Any program seeking to provide antiretroviral treatment (ART) to the many patients in need is bound to confront ethical dilemmas. Dilemmas, as we know, are situations in which decisionmakers are faced with a choice between equally unsatisfactory alternatives. Yet those in charge must make a decision or establish a policy that takes one pathway to the exclusion of another. Reasonable people may disagree over the choice, arguing that an alternative selection would have been ethically superior.
At the end of their fascinating account of the work of Médecins Sans Frontières (MSF) in Khayelitsha, Renée Fox and Eric Goemaere identify six items they refer to as “persistent ethical questions” that continue to plague the MSF program. In what follows, I reframe four of these questions in the form of ethical dilemmas, provide an analysis of the nature of the dilemmas, and suggest where tentative solutions might be found. The remaining two ethical questions are more easily resolved and do not take the form of genuine dilemmas.
MSF's Social Selection Criteria and Policies
Let me begin with the two questions I claim are more readily resolved than the four dilemmas. They arise out of two of the policies MSF had established for its treatment program. The first involves one of the social criteria for the selection of patients to receive ART: the initial requirement that those accepted for treatment be permanent residents of Khayelitsha township. Now that ART is more widely available in South Africa, MSF has abandoned this requirement and accepts patients for treatment who can be transferred to another treatment center in a different geographical location. Fox and Goemaere ask whether this change in the original residency requirement is justified, given the current, much wider availability of ART in South Africa.
The second question relates to MSF's policy of initiating ART for pregnant women to prevent mother-to-child transmission of HIV, and then discontinuing the treatment after delivery of the baby if the woman wishes to return to the rural interior of the country. Fox and Goemaere ask whether this policy is justifiable, given the wider availability of ART in the geographical area to which the woman plans to return. This question admits of two possibilities: Either the MSF program could refuse to accept such pregnant women in the first place, seeking instead to ensure that ARV treatment begins in their home territory, or the program could provide initial treatment for the women, as before, and make similar arrangements for continuation of treatment in a treatment center close to her home.
Both of these policies were established at a time before the scale-up of ART in South Africa, when the MSF program was one of the very few providing antiretroviral drugs. With the increasing availability of ARVs in South Africa and elsewhere in sub-Saharan Africa, policies developed at an earlier time may no longer be relevant. The AIDS pandemic is a fast-moving target, necessitating appropriate changes in policies and practices where they are needed. In the cases of both the social criterion requiring permanent residency in Khayelitsha and the practice of discontinuing ART in pregnant women after delivery of the baby, changing circumstances justify a change in policy.
There are additional factors to consider in the changing environment of HIV/AIDS. It is more than a year since the “Three Ones” principles were established and endorsed at a consultation that included major industrialized-country donors and developing countries affected by the AIDS pandemic, including South Africa.1
The Three Ones principles were launched in Washington on April 25, 2004, by UNAIDS in collaboration with national HIV/AIDS programs, bilateral donors, and the Global Fund. The purpose was to reinforce the commitment by international stakeholders to harmonize the response to the HIV/AIDS epidemic.
The various contributors to the scale-up of ART in South Africa and elsewhere include the Global Fund to Fight AIDS, Tuberculosis and Malaria,2
The Fund is a financial instrument created to increase resources to fight these three devastating diseases. It operates as a partnership between governments, civil society, the private sector, and affected communities. Available at: http://www.theglobalfund.org/en/about/how/ (accessed June 29, 2005).
WHO/UNAIDS. Treating 3 Million by 2005: Making it happen. The global initiative of the World Health Organization and UNAIDS to provide anti-retroviral treatment to 3 million people by the end of 2005. The initiative was launched on December 1, 2003. Available at: http://www.unaids.org/en/treat3millionby2005initiative.asp (accessed June 29, 2005).
A U.S. bilateral program providing US$15 million over a five-year period for prevention, treatment, and care in 15 developing countries. Available at: http://www.usaid.gov/our_work/global_health/aids/pepfarfact.html (accessed June 29, 2005).
