Gender and Poverty in South Africa in the Era of HIV/AIDS: A Quantitative Study

Abstract

Background:

Recent research identifies gender inequality as a driver of the HIV/AIDS epidemic. The feminization of poverty is also increasingly apparent, as is the disproportionate vulnerability of members of female-headed households. We sought to examine the relationships among sex, gender, age, HIV status, and socioeconomic characteristics, focusing on heads and nonheads of households.

Methods:

We interviewed 6,338 men and 10,057 women.

Results:

Significantly more males (51.4%) than females (34.8%) indicated that they were heads of households (p < 0.001). Female heads of households were significantly more likely to be infected with HIV than their male counterparts (17.9% vs. 13.1%). Among 15–24-year-old males, those who are often without cash are more likely to be infected with HIV than those who are never without cash (OR = 3.33, 95% CI 1.17-9.49). Similar results were observed among females, who sometimes had no cash (OR = 1.86, 95% CI 1.22-2.82), and among adults aged ≥25 years. Results confirmed that age and gender are related to HIV infection in South Africa and that poverty is a social determinant for HIV infection across all age groups. However, sex is a determinant only among the younger age groups. Young female heads of household are more likely to be poor and are more likely to be HIV positive.

Conclusions:

The results indicate that the HIV/AIDS epidemic in South Africa is characterized by gender inequalities. Young women are more likely to be HIV infected, especially heads of households. Young women are also more likely to live in poverty, although this study cannot establish the directionality of a causative relationship between poverty and risk of HIV. Greater attention must be paid to young women, especially those who head households, in terms of treatment, prevention, and poverty alleviation.

Introduction

There is growing recognition that the feminization of the HIV and AIDS epidemic reflects a complex intersection of biological, sociocultural, and economic factors.^1

–8 This article builds on that foundation by suggesting that there are differences in levels of poverty between men and women in South Africa and that these differences are associated with higher risk of HIV. Moreover, there is a relationship among age, gender, and poverty in South Africa.

In examining the relationship among poverty, HIV, sex, and gender in South Africa, we focus attention on differences between women who head households and those who do not. One reason for conducting a gender-based analysis of heads of household is that the HIV/AIDS epidemic in South Africa is concurrent with an increase in the number of single-adult households, a rapid decline in marriage rates, and gendered changes in family structure because of increased participation of women in economic life.^9

–14 Previous studies of gendered household poverty have identified a relationship between female-headed households (FHHs) and poverty,^3,15,16 yet studies specific to sub-Saharan Africa report more varied findings. For example, several studies from Tanzania have found that FHHs are poorer than male-headed households,^17,18 whereas data from Botswana indicate that sex of the household head has little bearing on the degree of poverty within a household.¹⁹

Evidence also indicates that socioeconomic inequality is linked to the relationship among gender, educational and economic attainment, and vulnerability to HIV/AIDS in South Africa. Based on an analysis of the 1998 South African Demographic and Health Survey, Booysen and Summerton²⁰ found that poor women are less likely to be knowledgeable about HIV/AIDS and are also more likely than comparatively affluent women to have engaged in risky sexual practices with a recent sexual partner. Kalichman et al.²¹ found a positive correlation between risk of HIV transmission and perceived lack of basic needs and services among township residents in South Africa.

There is also evidence suggesting that this relationship between poverty and HIV has deleterious effects on the health of women in particular. Researchers have postulated that in many disadvantaged contexts, poverty and gender inequality together create conditions where high-risk sexual behaviors become prevalent.^22,23 For example, in contexts where gendered inequalities influence access to needed resources, poor women can increase their access to resources through “sexual networking” with men.^{10,24

–29} Also, biological differences between the male and female reproductive tracts mean vaginal intercourse is a riskier activity for women than it is for men. This risk is exacerbated by the presence of ulcerative sexually transmitted infections (STIs), which are more likely to go untreated among the poor.⁸

This article examines the relationship between HIV and poverty, devoting special attention to heads of household. We also flesh out the underresearched relationship between HIV and gender by analyzing gendered differences in access to basic resources, taking into account head of household status. To that end, this article constitutes a quantitative analysis of poverty and gendered inequalities among households in South Africa in the era of HIV/AIDS.

Materials and Methods

Survey design and data collection

We use a gender-based analytical approach to examine aspects of the 2005 South African National HIV Prevalence, HIV Incidence, Behaviour, and Communication Survey. The survey was commissioned by the Nelson Mandela Foundation and implemented through the Human Sciences Research Council (HSRC), with the assistance of the Medical Research Council and Centre for AIDS Development and Research and Evaluation (CADRE).

