Skip to content

Works Not Faith

A forgotten strategy in the fight against HIV in Africa

In the early 1990s, I worked as a biochemist on an HIV vaccine research project in Uganda. My lab was in a ramshackle hospital in Kampala, then a dusty city of about a million people surrounded by forested hills a few miles north of Lake Victoria. Roughly one in seven Ugandan adults was HIV-positive, and there was still no treatment for AIDS. Everyone I met was affected by it. I’ll never forget sitting with a work colleague in her small house and catching a glimpse of her once-strong thirty-year-old son lying still on the bed in the next room, or coming upon a thin, listless child on the concrete veranda outside the lab.

Like most countries devastated by HIV, Uganda was dependent upon foreign assistance. The response began modestly. Catholic and Protestant churches had been running schools and hospitals since colonial times and were the first to offer help. They built upon traditional Ugandan systems of mutual aid by helping families care for the sick and counseling the young in ways that were, for the most part, compassionate, discreet, and un-preachy on sexual matters. In 1994, the U.S. Agency for International Development (USAID) began supporting programs that promoted condom use across the continent. Cheap, subsidized condoms, including ribbed and multicolored varieties, appeared in shops and were advertised on billboards as the ultimate sexual experience. At village gatherings, community health workers demonstrated how to use them by putting them on bananas and other objects. But even where these programs were diligently implemented and well-funded, the HIV rate continued to soar for reasons that would only become clear later. A turning point came in 1996, when researchers discovered that a combination of antiretroviral drugs could slow or even halt AIDS progression in HIV-positive people; that year, the Joint United Nations Program on HIV/AIDS (UNAIDS) was created to coordinate an expanded international response.

It was clear that the United States needed to contribute more aid, but by the time it finally arrived in the new millennium, HIV had spread to virtually every country on earth. The epidemic remained most severe in eastern and southern Africa. In Botswana, Swaziland (now Eswatini), Lesotho, southern Mozambique, and eastern South Africa, some 20 percent of young women were infected—a rate over a hundred times higher than in the United States. Things weren’t much better in Zimbabwe, Zambia, Tanzania, Kenya, Malawi, Rwanda, Namibia, and Uganda. Then, in the same 2003 State of the Union address in which President George W. Bush rallied support for invading Iraq on what would turn out to be specious grounds—leading to hundreds of thousands of needless civilian deaths—he stressed the compassionate nature of his administration by announcing a number of new charitable initiatives, including the President’s Emergency Plan for AIDS Relief, or PEPFAR, a multibillion-dollar program for AIDS prevention, treatment, and care in developing countries. In the twenty years since, PEPFAR has spent over $110 billion in some fifty countries, mostly in Africa. Roughly twenty-five million people with HIV are alive today as a result. I’m not always proud to be an American in Africa, but PEPFAR has had spectacular results—at least when it comes to treatment.

When it comes to HIV prevention, however, PEPFAR has had a rockier history. The disappointing results of USAID’s condom promotion programs enabled Bush to reward his evangelical supporters by making abstinence PEPFAR’s central prevention strategy. As soon as the initiative was up and running, USAID—now a major conduit of PEPFAR funds—was ordered to pull down the condom billboards and replace them with images of young Africans declaring their pride in being virgins. One third of PEPFAR’s initial HIV prevention budget was earmarked for school-based abstinence education, prayer meetings, parades, billboards, radio and TV spots, and lectures by chastity evangelists. By 2008, PEPFAR was spending $250 million a year on abstinence programs, but they were having virtually no effect—and risked turning the entire program into a laughingstock.

