South African HIV prevalence steadily rising; researchers investigate why some communities are harder hit

Three studies presented last Wednesday at the 2nd South African AIDS Conference Durban show that there is room for improvement in the country’s “HIV intelligence gathering” (HIV surveillance). Better, more current data that includes information on risk behaviours could help the country develop more targeted and effective prevention.

One study reported an increasing incidence of HIV among rape victims in the Transkei region (Eastern Cape). Another found an appallingly high prevalence in the rural midlands of KwaZulu-Natal — with up to two-thirds of the pregnant women aged 25-29 HIV-infected. Finally, the third study showed that prevalence varied markedly from community to community in Limpopo Province — and looked more closely at the reasons why.

Sexual assault has been on the increase in tandem with the spread of HIV in South Africa, at least partly because of a folk belief that sex with a virgin can cleanse one of the virus. It is estimated that about one million women or girls are raped in South Africa each year, and an increasing number of them are under 15 years of age. However, there has been relatively little data about how many of the victims become HIV-infected.

But according to findings of a study presented by Dr. Banwari Meel of the University of Transkei, the incidence of HIV infection among sexually assaulted women and children has been increasing over the last few years in the Transkei.

The retrospective study looked at HIV incidence among 831 victims of sexual assault who attended a clinic in the Transkei from 2000-2004. Over the 5-years study period, the clinic saw a two and a half fold increase in sexual assaults and a four and a half fold increase in incidence of HIV (See Table 1). Adult women were significantly more likely to test HIV-positive in this study, but a significant number of victims under 15 years of age tested positive as well.

Incidence of HIV seropositivity among victims of sexual assault at the time of incident reporting, during the five years (2000-2004) in Transkei Region, South Africa 

In general, early sexual activity has been associated with an increased HIV prevalence in a number of studies. A survey in Vulindlela, KwaZulu-Natal, conducted since 2001 by the Centre for the AIDS Programme of Research in South Africa (CAPRISA), has found a high rate of teenage pregnancy associated with shockingly high HIV infection rates (see Table 2).

Age specific HIV prevalence among ANC attendees in Vulindlela 2001-2003

In 2004, said Dr. Ayesha Kharsany of CAPRISA, “we wanted to identify behavioural factors that could be contributing to the high prevalence of HIV infection in the district.”

Vulindlela is a large rural district in the midlands of KwaZulu Natal situated about 170km west of Durban and about 70km from the regional hospitals around the town of Pietermaritzburg. The district contains many different communities and informal settlements with a total population of around 400,000. Many of the inhabitants live in poverty. What few work opportunities exist are in the forestry industry or in the neighbouring manufacturing and affluent residential towns.

Several clinics in the district provide comprehensive primary health care but “whilst the clinics are major providers of antenatal and family planning services,” said Dr. Kharsany, “ they are not included in the DOH annual HIV seroprevalence surveys.”

So in order to compare the data from Vulindlela to DOH figures, from October to December 2004, the researchers surveyed all (n=552) the pregnant women making their first antenatal visit to the various primary health care clinics. In addition to being tested for HIV, the women were questioned about their age, their current partners’ age, history of pregnancy and the year of previous pregnancy.

The mean age of the women was 23 years (ranging from 14 to 43 years). The mean age of their current sexual partner was 27 years.

“In the survey of 2004, young women less than 20 years accounted for 37% of cases,” said Dr. Kharsany. “In contrast, the DOH, for its surveys from 2001 to 2003, has consistently reported less than 20% of women in the less than 20-year age group.” (See Table 3.)

Age Distribution of ANC Attendees  (Vulindlela Vs. DOH Survey)

The overall HIV prevalence among antenatal clinic attendees was 43% — and when standardised to the DOH antenatal population, the prevalence was 47%. This is in sharp contract to the KZN provincial prevalence reported by the DOH: 37.5%, and the national prevalence: 27.9%.

The prevalence by age also appears to be steadily increasing — and cumulative (see Table 4).

Age specific HIV prevalence among ANC attendees in Vulindlela – 2004

The researchers found that having an older partner was a risk factor for HIV infection. For women under 20 years old, the risk of HIV infection was nearly doubled if their current sexual partner was aged between 20-24 years old, and if their partner was 25 years or older, the risk of HIV infection increased by 6 fold (OR 5.9 95% CI:1.9-18.4; p=0.004).

“These results,” said Dr. Kharsany “underscore the need to understand these sexual networks and target young women at high risk for HIV infection.”

A large study by the Rural AIDS & Development Action Research Programme (RADAR), of the University of the Witwatersrand further analysed the reasons why some communities are harder hit that others.

This time, the survey was in eight villages within a twenty kilometer radius of each other in the rural district of Sekhukhuneland, Limpopo province. Although the overall prevalence was much lower (10.8%) than in Vulindlela, it ranged up to 46.2% in some of the neighbourhoods under study.

According to Dr. Rico Euripidou, research manager for RADAR and who presented the study on behalf of the study’s lead author, Dr. Paul Pronyk, studies in other countries have shown that it is “not uncommon to find a 4-10 fold variation [in prevalence] over a 20km radius.” These differences cannot be attributed to differences in HIV subtype (within the same region or country), nor can they be totally explained by differences in sexual behaviour.

“Most often, differences in prevalence are not explored,” said Euripidou “or they may be broadly attributed to rural-urban geography or proximity to a trading centre.”

But there may be other structural factors (such as access to health services, circumcision rates, cultural practices or social norms, development, legal and policy issues) that put residents of some communities at greater risk than others.

So the researchers collected socio-demographic, mobility, sexual behaviour and HIV prevalence data on 2488 randomly selected people between the ages of 14-35 years to determine structural factors that could influence a community’s HIV prevalence.

They decided the following structural factors merited closer attention:

• Levels of bar (or “shebeen”) activity = numbers of bars + average number of clients + average monthly alcohol intake.
• Local availability of sex workers = total number of sex workers per village at all bars on average weekend evening.
• Ease of access mine/main town = distance + tar road + public transport (by village).
• Mobility (per subsection of each village): Percentage of travel to a city = proportion of residents who stayed overnight in a major city in the past year, and percentage of out migrants = proportion residents listed as permanent residents who are currently sleeping at home.
• Population stability = proportion resident for less than ten years.
• Wealth = principle components analysis score form value of selected household assets, quality of housing conditions, income/human capital, food security and social status.
• Education level = proportion of 10 to 35 year-olds who have completed secondary school.
• Social capital = index based on social network membership + response to questions on: levels of trust, reciprocity, solidarity in time of crisis, collective action (positive (marches/rallies) and negative (crime rate = police statistics of the total number of serious and violent crimes in an entire village over a twelve month period)

Based on the findings of the first questionnaire, a second survey was constructed and conducted on 825 community women. In addition, communities were profiled — which included surveys of the local bar with owner and patron interviews.

The study found a number of individual risk factors for higher HIV risk including older age for both sexes, sex before age 16 for males (p=<0.001), and higher lifetime numbers of partners for females (p=<0.001).

Among structural factors, easier access to a trading centre (p=0.02), higher proportions of short-term residents (p=<0.001), and lower levels of social capital (p<0.001 men, p=0.02 women) were associated with higher HIV prevalence in both sexes. Among males only, higher HIV prevalence was seen among communities with easier access to a local mine (p=0.05), a higher density and activity of local bars (p=0.004), a higher numbers of sex workers per village (p=<0.001), and lower proportions of out-migrants (p=0.002).

“Efforts to alter structural factors have the potential to change the vulnerability of whole populations to HIV", concluded Dr. Euripidou, "which is critical in a generalized epidemic.”