In February 2000, after more than a year of discussion, the Joint United Nations Program on AIDS (UNAIDS) issued its guidance document, Ethical Considerations in HIV Preventive Vaccine Research. This is, then, an appropriate moment to look back at the effort at consensus development on this matter of critical importance. An analysis of the conflicts that characterized the 3-year effort at achieving international agreement will underscore the extent to which profound and important disagreements continue to characterize discussion of key policy questions surrounding vaccine development.
Early in 1998, Barry Bloom, chair of the UNAIDS vaccine advisory committee, made clear the challenge posed by the worldwide epidemiology of HIV infection. Sixteen thousand individuals were becoming infected each day, 5.8 million per year. Ninety percent of these infections were in less developed countries. The dramatic advances in the management of HIV disease in the world's economically advanced nations – reliance on highly active antiretroviral therapies – was beyond the reach of the vast majority of those in need. In light of this grim picture, Bloom concluded, `the scientific community has increasingly come to believe that the best hope for stemming the global epidemic is the development of a preventive vaccine' . The challenge is daunting as failed hopes and promises over the past decade make clear. Writing in mid-1997, Margaret Johnston, then of the International AIDS Vaccine Initiative, summarized the state of affairs.
Whereas the prediction was made in 1983 that developing an effective AIDS vaccine would take only a few years, we no longer expect one by the year 2000 and probably well after that.  (emphasis supplied).
It is in the face of social urgency and scientific limits that research has gone forward, provoking a host of critical ethical questions. Analyses of such questions go back for more than a decade, the most skilful of which have underscored the central challenges in a way that remains pertinent to the current state of affairs [3,4]. Among the most complex matters has been the relationship between the scientific imperatives of vaccine trials and the role of behavioral interventions designed to reduce the incidence of HIV infection. Very early on, it became clear that a tension existed between the goal of determining the efficacy of candidate vaccines and the ethical obligation to protect trial subjects from the enhanced risk of becoming infected with HIV because of assumptions about the protective effect of the vaccines under study. Both science and economics would have dictated creating trials to demonstrate clearly and efficiently the relative efficacy of the vaccines under investigation. The ethics of research, on the other hand, required that trial participants be provided with counseling about the need to protect themselves from HIV both because candidate vaccines could prove ineffective and because of the possibility that one was being given a placebo. To the extent that such interventions were effective not only in reducing the threat of enhanced risk but in reducing the incidence of infection below that which prevailed in the community from which subjects were drawn, it would have been even more difficult, costly and time consuming to distinguish the effects of a candidate vaccine under investigation.
Despite this important tension, an international ethical consensus emerged that no vaccine trial could be conducted in the absence of protective behavioral interventions. Less clear, and still a matter of contention, is the yardstick by which to judge the adequacy of such interventions. Should they reflect that which is considered standard or even optimal in sponsoring countries despite the costs involved? Should they mirror the best that is available in countries similar to the host country? Should they reflect prevailing practice in the host country even if such practice falls below that which is recognized to be optimal, or even appropriate, in countries of similar socio-economic status? At stake in this dispute are matters that touch not only on the cost of trials but the extent to which the efficacy of vaccines ought to be measured against varying epidemiological, social, and economic conditions in differing nations.
These matters, so central to the conduct of vaccine trials, set the context for the discussion that follows. Rather than providing a summary of the full range of ethical issues that have emerged over the past decade, I want to focus on the remarkable 3-year effort at international collaboration (1997–2000) designed to forge a worldwide consensus that would both protect subjects enrolled in vaccine research while paving the way for scientific developments, a process that culminated in the issue of the UNAIDS Guidance. This discussion will focus on three broad questions. (1) What role should the principle of informed consent play in protecting people from research they choose not to be involved in and in assuring them access to trials in which they choose to participate? (2) What level of care is due to research participants who become infected with HIV during vaccine trials? (3) What ought to be the relationship between ethical review of proposed research in sponsoring and host nations?
