In a recent New York Times article, Celia Dugger reports on encouraging results from two studies on interventions that help women in South Africa and Malawi reduce their risk of HIV infection. The first study found that women using a vaginal microbicidal gel were 39% less likely to contract HIV than those using a placebo. It has been published in Science, and its authors write in their abstract:

The CAPRISA 004 trial assessed effectiveness and safety of a 1% vaginal gel formulation of tenofovir, a nucleotide reverse transcriptase inhibitor, for the prevention of HIV acquisition in women. A double-blind, randomized controlled trial was conducted comparing tenofovir gel (n = 445) with placebo gel (n = 444) in sexually active, HIVuninfected 18 to 40 year-old women in urban and rural KwaZulu-Natal, South Africa HIV serostatus, safety, sexual behavior and gel and condom use were assessed at monthly follow-up visits for 30 months. HIV incidence in the tenofovir gel arm was 5.6 per 100 women-years, i.e. person time of study observation, (38/680.6 women-years) compared to 9.1 per 100 women-years (60/660.7 womenyears) in the placebo gel arm (incidence rate ratio = 0.61; P = 0.017). In high adherers (gel adherence > 80%), HIV incidence was 54% lower (P = 0.025) in the tenofovir gel arm. In intermediate adherers (gel adherence 50 to 80%) and low adherers (gel adherence < 50%) the HIV incidence reduction was 38% and 28% respectively. Tenofovir gel reduced HIV acquisition by an estimated 39% overall, and by 54% in women with high gel adherence.

Dugger notes that because the trial was relatively small and the gel far from 100% effective, another trial with similar results will be necessary before health advocates and donors are ready to invest in a push for widespread use. Still, this study is a bright spot in a field of research that’s seen a slew of failures.

The other encouraging research results that Dugger highlights come from a World Bank-sponsored program in Malawi that gives small monthly payments to poor schoolgirls and their families with the goal of reducing girls’ risky sexual encounters. The study description on the World Bank website summarizes the results and the forces that seem to be at work:

[The] Malawi program gave girls ages 13 to 22 and their parents as much as $15 each month if the girls attended school regularly. A control group, however, didn’t get any cash reward for schooling. In total, the study enrolled 3,796 never-married schoolgirls in Zomba, a district in southern Malawi. A year later … more schoolgirls receiving cash (95%) stayed in school than the control group (89%).

But there’s a surprise finding: 18 months after the program began in January 2008, biomarker data show that HIV infection rates among girls who received cash was 1.2% versus the control group’s 3%. This translates to 60% lower prevalence. Girls in the cash group also had a lower infection rate of herpes simplex virus type 2, the common cause of genital herpes (0.7% vs. 3%). Those findings hold even for a third group of girls who got cash without any schooling or other strings attached.

How did it happen? The key seems to be an “income effect” on the sexual behaviors of young women receiving cash payments. A year after the program started, girls who received payments not only had less sex, but when they did, they tended to choose safer partners, says Berk Özler, a senior economist at the Development Research Group who conducted the study with Sarah Baird of George Washington University and Craig McIntosh at the University of California, San Diego. In fact, the infection rate among those partners is estimated to be half of that of partners of the control group.

The cash transfers may have led to a drop in the so-called “transactional sex.” At the beginning of the study, a quarter of sexually-active participants said they started relationships because they “needed his assistance” or “wanted gifts/money.” Meanwhile, among the sexually-active schoolgirls in the control group, 90% said they received an average of US$6.50 a month in gifts or cash from their partners. Such “gifts” are significant, given the country’s GDP per capita was $287.5 in 2008.

After a year, schoolgirls receiving payments from the cash-transfer program seemed to avoid older men, who tend to be wealthier and are much more likely to be HIV positive than schoolboys. The sexual partners were two years older on average than the girls, compared with three years for the control group.

When I saw the description of this study in Dugger’s article, my first thought was that the intervention worked by keeping girls in school. The study was evidently designed to test this hypothesis, though, because the researchers had two intervention groups, one in which the payments were conditional on school attendance and one in which they were unconditional (assignment to treatment group was by enumeration area). No significant differences were found between the two treatment groups – in other words, girls who received payments regardless of whether they stayed in school showed the same lower infection rate as those who received the payments conditionally.

Conditional cash transfer (CCT) programs that provide regular payments to poor families if their children stay in school (or get vaccinations, or engage in other desirable behaviors) have become a favored anti-poverty intervention in low- and middle-income countries, and a new Economist report on CCTs gives a concise explanation for the reason:

The programmes have spread because they work. They cut poverty. They improve income distribution. And they do so cheaply.

By contrast, many behavioral interventions that aim to reduce HIV transmission through education, motivational efforts, or improving access to needed products or services have disappointing results where subjects face significant barriers to adopting desired behaviors. An executive summary of the Malawi study (posted on the World Bank site, not from a journal) quotes a recent review of behavior change interventions (McCoy, Kangwende, and Padian, 2009 – probably this one) that reported “current behavior change interventions, by themselves, have been limited in their ability to control HIV infection in women and girls in low- and middle-income countries.”

The Malawi study suggests that, at least in the case of behaviors related to HIV, payments may not even need to be conditional if the money reduces pressures that lead to unhealthy behaviors. (Girls’ choice of sexual partners may or may not be based on HIV considerations, but they apparently prefer sexual partners closer to their own ages, who are less likely to be HIV positive.)

Both the microbicidal gel and cash-payment interventions have an important aspect in common: they put decisions about safer sex in women’s hands. It’s important that the gel, unlike condoms, can be used by women without men’s consent or knowledge – study participants were instructed to use pre-filled gel applicators within 12 hours before sex and as soon as possible within 12 hours following it.

Halting the spread of HIV will require a combination of interventions, both existing and new. These studies suggest that many women in low- and middle-income countries want to have sex that is safer. We need to make options for safer sex more readily available to them.


  1. #1 Katherine
    August 10, 2010

    1st comment is spam.

  2. #2 uk visa lawyer
    August 10, 2010

    First of all, and I understand the necessity, it’s tragic that girls might be healthy non-HIV were it not for the fact they were in the placebo group.
    And the second tragedy the monetary aspect – US$6.50 a month in gifts or cash – is so little compared to the cost of a young girl’s life.
    It makes it very difficult to believe in a benevolent God.

  3. #3 Vicki
    August 12, 2010

    It’s easy after the fact to say it’s “tragedy” that some of the girls were in the placebo group. But there’s a long history of placebo-controlled trials (including of AIDS prevention and treatment) in which the placebo group fared as well as the treatment group: is it a tragedy that some people dealt with possibly serious side effects? What of medical studies in which the treatment group fared worse than the placebo group, to the extent that the study was ended early?

  4. #4 Lance
    August 15, 2010

    Interesting and important discussion but one note; The science blogs link I followed here was titled,

    “To Beat AIDS, Help Women Choose Safer Sex: The pump Handle”

    Wow. While I guess you can’t get AIDS from a pump handle it seems that there would be other risks involved.

  5. #5 Rich
    August 17, 2010

    Sadly, cash transfer programs are unlikely to be sustainable without external funding. Note also that the difference in school attendance was rather small.

    For all the buzz about microbicides, the recent trials only did a little better than previous trials that had been viewed as failures. The effects of the most recent trial were very much related to adherence and adherence is only part of the story. Microbicide researchers have not found a good consistent delivery system. Partial protection in high risk populations and circumstances, absent the supportive aspects of a trial (prevention counseling, etc.) may have limited public health impact.