region-specific barriers and enablers to achieving SDG 3 targets, using Italy as a case study. To this end, we apply spectral bi-clustering, an innovative data-mining technique, to regional SDG 3 indicators compiled by ISTAT for the years 2013-2019. The analysis pursues two objectives: (1) to identify clusters of Italian regions with similar SDG 3 profiles; and (2) to determine which indicators are most salient within each cluster and how they diverge from national benchmarks, deriving policy implications tailored to each group of regions. Our findings reveal three distinct regional clusters: the analysis demonstrates that certain health indicators are more relevant within specific regional contexts, pointing to structural and systemic variations in healthcare provision and outcomes. These results underscore the inadequacy of uniform policy approaches and highlight the need for regionally differentiated strategies. This study provides one of the first applications of spectral bi-clustering to health equity analysis at the subnational level, offering actionable insights for policymakers seeking to localize SDG 3 implementation and bridge health gaps across regions.