Confronting the challenges: first steps toward consensus
In September 1997, UNAIDS convened a meeting in Geneva, which brought together vaccine researchers, ethicists and social scientists to begin the process of examining the ethical issues that would arise in the conduct of AIDS vaccine trials in less developed countries. (Ethical Considerations in Clinical Trials of Preventive HIV Vaccines: a Preparatory Meeting, 23–24 September 1997, Geneva; unpublished report.) The focus on the world's poorest nations was dictated both by the epidemiology of HIV and the recognition that those nations were most vulnerable to exploitation and were less well positioned to put forward standards to protect their own populations from the risks associated with research. A broad array of ethical issues was identified. Those were to be the subject of further analysis at special regional meetings to be held in Latin America (Brazil), Asia (Thailand) and Africa (Uganda). The discussion that took place at those sessions revealed important fissures within and among the world's poorest nations on critical matters touching on vaccine trials. (UNAIDS-Sponsored Regional Workshops to Discuss Ethical Issues in Preventive HIV Vaccine Trials. Unpublished final report.)
At each of the sessions there was consensus that, despite vast differences in cultures, the bedrock of the ethics of research was the requirement that each participant give his or her informed consent to participation. That consensus was, in and of itself, not surprising given the international commitment to the principle of consent dating back more than 50 years. From the Nuremberg Code of 1947, which, in the wake of Nazi atrocities, underscored the importance of `voluntary consent' to research, to the International Covenant on Civil and Political Rights of 1966, which stated that `no one shall be subjected to torture or to cruel, inhuman or degrading treatment or punishment – in particular, no one shall be subjected without his (her) free consent to medical or scientific experimentation' , the protection of potential subjects from the risks or burdens of research had been the central mission of research ethics.
But the enunciation of a broad principle of informed consent only set the stage for further and more complex considerations. First were those that entailed providing subjects with information that would permit them to understand the nature of the trials to which they were being recruited. Involved were not only generic issues – how does one explain to individuals of radically different cultural backgrounds and with vastly different levels of education the meaning of randomization, and the nature of a placebo? – but those associated particularly with HIV vaccines: the fact that a vaccine might be only partially effective; that adherence to safe sexual norms remained crucial; that involvement in one trial might preclude participation in a later trial; that participants might be exposed to discrimination.
Second, there were troubling questions about the extent to which potential subjects might fully appreciate the extent to which they, in fact, retained the right not to participate in a trial or to withdraw from a trial at any moment. This was a matter of particular significance given the relationship between incidence of infection and social vulnerability. If the demands of design dictated the selection of research subjects from populations characterized by relatively high incidence levels of HIV, and if high incidence were associated with greater social vulnerability, then the demands of research might dictate the targeting of those from whom it might be most difficult to obtain truly informed consent.
The Latin American nations that convened in Brazil underscored the ethical issues involved; their understanding was reflected in the Asian and African sessions as well.
It is not ethical to conduct a trial involving groups or individuals in which there are significant problems in obtaining [informed consent]. When obtaining individual informed consent is difficult, a sponsor might consider initiating activities which will aid the individual, the population, or the culture to develop the capability to give individual informed consent.
In recognizing that informed consent might be difficult to elicit and in calling for interventions that might enhance the prospect for obtaining consent, those who considered this issue had demonstrated a sophistication too often missing from discussions of this matter. Absent, however, was a forthright recognition of the price that the proposed standard could exact in terms of the possibility of constructing trials among those who evidenced especially high levels of HIV infection.
Almost exclusively, the focus of attention at each of the working sessions was on situations that might entail risks of coercion to participate in trials. It was in that context that cognizance was given to situations in which community leaders might be expected to give consent for research before individuals could be approached for individual consent. However important such communal consent might be, it could not substitute for obtaining the agreement of each individual who was to participate. But in accepting such socio-cultural standards and practices, no attention was given to the possibility that individuals who might want to participate in trials could be precluded from so doing were community leaders to withhold approval. In so doing, those who met to consider these issues reflected the conventional perspective of research ethics, which has given far more attention to protecting individuals from research than upon protecting the rights of those who would choose to be research subjects. This is especially noteworthy since one of the remarkable changes that has occurred in advanced industrial nations in the years since the onset of the AIDS epidemic has been a rethinking of the protective impulse of research regulation, premised on the assumption that involvement in research represents a burden rather than a benefit.
