Spatial Analysis

Healthcare access barriers showed major regional heterogeneity rather than a uniform national pattern. Spatial analysis included prefecture prevalence, interpolation, Moran’s I, and Kulldorff cluster detection. HIV prevalence was significantly positively autocorrelated in both survey years. Moran’s I assessed spatial autocorrelation using a 10-nearest-neighbour geographic-distance matrix. Raw HIV prevalence was calculated for each of Guinea’s 33 prefectures using HIV positives divided by people tested. Continuous prevalence surfaces were generated with Gaussian kernel density estimation and adaptive bandwidths. Spatial analysis was conducted at the regional administrative level using weighted proportions of healthcare access barriers. Getis-Ord Gi* analysis identified significant hotspots in the central interior and north-central regions. DHS coordinate displacement protects respondent confidentiality but reduces geographic precision. The highest prevalence of barriers was found in Togdheer and a south-central cluster including Bay, Bakool, and Hiiraan. Coldspots were identified in Bari, Nugaal, and Lower Juba, indicating lower spatial clustering of barriers relative to neighborin…