The survey design was a multistage disproportionate stratified sampling using the 2001 census as the sampling frame and the census enumeration area (EA) as the primary sampling unit. In total, 1000 EAs were used, and 22 households were randomly selected from each EA. The survey was stratified by province and locality type (formal urban, formal rural, tribal authority area, and informal areas) to ensure the representativeness of the sample. In formal areas, race was also used as a stratum. This allowed for oversampling of formal urban areas, where whites and Indians mostly live. Because of these varying probabilities of selection and unequal sampling probabilities, the sample was weighted to correct for possible selection bias. The secondary sampling unit was the visiting point or household from which individuals were selected. One person in each of the following age groups in a household was selected: child 2–14 years, youth aged 15–24 years, and adult aged ≥25 years. For the purpose of this study, only those aged ≥15 years were studied.

Data were collected using a combination of the household questionnaire and the individual questionnaire, with a focus on demographic characteristics of the individuals within the households and poverty. Poverty was measured using several variables, including regularity of having a cash income, having sufficient food and water, having sufficient fuel to heat or cook, and having adequate access to medicine or medical treatment. Dried blood spot (DBS) specimens were collected and analyzed following methodology reported elsewhere.³⁰ The questionnaire data were anonymously linked with blood specimens using a bar code, which was in line with ethical plans approved by the Ethics Committee of the HSRC. The staff was thoroughly trained to obtain informed consent while observing confidentiality and ensuring the respect of participants, given the sensitive nature of questions asked.

Statistical analysis

The data reported in this article were primarily analyzed using STATA, version 8. Because of the multistage sample design, the STATA survey command (svy command) was used extensively to incorporate the complex design aspect of the survey. The data were analyzed with a focus on heads of household by age, sex, and HIV status. The relationships among poverty, HIV status, and sex of household head were examined by focusing on the following variables: level of formal education, source of household income, marital status, and access to sufficient clean water, food, and fuel for cooking and heating the home. The analysis was mainly bivariate. Odds ratios (OR) are reported to indicate the direction and strength of the association; 95% confidence intervals (CI) for the OR are reported. A p value ≤5% indicates statistical significance.

Results

A total of 6,338 men and 10,057 women were interviewed. Among those interviewed, 2,535 men and 3,171 women were aged between 15 and 24 years, and 3,801 men and 6,886 women were ≥25 years. Overall, 56.9% of males and 60.4% of females self-identified racially as African, 12.2% and 11.4% as white, 19.2% and 17.9% as colored, and 11.7% and 10.3% as Indian male and female, respectively. Significantly more males (51.4%) than females (34.8%) indicated that they were heads of households (p < 0.001).

The study findings showed that the relationship between poverty and HIV was complex and was moderated by gender, age, and the role within the household. Females who identified themselves as heads of households were significantly more likely to be infected with HIV (17.9%) compared with their male counterparts (13.1%) (p = 0.005). However, females were generally more likely to be infected with HIV irrespective of their head of household status.

Youth aged 15–24 years

The results in Table 1 show that among young males aged 15–24, those who were often without cash were significantly (p = 0.025) more likely to be infected with HIV compared with those who were never without cash (OR = 3.33, 95% CI 1.17-9.49). Similarly, among females of the same age group, those who often went without cash were more likely (OR = 1.64) to be infected with HIV than those who were never without cash, but this was not statistically significant. Comparing those who often went without cash across males and females, there was no statistically significant difference between males and females even though females who often went without cash were two times more likely than males to be infected with HIV.

Table 1.

HIV Prevalence by Access to Cash Income by Sex, Ages 15–24

	Overall 15–24 years
	Male			Female
How often without cash income	n	HIV+ (%)	OR (95%CI)	n	HIV+ (%)	OR (95%CI)
Often	300	9.4	3.33 (1.17-9.49)	374	19.0	1.64 (0.97-2.79)
Sometimes	477	3.1	1.02 (0.33-3.09)	676	21.0	1.86 (1.22-2.82)
Rarely	88	6.3	2.15 (0.62-7.51)	129	20.0	1.74 (0.81-3.75)
Never	886	3.0	1	1048	12.5	1
	Nonheads of household
	Male 15-24 years			Female 15-24 years
Often	276	10.0	4.40 (1.49-12.93)	348	17.9	1.61 (0.92-2.82)
Sometimes	437	3.3	1.33 (0.41-4.33)	619	20.2	1.86 (1.20-2.89)
Rarely	84	6.2	2.63 (0.70-9.78)	119	20.1	1.86 (0.84-4.12)
Never	797	2.5	1	1048	11.9	1

Among those who were nonheads of households, the results show that young males who often went without cash were significantly more likely to be infected with HIV than those who never went without cash (OR = 4.40, 95% CI 1.49-12.93). Similar results were observed among females, but this was not statistically significant (Table 1). However, females who sometimes went without cash were significantly more likely to be infected with HIV than those who never went without cash (OR = 1.86, 95% CI 1.22-2.82). Comparisons between those who often went without cash showed no significant differences between males and females who are nonheads of households.