The centrality and limits of the principle of consent were most dramatically underscored in the discussion which centered on the participation of women in general and, more specifically, those who were pregnant. Each of the consultative sessions endorsed, in principle, the right of women to be research subjects, although the African meeting noted without criticism that there might be circumstances in which women could not give consent without first obtaining the approval of either their husbands or their families. Far more controversial was the issue of pregnancy. At the center of the debate was whether the potential for fetal toxicities should be presented to women who could then choose whether to take such risks or whether they should serve as a basis for excluding women from vaccine trials. Although a minority of those who met in Brazil asserted that fully informed pregnant women should be permitted to enroll in trials, a majority held that `paternalism is justified for the protection of the fetus'. Those who met in Thailand adopted a far more restrictive posture. A majority held that `those who are pregnant, likely to become pregnant, or breast feeding should be barred from any initial trial and enrolled only following phase 3 trials in other adults and with adequate animal data on teratogenicity'. Within the Ugandan session there was division on this matter as well.
Access to treatment
Far more controversial than the issue of informed consent was the question of the level of care vaccine trial sponsors should be obligated to provide to those who become infected during the course of study. Here the terrain of conflict shifted from that which centered on the rights of the individual to that which involved the socio-economic inequality within which the HIV epidemic is embedded. Although the controversy involved elements that had been considered in discussions about the kind of behavioral interventions that should accompany vaccine trials, it was another conflict that shaped the discussion – the furious international debate that had just been waged on the ethics of placebo-controlled clinical trials designed to find an inexpensive regimen to reduce the risk of HIV transmission from pregnant women to their fetuses .
That trial was sparked by the 1994 finding in clinical trial 076 that the provision of zidovudine to pregnant women during their second and third trimesters and to their babies just after birth could radically reduce the risk of vertical HIV transmission . But that regimen, costing about $900 per pregnant woman, was beyond the reach of the nations where pediatric AIDS represented such a terrible burden. The placebo-controlled trials designed to find less expensive interventions clearly would have been unethical in the advanced industrial world where the 076 regimen was affordable. But were they acceptable in poor nations? Did such research represent a double standard for the protection of research subjects? Among the most bitter critics of the placebo-controlled trials was Marcia Angell, Executive Editor of the New England Journal of Medicine, who argued that international ethical standards barred the use of placebos when a known effective intervention was available:
only when there is no known effective treatment is it ethical to compare a potential new treatment with a placebo. When effective treatment exists, a placebo may not be used. Instead, subjects in the control group of the study must receive the best known treatment .
Given this premise, Angell rejected as irrelevant the fact that health care available in most poor nations provided nothing like the health care available in industrialized countries. Citing the Declaration of Helsinki for authority, she noted that the control groups must be provided with the best current therapy, not simply that which was available locally.
The shift in wording between `best' and `local' may be slight, but the implications are profound. Acceptance of this ethical relativism could result in widespread exploitation of vulnerable Third World populations for research programs that could not be carried out in the sponsor country .
Those who supported the trials argued that they were dictated by the very poverty of the nations within which they occurred. To demand that the 076 regimen be used as the standard against which to test the less costly interventions ignored the need to determine whether less costly interventions could provide some benefit .
The questions posed by vaccine trials were not precisely the same, but they did pose the question of whether the conduct of research should differ between rich and poor nations, and whether such differences represented the application of differential standards of protection. Were an individual in an AIDS vaccine trial in the United States to become infected with HIV, there is no question but that the individual would be offered highly active antiretroviral therapy. Given that standard, were subjects in poor nations, where such therapy was unaffordable, entitled to similar care?
