The concept of exploitation is not explicitly included in the four canonical principles of bioethics—autonomy (or respect for persons), non-maleficence, beneficence, and justice.¹ Nonetheless, bioethicists have appealed to the language of exploitation in a variety of contexts. It is frequently claimed that the sale of kidneys for transplants exploits the sellers. It is said that charging high prices for anti-retroviral therapies exploits those with AIDS. It is said that psychotherapists who have sexual relations with their patients are engaged in exploitation. It is said that commercial surrogacy exploits the surrogate mothers. It has been argued that stem cell research treats “nascent human life as raw material to be exploited as a mere natural resource.”²

The concept of exploitation has also come to play an especially important role in discourse about the ethics of clinical research. Indeed, some bioethicists have argued that the principle of non-exploitation is a rationale—even the underlying rationale—for many of the oft-mentioned principles of ethical research such as social value, scientific validity, informed consent, fair participant selection, and favorable risk-benefit ratio.³ On this view, “the fundamental ethical challenge of all research with humans is to avoid exploitation.”⁴ It may be said that research that has no social value or is not scientifically valid (p. 192 ) exploits research participants because they are being “used” for no good end. To use people for research without their informed consent is to exploit or take unfair advantage of them. To subject subjects to unreasonable risks is to exploit them, and so on.

Accusations of wrongful exploitation in research are most frequently invoked with respect to research with vulnerable populations such as prisoners, other institutionalized persons, low-income persons, and the desperately ill. The charge seems particularly poignant in the international context when poor citizens in less developed countries (LDCs) are used as subjects in research that is primarily designed to benefit those in developed countries—“the specter of exploitation is the most serious ethical issue in multinational clinical research.”⁵ Accusations of exploitation are sometimes directed at research conducted by academic or non-profit organizations, but are especially applied to the growing practice in which pharmaceutical corporations “outsource” medical research to contract research organizations that, in turn, conduct many of their studies in LDCs—“A huge population with a diversity of diseases that are untreated—yes, that is the ‘India Advantage’” (Igate Clinical Research, Inc.).⁶

Such accusations reached fever pitch when investigators conducted placebo controlled trials in LDCs when proven effective treatment was available in developed countries. It was argued that whereas it may be legitimate to recruit subjects into a placebo controlled trial under clinical equipoise, when medical science has no reason to think that the intervention is superior to no treatment (all things considered), it is grossly unethical to recruit subjects into a placebo controlled trial of an intervention when an available intervention is clearly superior to no treatment. To set the stage for the discussion, consider two examples.

The Short-Course ART Trial (SCAT)

Placebo-controlled trials had unequivocally established the efficacy of a “long course” (ACTG 076) use of the antiretroviral drug zidovudine for reducing maternal-fetal transmission of HIV.⁷ The protocol involved administering the drug orally to HIV positive women during pregnancy, administering the drug intravenously during labor, and subsequently administering the drug to the newborn infant. When studied in developed countries, the regimen promised to save the lives of one in seven infants born to HIV positive mothers. Unfortunately, the efficacy and use of the 076 regime could not be confidently (p. 193 ) extrapolated to LDCs. First, the drug might not be as efficacious in LDCs due to differences in immune status and breastfeeding practices. Second, even if the 076 regime proved to be efficacious in LDCs, many thought that its use was not administratively or economically feasible. It would prove too expensive, compliance with the regime would prove virtually impossible for many women, and many LDCs lacked the medical infrastructure to support its administration. Given these facts on the ground, investigators wanted to determine whether a cheaper and simpler “short course” use of zidovudine would be at least reasonably effective in reducing maternal-fetal transmission of HIV even if it would not be as effective as the 076 regimen.

To test the short course regimen against the long course regimen in a superiority study or active controlled trial (ACT) would not answer the scientific question. If the 076 regimen were found to be better than short-course anti-retroviral treatment (ART), it would be unclear if and to what extent it was better than no treatment, and that was what the researchers wanted to know. It would have been impossible to conduct a placebo controlled trial (PCT) of the short course regimen in a developed nation where the local standard of care would have included the 076 regimen. It would not be approved by an IRB and even if it were approved, women would not consent to participate if the long course were available. By contrast, it was feasible to recruit subjects to a PCT of the short course treatment in a LDC where the local standard of care was to receive no treatment at all. Half a loaf is better than none. The placebo controlled trials were conducted but were widely condemned as unethical and exploitative because the investigators deliberately withheld a proven effective intervention from those in the control group.

The Surfaxin Trial⁸

Respiratory distress syndrome (RDS) is a common and potentially fatal disease in premature infants, caused by insufficient surfactant in the lungs. Surfactant is a protein fluid that reduces alveolar surface tension, enabling proper lung inflation and aeration. Several replacement surfactants such as Exosurf had been approved by the U.S. Food and Drug Administration (FDA). The use of surfactant replacement therapy is the standard treatment for RDS in the developed world. It is not a viable option in many LDCs because of its high cost (over $1,000 per child).