Adults aged ≥25 years

In Table 2, a similar analysis to that conducted among the youth is shown for adults aged ≥25 years. The study results showed that adult men who were often without cash were significantly (OR = 2.41, 95% CI 1.48-12.93) more likely than those who were never without cash to be infected with HIV. Similar results were observed among males who sometimes went without cash (OR = 1.69, 95% CI 1.09-2.65) compared with males who were never without cash.

Table 2.

HIV Prevalence by Access to Cash Income by Sex, Ages ≥25

	Total males age ≥25			Total females aged ≥25
How often without cash income	n	HIV+ (%)	OR (95%CI)	n	HIV+ (%)	OR (95%CI)
Often	361	20.9	2.41 (1.48-3.92)	762	22.3	1.99 (1.44-2.74)
Sometimes	669	15.7	1.69 (1.09-2.65)	762	21.2	1.86 (1.41-2.47)
Rarely	154	17.0	1.87 (0.98-3.57)	303	14.6	1.19 (0.71-2.02)
Never	1510	10.0	1	2582	12.6	1

Among females, the results showed that those who were often (OR = 1.99, p < 0.001) or sometimes (OR = 1.86, p < 0.001) without cash income were significantly more likely to be infected with HIV than those who were never without cash. The poorer the woman, the higher is the likelihood of being infected with HIV (Table 2). Analysis of the relationship between poverty and sex of adults did not show statistically significant differences in HIV prevalence in the two groups.

Table 3 shows results of HIV status of adult males and females by various indicators of poverty: adequacy of clean water to cook and drink, having enough food to eat, adequacy of fuel to heat and cook. The analysis is stratified by head of household status. Among adult males having or not having adequate clean water was not significantly (p = 0.956) associated with HIV. Similar results were observed among adult women. Comparison between adult males and adult females with respect to adequacy of clean water to cook and drink did not reveal any association with HIV status. However, there was a significant association between HIV infection and being without enough food to eat for both males (p = 0.017) and females (p = 0.012). The odds of being infected with HIV for males who were often without food to eat were 2.76 times higher than the odds for adult males who were never without food to eat (p = 0.003).

Table 3.

HIV Prevalence by Sex; Various Indications of Poverty Ages 25 and Older

	Total: Males age 25+			Total: Females age 25+
	n	HIV+ (%)	OR (95%CI)	n	HIV+ (%)	OR (95%CI)
Was without clean water to cook and drink
Often	78	14.0	1.06 (0.43-2.57)	204	17.5	1.07 (0.66-1.75)
Sometimes	167	14.0	1.06 (0.61-1.84)	384	17.4	1.09 (0.71-1.68)
Rarely	81	17.2	1.35 (0.46-3.98)	193	20.1	1.27 (0.80-2.00)
Never	2370	13.3	1	4335	16.5	1
Was without enough food to eat
Often	120	26.8	2.76 (1.42-5.35)	315	20.9	1.53 (1.02-2.29)
Sometimes	432	14.8	1.31 (0.84-2.04)	907	21.7	1.60 (1.16-2.20)
Rarely	146	19.1	1.78 (0.79-4.01)	306	19.8	1.42 (0.88-2.29)
Never	2275	11.7	1	3578	14.8	1
Was without enough fuel to heat and cook
Often	65	35.4	4.03 (1.74-9.35)	200	22.1	1.49 (0.94-2.39)
Sometimes	272	19.0	1.73 (1.09-2.72)	540	21.1	1.42 (0.92-2.18)
Rarely	85	16.1	1.41 (0.51-3.92)	210	18.7	1.22 (0.76-1.95)
Never	2275	12.0	1	4165	15.9	1
	Heads of household
	Males 25+			Females 25+
Was without clean water to cook and drink
Often	65	10.7	0.79 (0.29-2.17)	135	16.4	0.94 (0.49-1.77)
Sometimes	135	14.5	1.12 (0.63-2.00)	205	16.9	0.97 (0.51-1.85)
Rarely	67	21.5	1.81 (0.74-4.43)	111	20.3	1.22 (0.66-2.22)
Never	1842	13.2	1	1825	17.3	1
Was without enough fuel to heat and cook
Often	103	27.2	2.93 (1.42-6.06)	200	18.0	1.02 (0.59-1.78)
Sometimes	350	14.7	1.35 (0.79-2.28)	508	16.6	0.93 (0.59-1.46)
Rarely	102	22.1	2.22 (0.95-5.23)	160	16.5	0.92 (0.44-1.94)
Never	1550	11.3	1	1401	17.7	1
Was without enough food to eat
Often	57	38.6	4.95 (2.04-11.97)	133	19.3	1.12 (0.60-2.06)
Sometimes	217	19.5	1.91 (1.13-3.23)	296	14.2	0.76 (0.46-1.31)
Rarely	62	21.9	2.21 (0.97-5.01)	121	20.1	1.18 (0.62-2.27)
Never	1774	11.3	1	1725	17.6	1