Those who convened in Brazil were most sympathetic to the idea that those who became infected `should be provided [care] at the level of that offered in the sponsoring country, [that it] should continue at least for the duration of the trial, and [that] further provision should be negotiated'. In addition, there was considerable support for the proposition that `the need to measure secondary endpoints was not . . . a valid rationale for withholding early treatment if treatment was proven to be effective'. Only a minority of those who attended the session were willing to acknowledge that there could be any exception to the duty to provide the `best proven' therapy. And this they did reluctantly. In sharp contrast, both the Thai meeting and that which was held in Uganda rejected the notion that trial participants were due the type of care provided in sponsoring nations.
For those who contract HIV infection during the course of the trial but not as a result of the trial, treatment should be provided at a level consistent with that available in the host country. There is no imperative to provide a level of care consistent with that in the sponsoring country, or with the highest available in the world.
The clash, which touched on fundamental matters of principle, conceptions of international social equity, and the dictates of harsh realities, also reflected the commitments of the nations that hosted the consultative sessions. Both Thailand and Uganda had been eager participants in the trials designed to find markedly less costly alternatives to the regimen established by clinical trial 076 to reduce maternal–fetal transmission of HIV and had rejected the criticism of those who viewed them as profoundly flawed. Brazil, on the other hand, had made an extraordinary budgetary commitment to provide antiretroviral therapy to all of its HIV-infected citizens. From the Brazilian viewpoint, the claims that the local standard of care should provide the yardstick for judging the care given to those in vaccine trials seemed utterly regressive, a failure on the part of sponsoring nations and the international community to make the kind of commitment that it believed the ethics of research required.
The ethics of review
Finally, each of the consultative meetings underscored the importance of strengthening the ability of host nations to conduct ethical review of proposed vaccine research. In so doing, they acknowledged that just as the capacity for scientific review might be limited, the skill to undertake careful analysis of the ethical issues posed by vaccine research might require capacity building. Thus the Thai meeting concluded that `an effective ethical review mechanism [had to be present] before a vaccine trial may proceed', and that either sponsoring countries or independent international bodies `should provide support in this area'.
The substance of the discussions in the Asian, African, and Latin American sessions took as a starting point but did not challenge a principle enunciated in the Council for International Organizations of Medical Sciences (CIOMS) International Ethical Guidelines for Biomedical Research Involving Human Subjects : that externally sponsored research had to be subject to review in both sponsoring and host countries; that the ethical standards applied by the former to the latter had to be `no less exacting than they would be in the case of research carried out in that country' . An issue that was never addressed was one that was first brought into bold relief in the controversy during the international trials to reduce the risk of maternal/fetal HIV transmission and that will most certainly arise in the conduct of vaccine trials, namely: should host countries be permitted to go forward with research which they take to be ethically acceptable but which has failed to receive approval by the appropriate ethical review body in the sponsoring nations?
The roots of dual review are to be found in the concern of the architects of the CIOMS guidelines about the potential for exploitation of vulnerable populations in less developed nations. Only by holding research to the standards of sponsoring nations would it be possible to avoid the selection of research sites that might tolerate investigations that could not be undertaken under the more exacting standards that applied in sponsoring nations. Such a protective posture is now viewed, by some, as a vestige of an out-dated paternalism at best, of a neo-colonialist outlook at worst. Why, critics assert, should a democratic nation with a commitment to human rights be held to standards that reflect the needs of wealthy and more powerful nations, no matter how well intended? (H.M. Coovadia, personal communication.)
Striving for consensus
When the UNAIDS advisory committee reconvened in July 1998, both the agreements and the divisions that had been expressed in the three regional meetings were mirrored in its sessions . Although consensus was reached on the importance of monitoring the validity of the informed consent process and on a modification of the requirement that vaccines be tested first in sponsoring nations, sharp conflict arose on the question of the level of care to be provided to those who would become infected during the course of vaccine trials.