HIV prevention is not impossible. Culturally appropriate, scientifically accurate information about sexual risk, tactfully conveyed, can quickly reduce the incidence of sexually transmitted infections in young people. For a very brief interval, PEPFAR supported a small number of such programs, but in 2012 it abruptly began shutting them down. Since then, the vast majority of PEPFAR funding has gone to medical programs, mostly for HIV treatment and drug-based prevention. Medicine is crucial for fighting HIV, and it’s also a moral imperative. But HIV medications are not a cure, they don’t work for everyone, and they can have debilitating side effects. Without focused, appropriate AIDS education, PEPFAR’s goal of ending the epidemic will not be met by its declared deadline of 2030, if ever.

The Concurrency Hypothesis

By the time Bush launched PEPFAR, I’d become a public health consultant, traveling around Africa reporting on HIV for UN agencies, NGOs, magazines, and foundations. One thing that had always confused me about the epidemic was its severity on the continent compared to the rest of the world. Everywhere, men who had sex with men and sex workers tended to have larger numbers of casual partners, so their vulnerability to HIV was well understood. But promiscuity couldn’t explain what was going on in Africa. In Uganda, the typical HIV positive person was a churchgoing man or woman who worked as a teacher, lab worker, journalist, government official, shopkeeper, politician, nurse, orderly, or doctor. On surveys, they consistently reported they’d had fewer sexual partners than the average American, homosexual or heterosexual. Intravenous drug use was virtually unknown. Other explanations for the epidemic, such as especially aggressive viral strains or some feature of African physiology that increased susceptibility to infection, didn’t hold up. This had to have something to do with sex, but what?

Promiscuity couldn't explain what was going on in Africa.

Had our experimental vaccine worked, I would not have cared, but it was a flop. Even today a vaccine remains elusive despite billions of dollars invested in research. So I was intrigued when, in 2003, I came across the work of Martina Morris, a mathematical sociologist at the University of Washington who’d developed a theory to explain why HIV spreads more rapidly in some communities than others. Her findings, drawn from computer simulations of the epidemic using data from sexual behavior surveys conducted around the world, not only helped explain what was going on in Uganda and other severely affected countries but also had profound implications for HIV prevention.

In 1993, she was giving a lecture about her project to some Ugandan colleagues when one of them raised his hand.

“Can your computer handle cases where people have more than one sexual relationship at a time?” he asked.

“No,” she replied. The mathematical tools for that didn’t exist. The man walked out of the room. Afterward, the other Ugandans told her that her computer model would be useless if she didn’t account for these overlapping sexual partnerships, which Morris and the rest of the public health community eventually came to call “concurrent.”

In the years that followed, Morris and other researchers confirmed what the Ugandans had told her. The sex trade in Sub-Saharan Africa was much smaller than it was in Asian countries like India and Thailand, which nevertheless had much lower HIV rates. But in eastern and southern Africa especially, people were much more likely to have a small number—perhaps two or three—long-term relationships that overlapped for months or years. A man might have two wives, or a wife and a girlfriend, while the girlfriend or wife might have another partner whom she kept secretly, and so on. Using real data from countries around the world, Morris and other researchers were able to demonstrate that if enough people in a community did this, a sexual network would emerge that could serve as a superhighway for HIV, putting large numbers of people at risk, including those whose behavior would otherwise be considered “safe,” such as women and men with only one sexual partner. If that one partner had other partners, they’d all be on the HIV superhighway.

HIV plays another trick that makes concurrent relationships very risky. People are much more likely to transmit HIV in the month or so after they’ve been infected. If someone with two ongoing relationships contracts HIV from one partner, she’s very likely to pass it on to the other one. But if a serial monogamist is infected by his one partner, he won’t pass it on until he links up with someone else, which could be months or years in the future. By then, he’ll be much less infectious and might not transmit at all. Concurrency also helps explain why HIV rates are relatively low in the Muslim societies of West Africa, even though many men there have multiple wives. Women in those societies are under more intense surveillance than those in the largely Christian societies of eastern and southern Africa, so the polygamous unions aren’t linked up. (Muslim societies also practice near-universal male circumcision, which offers some protection against HIV infection because the foreskin contains Langerhans cells which have HIV receptors.)