Although Fox and Goemaere do not raise questions about the ethics of MSF's other social selection criteria, the requirement that a patient have a “treatment assistant” and the use of alcohol or other substances as a contraindication for initiating ART could be ethically problematic. It is true that evidence exists that some individuals who habitually use intoxicating substances have a lessened capability for adherence, but other evidence from harm reduction programs has shown that drug injection users can be highly adherent to ART. Selecting individuals who have family or other social supports could discriminate against unmarried individuals or those lacking social supports, especially members of marginalized groups. At the same time, application of this eligibility criterion could result in excluding people antecedently suspected of being nonadherent (such as those who use drugs or alcohol), even though their family or other social supports would be likely to improve their prospects for adherence.
Nevertheless, one of the main points Fox and Goemaere make is that individuals who do not meet all the selection criteria are admitted into the ARV program anyway. In that case, the stated criteria do not have the effect of discriminating against any individuals or groups. It is reasonable to wonder, then, why MSF continues to maintain a set of social and adherence selection criteria for acceptance into the ARV treatment program. If, de facto, the system is “first come, first served” (modified only by “jumping the queue”), it could be more honest to abandon the pretense of having such selection criteria.
Four Ethical Dilemmas
The four ethical questions cited by Fox and Gomaere, which I have reframed as ethical dilemmas, are as follows: (1) the need to choose between the quantity of persons treated versus quality of treatment provided; (2) whether to adhere to the principle, “first come, first served” or make exceptions that enable some individuals to “jump the queue”; (3) whether to require patients to have a “treatment assistant,” thereby necessitating disclosure of HIV status, or adhere to the ethical obligation to preserve confidentiality; (4) whether to stop ART for nonadherent patients or take the public health risk of development of drug-resistant strains of the virus.
Quantity of Persons Treated versus Quality of Treatment
The authors point out that the scarce resource now (in contrast to the past) is not antiretroviral medications but staff. Therefore, the dilemma is whether to treat as many people as possible or to treat fewer people and ensure both that the care of patients is adequate and that drug-resistant strains do not develop. The conflict between these two options does not arise out of a conflict between two different ethical principles. Rather, it stems from two different interpretations of the utilitarian principle: Maximize the beneficial outcomes of a policy or practice and minimize the harmful outcomes.
One way of interpreting the obligation to maximize benefits and minimize harms is to confer the benefits on the largest number of individuals. The choice to treat as many people as possible is supported by this interpretation. However, the actual benefits achieved—both to the individuals treated and also from a public health standpoint—may be considerably less if everyone is treated. If failure to have adequate follow-up care results in cases of serious, drug-related side effects that are not properly managed, the harm that results can lessen the overall health benefits realized from treating more people. Similarly, the harmful public health consequences of the emergence of drug-resistant forms of HIV must be weighed against the health benefits of treating more individuals.
Although it is impossible to make precise calculations of these benefits and harms in each scenario, it is important to recognize that both options can be justified by appealing to the principle of utility. It requires both medical judgments and public health expertise to predict what would be the degree of harm resulting from the choice to treat as many people as possible. If very accurate predictions could be made of the number and degree of serious side effects and the probability of nonadherence resulting in drug-resistant strains, those harms could then be compared with the harm of not treating more patients in need of ARVs. The difficulty of choosing between these options arises not from an ethical conflict but from an inability to predict the consequences.
“First Come, First Served” versus “Jumping the Queue”
Fox and Goemaere characterize this question as follows: “Under what circumstances, if any, is it justifiable to disrupt an equitable ‘first come, first served’ ARV treatment policy… ?” The first thing to note is that “first come, first served” is only one among several different principles of equity. There are competing principles often put forward as the “correct” principle of equity, and the most appropriate one may depend on the circumstances. If everyone arrives at the same time for a limited resource, there is no “first come, first served.” In that case, the most just method of choosing may be a lottery. Although that situation would rarely occur in a medical context, it is clearly applicable in some situations, such as a raffle. It could occur in a natural disaster or a terrorist act, where many people are seriously injured simultaneously. In those cases, the most likely response would be a classic triage rather than devising some sort of lottery scheme.