Women who were often or sometimes without enough food to eat were significantly (OR = 1.53, 95% CI 1.02-2.29 and OR = 1.6, 95% CI 1.16-2.20, respectively) more likely to have a higher prevalence of HIV infection than those women who were never without enough food (Table 3). Adult men who were often or sometimes without enough fuel to heat and cook were statistically significantly (OR = 4.03, p = 0.001 and OR = 1.37, p = 0.019, respectively) more likely to be infected with HIV compared with adult males who were never without. For females, having enough fuel to heat and cook was not associated with HIV status. There were no differences across the two sexes in these outcomes.

The next level of analysis only considered heads of households and stratified by sex for adults aged ≥25 years. The analysis used having inadequate clean water to cook and drink as a measure of poverty. The results showed that there were no significant differences in HIV status for different levels of poverty among adult males. The same was observed among females. When comparison was made between men and women with respect to HIV status by level of poverty using the indicator of adequacy of water to cook and clean there were no significant differences (Table 3). Among heads of household males aged ≥25 years, there was a significant association between HIV and having enough fuel to heat and cook (p = 0.016). In this group, males who were often without fuel to cook or heat were 2.93 (95% CI 1.42-6.06) times more likely to be infected with HIV than those who were never without. Among FHHs, there was no significant association.

Further analysis was done among the youth using a variety of indicators of poverty: being without clean water to cook and drink, being without enough food to eat, and being without enough fuel to heat and cook. The results in Table 4 show that all the expected associations between poverty and HIV were not statistically significant for males and females, except that those who rarely were without water to cook and clean had a significantly (OR = 5.88, 95% CI 1.79-19.27) higher HIV prevalence than those who never had this problem. This finding is probably a variation due to a very small sample size in the cells. Separating these data by heads of households led to even smaller samples (results not shown), yielding unstable results.

Table 4.

HIV Prevalence by Sex and Poverty Indicators, Ages 15–24

	Males aged 15–24			Females aged 15–24
	n	HIV+ (%)	OR (95%CI)	n	HIV+ (%)	OR (95%CI)
Was without clean water to cook and drink
Often	62	3.8	1.25 (0.41-3.79)	91	20.9	1.36 (0.55-3.35)
Sometimes	139	8.8	3.02 (0.55-16.68)	185	19.8	1.27 (0.66-2.42)
Rarely	74	15.8	5.88 (1.79-19.27)	81	19.5	1.24 (0.62-2.49)
Never	1476	3.1	1	1948	16.3	1
Was without enough food to eat
Often	920	17.5	7.80 (1.96-31.11)	138	20.4	1.42 (0.74-2.69)
Sometimes	279	2.9	1.09 (0.54-2.22)	384	19.5	1.34 (0.85-2.09)
Rarely	996	12.9	5.42 (1.55-19.04)	144	24.4	1.79 (0.95-3.37)
Never	1281	2.7	1	1636	15.3	1
Was without enough fuel to heat and cook
Often	49	14.1	3.72 (0.60-22.84)	81	26.1	1.82 (0.77-4.29)
Sometimes	174	4.2	0.99 (0.41-2.42)	168	18.5	1.16 (0.71-1.91)
Rarely	60	0.7	0.17 (0.03-0.93)	99	17.3	1.07 (0.53-2.19)
Never	1470	4.2	1	1805	16.3	1

A potential confounder of the relationship between HIV status and head of household status is educational attainment, although analysis did not show any significant overall differences between those who had schooling and those without schooling in HIV status of males and females of all age groups. Stratifying data by age and household head status showed no significant differences in HIV status for males of all ages. For women, the situation was the same, except that young female nonheads of household who had no or little education had significantly higher HIV status than those with high school or university education. Young females who were nonheads of household had a significantly higher HIV prevalence than their male counterparts. Young females who had attained only secondary education had an HIV prevalence three times higher than that of their male counterparts. However, these comparisons are limited because of the cross-sectional design of this study, as the highest educational level might have been realized after HIV infection.

