Gender and Access to Antiretroviral Treatment in South Africa

Abstract

This paper explores the gender dimensions of access to highly active antiretroviral therapy (HAART) in South Africa. It shows that women are more vulnerable to HIV infection than men, but that women access HAART in disproportionately large numbers. Regression analysis on data from the South African Demographic and Health Survey suggests that men in general access health services less readily than women. This ‘masculinity factor’ accounts for most of the difference between men and women when it comes to accessing HAART. Although men were more likely to favor traditional medicine than women, this was not a statistically significant factor, and it appears that visiting a traditional healer is complementary to, rather than a substitute for, accessing HAART. In short, it seems that gendered norms that make it difficult for men to admit weakness and seek medical attention are the main probable cause for the low proportions of men accessing HAART.

Keywords:

• AIDS
• antiretroviral
• South Africa
• gender
• healthcare
• JEL Codes: I, J

Notes

According to data from UNAIDS, 59 percent of HIV infections in sub-Saharan Africa are among women (see http://data.unaids.org/pub/GlobalReport/2006/Annex2_Data_en.xls, and discussion in Nicoli Nattrass (2006d).

See, for example, Carolyn Baylies and Janet Bujra 2001: 1–24; Liz Walker and Leah Gilbert 2002; Carol Vlassoff and Claudia Garcia Moreno 2002; Ida Susser 2002; Kristin Dunkle, Rachel K. Jewkes, Heather C. Brown, Glenda E. Gray, James A. McIntyre, and Siobán D. Harlow 2004; UNAIDS, UNFPA, and UNIFEM 2004; Stephen Lewis 2005; Eileen Stillwaggon 2006.

The Global Coalition on Women and AIDS has identified seven action areas to address women's vulnerability to HIV: improving reproductive care; reducing violence against women; protecting property and inheritance rights of women; ensuring equal access for females to treatment and care; supporting efforts to provide universal education for girls; supporting improved community care with a special focus on women; and promoting safe sex technologies that are controlled by women, such as the female condom and microbicides (2005: 2).

People who have HIV do not need antiretroviral treatment until they are diagnosed as having full-blown AIDS. This usually occurs within eight to ten years of infection after the body's immune system has been so weakened that the patient is vulnerable to increasing numbers of opportunistic infections.

This study draws this estimate from the ASSA 2003 demographic model, which is available on http://www.assa.org.za. See also Nattrass (2006) for a discussion of what we can learn from this model in this regard.

Nachega et al. (2006: 80) reported that 60.5 percent of 6,288 patients enrolled in a South African private sector AIDS management program were women. This is greater than would be expected, given that only 33 percent of workers in the formal sector are female (see http://www.statssa.gov.za/publications/P0210/P0210September2005.pdf).

HAART has been shown to reduce AIDS-related mortality and morbidity (see review of scientific and clinical studies in Nicoli Nattrass (2007: Chapter 2).

Data drawn from a survey of Khayelitsha in 2004 by the University of Cape Town (see Jeremy Magruder and Nicoli Nattrass (2006) for a discussion of this panel study, which started in 2000).

Andrew Boulle, Des Michaels, and Katherine Hildebrand 2004: 7; David Coetzee, Katherine Hildebrand, Andrew Boulle, Gary Maartens, Francoise Louis, Veliswa Labatala, Hermann Reuter, Nonthutuzelo Ntwana, and Eric Goemare 2004: 891; Nachega et al. 2006.

I once interviewed an HIV-positive man who was living in close proximity to one of the best HAART clinics in the Western Cape, South Africa. He had TB, and I advised him to go to the clinic to investigate whether he needed to go onto HAART. He refused saying that the clinic was for “women and children.” Initially I thought he was demonstrating ignorance but subsequently came to understand that this perception was rooted in local understandings and that for him to go to the clinic was to threaten his identity as a man.

According to UNAIDS estimates for 2006, between 4.8 and 5.8 million HIV positive people live in South Africa and between 3.4 and 9.3 million live in India. See: http://data.unaids.org/pub/GlobalReport/2006/Annex2_Data_en.xls.

The Khayelitsha HAART project demonstrated very good adherence to treatment regimens and excellent clinical outcomes (Coetzee et al. 2004). This, together with similar success stories from rural Haiti (Paul Farmer, Fernet Léandre, Joia S. Mukherjee, Marie S. Claude, Patrice Nevil, Mary C. Smith-Fawzi, Serena P. Koenig, Arachu Castro, Mercedes C. Becerra, Jeffrey Sachs, Amir Attaran, and Jim Yong Kim 2001) helped swing international opinion in favor of supporting efforts to increase access to HAART in developing countries (Paul Farmer 2005).

Only 4.3 percent of Africans (and 3.5 percent of the total population) reported that they had visited such an alternative healer in the past month.

36% = (0.75 ∗ 0.43)/((0.75 ∗ 0.43) + 0.57).

Information provided by Marta Darda and Toby Kaspar.

This survey started off life as a random survey of Khayelitsha households in 2000. Respondents were then re-interviewed in 2004. Although there was some attrition between the two surveys, it had no statistically significant impact on labor-market outcomes (Magruder and Nattrass 2006).

First, the two data sets were appended and a new variable created taking the value of 1 if the respondent was in the HAART data set and a value of 0 if the respondent was in the Khayelitsha data set. A probit regression was then run on this variable using age, gender, and education as independent variables. This regression produced a predicted value for all respondents (HAART and Khayelitsha) – that is, an estimated probability for being in the HAART data set given the model based on education, age, and gender. For each HAART respondent, the closest match was selected (in terms of estimated predicted probability) from the sample of Khayelitsha respondents. I am very grateful to Chris Udry who designed the first programme to help us draw the matched data set and to Jeremy Magruder who helped refine it, but the final program including mistakes is my responsibility.

For the HAART sample, 4.1 percent of men and 1.0 percent of women agreed that traditional African medicine could help fight AIDS, but this was not a statistically significant difference (the t-test result was 1.494 (Pr = 0.1365), and the standard errors around the means were 2.9 percent and 0.8 percent, respectively). For the Khayelitsha matched sample, the percentages were 22.7 percent and 17.1 percent, respectively. This was also not a statistically significant difference (the t-test result was 0.856 (Pr 0.3927), and the standard errors around the means were 6.4 percent and 2.7 percent, respectively).