In 2000, a private U.S. drug company, Discovery Labs (DL), planned a study to demonstrate the efficacy of a new synthetic surfactant, Surfaxin, (p. 194 ) which was thought to be easier and cheaper to manufacture and which might have fewer side effects than surfactants derived from animal tissue. The principal target market for Surfaxin was the United States and Europe, as it would prove unaffordable in LDCs. The FDA does not require that a product be shown to be superior to other interventions. It requires only that it is safe and effective. For several reasons, which need not concern us here, DL proposed to conduct a multi-centered, double-blinded, randomized, two-arm placebo controlled trial involving 650 premature infants with RDS in Bolivia and other sites in South America. In the original proposal, DL agreed to provide endotracheal tubs, ventilators, and antibiotics for all study participants. Parents of infants with RDS would be asked to give consent for their babies to participate. After intubation with an endotracheal tube, half the infants would receive air suffused with Surfaxin and half the infants would receive (“sham”) air without any drug, although they would not be denied a treatment to which they would otherwise have had access. Indeed, the ventilator support with sham air was known to improve survival and was superior to the treatment generally available to Bolivian infants with RDS.

Although the short-course HIV trial received more extensive discussion, the Surfaxin Trial (ST) was criticized on similar grounds. Both protocols involved withholding effective therapy from participants in the control group. Moreover, unlike SCAT, which was designed to test an intervention for use in LDCs, ST would be using infants in a poor nation to test the efficacy of a drug that would be primarily used in wealthy nations. In the face of intense criticism, DL withdrew its proposal to conduct a PCT in Bolivia and, instead, conducted a multi-centered ACT at sites where ventilation and surfactant therapy were ordinarily available.

Because my purpose is to use these cases as examples of alleged exploitation, references to ST in this chapter will refer to the original study design for a PCT in Bolivia (and other LDCs) and not the actual ACT that was eventually conducted.

The Exploitation Complaint

As I indicated above, many bioethicists have claimed that much clinical research (such as SCAT and ST) in underdeveloped societies exemplifies wrongful exploitation. Although these charges have considerable rhetorical force and their intuitive appeal is difficult to deny, they are not based on a careful examination as to what constitutes exploitation or its moral force. (p. 195 ) Here are some typical statements, accompanied in italics by some questions that they raise.

1. 1. “Unless the interventions being tested will actually be made available to the impoverished populations that are being used as research subjects, developed countries are simply exploiting them in order to quickly use the knowledge gained from the clinical trials for the developed countries’ own benefit.”⁹ Questions: would there be no exploitation of the research subjects if the intervention were made available to impoverished population? Who are the victims of exploitation? The subjects? The countries?

2. 2. “… the placebo-controlled trials are exploitative of poor people, who are being manipulated into serving the interests of those who live in wealthy nations…”¹⁰ Questions: what is manipulation? Would such trials be exploitative if the subjects gave competent, uncoerced, and rational consent?

3. 3. “… there is always the nagging possibility that the assurances of such benefits may offer inordinate inducements to poor and impoverished populations and thus represent another form of exploitation.”¹¹ Questions: what makes an inducement “inordinate”? Would people be less exploited if there were no inducements? Are such inducements inordinate only if they compromise consent? Is it less exploitative to provide a small inducement than a large inducement? Is the possibility of receiving life-saving treatment an undue inducement?

4. 4. “If the knowledge gained from the research in such a country is used primarily for the benefit of populations that can afford the tested product, the research may rightly be characterized as exploitative and therefore, unethical.”¹² Questions: Are subjects exploited just because the tested product is used for wealthy nations? Are workers in developing societies who produce athletic shoes exploited just because the product is typically unaffordable in their own society?

5. 5. “If the results of a clinical trial are not made reasonably available in a timely manner to study participants and other inhabitants of a host country, the researchers might be justly accused of exploiting poor, undereducated subjects for the (p. 196 ) benefit of more affluent populations of the sponsoring countries.¹³ Questions: What if the trial is unsuccessful? If the results are not available in any society, does it mean that no one is exploited? Why do “other inhabitants” of a host country have a claim to be benefit from the efforts of participants?

6. 6. “Residents of impoverished, postcolonial countries, the majority of whom are people of color, must be protected from potential exploitation in research. Otherwise, the abominable state of health care in these countries can be used to justify studies that could never pass ethical muster in the sponsoring country.”¹⁴ Questions: do the principles of ethical research require that if a study design were unethical in one society (given its conditions) that the study must be unethical in another society (given its conditions)? If $10 per hour is a fair wage in the United States for a manufacturing job, is it unethical to pay people less in a country where the average wage is $1 per hour?