Another potential confounder of the relationship between HIV status and head of household status is marital status. Our data revealed that among males of all ages, marital status was not related to HIV status, regardless of the head of household status. Among adult females (≥25 years), however, single women had a significantly higher HIV prevalence than married, divorced, or widowed women. Differences between the sexes were evident between males and females aged ≥25 years; single males had significantly lower HIV prevalence than their female counterparts, but there was no difference in HIV prevalence for married/cohabiting, widowed, divorced, or separated males or females. Most young people in our sample were unmarried, and, therefore, gender-based analysis of marital status did not yield useful findings.

Discussion

Although debates about the link between HIV/AIDS and poverty abound^31
–33 and some data suggest a negative relationship between poverty and HIV,³⁴ there is increasing evidence of a link between HIV and poverty, especially in South Africa. For example, national household data suggest that it is the poor in urban informal settlements that are hard hit by HIV, not the rich.³⁰ Incidence levels are a testimony to this; urban informal settlements have an annual incidence rate of 7% compared to a 1.8% in urban formal areas, 2.8% in rural formal areas, and 2.7% in rural informal areas.³⁰ Gendered and intergenerational aspects of poverty further add more complicated dimensions to the relationship between HIV/AIDS and poverty. Data abound that link gender inequality with HIV infection.^4,35
–37

Our results support the hypothesis that differences in risk of HIV prevalence between men and women are linked to levels of poverty. It is widely accepted that women are at risk for HIV because of their gender and social standing. In many societies, women tend to be dependent on men, making it difficult for them to protect themselves from HIV infection.^2,26 FHHs are at especially high risk, as are young people more generally. Our finding that HIV risk is especially high among young women aged 15–24 is consistent with the literature.^1,2,4

–7 We are surprised, however, to find no significant differences in HIV infection rates between men and women in the ≥25-year cohort for any variable of our analysis, suggesting that risk of HIV prevalence in South Africa may be gendered only among youth. Accordingly, our results suggest that further energies must be devoted to examining and addressing gendered aspects of risk among South African young people.

Our findings also indicate that HIV prevalence is highest among the poorest segments of society, although many nonpoor are affected by HIV. For example, regarding access to food, we found that young female nonheads of household who had often gone without food were seven times more frequent than young men who had often gone without, yet those who had never gone without were still at one and a half times more frequent than young men who had never gone without food. Our data indicate that the association between poverty and likelihood of HIV infection is more evident among young women than young men.

These findings indicate that poverty and HIV/AIDS are concurrent challenges for many South Africans and must be addressed as such. Our findings may also support the position that poor women are exposed to greater HIV risk through sexual networking.^{10,24

–29} Our findings that young FHHs are 1.5 times more likely that their male counterparts to be infected is particularly indicative in this regard. Moreover, although many poor young women may have higher numbers of sexual partners as a result of networking, it is possible that poor young men use condoms with their sexual partners. Nevertheless, the fact that nonpoor women are still more likely to be HIV positive indicates that gendered social inequalities are not restricted to contexts of poverty. Further attention to gendered social dynamics is critical.

The question of access to higher education is an important one, as our results show no significant differences in the prevalence of HIV infection among young men and women with tertiary levels of education, yet our findings for the 15–24-year-old cohort show that having achieved only primary or secondary qualifications correlates with higher HIV prevalence among women. HIV prevalence is highest among young people with no schooling, regardless of sex. These findings also echo the findings of a recent national study conducted among educators in South Africa, where it was observed that the higher the qualification, the lower the HIV prevalence.³⁸ A study done in Mozambique found that low educational levels were associated with a lower degree of knowledge about preventive methods and transmission routes and less condom use. The study also found that the probability of not using a condom was higher among females and among people living in less privileged districts, with no education.³⁹ These findings suggest young women's vulnerability to HIV/AIDS infection could be reduced through greater emphasis on continuing education for girls and young women and that access to education remains an important tool in the fight against HIV in South Africa. Other studies also have reported that education provides a protective effect.^40,41 For example, in a study conducted in Vietnam,⁴¹ having more education was also beneficial in preventive knowledge. Those with tertiary education compared with those without were more than 6 times as likely to be aware of HIV preventive benefits of condom use, 4 times as likely to know the benefits of having one sex partner, 34 times as likely to know where to get condoms, and 26 times as likely to know how to use a condom. The findings from these studies show that being poor and having less education increases vulnerability to HIV.

