Understanding functional diversity in public primary health care: a cluster analysis of utilization patterns.

BACKGROUND

Primary health care (PHC) systems are typically organized using uniform administrative models that assume functional similarity across facilities. In practice, however, PHC units differ substantially in how services are delivered. In Greece, despite ongoing reforms, empirical evidence on the functional heterogeneity of public PHC units remains limited. This study aims to classify PHC units based on observed service utilization patterns and to identify the organizational and geographical determinants shaping these functional typologies.

METHODS

This cross-sectional study analyzed administrative data from 649 public PHC units in Greece (2024). Visit compositions (regular, emergency, prescription, and other) were analyzed using compositional data analysis with centered log-ratio transformation. K-means clustering was applied to identify utilization-based functional typologies. Associations between cluster membership and unit characteristics (size, organizational type, and geography) were examined using ANOVA and chi-square tests.

RESULTS

Three distinct functional typologies emerged. Cluster 1 ("balanced, generalist units", n = 308) exhibited a mixed service profile across visit types. Cluster 2 ("routine care-focused units", n = 66) was dominated by regular visits (82.4%) and minimal prescription activity, but showed substantial within-cluster variability in emergency care. Cluster 3 ("mixed, higher-intensity units", n = 275) displayed a heterogeneous case-mix and a greater share of complex or administrative activity. Cluster membership was significantly associated with unit size, organizational type, and geographical location (p < 0.001). Cluster 1 units were significantly larger, followed by Cluster 3, while Cluster 2 units were the smallest. Organizationally, Cluster 1 was dominated by Rural Health Centers and Local Health Units, Cluster 2 by Local Medical Practices and Urban Health Centers, and Cluster 3 by Multipurpose Regional Medical Practices and Local Health Units. Spatially, higher-intensity units (Cluster 3) were predominantly located in major urban centers.

CONCLUSION

PHC is functionally heterogeneous, with distinct utilization profiles that are not captured by existing administrative classifications. These findings reveal a structural misalignment between formal organizational labels and actual service delivery. A utilization-based typology provides a practical tool for health policy and management, supporting more targeted approaches to resource allocation, workforce planning, and regional service design. Incorporating functional differentiation into PHC governance can enhance efficiency, equity, and system responsiveness.