Responses to the global HIV/AIDS epidemic are often driven not by evidence but by ideology, stereotypes, and false assumptions. Referring to the hyperepidemics of Africa, an article in The Lancet this fall named “ten myths” that impede prevention efforts”including “Poverty and discrimination are the problem,” “Condoms are the answer,” and “Sexual behavior will not change.” Yet such myths are held as self-evident truths by many in the AIDS establishment. And they result in efforts that are at best ineffective and at worst harmful, while the AIDS epidemic continues to spread and exact a devastating toll in human lives.

Consider this fact: In every African country in which HIV infections have declined, this decline has been associated with a decrease in the proportion of men and women reporting more than one sex partner over the course of a year”which is exactly what fidelity programs promote. The same association with HIV decline cannot be said for condom use, coverage of HIV testing, treatment for curable sexually transmitted infections, provision of antiretroviral drugs, or any other intervention or behavior. The other behavior that has often been associated with a decline in HIV prevalence is a decrease in premarital sex among young people.

If AIDS prevention is to be based on evidence rather than ideology or bias, then fidelity and abstinence programs need to be at the center of programs for general populations. Outside Uganda, we have few good models of how to promote fidelity, since attempts to advocate deep changes in behavior have been almost entirely absent from programs supported by the major Western donors and by AIDS celebrities. Yet Christian churches”indeed, most faith communities”have a comparative advantage in promoting the needed types of behavior change, since these behaviors conform to their moral, ethical, and scriptural teachings. What the churches are inclined to do anyway turns out to be what works best in AIDS prevention.

This good news is often lost on organizations that purport to represent churches and the faith-based response to AIDS. The Berkley Center at Georgetown University, for instance, issued a report late last year called Faith Communities Engage the HIV/AIDS Crisis . The report is worth taking seriously, as it reflects the thinking of many international organizations, including many of the faith-based organizations that respond to AIDS. This thinking is often drastically out of sync with the culture and values of the beneficiaries. The Georgetown report claims to explore “development issues from the perspective of faith institutions,” but in fact the report betrays a deep ambivalence about whether faith communities, particularly Christian churches, are part of the problem or part of the solution to AIDS.

Katherine Marshall and Lucy Keough, lead authors of the report, are clearly uncomfortable with approaches to HIV prevention that emphasize sexual responsibility, behavior change, and morally based messages. They praise the work and compassion of faith communities in treating and caring for people ­living with AIDS and their families, yet harshly ­criticize the messages of faith communities for increasing the stigma of AIDS. Their discomfort with attempts to change sexual behavior is evident early in the report, when, for example, they muse: “Should the focus be on changing the behaviors that contribute to HIV/AIDS? (Is that possible? Desirable ? How? With what assurance?)”

If Marshall and Keough are undecided as to whether changing sexual behavior is even desirable in the context of an epidemic driven by people who have more than one sex partner, they then need to become educated in the basic epidemiology of HIV transmission. One must ask whether they are more concerned with upholding a Western notion of sexual freedom or with saving lives. Their concern over any prevention approach that might be “moralistic” causes them to miss entirely the evidence for the remarkable success of sexual-behavior change in reducing HIV infections. They miss, as well, the crucial contribution of faith communities to HIV prevention, even while they are producing a report on the role of faith communities in the HIV crisis.

Marshall and Keough reflect conventional wisdom when they blame poverty, gender inequality, powerlessness, and social instability for the spread of AIDS. Yet epidemiological evidence is increasingly challenging this wisdom. In Africa, for instance, the wealthy are more likely to be HIV-infected (as a 2007 study in AIDS and a 2005 report in The Lancet have both noted). The countries of southern Africa are both the wealthiest on the continent and the worst affected. Meanwhile, within many countries, the wealthy are most likely to be HIV-infected”and, surprisingly, it is often among women that the greatest difference in HIV prevalence between poor and wealthy is seen. For instance, in Tanzania, women in the wealthiest quintile of the population are more than four times more likely to be infected than women in the poorest quintile. Poverty may make some individuals prone to risky sexual behaviors that can spread HIV; yet wealth can facilitate lifestyle choices that increase HIV risk, such as living in an urban area, abusing alcohol, and having the mobility and opportunity to acquire extramarital sexual partners.

While gender inequality may severely circumscribe a woman’s right to choose or refuse sex, and while faithful women can be and are infected by their husbands, new data are showing that women also bring HIV into marriage, putting husbands at risk. Last year the researcher Damien de Walque showed that, for 30 to 40 percent of infected couples in five African countries, the woman alone was infected. Vinod Mishra similarly reported that in some African countries, among couples in which one partner was infected and the other was not, the woman, not the man, was infected in more than half of couples. Both studies conclude that women’s extramarital sex must be the predominant factor behind these surprisingly high rates of female-­discordant couples”and thus “be faithful” messages must be targeted to women as well as to men.

