Geographic Clustering of Immunization Risk in Sierra Leone: A Facility-Based Geospatial and Health-Systems Analysis Across 16 Districts

Background: Despite high national childhood immunisation coverage in Sierra Leone, subnational disparities continue to leave communities at risk of under-immunisation and zero-dose status. Understanding the geographic distribution of immunisation risk and the health-system factors associated with these disparities is essential for designing targeted interventions. This study assessed the spatial clustering of facility-level immunisation risk and compared programme and health-system characteristics between hotspot and non-hotspot districts across Sierra Leone.

Methods: A nationwide cross-sectional geospatial and health-systems analysis was conducted using supportive supervision data from 249 routine immunisation facilities across all 16 districts of Sierra Leone. A composite immunisation risk score (0-9) was constructed from three indicators of under-immunisation risk and categorised into low, medium, and high risk. Global Moran's I and Local Indicators of Spatial Association (LISA) identified spatial clustering and hotspot districts. Descriptive statistics summarised programme characteristics, while Pearson's Chi-square and Fisher's exact tests compared health-system indicators between hotspot and non-hotspot districts at a significance level of p<0.05.

Results: Of the 249 facilities, 68.7% were classified as low risk, 18.1% as medium risk, and 13.3% as high risk. Immunisation risk demonstrated significant spatial clustering (Global Moran's I=0.223; permutation p=0.001). Western Area Urban (mean risk score=3.7; 46.7% high-risk facilities) and Bonthe (2.5; 43.8%) exhibited the highest risk, with Western Area Rural (2.1; 26.7%) also showing elevated risk. Hotspot districts had significantly higher availability of updated routine immunisation micro plans (48.4% vs. 13.3%, p<0.001), recent funding (90.3% vs. 69.7%, p<0.001), and supervisory visits (77.4% vs. 56.0%, p<0.001), but were more likely to report insufficient funding (96.4% vs. 85.5%, p=0.02) and lower community health worker assignment (83.9% vs. 95.0%, p=0.002).

Conclusion: Immunisation risk in Sierra Leone is geographically clustered rather than randomly distributed, with distinct hotspots concentrated in Western Area Urban, Western Area Rural, and Bonthe. Integrating geospatial analysis into routine immunisation monitoring can facilitate precision public health approaches by enabling targeted resource allocation, strengthening community outreach, and reducing geographical inequities in immunisation coverage.