A different principle of equity is “to each according to his or her need.” If need is narrowly defined as medical need, or better, as urgent medical need, then this interpretation of the equity principle can justify jumping the queue. Urgent need for ART would be greater than the need of others who are eligible for treatment but not in immediate need of ARVs. However, if the medical conditions of those others will deteriorate while they await treatment, that could put even more people in “urgent need” of treatment. Moreover, the very sickest AIDS patients may have a poorer chance of doing well on treatment than the less sick who are medically eligible.
This dilemma is known as the “best outcomes” versus “fair chances” problem.5
Brock DW. Fairness and health. In: Murray CJL, Salomon J, Mathers C, Lopez A, eds. Summary Measures of Population Health. Geneva: WHO; 2002.
Some years ago, I heard transplant surgeons invoke a different ethical principle to justify moving patients awaiting a transplant up the waiting list. The surgeons referred to their obligation to move a patient in urgent need of an organ up the list as stemming from the “principle of fidelity.” Even if a surgeon's patient has been on the list a shorter time than other patients, the surgeon's “fidelity” to that patient calls for jumping the queue to ensure receipt of the needed organ. One surgeon argued against that view, contending that the sicker patient who was moved up might have already deteriorated sufficiently so that it would be a waste of a scarce organ. That is a utilitarian argument.6
Macklin R. Comment: Should Selected Patients Ever Be Moved Up? Transplantation Proceedings 1989;21:3395–6.
What this analysis shows is that there are different interpretations of what equity requires, that a principle of equity can be in conflict with the principle of utility, and that there is no clear, agreed upon method for choosing one over another of these principles. Here again, as in the previous dilemma, accurate empirical data might help to determine how much good or harm could result by choosing between “best outcomes” and “fair chances.” But no amount of empirical data can determine which ethical principle to choose when two respected principles are in direct conflict. Reasonable people disagree on which principle should take precedence in such cases.
Disclosure of HIV Status
Steps MSF has taken to help ensure adherence to the medication regimen include disclosure of HIV status to at least one family member who will act as a “treatment assistant,” allowing a home visit by a counselor or nurse, and attendance at a support group. Fox and Goemaere note that these steps may also “contribute to the overall destigmatization of HIV/AIDS.” Although efforts to destigmatize HIV/AIDS are surely laudable, no evidence is provided to suggest that these steps can contribute to that worthy goal. The ethical problem, however, is the required disclosure of HIV status, both within the family and to outside persons, thereby breaching confidentiality regarding highly sensitive medical information.
This dilemma requires a delicate balancing act: seeking to ensure maximum compliance with the medication regimen versus the obligation to maintain confidentiality and abstain from forced disclosure of HIV status. A policy that would restrict eligibility for ART to individuals who agree to the disclosure, home visits, and support group attendance appears to be overly strict. Because the majority of treatment programs do not include these additional steps, it is reasonable to assume that compliance with the medical regimen can be achieved without them. Adherence never achieves 100% of all persons treated, however. Furthermore, these additional steps may help to ensure compliance, but there is no guarantee of that either. The family member acting as a treatment assistant could not reasonably force the nonadherent relative to take the medication if the patient were to resist for whatever reason.
It is likely that in most situations, at least some family members will be aware of a patient's HIV status and be willing to cooperate in any way. Participation in a support group outside the home may be less acceptable to some patients and could also have logistical problems. Most important, there can be no guarantee of confidentiality in such situations, because any group member could reveal another member's HIV status to someone outside the group. In sum, these additional eligibility requirements are burdensome and overly restrictive. They violate the ethical principle that requires choosing “the least restrictive alternative” when it comes to limitations on an individual's freedom or privacy. Liberty and privacy rights are compromised by these eligibility criteria, and these ethical shortcomings are not outweighed by the likelihood of achieving significantly greater adherence to the medication regimen.