Although turmoil and instability may make people more vulnerable to HIV, it does not follow that an HIV-prevention strategy aimed at changing sexual behavior is doomed in circumstances of turmoil and instability. Many of the greatest successes in HIV prevention have been in situations of social, political, and economic turmoil, such as Uganda in the late 1980s and Zimbabwe in the early 2000s. Experts predicted that the HIV epidemic would explode in Rwanda, but it did not, in spite of extreme violence and instability and tremendous numbers of rapes. Sexual behavior in Rwanda has remained conservative, and, at 3 percent, HIV prevalence is low for the region.

Of course, many other reports”and more alarmingly, peer-reviewed articles”make the same mistake of repeating conventional wisdom that does not stand up to scientific scrutiny. But the report from Georgetown is guilty not only of poor epidemiology but also of ignoring the perspectives of faith institutions that it claims to put forth. Fortunately, faith communities seem to be going forward with what they can address”influencing sexual behaviors and norms in their own parishes and communities”and not heeding the warnings of experts that such efforts are doomed as long as poverty, gender inequality, and less-than-ideal political and economic conditions persist. But the blessing and backing of the AIDS establishment would surely energize this work.

Uganda provides an illustrative example of the central role of faith communities (among others) in bringing about behavior change. In a sidebar in Faith Communities Engage the HIV/AIDS Crisis , Marshall and Keough give credit to the work of faith communities in Uganda, but they get most of the story wrong. Their account emphasizes the role of increased condom use in bringing down Uganda’s HIV rates and downplays the dramatic increases in the number of people reporting abstinence and faithfulness behaviors. In making their case, Marshall and Keough cite a little-known (and non-peer-reviewed) World Bank report written by Keough herself, and they ignore the wealth of peer-reviewed literature showing that the critical factor in Uganda was not increased condom use but reductions in the number of sexual partners.

The list of countries that have seen both changes in sexual behaviors and declining HIV prevalence is growing and now includes Uganda, Kenya, Haiti, Zimbabwe, Thailand, and Cambodia, as well as urban areas of Ivory Coast, Ethiopia, Zambia, and Malawi. Many countries that have not seen declines in HIV have seen increases in condom use, but in every country worldwide in which HIV has declined there have been increases in levels of faithfulness and usually abstinence as well.

Arguably, every community and institution has been guilty of some fear, stigma, discrimination, and marginalization of those living with HIV. No faith community, including the Catholic Church, should claim to be immune, and, where stigma and fear exist, they should be openly admitted and confronted. Yet the Georgetown report treats faith communities particularly harshly, claiming that churches impose “retribution for ‘sinful behavior’” and that “religion has been used to foster stigma, exclusion, and marginalization related to HIV/AIDS.” Indeed, the report continues, “faith hierarchies, leaders, and communities have in the past often been promoters of stigma associated with HIV and AIDS, partly because of their difficulty in confronting aspects of human sexuality and partly because they often assume a link between AIDS and what they regard as sinful activities.”

Faith communities are, in fact, facing the challenge of upholding orthodox beliefs about sexuality without contributing to stigma. Rather than accurately reporting this, however, Marshall and Keough offer only their own perspective, insisting that religious beliefs about sexuality are “values structures” that “have tended to perpetuate stigmatization.”

This language is reminiscent of the campaign that appeared immediately after the Fourteenth International AIDS Conference in Barcelona in 2002. Such comments as “Religion kills” and “The only good priest is the priest who distributes condoms” flooded many of the more ideologically driven HIV/AIDS email listservs and online discussion groups. Within the international community, a religious group’s willingness to promote condoms was the unsubtle litmus test for funding in AIDS prevention until the United States Congress changed the discriminatory practice by law in 2003.

In Faith Communities Engage the HIV/AIDS ­Crisis , Marshall and Keough make a particular effort to discredit the ABC approach for preventing the sexual transmission of HIV ( A bstain, B e faithful, or use C ondoms). They write, “Many faith-based groups, like many governments, have been attracted to an approach to HIV/AIDS prevention, first articulated in Uganda, that has come to be known as the ABC model . . . . While aspects of this approach are incontrovertibly effective in reducing the spread of HIV/AIDS, the current consensus is that it does not go far enough.”

Whose consensus, one must ask? Are the authors truly representing the consensus of the world’s faith communities, or rather the consensus of a public-health community that is deeply uncomfortable with an approach that calls, in a simple and straightforward manner, for sexual responsibility? A more cynical view is that simple behavior changes such as mutual fidelity do little to contribute to a robust and ever-expanding multibillion-dollar “risk-reduction” AIDS industry focused on medical services, drugs, and devices such as condoms while leaving the true driver of the pandemic, sexual behavior, alone.