Among heterosexuals in the United States, the HIV rate remains low because they tend to have their partners sequentially; this prevents the formation of stable, overlapping sexual networks and slows or even stops viral transmission completely. Obviously, I oversimplify. Make no mistake, many people in eastern and southern Africa have serial relationships, and many American heterosexuals have concurrent ones, but careful surveys of sexual behavior conducted in the 1990s when HIV was spreading rapidly found rates of concurrency correlate with HIV prevalence in a range of communities in Africa, Asia, and within the United States.

Morris’s hypothesis fascinated me because it helped explain some epidemiological mysteries, including the failure of USAID’s early condom programs. Condoms are mainly used with sex workers or in one-night stands. In eastern and southern Africa, concurrent relationships tend to be long-term and intimate; children are often desired, so condoms are seldom, if ever, used. The concurrency hypothesis also helped explain why, in Africa, more women than men are HIV-positive, whereas the reverse is true virtually everywhere else in the world. In a concurrency network, a woman’s risk depends not only on her own behavior but also on that of her partner. Men tend to have more concurrent partners, and the size of the network they’re linked to determines the risk for all the women they’re having sex with.

Sexual systems vary around the world. There is no ideal one, just as there is no ideal cuisine or musical style. In the 1970s, British anthropologist Jack Goody theorized that nonmonogamous sexual systems suited African populations—and probably all human groups in the past—because they maximized fertility. It was only when Eurasian societies came to value material wealth over kinship that lifelong monogamous marriage came to be seen as the natural order of things. Of course, lifelong monogamy isn’t the natural order of things—hence the prevalence of divorce, infidelity, and domestic violence. However, enforcement of monogamy may have arisen as a norm, especially for women, because it helped reduce battles over patrimony and inheritance as families accumulated wealth worth fighting over.

Ugandan acquaintances smiled when I explained Morris’s concurrency theory to them. They knew what I was talking about; they also understood why it made people vulnerable to HIV and agreed that the danger of concurrency networks was something young people needed to know about. But trying to change anyone’s culture, particularly their sexual culture, is a fool’s errand, as the failure of PEPFAR’s abstinence programming made clear. As my colleagues and I saw it, educating people about concurrency needn’t be prescriptive. The goal should be to promote frank, compassionate, and open discussions of HIV risk in general so people would know where their risks came from. Then they could figure out for themselves how to make their relationships safer—either through mutual fidelity, more consistent condom use, or other forms of safer sex, including no penetrative sex if it suited them.

There was historical evidence that such campaigns might work. During the late 1980s, years before scientists like Martina Morris began researching concurrency, the Ugandan health ministry commissioned murals around the country that carried implicit warnings about concurrency. The slogan “Zero Grazing”—an agricultural term meaning, roughly, “Keep your animals tied up”—was plastered on buildings everywhere; everyone understood what it meant. If they didn’t, posters urging Ugandans to “Love Carefully” and “Love Faithfully” were erected too. Coincidentally or not, the HIV rate in young pregnant Ugandan women fell by around 60 percent in the years after the campaign commenced, even as infections elsewhere on the continent soared. Obviously, these women were not having safe sex since they were pregnant, and condom use wasn’t widespread. It must have been that they and/or their partners were heeding the Zero Grazing message and having fewer partners “in the next paddock,” as a Ugandan health official explained it to me in 2010. Numerous surveys of sexual behavior carried out at the time demonstrate that this is in fact what Ugandans were doing.

The Best Medicine

With the concurrency hypothesis in mind, it did not surprise me when PEPFAR’s abstinence campaigns became an object of ridicule. In Uganda, preachers pitched revival tents on nearly every street corner to vie for Bush’s abstinence cash. A condom-burning ceremony presided over by one bellowing PEPFAR-funded Ugandan preacher even became the subject of a Congressional hearing. American comedians also got into the act: for a fictional book entitled Savin’ It, former senator Al Franken sent letters to dozens of conservative leaders, including National Security Advisor Condoleezza Rice, former Attorney General John Ashcroft, and Secretary of Defense Donald H. Rumsfeld, asking them to “share a moment when you were tempted to have sex, but were able to overcome your urges through willpower and strength of character.” None did so.