Available evidence on the relationship between marital status and HIV is contradictory, as some studies have found lower HIV prevalence among married people,⁴² whereas others have found marriage to be a risk factor.^43,44 In this study among adults ≥25 years, we found that marital status had no significant association with HIV prevalence, regardless of sex or head of household status. As with other variables of analysis, we found significant results only among the 15–24-year-old group. Among these young people, the likelihood of HIV infection among single women was found to be three times higher compared to single men; however, no significant differences were found among married or cohabiting young people. This may suggest some degree of protection for people cohabiting or married. However, data on the incidence of HIV infection in stable sexual partnerships suggest that these relationships are not as protective as many believe, especially when one partner is already infected.⁴⁵ It was found that individuals living with an HIV-positive partner were more likely to be HIV-positive at baseline (women: OR = 75.7, 95% CI 33.4-172; men: OR = 62.4, 95% CI 28.5-137). Seroincidence rates in discordant couples were 10/100 person-years and 5/100 person-years for women and men, respectively (rate ratio [RR] = 2.0, 95% CI 0.28-22.1). In concordant-negative couples, seroincidence rates were 0.17/100 person-years in women and 0.45/100 person-years in men (RR = 0.38, 95% CI 0.12-1.04). Individuals living in discordant relationships were at a greatly increased risk of infection compared with individuals in concordant-negative relationships (RR = 57.9, 95% CI 12.0-244 for women; RR = 11.0, 95% CI 1.2-47.5 for men).⁴⁵

The study by Hugonnet et al.⁴⁵ concluded that men were more likely than women to introduce HIV infection in concordant-negative partnerships. Among discordant couples, the incidence of HIV-negative women was twice as high as in men. HIV-negative individuals in discordant partnerships are at high risk of infection, and preventive interventions targeted at such individuals are urgently needed. Unfortunately, data in the categories of widowed and divorced were inconclusive because of the small sample size. Nonetheless, these two groups are important to understand, as they are likely to be exposed to new sexual partnerships and may need to navigate a new terrain of sex and HIV.^46,47

Among young heads of household, we found that young FHHs are eight times more likely to be infected compared with young male household heads. Given that young women who do not head households are at three times the risk of their male peers, our results clearly indicate that being single puts young South African women in an extremely vulnerable position. Our study does not allow us to identify why young FHHs have higher HIV prevalence than their male counterparts, yet some speculations are possible. On one hand, studies of female-headed households suggest that these families tend to dominate in the poorer sections of the community and that these women's access to resources is limited by cultural, social, and economic influences.^14
–16 Furthermore, if these FHHs are poor, some of them may be exposed to increased risk through sexual networking.^24

–29 Pressure to engage in such relationships may be further compounded if a young woman has dependents. It is also possible that these women may be more vulnerable to exploitation and abuse.⁴⁸ The household could be single headed as a result of HIV infection through the death of or abandonment by a partner.

There are several limitations to this study. Although we can establish a causative relationship among age, sex, and risk of HIV infection, our data and methods do not allow us to establish a causative relationship between risk of HIV infection and poverty. Although poor people may be more vulnerable to infection because of, for instance, sexual networking, it may also be that HIV-positive people become poorer due to the social and economic constraints of HIV/AIDS. Further research is required to understand these dynamics. Furthermore, the category of female-headed household is hugely diverse in terms of residency pattern and socioeconomic status. Our data do not allow us to distinguish among wealthy widows, female tertiary-level students who are single, and young single mothers with limited access to resources. Given that our data demonstrate a relationship among head of household status, sex, and poverty, however, it is reasonable to speculate that many FHHs are economically compromised. Confounding is likely to have played a role in the analysis of these data.

Conclusions

This study confirms that sex and gender play a significant role in determining vulnerability to HIV infection in South Africa. Our findings confirm that age and gender are related to HIV infection in South Africa, and poverty is a social determinant for HIV infection across all age groups. Moreover, this correlation between poverty and HIV infection is much more pronounced among women than men. We have found that higher educational achievement correlates with lower risk of HIV infection, indicating that continued support for furthering educational attainment among South African women may be an important tool in the fight against HIV/AIDS.