Since the beginning of the global epidemic, most AIDS programs have been designed solely with high-risk groups in mind. Risk reduction seems to have had some success among high-risk groups. (Although, in certain groups, such as American gay men, HIV is once again rising.) But a risk-reduction approach ignores a central epidemiological fact: The great majority of people worldwide are not at much risk for HIV infection, which in fact does not occur easily. Thus, encouraging the majority to maintain low-risk behaviors is the great missing piece of AIDS prevention.

The criticisms that Faith Communities Engage the HIV/AIDS Crisis levies against the ABC approach are hardly original and do not face up to the evidence that this approach has proved effective in various settings”so much so that it was endorsed by a landmark 2004 statement in The Lancet signed by more than 150 public health experts and leaders from around the world. Marshall and Keough claim that an ABC approach is insufficient because it does not ­recognize the role of voluntary counseling and testing (a measure that has been shown to have no effect in preventing new HIV infections, however important it is as a gateway to treatment); does not address prevention of mother-to-child transmission (a matter that the ABC approach, which targets sexual transmission, makes no claims to address); does not address the care of orphans and vulnerable children (clearly also beyond the scope of a prevention approach); and does not address women’s risk of becoming infected even if they do practice faithfulness. This is akin to criticizing smoking-cessation programs because they do not ­provide chemotherapy for those suffering from lung cancer or do not impose regulations on secondhand smoke and air pollution.

The Georgetown report clearly gets it wrong when it states that, for the ABC approach “to be effective, abstinence and fidelity must be practiced by both partners.” In fact, abstinence is always 100 percent effective in preventing sexual transmission when practiced by an individual . As for fidelity, it is certainly true that sexually faithful people may be infected by unfaithful partners”but this is true for men as well as for women. Proponents of the ABC approach do not claim that it confers total protection”for one thing, even consistent condom use reduces risk by, at best, 80 to 90 percent. Yet people (even women whose husbands are unfaithful) can reduce their own risk by choosing to practice faithfulness. More important, when ABC behaviors are promoted at a population level, risky ­sexual behaviors (particularly multipartner sex) are reduced, and a population-level decline in HIV infections is seen.

Marshall and Keough promote the SAVE approach, developed by ANERELA+, a network of African clergy led by Gideon Byamugisha. ( SAVE stands for S afe sexual practices, A ccess to treatment, V oluntary counseling and testing, and E mpowerment.) “The objective in developing such a new approach,” the authors explain, “is to move away from judgmental, moralizing stigma, and towards a more positive approach.” The problem with SAVE, however, is that three of the four components have already been demonstrated to have no effect on reducing new HIV infections. Only the S , safe sexual practices, truly addresses prevention”and in a sufficiently vague way that it provides no clear call for changes in sexual behavior that will actually reduce transmission. Moreover, in the AIDS world, “safe sex” is understood to mean condom use. Criticizing the ABC approach has evidently been something of a crusade for Byamugisha, an Anglican priest, as he has made clear in multiple public statements. Byamugisha does not represent the views of most Ugandan or African clergy, and the SAVE approach is more a political statement than a guide to AIDS prevention.

The Georgetown report tells us: “While the ‘mainstream’ HIV/AIDS program and global communities accept that widespread availability of condoms and promotion of condom use are major elements in successful HIV/AIDS prevention strategies, a focus on condoms is contentious for some religious communities because it contradicts the core recommended strategy of abstinence before marriage and faithfulness within marriage.”

In fact, the mainstream HIV/AIDS community has continued to champion condom use as critical in all types of HIV epidemics, in spite of the evidence. While high rates of condom use have contributed to fewer infections in some high-risk populations (prostitutes in concentrated epidemics, for instance), the situation among Africa’s general populations remains much different. It has been clearly established that few people outside a handful of high-risk groups use condoms consistently, no matter how vigorously condoms are promoted. Inconsistent condom usage is ineffective”and actually associated with higher HIV infection rates due to “risk compensation,” the tendency to take more sexual risks out of a false sense of personal safety that comes with using condoms some of the time. A UNAIDS-commissioned 2004 review of evidence for condom use concluded, “There are no definite examples yet of generalized epidemics that have been turned back by prevention programs based primarily on ­condom promotion.” A 2000 article in The Lancet similarly stated, “Massive increases in condom use world-wide have not translated into demonstrably improved HIV control in the great majority of countries where they have occurred.”

Faith communities are not shutting their eyes to evidence when they choose to emphasize the “core recommended strategy of abstinence before marriage and faithfulness within marriage.” These behaviors have, in fact, proved far more effective than condom use in curbing HIV transmission for the vast majority of any population. A 2001 study of condom use in rural Uganda found that only 4.4 percent of the population reported consistent usage in the previous year, a rate that is probably typical of much of Africa. In contrast to the estimated 95 percent or more of Africans who did not practice consistent condom use in the past year, studies from all over Africa show a solid majority of men and women reporting fidelity over the past year, with a majority of unmarried young men and women reporting abstinence.