Without focused, appropriate AIDS education, PEPFAR’s goal of ending the epidemic will not be met by its declared deadline of 2030, if ever.

No doubt concerned about the program’s effectiveness, let alone its reputation, PEPFAR officials began quietly funding a small number of pilot concurrency education programs in 2005, which they managed to shoehorn under the abstinence funding mandate. Most took the form of social marketing—Madison Avenue-style radio and TV spots and billboards—that also implicitly spread the American creed of consumerism. Although I was glad to see that concurrency was being taken seriously (and served as an advisor on some of these projects), not all were as tactful as I would have liked. In Eswatini, PEPFAR billboards erected in 2006 declared, “Your secret lover can kill you.” Horrified Swazis living with HIV marched on the capital demanding that the insulting posters be taken down, and within two weeks, they were. In Uganda, billboards proclaimed, “Get off the sexual network!” This was better, and a USAID evaluation found that women who were aware of the campaign were about 30 percent more likely to have had only one sexual partner in the past twelve months than women who had not. But the campaign didn’t seem to influence the behavior of men. It also missed the mark because many impoverished young women and girls are drawn into sexual networks not by choice but through seduction, coercion, trickery, or force. At any rate, acquaintances told me that “How’s your sexual network?” soon became a popular pickup line in Uganda.

“This is the medicine. You have found it.”

I had to admit that PEPFAR’s concurrency billboards came across like unwelcome advice from a clueless relative. What makes talking about HIV so difficult is the assumption that the virus is invariably associated with immoral behavior, a view firmly held by President Bush’s evangelical supporters. There had to be a more constructive way to help young people understand the complex role concurrency played in the epidemic and in their lives. Rather than conversations about who was “responsible” for AIDS—or who was “killing” whom—I had hoped that PEPFAR’s concurrency campaigns would promote discussions about the cruelty of HIV-related stigma and the importance of solidarity and compassion. After all, anyone in a sexual system involving concurrency can become infected.

Many schools in the region had allocated time for AIDS education, but teachers were often flummoxed about how to fill it and, in particular, what to tell their students about how HIV spread from person to person. (In Zambia, for example, one teaching manual instructed them to say that you could get HIV from pencil sharpeners.) So in 2013, with funding from PEPFAR, my colleagues and I began conducting seminars introducing the concurrency hypothesis to teachers at a handful of schools in Zambia to see if they thought it would be acceptable in their classrooms and communities. When we showed them simple diagrams illustrating how concurrency networks facilitated the spread of HIV, they, like my Ugandan friends, recognized at once how important the theory was. We discussed the implications for HIV-related stigma, sexual coercion of girls, and violence against women, who were often accused by their partners of “bringing HIV into the house”—even when both partners were positive and there was no way of knowing who had been infected first. These meetings lasted hours, sometimes well into the evening. No one left early or asked to break for tea. Similar community discussion programs were piloted in other countries. “This is the medicine,” said a Kenyan participant in a concurrency exercise designed by Martina Morris. “You have found it.”

For any intervention to move beyond the pilot phase, it’s necessary to measure its effects experimentally to see if people who receive it really are less likely to contract HIV than those who don’t. In 2009, Morris applied to the National Institutes of Health, which handled PEPFAR-related research, for funding to carry out such an experiment, but her application was rejected without explanation. Because PEPFAR is so huge, it dominates global funding for HIV related activities worldwide. Other donors may fund small projects, but proving that concurrency education really works is expensive—large numbers of people must be tested regularly for HIV. Only PEPFAR could afford to pay for such a trial, so when Morris was turned down, she had nowhere else to go.