7. 7. “… it is a fundamental ethical principle that those involved in research in developing countries… should not take advantage of the vulnerabilities created by poverty or a lack of infrastructure and resources.”¹⁵ Questions: does one exploit another just because one takes advantage of their vulnerabilities? Do roofers exploit home owners when they take advantage of the vulnerabilities created by leaky roofs? Does it matter whether the vulnerable person benefits from or consents to participation?

As these statements suggest—and one can find numerous additional similar statements in the literature—many commentators seem to accept what I call The Exploitation Argument. Reduced to its essentials, that argument maintains something like this:

1. (1) If a practice is exploitative, it should not be permitted.

2. (2) Placebo controlled trials (PCTs) that withhold proven effective therapy (such as ST) are exploitative.

3. (3) Therefore, such PCTs (or other exploitative practices) should not be permitted.

As it stands, this argument moves much too quickly. First, despite the ease with which accusations of exploitation are advanced, it is not clear when a practice or transaction is properly described as exploitative (2). (p. 197 ) Indeed, it is not clear who are thought to be the victims of exploitation—the research subjects, the community, the nation? Until we have a better grip on the concept of exploitation, we cannot say with confidence that (2) is true. Second, it is not clear whether we should accept (1). There is an important and neglected distinction between the ethics of interaction between researchers and subjects and the ethics of intervention with the interaction between researchers and subjects. I shall argue that whereas it is wrong for researchers to exploit subjects, it simply does not follow that we should seek to prevent such exploitation.

My Plan

The major purpose of this chapter is to attempt to shed some philosophical light on a topic that has generated too much rhetorical heat and too many question-begging arguments. Here, once again, we need to widen the lens. To properly assess these claims about the exploitative character of much clinical research, we must step back and examine the concept of exploitation with care. The claim that clinical research is exploitative is sometimes true. But it is important to determine just when it is true and to get clear about the moral force of that truth. Drawing on some earlier work, I first develop some analytical distinctions about the concept of exploitation and about the forms it can take. In particular, I will distinguish between harmful and nonconsensual exploitation and mutually advantageous and consensual exploitation. I also seek to identify the essential elements of exploitation.¹⁶ Second, I explore the moral force of exploitation and, in particular, the arguments for prohibiting mutually advantageous consensual exploitation. After all, the question for research ethics is not only whether a practice is properly described as exploitative, but what we should do, given its wrongfulness. Third, I try to bring those analyses to bear on the ethical issues that arise with respect to use of placebo-controlled trials in cases such as SCAT and ST. Although my analysis of exploitation also applies to other issues, such as the provision of post-trial treatment to subjects, I will mention them only briefly in this chapter because I consider such issues in more detail in Chapter 6. In considering the alleged exploitativeness of PCTs, I focus on ST rather than SCAT because it presents a greater challenge to my argument. Although both trials withhold proven effective treatment from the control group, the aim of SCAT was to develop an intervention that would benefit people in LDCs. If it was exploiting the poor, it was exploiting the poor for the benefit of the (p. 198 ) poor. If, as I shall argue, it is actually quite hard to justify interference with such trials when they benefit people in wealthy nations, there is certainly less reason to interfere when they would benefit people in LDCs.

A caveat. I will mostly set aside questions as to whether allegedly exploitative practices are compatible with the principles articulated in any of the canonical codes of research ethics such as the Declaration of Helsinki, the CIOMS Ethical Guidelines for Biomedical Research Involving Human Subjects, The Belmont Report, or The Common Rule. For example, the Declaration of Helsinki states that: “The benefits, risks, burdens and effectiveness of a new method should be tested against those of the best current prophylactic, diagnostic, and therapeutic methods.”¹⁷ If researchers were to comply with this passage, it would seem that they could not conduct PCTs such as SCAT or ST. I will, however, not be concerned as to whether ST can be rendered compatible with this or any other similar provision, say by claiming that “best current…method” could be understood in local terms, where the “best current… method” might be no treatment at all. The question here is whether a PCT such as ST is wrongfully exploitative and whether it should be prohibited for that reason. If the best answers to those questions are incompatible with the provisions of any particular document, so be it.

The word “exploitation” can be used in a non-moral or non-derisive sense, where it means simply to “use” or “take advantage of.” So we might say, “The basketball player exploited his great jumping ability.” We are interested in wrongful exploitation. Although one finds multiple and quite different accounts of wrongful exploitation in philosophical literature, most—although not all—are compatible with the claim that A exploits B when A takes unfair advantage of B.¹⁸ One problem with such a broad account is that there will “be as many competing conceptions of exploitation as theories of what persons owe to each other by way of fair treatment.”¹⁹ Consider a few examples which might be thought to involve exploitation.

