Primary Health Care (PHC) in Brazil's Unified Health System (SUS) requires territorial planning tools that address social inequalities and support the Family Health Strategy (ESF). This study proposes a spatial decision-support framework that integrates a multidimensional Health Vulnerability Index (HVI) with spatially constrained clustering at the census-sector level. The HVI is constructed from variables in the 2022 Brazilian Demographic Census across the domains of Urban Infrastructure, Human Capital, Income and Employment, and Demographic Vulnerability. The resulting HVI is then used in a spatially constrained hierarchical clustering algorithm that incorporates queen contiguity and ESF-aligned population thresholds to generate homogeneous and contiguous PHC planning regions. Results show a 77.8% reduction in the number of planning units nationally while preserving spatial coherence and internal homogeneity. The proposed framework transforms fine-grained vulnerability data into operationally meaningful planning geographies, providing a scalable and reproducible basis for PHC needs-based territorial planning and future allocation models within SUS.