However, during concurrency’s brief heyday at PEPFAR, a single experimental trial was carried out, not in Africa but in the southern United States, where young women are at low risk of HIV but contract other, more transmissible infections—herpes, human papilloma virus, syphilis, gonorrhea, and chlamydia—pretty frequently. In 2004, researchers at Emory University and the University of South Carolina invited hundreds of unmarried, sexually active, and mostly low-income young women to join the study. They divided them into two groups: one attended two four-hour educational sessions about the risks of concurrent sexual partnerships while the other attended two four-hour sessions about nutrition. Over the following year, women in the concurrency education group were about half as likely to contract all measured sexually transmitted infections. I’ve been unable to identify any other behavioral education program—including condom and abstinence programs—that had such a powerful effect on infection risk.

Normally, such a result would trigger funding for follow-up studies to determine if similar or even better results could be obtained in Africa with HIV. Why PEPFAR officials declined to support such studies frustrates me to this day. It’s as though they didn’t want to know whether concurrency programs worked or not. I wasn’t entirely surprised. Between 2009 and 2011, the Lancet, the Journal of the International AIDS Society, and other prominent medical periodicals published several articles purporting to pour water on the concurrency hypothesis. The articles all had serious flaws, and their vehement tone made me wonder, though I could never prove it, whether a concerted effort was underway to discredit the theory. It reminded me of a 1961 New England Journal of Medicine editorial sowing doubt about the—by then incontrovertible—evidence linking smoking and lung cancer, and, later, Washington’s role in the flurry of news stories about Iraq’s weapons of mass destruction.

Two of the anti-concurrency articles took the form of reviews and weren’t based on new research. One pair of authors claimed concurrency wasn’t clearly defined (not true) and that there was no evidence that concurrency was more common in Africa (there was). Another pair of authors claimed concurrency education would divert resources from other, presumably better programs, which they didn’t describe, and called for an end to all research on the subject. This pair of authors attributed Africa’s high HIV rates instead to malaria, malnutrition, other sexually transmitted infections, recreational drug use, homosexuality, “numerous forms of blood exposures,” and other factors for which there is little to no evidence. A third paper claimed that polygyny—a common form of concurrency—wasn’t a risk factor for HIV in Africa, but neglected to note that seven other studies based on data from different African communities had found that it was. These authors also controlled for extramarital sex, without explaining how else the virus would get into these relationships.

Perhaps the oddest paper appeared in the Lancet in 2011. It purported to show that concurrency couldn’t explain the spread of HIV in a rural area of South Africa where the infection rate was very high. The authors had been studying the community intensively for many years and knew which women had been recently infected. Using complex statistical method, they showed that the fraction of men practicing concurrency who lived within a three-kilometer radius of these recently infected women was no higher than it was within a three-kilometer radius of women who had not been recently infected. The authors had no information about who was actually sleeping with whom. Rather, they assumed that women in this community were most likely to have sex with their immediate neighbors, increasingly less likely to have sex with men living somewhat farther away, and never had sex with men who lived more than four kilometers away—even though their own data showed that some 40 percent of these women had had no partners within that area—meaning they most likely had partners outside of it.

These papers were widely shared among AIDS researchers and African officials, few if any of whom would have understood the technical details. Martina Morris and I, along with other concurrency researchers, wrote articles and letters refuting the concurrency critics. They were published after long delays and had no impact on PEPFAR’s decision to wind down funding for concurrency education, which it began to do in 2012—the same year officials announced a mere 3 percent of total PEPFAR funding would go toward behavioral HIV prevention of any kind. Although concurrency remains in the official AIDS education curriculum of Zambia and a few other countries, without PEPFAR backing, it’s unlikely to be emphasized.