The response by MSF to these observations may very well be what Fox and Goemaere stress in their article: “in almost every instance, the patient is ‘accepted’ for treatment … even when a patient does not meet certain of the social and adherence criteria.” If that is, in fact, the case, there is little justification for having criteria for selection “on paper,” especially when such criteria pose potential ethical problems.
Stopping ART for Nonadherent Patients
The rationale for suspending ARV treatment of a patient who is not regularly taking medication is to avoid the emergence of a drug-resistant strain of the virus. There is a prior scientific question that requires an answer: Would stopping the medication for patients who have already been on the regimen similarly be likely to lead to the development of drug-resistant strains? If so, little is gained by stopping cold. It almost seems as if the “strategy” of suspending treatment is to scare the patient into becoming more adherent. That would be an ethically questionable strategy. Patients are nonadherent for a wide variety of reasons, and as the case of Patient L. demonstrates, this strategy may not work. However, if ample scientific evidence exists that there is considerable likelihood that intermittent adherence to medication leads to drug-resistant strains, and sufficient evidence that stopping the treatment regimen altogether does not lead to drug-resistant strains, then the policy is ethically acceptable. However, patients must be told this in advance and warned repeatedly that nonadherence will result in their treatment being suspended.
Selecting Patients when Resources Are Scarce
Fox and Goemaere conclude their article by noting that as scaling up of ART continues in South Africa, the problem of a dearth of physicians and nurses will not only persist, but may even be amplified. Phasing in the South African government's national HIV/AIDS plan will take place incrementally, necessitating hard choices. These choices will not only include the order in which treatment centers are established and where they will be set up, inevitably, it will require selection of patients who will be treated first.
As noted earlier, “first-come, first-served” is only one of several principles of equity that are potentially applicable. In addition to principles that strive for an equitable distribution of a resource, there is also the ethical principle of utility, which aims at maximizing benefits and minimizing harms. Potentially applicable principles in the context of selecting patients or groups of patients to receive priority for ART are the following.7
The sections that follow are adapted from my “Ethics and Equity in Access to HIV Treatment—3 by 5 Initiative.” Background paper for the Consultation on Equitable Access to Treatment and Care for HIV/AIDS, Geneva, Switzerland, 26–27 January 2004. WHO/SDE/ETH/2004.1 Available at: www.who.int/ethics/en/background-macklin.pdf (accessed June 29, 2005).
Utilitarian Ethical Principle
Utilitarian principles can take several forms. In its most general form, the utilitarian principle calls for maximizing overall societal benefits. Utility is sometimes understood as efficiency in reaching the goal of maximizing societal benefits. Applied specifically to health policy, the principle aims at maximizing health benefits for the society as a whole. The best policy would be one that embodies a mix of healthcare services that produces the greatest overall health effects. Although the utilitarian principle has been understood to embrace considerations of justice, it does not, strictly speaking, provide a basis for choosing among eligible groups in the most equitable manner.
In the context of scaling up ART, a utilitarian approach could call for treating the greatest number of people, even if some (e.g., the sickest) could benefit only temporarily before their overall health status worsens. In this view, health benefits would be maximized even though the sickest people would receive a small amount of benefit. On a different and more widely accepted interpretation, maximizing health benefits would mean giving priority to people whose medical condition is such that they respond better to treatment and are likely to survive for the longest time. Adhering to this second interpretation of the utilitarian principle would therefore exclude patients whose HIV disease has progressed to a point where only a temporary health benefit can be expected. This is an example of the “fair chances versus best outcome” problem mentioned earlier. Some would argue that equity requires giving everyone a fair chance at treatment, even if that option would not result in the best overall health outcome.
The implications of using the utilitarian principle are, first, that exclusive use of this principle could result in ignoring specific claims for priority of access on the part of vulnerable or marginalized groups and, second, that it would involve the use of a triage system designed to maximize medical utility.