The Georgetown report devotes several paragraphs to the position of the Catholic Church on condom usage and the apparent “nuance” within Catholic communities on the issue. The report seems to imply that the Church’s teaching on condom usage is detrimental to the fight against AIDS, while recognizing the Church’s contribution to prevention through promotion of abstinence and faithfulness. (For instance, the authors note that Pope John Paul II chose to emphasize abstinence and faithfulness rather than directly criticizing condom use.)

The report also erroneously claims that Protestant evangelicals are “among the staunchest supporters of the U.S. Government PEPFAR (President’s Emergency Plan for AIDS Relief) earmark for ‘abstinence only’ prevention programs.” This is mistaken. There is no such “abstinence only” earmark within PEPFAR, nor are the great majority of Protestant groups who receive PEPFAR funds implementing abstinence-only programs. Current PEPFAR guidance recommends that two-thirds of funds for the prevention of sexual transmission of HIV be allocated to abstinence-until-marriage and faithfulness or partner-reduction programs. This amounts to less than 7 percent of PEPFAR funds. Among recipients of these funds, faith-based organizations such as World Vision, World Relief, and Samaritan’s Purse implement programs that emphasize abstinence and faithfulness but also include accurate information on condoms”in other words, a comprehensive ABC approach, the approach known to work best.

Marshall and Keough are right to call faith communities to action in defending the rights of women and protecting women and girls from violence, coercion, and exploitation. Yet the presence of gender inequality does not negate the need for, and effectiveness of, approaches that focus on sexual responsibility and behavior change. On the contrary, central to faithfulness interventions”as stated clearly in the PEPFAR Guidance document for implementing “B” programs within a context of ABC”is the focus on changing male behavior in particular.

If protecting highly vulnerable women and girls in patriarchal societies is a genuine goal rather than a political posture, then there must be explicit strategies for discouraging men from sexual abuse, rape, infidelity, and seduction of minor females. Furthermore, women must be empowered to refuse unwanted sex (as one of us, Edward Green, has been arguing in publications since 1988), not simply to “negotiate condom use.”

Thus far, research has produced no evidence that condom promotion”or indeed any of the range of risk-reduction interventions popular with donors”has had the desired impact on HIV-infection rates at a population level in high-prevalence generalized epidemics. This is true for treatment of sexually ­transmitted infections, voluntary counseling and ­testing, diaphragm use, use of experimental vaginal microbicides, safer-sex counseling, and even income-­generation projects. The interventions relying on these measures have failed to decrease HIV-infection rates, whether implemented singly or as a package. One recent randomized, controlled trial in Zimbabwe found that even possible synergies that might be achieved through “integrated implementation” of “control strategies” had no impact in slowing new infections at the population level. In fact, in this trial there was a somewhat higher rate of new infections in the intervention group compared to the control group.

The one medical intervention that has now been proven effective according to the highest standards of scientific research is male circumcision, which reduces a man’s risk of HIV transmission by more than half. Lack of male circumcision, along with high rates of long-term concurrent sexual partnerships, likely accounts for the hyperepidemics of southern Africa. But even many advocates of male circumcision believe that it needs to be promoted along with partner ­reduction.

Meanwhile, the other interventions that have generally been called “best practices” simply do not seem to work in generalized epidemics, even though they are still applauded loudly at global AIDS conferences, while mention of fidelity and abstinence is received by booing, as Bill Gates discovered at the International AIDS Conference in Toronto in 2006. If we are to progress beyond science-by-popular-acclaim, we must accept that the evidence is much stronger for fidelity or partner reduction than for any of the standard-package HIV-prevention measures”in Africa at least”and so we need to rethink and reprogram AIDS-prevention interventions.

Admittedly, changing direction is hard when there has been massive investment in these “best practices.” It is not in the interest of a multibillion-dollar global AIDS industry to endorse interventions that are low-cost and homegrown and that rely on simple behavior change rather than medical products or services provided by outside experts. And so the major donors of AIDS programs continue to do the same things, expecting different results. The authors of the Georgetown report reflect this popular but misguided opinion, despite mounting evidence to the contrary.

That’s a shame, for a report like Faith Communities Engage the HIV/AIDS Crisis offered an opportunity to rethink the failing group consensus and to point toward the central fact that has emerged from all the recent studies of the HIV epidemic: What the churches are called to do by their theology turns out to be what works best in AIDS prevention.

Edward C. Green is the director of the AIDS Prevention Research Project at the Harvard Center for Population and Development Studies, where Allison Herling Ruark is a research fellow.