Abstract Body

Background: In order to develop effective control strategies for HIV epidemics, it is essential to know how many individuals are infected with HIV and their location; i.e. to construct a density of infection (DoI) map. Here we show how to construct such a map for Malawi where the prevalence of HIV is ~11%.

Methods: We used geo-referenced data from the 2010 Malawi Demographic and Health Survey (MDHS), a nationally representative survey that included HIV testing. An individual’s demographic characteristics are linked with their test results. We used data from 7,091 women and 6,497 men aged 15-49 years old. We used these geo-referenced data and applied spatial interpolation techniques to construct a surface prevalence map of the HIV epidemic for the entire country. We then used WorldPop data to construct a demographic map showing the geographic distribution of the population and the population density. We then constructed the DoI map for Malawi by combining the surface prevalence and demographic maps using raster multiplication.

Results: The surface prevalence maps shows that there is significant geographic variation in HIV prevalence throughout the country. In the rural northern region prevalence is ~7%. In the central region, which is semi-urban (i.e. a mix of rural and urban) prevalence is only slightly higher, ~8%. Notably, HIV prevalence is substantially higher (~15%) in the southern urban region. In all regions prevalence is higher in women than in men: ~8% vs. ~5% (northern region), ~9% vs. ~6% (central region), and ~18% vs. ~11% .The demographic map shows that there is significant spatial clustering of the population: 10% of the country contains 48% of the population. In addition 13% of the population lives in the northern region, 42% lives in the central region, and 45% in the southern region. Overall, over 60% of the population lives in urban centers. Our DoI map shows that although the HIV epidemic in Malawi is generalized, the majority of the infected individuals live in urban centers. Using our map we estimate that approximately 690,000 individuals aged 15-49 are infected.

Conclusions: The new methodology that we have developed enables us to identify, and locate, both diagnosed and undiagnosed individuals who are infected with HIV. The DoI maps can be used to identify the areas in greatest need of treatment and prevention programs. Notably, the methodology that we have developed can be used to construct DoI maps for 23 other sub-Saharan African countries.