Mary Lou Clements-Mann made the case for a differential approach on the basis of the scientific issues at stake. In nations where antiretroviral treatment was unaffordable, the function of a vaccine would be, she asserted, to affect viral load and disease progression. But if the treatments available in wealthy nations that reduced viral load and slowed disease progression had to be provided to those in poor nations, `it would be impossible to design a scientifically valid vaccine trial'. A participant from Thailand argued that the provision of the level of care accorded in industrial nations would not be sustainable at the end of the trial. It would, said Dwip Kitayoporn, be `like leaving a Cadillac or Rolls Royce in our country, but no one can afford to drive or even repair it'. A Ugandan delegate warned that promising such care, even for the duration of the trial, would serve to draw participants in a way that would almost certainly violate the ethical stricture against undue inducements. Finally, as was to be expected, the coordinator of an AIDS vaccine center in Brazil, Dirceu Gerco, asserted that providing inferior care to research participants in poor nations would set the stage for a dangerous, slippery slope. `When you put the level of ethics below the maximum, it's very easy to lower it more.'
Although a majority of those in attendance rejected the Brazilian position, and although advisory committee chair, Barry Bloom, had long been associated with support for the `highest attainable standard', consensus eluded the session. Reporting to the XII World AIDS Conference in Geneva just days later, Ruth Macklin, an ethicist member of the committee, was compelled to acknowledge that in the absence of substantive agreement, the meeting's participants could only agree on a procedural approach that would allow for the equitable negotiation of this issue in each research setting (R. Macklin, personal communication).
In November 1998, UNAIDS issued a draft Guidance Document on Ethical Considerations in International Trials on HIV Preventive Vaccines which sought to balance the `ethical imperative to develop effective vaccines' with the `imperative to safeguard the rights and well being of the human subjects who participate in trials'. Most striking was not the extent to which the draft sought to apply prevailing international standards on the ethics of research but, rather, the open recognition that the demands of vaccine development in the context of an epidemic ravaging poor nations required revision in some prevailing norms. Most boldly, the draft declared:
[The CIOMS] guidelines . . . are grounded in an assumption that the so-called `developing' countries are vulnerable in morally relevant aspects, and, for that reason, focus on protecting such countries and their residents from harm and exploitation. This protectionist attitude, which prevailed at the time that those Guidelines were written, was justified by some early experiences. The current Guidance Document reflects a widely supported movement away from excessive paternalism of the recent past to a recognition of the entitlement of all countries, including the more vulnerable, as well as the less vulnerable, to be self-determining.
Despite that broad assertion, the Guidance Document never addressed the question of whether countries judged to have the capacity for ethical review ought to be able to proceed with vaccine trials that had not passed ethical review in sponsoring nations. In that regard, the document remained loyal to the protectionist ethos that informs the process of dual review.
The antipaternalism of the draft Guidance Document is most clear in its forthright rejection of conventional protectionist and exclusionary standards regarding pregnant women, so much in evidence in the three regional meetings.
Pregnant women, as a class of persons, should not be excluded from phase II and III . . . trials . . .. Pregnant women are autonomous decision makers and should not be treated as vulnerable or impaired because of their condition.
Although stipulating the importance of conducting mutogenicity and teratogenicity studies before pregnant women were enrolled in trials, the draft makes clear that such information was to be used as part of the informed consent process rather than as a basis for barring pregnant women from studies.
On the bitterly disputed question of the level of care due to those who might become infected during the course of vaccine trials, the draft Guidance Document clearly rejected the `Brazilian position' as well as those readings of the CIOMS guidelines that would have required the provision of `best proven' therapeutic interventions. Instead, the document stipulated `the level at which such treatment should be provided should be negotiated and agreed among and between all concerned parties before the vaccine trial is initiated'. Only in so far as it insisted that care not fall below the `highest practically attainable' local standard did the draft stipulate a substantive principle.