Our results showed that that among males, marital status was not related to HIV status, regardless of the whether or not the man identified as head of household. Among adult females (≥25 years), however, single women had a significantly higher HIV prevalence than married, divorced, or widowed women. Further research is needed to understand why single women who are neither divorced nor widowed are at higher risk of HIV infection relative to other women. Finally, our findings showed that young women who head households are extremely vulnerable relative to other women. Greater research attention is needed to understand the social determinants of vulnerability among FHHs.

Footnotes

Acknowledgments

We acknowledge the financial support of the Nelson Mandela Foundation, the Swiss Agency for Development and Cooperation, and the Human Sciences Research Council. We also wish to acknowledge financial and technical support of the U.S. Centers for Disease Control and Prevention. Finally we thank all the South Africans who provided the blood samples and personal information that informed this study.

Disclosure Statement

The authors have no conflicts of interest to report.

References

Gilbert

, Walker

. Treading the path of least resistance: HIV/AIDS and social inequalities—A South African case study. Soc Sci Med, 2002; 54:1093–1110.

Hallman

Gendered Socio-economic conditions and HIV risk behaviours among young people in South Africa. Afr J AIDS Res, 2005; 4:37–50.

Buvini

, Gupta

. Female-headed households and female-maintained families: Are they worth targeting to reduce poverty in developing countries? Econ Dev Cult Change, 1997; 45:259–280.

Gregson

, Nyamukapa

, Garnett

et al. Sexual mixing patterns and age differentials in teenage exposure to HIV infection in rural Zimbabwe. Lancet, 2002; 359:1896–1903.

Mabala

. From HIV prevention to HIV protection: Addressing the vulnerability of girls and young women in urban areas. Environ Urban, 2006; 18:407–432.

Obi

, Onabolu

, Momba

MNB

et al. The interesting cross-paths of HIV/AIDS and water in Southern Africa with special reference to South Africa. Water SA, 2006; 32:323–343.

Pettifor

, Measham

, Rees

, Padian

. Sexual power and HIV risk, South Africa. Emerg Infect Dis, 2004; 10:1996–2004.

Stillwaggon

. AIDS and the ecology of poverty. London: Oxford University Press, 2006.

Bakilana

. Age at sexual debut in South Africa. Afr J AIDS Res, 2005; 4:1–6.

10.

Hunter

. The materiality of everyday sex: Thinking beyond “prostitution.” Afr Stud, 2002; 61:99–120.

11.

Hunter

. The changing political economy of sex in South Africa: The significance of unemployment and inequalities to the scale of the AIDS pandemic. Soc Sci Med, 2007; 64:689–700.

12.

Montgomery

, Hosegood

, Busza

, Timaeus

. Men's involvement in the South African family: Engendering change in the AIDS era. Soc Sci Med, 2006; 62:2411–2419.

13.

Preston-Whyte

. Women who are not married: Fertility, “illegitimacy,” and the nature of households and domestic groups among single African women in Durban. S Afr Tydskr Sosiol, 2002; 24:63–71.

14.

Smit

. The changing role of the husband/father in the dual-earner family in South Africa. J Comp Fam Stud, 2002; 33:401–415.

15.

Barros

, Fox

, Mendonca

. Female-headed household, poverty and the welfare of children in urban Brazil. Econ Dev Cult Change, 1997; 45:231–257.

16.

Goldberg

, Cremen

. The feminization of poverty: Only in America? New York: Praeger1990.

17.

Katapa

. A comparison of female and male-headed households in Tanzania and poverty implications. J Biosoc Sci, 2006; 38:327–339.

18.

van Vuuren

. Female headed households: Their survival strategies in Tanzania Leiden. African Studies CentreASC Working Paper No. 44, 2002.

19.

International Fund for Agricultural Development. Botswana—Female headed households. www.ifad.org/gender/learning/challenges/women/bt_6_2htm. 2009 April 5.

20.

Booysen

FLR

, Summerton

. Poverty, risky sexual behaviour, and vulnerability to HIV infection: Evidence from South Africa. J Health Popul Nutr, 2002; 20:285–288.

21.

Kalichman

, Simbayi

, Jooste

, Cherry

, Caim

. Poverty-related stressors and HIV/AIDS transmission risks in two South African communities. J Urban Health, 2005; 82:237–249.

22.

David

, Mercado

, Becker

, Edmundo

, Mugisha

. The prevention and control of HIV/AIDS, TB, and vector-borne diseases in informal settlements: Challenges, opportunities, and insights. J Urban Health, 2007; 84:i66–i74.

23.

Masanjala

. The poverty-HIV/AIDS nexus in Africa: A livelihood approach. Soc Sci Med, 2007; 64:1032–1041.

24.

