Background:  Lesotho is one of the most resource-poor Sub-Saharan African nations. It has one of the highest HIV prevalence rates among adults (as many as 40% among those aged 30 to 40 in 2007). Previous studies have implicitly assumed spatial homogeneity in the spread of HIV. But spatial data in the form of administrative regions and ecological zones are either too broad or too arbitrary to estimate the true geospatial extent of the HIV epidemic, which can be highly clustered or may extend across district boundaries. In the case of Lesotho, difficult terrain, inadequate transportation networks, and isolated populations all pose serious challenges to the assumption of spatial homogeneity.

Methods:  We base our analysis on the 2004 Lesotho Demographic and Health Survey (DHS), which we cross-referenced with the geographic positioning system (GPS) coordinates of the sample clusters in the survey. The analytic sample consists of 3345 females and 2648 males distributed among 381 sample clusters. We used the spatial autocorrelation (kriging) and the “hot-spot” analysis features included in the ArcGIS software to determine hot-spots of infection and map the geospatial spread of HIV.

Results:  We found significant heterogeneity in the spatial distribution of Lesotho’s HIV epidemic. HIV prevalence ranges from 4 to 39% for men and 18 to 56% for women depending on the location of the sampled individuals. Kriging results show that high HIV prevalence is concentrated in the lowlands and in the urban areas of the north and the west, and it is also spatially heterogeneous within each of the 10 health care districts. There is also significant spatial clustering of men’s HIV as opposed to women’s:  high positive values of Getis-Ord General G statistics (cluster p <0.01) indicate acute male HIV infection clusters in the southern mountain region. No similar clustering was found for women. Clusters of HIV (general population’s and males’) seem to coincide with the locations of the transportation hubs.

Conclusions:  This study enables the construction of a spatial health policy map for Lesotho. The determination of gender differences in the spatial distribution of HIV provides a concrete basis for rolling out gender-specific prevention programs such as circumcision. Our results will not only serve to better understand the epidemiology of HIV in Lesotho, but also help devise effective delivery of care strategies.