UNAIDS issues its guidance
To the surprise of many who had labored to fashion the November 1998 statement for UNAIDS, but to the delight of those who had seen in that guidance an abject capitulation to the forces attempting to subvert the principles of international research ethics, the draft report met with the dissatisfaction of the UNAIDS leadership. Both the tone and substance of the draft were subject to withering internal criticism. Perhaps most important, there was grave concern that, as written, it would be viewed as a justification for lowering the standards for the ethical review in the poorest nations. It would take more than a year to redraft the statement so that it could meet the approval of Peter Piot, the Director of UNAIDS. (Ethical Considerations in HIV Preventive Vaccine Research: UNAIDS Guidance Document. Unpublished pre- publication document.)
Strongly worded to make clear the importance of protecting the most vulnerable from the threat of exploitation, the Guidance nevertheless deleted the November 1998 Draft language on the need to confront the `excessive paternalism' that inspired aspects of the CIOMS guidelines on research involving poor nations. Like the earlier document, the UNAIDS Guidance stressed the importance of ethical and scientific capacity building in host nations, but did not address the potential for clashes between host nations with the capacity for ethical review that decided to proceed with trials despite the failure to pass ethical review in the sponsoring nations. Like the earlier draft, the Guidance rejected the exclusion of pregnant women from vaccine trials and stressed the importance of assuring that only those who gave individual informed consent be enrolled as research participants. Where the final Guidance reflected a marked shift from the earlier draft was with regard to the much-contested issue of the standard of care due to those who became infected with HIV during the course of vaccine trials.
Acknowledging the absence of consensus on this matter, the Guidance nevertheless positioned itself much closer to those who had argued for a universal `best proven therapy' standard. Stressing the importance of negotiations between host nations and sponsoring parties, and the importance that agreed levels of care should not fall below `the highest level of care attainable in the host country', the Guidance made clear that the aim of negotiations should be `achieving, as closely as possible, the ideal of provision of the best proven therapy for trial participants, in light of relevant conditions and concerns'.
In commenting on this revision, Peter Lurie and Sydney Wolfe (of the US-based advocacy organization Public Citizen Health Research Group), who had so bitterly attacked the November 1998 draft and who had strongly opposed any deviations from the best proven therapy standard, noted:
[While] this is an improvement over the previous version, for it at least puts forth a standard of treatment to be aspired to should a trial participant acquire HIV . . . this [guidance] provides just enough wiggle room for researchers to provide no or inadequate treatment . . . in countries where antiretrovirals are not available. This is what many of the researchers were seeking. (Letter to Peter Piot, February 11, 2000.)
The encounter with the challenges of international vaccine trials at this desperate moment in the AIDS pandemic has compelled a rethinking of a number of propositions that had at one point been central to thinking about the ethics of human subjects research. Most significantly, in an epidemic in which the burden of disease and death is increasingly borne by the poorest nations, the encounter has required a careful evaluation of how research that can only be funded by the richest and most powerful nations or by international agencies can move forward in a way that is respectful of both the needs and rights of the most vulnerable. But, in the end, attention to the ethics of the conduct of research will represent nothing more than an interesting – albeit important – diversion if a commitment to the provision of potentially effective vaccines is not assured to those who will most need them. The UNAIDS Guidance appears to address this issue in a forthright manner.
Any preventive vaccine demonstrated to be safe and effective . . . should be made available as soon as possible to all participants in the trials in which it was tested as well as to other populations at high risk of HIV infections. Plans should be developed at the initial stages of HIV vaccine development to ensure such availability.
But unlike the relatively small cost implications of the majority of the recommendations put forward by UNAIDS in its Guidance (even those involving the standard of care for research participants), this proposition would ultimately entail substantial resource commitments. It is that which gives this standard a desperately aspirational tone. Lurie and Wolfe call it `toothless'.
There are too many examples of critically important and relatively inexpensive treatments that remain unaffordable to those who need them to assume that the issue of accessibility will take care of itself once a vaccine is proven effective. The most elemental conception of solidarity and fairness dictates that the international community commit itself to global provision to meet the global threat of AIDS. However, whether the international community will make the necessary commitments in deed as well as word remains utterly unclear.