Eaton

, Flisher

, Aaro

. Unsafe sexual behavior in South African youth. Soc Sci Med, 2003; 56:149–165.

25.

Hunter

. Cultural politics and masculinities: Multiple-partners in historical perspective in KwaZulu-Natal. Cult Health Sex, 2005; 7:389–403.

26.

Leclerc-Madlala

. Transactional sex and the pursuit of modernity. Soc Dyn, 2003; 29:1–21.

27.

Morris

, Kretzschmar

. Concurrent partnerships and the spread of HIV/AIDS. AIDS, 2007; 11:641–648.

28.

Selikow

, Zulu

, Cedras

. The ingara, the regte, and the cherry: HIV/AIDS and youth culture in contemporary urban townships. Agenda, 2002; 53:22–32.

29.

Vos

. Attitudes to sex and sexual behavior in rural Matabele-land, Zimbabwe. AIDS Care, 1994; 6:193–203.

30.

Shisana

, Rehle

, Simbayi

et al. South African National HIV Prevalence, HIV Incidence, Behaviour and Communication Survey, 2005. Cape Town: HSRC Press, 2005.

31.

Gillespie

, Kadiyala

, Greener

. Is poverty or wealth driving HIV transmission? AIDS, 2007; 21,Suppl 7:S5–S16.

32.

Greener

. HIV, growth, poverty and inequality in sub-Saharan Africa. Draft paper, UNAIDS, 2008.

33.

Greener

. HIV, growth, poverty and inequality in sub-Saharan Africa. Paper presented at the World Bank conference, Cape Town, June 2008.

34.

Mishra

, Assche

, Greener

et al. HIV infection does not disproportionately affect the poorer in sub-Saharan Africa. AIDS, 2007; 21,Suppl 7:s17–s28.

35.

Dunkle

, Jewkes

, Brown

, Gray

G E

, McIntyre

, Harlow

. Transactional sex among women in Soweto, South Africa: Prevalence, risk factors and association with HIV infection. Soc Sci Med, 2004; 59:1581–1592.

36.

Jewkes

, Levin

, Penn-Kekana

. Gender inequalities, intimate partner violence and HIV preventative practices: Finding of South Africa cross-sectional study. Soc Sci Med, 2003; 56:125–134.

37.

LaFraniere

. Sex abuse of girls is stubborn scourge in Africa. New York TimesDecember 13, 2006.2008 July 14.

38.

Zungu-Dirwayi

, Shisana

, Louw

, Dana

. Social determinants for HIV prevalence among South African educators. AIDS Care, 2007; 19:1296–1300.

39.

Font

, Puigpinós

, Chichango

, Cabrero

, Borrell

. AIDS-related knowledge and behaviors in Mozambique. Rev Epidemiol Sante Publique, 2006; 54:341–354.

40.

Barnighausen

, Tanser

, Mbizana

et al. High HIV incidence despite high prevalence in rural South Africa: Findings from a prospective population-based study. AIDS, 2008; 22:139–144.

41.

Bloom

, River Path

Associates

, Sevilla

. Health, wealth, AIDS, and poverty—The case of Cambodia, Asian Development Bank, 2001. www.adb.org/Documents/Reports/HW_Cambodia/HWCAM.pdf

42.

Shisana

, Zungu-Dirwayi

, Toefy

, Simbayi

, Malik

, Zuma

. Marital status, risk of HIV infection in South Africa. S Afr Med J, 2004; 94:537–543.

43.

Clark

. Early marriage and HIV risks in sub-Saharan Africa. Stud Fam Plann, 2004; 35:149–160.

44.

Newmann

, Sarin

, Kumarasamy

et al. Marriage, monogamy and HIV: A profile for HIV-infected women in South India. Int J STDs AIDS, 2000; 11:250–2503.

45.

Hugonnet

, Mosha

, Todd

et al. Incidence of HIV infection in stable sexual partnerships: A retrospective cohort study of 1802 couples in Mwanza Region, Tanzania. J Acquir Immune Defic Syndr, 2003; 32:349–351.

46.

Agot

, Agot

, Agingu

, Bukusi

. Recognizing and addressing tensions between culture and HIV prevention: Lessons from a study on widow inheritance. International Conference on AIDS, July 11–16, 200415: Abstract No. TuPpD2049.

47.

Okeyo

, Allen

. Influence of widow inheritance on the epidemiology of AIDS in Africa. Afr J Med Pract, 1994; 1:20–25.

48.

Jewkes

. Intimate partner violence: Causes and prevention. Lancet, 2002; 359:1423–1429.