Egalitarian Principle of Equity
Egalitarian theories call for distributing resources equally among persons or distributing goods, such as healthcare services, or health, equally among different groups. This could mean either that everyone should receive the same amount of resources for healthcare or that everyone should receive the same amount of healthcare services or that insofar as possible, health status disparities among different groups should be minimized. This principle is the basis for schemes that emphasize health equity over health maximization. The goal is to reduce disparities in health status among different groups or strata in society: the poor, women, people living in rural areas, ethnic or racial minorities, and others.
Within this approach there are also ambiguities and disagreements, primarily with regard to what should be equalized: resources for health, health status, or access to health care. Because health needs differ widely among people and groups, the choice has important policy consequences. Giving everyone the same resources will produce different outcomes because health care needs vary. This illustrates how providing equal amounts of resources can be inequitable. For example, groups that have TB or malaria in addition to HIV/AIDS have greater health needs than people living with HIV/AIDS without other diseases. Aiming at equalizing health status will potentially divert resources to interventions that produce very small benefits for people who require enormous resources to obtain any benefit at all.
This principle is relevant to providing equitable access to ART for groups that have been underserved by the health system or denied care altogether and therefore suffer generally worse health status. There is some overlap with the principle of “prioritarianism,” concern for the worst off (see below), because the economically least advantaged classes in every society have generally been underserved, have had less access to healthcare, and have a worse health status than more affluent groups. However, specific groups such as women and racial or ethnic minorities have been underserved in some societies even when they are not among the least advantaged economically. Moreover, in some societies, marginalized groups such as injection drug users (IDUs), sex workers, and men who have sex with men (MSMs) have either been denied access to ART or have had difficulty gaining access. This principle calls for ensuring that all groups and categories of persons have equal access to ARVs, but taken by itself, it does not imply that these groups should be given priority over others with the same medical needs. Additional considerations are needed to make the case for determining priority of access.
“Prioritarianism”
This infelicitous term denotes concern for the worst off or the least advantaged in society. The prioritarian principle holds that benefiting people matters more, from an ethical perspective, the worse off those people are. This does not imply giving strict priority to the least advantaged, but it does call for a strong presumption to benefit those who are worst off in some relevant respect. In the context of healthcare delivery, this is usually understood to refer to those who are worst off in terms of health status, but it could also apply to (a) the poorest, (b) the lowest socioeconomic class, (c) the most vulnerable (e.g., children, especially orphans), (d) groups that are marginalized or most discriminated against (in many societies, IDUs, sex workers, MSMs), or (e) those who live in remote areas with poor access to any type of health services.
In one interpretation, we can consider the least advantaged to be those who are the sickest, irrespective of how much they would benefit from ART. This would mean that if patients in an advanced stage of clinical AIDS can still benefit from ART, they should be among the first treated. The consequence of this would be to divert resources away from people who could potentially benefit more from treatment; the result would be in direct conflict with a health maximization principle. This is, again, the fair chances–best outcome problem described above.
In another interpretation, concern for the worst off would mean giving priority to those most in need, where need is understood as how badly off people would be if not treated (as noted earlier, this can be understood as urgent need). According to this view, preference should be given to patients in urgent need of ART on the basis of how much they can benefit individually, treating the most needy first. Using this principle would result in treatment of comparatively fewer patients than using a straightforward utilitarian principle. It would also justify moving patients up the queue.
There is no clear or obvious way of determining how much weight should be given to a principle that calls for giving priority to the sickest. On the assumption that giving first priority to the sickest is a reasonable ethical principle, the problem becomes how much aggregate health benefit should be sacrificed to treat the sickest first.8
Brock DW. Ethical issues in the use of cost effectiveness analysis for the prioritization of health care resources. In: Khusfh G, Englehardt HT Jr, eds. Bioethics: A Philosophical Overview. Dordrecht: Kluwer Publishers; 2004. Also in Anand S, Peter F, Sen A, eds. Ethical Foundations of Health Equity. Oxford: Oxford University Press; 2004.
See note 3, WHO/UNAIDS 2005:10.
Careful consideration of these alternative ethical principles and what they imply raises a serious question of whether “first come, first served” is the most just system of allocating a scarce medical resource. Reflecting on these principles can also demonstrate that there is no one right solution to this quandary.
