GIS SHOWS THAT HIV/AIDS RATES IN EASTERN AFRICA STAGNATE

The HIV/AIDS epidemic affects Africa more starkly than the rest of the world. Southern Africa in particular, which constitutes only 2% of the population, contains 30% of the people worldwide affected by HIV/AIDS. High concentrations of HIV/AIDS cases in Africa and the spread of the disease worldwide has led researchers on a search to understand the spread of HIV/AIDS, calling for intersections between the study of geography and epidemiology.

Researchers Dr. Ezekiel Kaipeni and Dr. Leo Zulu from University of Illinois and University of Michigan respectively, mapped incidences of HIV/AIDS on the African continent, looking to answer how HIV/AIDS progresses and what (if any) trajectory it follows. Dr. Kaipeni and Dr. Leo mapped data of both prevalence and time of HIV/AIDS incidence using U.S. Census Bureau HIV Surveillance Database from 1986-2003. A spatial interpolation was run to project the rates where data was not available. Spatial interpolation allows for a continuous representation of rates throughout the African continent, by predicting the rates of prevalence where no census data is available (figure 1). The map’s results indicate that the HIV/AIDS Prevalence Rate increased from 1986 to 2003 throughout Africa, with particular impact on Southern Africa. The epicenter, or highest rate of prevalence from which HIV/AIDS prevalence radiates out from, shifted from the Great Lakes region in 1986 to a secondary epicenter encompassing the nations of South Africa and Botswana by 1994. Between 1999-2003 the map shows that HIV/AIDS prevalence expanded dramatically through southern Africa, whereas the prevalence in Eastern Africa decentralized to what are called “pockets” of prevalence rather than epicenters.

Figure 1: Interpolated rates of HIV aids using IDW and Kriging GIS statistics.

Researchers used this data to project HIV/AIDS prevalence rates and spread throughout the years for 2004-2010, with the understanding that epidemics follow an S-shape curve. The S-shape pattern supposes that an epidemic’s rates begin with a slow, logarithmic increase, reach an exponential increase in which sharp incline of rates is present, and then stagnate to a slow rate of increase until a decline begins to occur. Data for the prevelance, by country, was graphically plotted using the UNAIDS EPP software which estimates, models, and projects epidemiological data by plotting incidence rates and population statistics into an S-shaped curve. This type of statistical analysis is called “curve fitting” and is a well used statistical estimate to graphically represent relationships, in this case, between time and rates of HIV/AIDS incidence (fig 2). GIS data from the previous interpolation was also plugged into this graphical representation, for countries with high incidences of HIV/AIDS. The results show that the prevalence of HIV/AIDs and the spread of the disease are regionally specific, meaning that each is characterized by their own S-shape model, and that projections using data between countries would thus be highly inaccurate. Furthermore, the figures extrapolate the rates of incidence between the years of 2004-2010, where it appears that Eastern African countries of Uganda, Burundi, Rwanda and Zimbabwe will reach relatively low to stagnant rates of HIV/AIDS prevalence in the near future. Nonetheless, Southern African countries such as Botswana, Lesotho, Namibia and Swaziland HIV/AIDS rates persist at 20 to 30/100 persons, indicating that these countries will still experience rate increases, necessitating increased attention at a medical and policy level.

Fig 2: S-shaped curve with exponential and logarithmic rates of HIV/AIDs in selected countries of southern Africa. Projections to 2005-2010 indicate high rates and continued increase of HIV/AIDS prevalence.

There are several limitations to the research that the researchers acknowledge. For one, the data from the US Census Bureau comes from pregnancy surveillance clinics which provides a relatively narrow representation of the HIV/AIDS prevalence amongst whole populations of males, females (pregnant and non-pregnant), or rates amongst demographics farther away from urban centers. These limitations reduce the legitimacy of the spatial interpolation, though useful as it may be to track incidence rates through space and time.

Kalipeni, E. and L. Zulu. 2008. Using GIS to Model and Forecast HIV/AIDS Rates in Africa, 1986-2010. The Professional Geographer 60(1): 33-53

Vanessa Toro