33 million people are infected with HIV worldwide. But what does that tell us? The world is big. 67% of those infected lives in Sub-Saharan Africa. Still, Africa is a very large continent. 35% of all those infected live in southern Africa and within South Africa, the KwaZulu Natal province has the highest prevalence of HIV infection; 39% of women who attend an antenatal clinic are HIV positive.

Knowing the HIV epidemic is fundamental when setting up interventions and my opening statement clearly shows that generalisation are not very useful or helpful. But what about KwaZulu Natal itself? Earlier this year, Frank Tanser and colleagues from the Africa Centre for Health and Population Study reported on the geographical patterns and clustering of HIV infections in a 438 km² area part of the Hlabissa sub-district (one of the five sub-districts of the Umkhanyakude district) about 250 km north of Durban in South Africa.

View Larger Map

This mostly rural area contains a population of less than 90,000 Zulu-speaking people who live in scattered homesteads. There is only one township (KwaMsane) and a series of settlements along the national road that links Durban in the south to Mozambique in the North. It is in this area that the researchers have conducted HIV surveillance studies and have mapped the results spatially.

Using advanced statistical analysis Tanser and colleagues identified clusters with a high number of HIV infection along the national road (Cluster 1, 2 and 3 in the figure below) but also clusters where the prevalence of HIV was very much lower (cluster 4, 5 and 6 below).

This is interesting for several reasons. The first is that more than 25 years into the epidemic and despite a high level of migration, HIV prevalence is still not homogeneous even in a country exposed to HIV for so long. This is particularly remarkable in this area of high HIV prevalence where some small pockets of people are more affected by HIV than other (and this despite the absence of identifiable village or town).

The second is that a cluster of low prevalence (cluster 5, mid bottom above) is adjacent to two clusters of high prevalence. This indicates a limited sexual mixing even at the scale of a kilometre.

Third, HIV prevalence is particularly high along the national road and in fact, the estimated density of people living with HIV is 15.7 times higher within 1 km of the road than it is in the rest of the area.

There are several reasons that can explain this distribution such as differences in age distribution between clusters, mobility and labour-related migration, and prevalence of sex work, but not the usual culprits of poverty, lack of education or unemployment; indeed, the researchers report that the area of high HIV prevalence were “characterised by higher aggregated levels of educations, household wealth and nearly double the rate of employment” but had lower level of marriage and proportion of migrants.

This small study also challenges the prevailing paradigm of a generalised rural epidemic of HIV in South Africa and as a consequence, “one size fits all” interventions. It also begs the question of the efficacy of costly HIV outreach programmes in “hard to reach” areas when these are based on “global” data. (note that this is not to say that nothing should be done for hard-to-reach area, but if they are hard to reach for prevention they may also be hard to reach for HIV).

Overall, this is another study emphasising the importance of knowing your epidemic and to take actions based on evidences, not on generalisations and assumptions.

Related posts brought to you by Yet Another Related Posts Plugin.