Integrated Pan-Continental Map of HIV Prevalence

Subnational HIV prevalence in 2000.

Subnational HIV prevalence in 2000.

The geographic heterogeneity of the burden of the HIV epidemic within countries is receiving increasing attention. A recent modelling study of HIV in Kenya has indicated that targeting intervention strategies to prevalence "hotspots" may have a substantially greater impact on future incidence than a spatially uniform intervention programme at the same cost. There is a need to better understand the subnational heterogeneity in HIV prevalence in other African countries.


This work supports the argument that spatially structured policy is of potentially high value, and also raises important open questions: Do these results hold when the modelling scope is expanded to Africa as a whole, irrespective of country boundaries? If so, what combinations of interventions should be applied in the different epidemic settings to optimise the impact and efficiency of future investments?

Study aims

We plan to develop an integrated pan-African model of HIV prevalence at the subnational level, which will enable us to target intervention strategies to the epidemiological characteristics of each province or county across the continent. Our hypothesis is that taking just one step down along the spatial scale, from national boundaries to top-level subnational divisions, can yield dividends in efficiency and cost-effectiveness of interventions.


The model used in Anderson et al. [1] will be extended to a "patch" framework of subnational units for all countries in sub-Saharan Africa. An integrated fitting of these patches will be carried out using data derived from the DHS Programme and antenatal clinics, with an application of Bayesian methods to incorporate uncertainty. We will then use this pan-continental model to perform cost-benefit analyses that explore the impact of optimising interventions by predicted subnational prevalence patterns. Our ultimate goals are to investigate the large-scale potential of spatially targeted intervention programmes and to contribute a modelling tool that will be instrumental in guiding allocation of future treatment and prevention resources across Africa.

[1] Anderson SJ, Cherutich P, Kilonzo N, Cremin I, Fecht D, Kimanga D, Harper M, Masha RL, Bahati P, Maina W, Dybul M, Hallett TB. Maximising the effect of combination HIV prevention through prioritisation of the people and places in greatest need: a modelling study. The Lancet 2014; 384(9939): 249 - 256.