PEPFAR didn’t give up on HIV prevention entirely. Since 2014, it has supported initiatives like DREAMS—which stands for Determined, Resilient, Empowered, AIDS-free, Mentored, and Safe—a collection of HIV prevention programs for young African women and girls that includes medication to treat and prevent infection, information about condoms and HIV testing, community discussions about gender violence prevention, and a modest number of scholarships to enable girls to stay in school and other forms of economic support. These programs aim to address factors like violence and poverty that put young women at risk of HIV and are valuable in their own right. But not even PEPFAR is rich enough to undertake the social and economic transformation of the continent in just seven years. In any case, evidence that the various interventions in the DREAMS package actually reduce HIV risk is weak.

Over the last decade, though, the majority of prevention dollars have gone to a new medical behemoth known as “Test and Treat,” which works like this: antiretroviral drugs delay the onset of AIDS by reducing the amount of virus circulating in blood and genital fluids. This also makes HIV-positive people undergoing treatment less likely to infect their partners. Because most HIV-positive people live for years with no symptoms, PEPFAR’s early programs administered medication only to patients whose blood tests showed their immune system was beginning to falter. But after studies demonstrated the ability of the drugs to suppress HIV blood levels, PEPFAR officials theorized that they could slow the epidemic by testing everyone regularly—testing teams even went door to door—and recruiting anyone who was positive into treatment immediately, even if they were still otherwise healthy. HIV-negative people are also encouraged to take the drugs prophylactically, in order to prevent infection in case of exposure. Thus, HIV prevention was to be accomplished almost entirely by drugs alone.

Trick and Treat

In 2013, PEPFAR launched a study comparing Test and Treat to treatment-as-usual in various communities in southern Africa. Some researchers, including James D. Shelton of USAID, Martina Morris, and myself, warned that mass HIV treatment might not work where concurrency was common. Ordinary HIV tests measure antibodies, which can take weeks or months to appear post-infection, but people with concurrent partners tend to pass the virus on within days or weeks of becoming infected themselves. The testing programs were also potentially coercive. In one trial community, reluctant research subjects had fled their homes when they saw the researchers’ van approaching, and one man reportedly died by suicide after being forcibly tested for HIV.

We’ll never know if more people would have escaped infection had programs to educate them about concurrency received more support.

In 2019, the results of the study were published in the New England Journal of Medicine: early HIV treatment had virtually no effect on HIV transmission, partly for the reasons we’d warned about, and also perhaps because HIV incidence was falling as general awareness spread. Nevertheless, HIV infection in many parts of eastern and southern Africa remains very common: some five hundred thousand people contracted HIV there last year. In South Africa’s Umkhanyakude district, where intensive PEPFAR activities—including Test and Treat, condom promotion, and AIDS education (minus concurrency)—have been carried out, over 4 percent of young women were still getting infected each year as of 2019, the last time data was reported. We’ll never know if more people would have escaped infection had programs to educate them about concurrency received more support.

I fell in love with public health because it helps people avoid doctors. My heroes cleaned up London’s cholera-infested drinking water, fought the lead paint companies, and educated mothers about breast feeding. This is real empowerment. It was therefore dispiriting to see the fight against AIDS— which in its early years had done so much to confront stigma, shame, cruelty, and despair—disintegrate into a mechanism for pushing pharmaceuticals, beneficial as they are, as the sole solution to everything.

It may become a moot point. Twenty years from its genesis, the bipartisan consensus that’s kept PEPFAR afloat is fraying: when it came up for reauthorization in September, far right Republicans in Congress approved it for just one year, rather than the usual five. Pro-life lobbyists, angry about new Biden administration rules allowing funding for groups that use funds from other sources for work on family planning and abortion, have called PEPFAR “a massive slush fund for abortion and LGBT advocacy.” This is, by any stretch of the imagination, a fallacy; the program is vital to the survival of tens of millions of people. But in the fluid reality in which we all now live, PEPFAR can be anything those with control over its budget want it to be. That’s not a good thing.