Student Athletes. A, a major university, provides B with a scholarship to play on its football team. A gains considerable revenue, but, like most athletes, B gets little education, does not graduate, and does not go on to play professional football.²⁰

(p. 199 )

Kidneys. A, who is affluent, offers to pay B $25,000 for one of his kidneys for purposes of transplantation. B, who is poor, agrees in order to better provide for her family.²¹

Nazi Trials. A, a Nazi medical scientist, wishes to discover how long a person can live in freezing water. He places B, who has been placed in a death camp, in freezing water. B dies within an hour.

Surrogacy. A pays B $10,000 to become impregnated with A’s sperm (through artificial insemination) and to waive her rights to the child after birth.²²

Psychotherapy. A, a psychotherapist, proposes to B, his patient, that they have sexual relations. B, who is infatuated with her therapist, agrees.²³

Lumber. There has been a hurricane in Florida. A, a lumber retailer, triples his price for lumber. B, who needs lumber to rebuild, pays A’s price.

Lecherous Millionaire. B’s child will die unless she receives expensive surgery for which the state will not pay. A, a millionaire, proposes to pay for the surgery if B will agree to become his mistress. B agrees.

Rescue. B’s car is stuck in a snow drift late at night on a rural road. A comes along in his pick-up truck and offers to pull B out for $200. It will take A two minutes to complete the job. B agrees.

Let us refer to a statement that A’s interaction with B is wrongfully exploitative as an exploitation claim. The first task of a theory of exploitation is to provide the truth conditions for an exploitation claim. We want to know whether it is true that A exploits B in Lumber or Rescue. Yet the truth of an exploitation claim settles less than is often thought. For even if A wrongly exploits B in Lumber or Rescue, it does not follow as a matter of moral logic that we should prohibit or regulate the transaction between A and B. After all, in Rescue, B might say, “I know that he is exploiting me, but I’d rather be exploited than left here.” And so the second task of a theory of exploitation is to provide an account of its moral force. In particular, we must determine when and why we are justified in prohibiting or regulating or interfering with exploitative transaction.

Let us first attend to the first question. What are the truth conditions of an exploitation claim? When is a transaction exploitative? Consider the following proposed accounts of exploitation.

(p. 200 )

1. 1. “[T]o exploit a person involves the harmful, merely instrumental utilization of him or his capacities, for one’s own advantage or for the sake of one’s own ends.”²⁴

2. 2. “It is the fact that the [capitalist’s] income is derived through forced, unpaid, surplus [wage] labor, the product of which the workers do not control, which makes [wage labor] exploitive.”²⁵

3. 3. “Exploitation necessarily involves benefits or gains of some kind to someone… Exploitation resembles a zero-sum game, viz what the exploiter gains, the exploitee loses; or, minimally, for the exploiter to gain, the exploitee must lose.”²⁶

4. 4. “Exploitation [in exchange] demands… that there is no reasonably eligible alternative [for the exploitee] and that the consideration or advantage received is incommensurate with the price paid. One is not exploited if one is offered what one desperately needs at a fair and reasonable price.”²⁷

5. 5. “Common to all exploitation of one person (B) by another (A)… is that A makes a profit or gain by turning some characteristic of B to his own advantage… exploitation… can occur in morally unsavory forms without harming the exploitee’s interests and… despite the exploitee’s fully voluntary consent to the exploitative behavior. .”²⁸

6. 6. “Persons are exploited if (1) others secure a benefit by (2) using them as a tool or resource so as (3) to cause them serious harm.”²⁹

7. 7. “An exploitative exchange is… an exchange in which the exploited party gets less than the exploiting party, who does better at the exploited party’s expense… [T]he exchange must result from social relations of unequal power… exploitation can be entered into voluntarily; and can even, in some sense, be advantageous to the exploited party.”³⁰

8. 8. “[E]xploitation is a psychological, rather than a social or an economic concept. For an offer to be exploitative, it must serve to create or to take advantage of some recognized psychological vulnerability which, in turn, disturbs the offeree’s ability to reason effectively.”³¹

(p. 201 )

9. 9. “Exploitation, at its heart, is not just about unfair shares or garden-variety degradation; it is about wrongfully exacting benefit from another’s actual vulnerability.”³²

All these accounts are compatible with the general claim that “A wrongfully exploits B when A takes unfair advantage of B,” but there are some important differences among them. Some accounts are technical definitions of exploitation specific to a Marxist approach (1). Some accounts invoke the Kantian notion that one wrongfully exploits when one treats another instrumentally or merely as a means (1, 6, 9). On some accounts, someone is exploited only if she is harmed (1, 2, 3, 6), whereas others allow that the exploited party may gain from the relationship (4, 5, 7). On some accounts, the exploited party must be coerced (2, 4) or exhibit a defect in the quality or validity of the consent (8), whereas other accounts maintain that exploitation can be fully voluntary or consensual (5, 7).

I think it best not to put rigid constraints on what counts as exploitation. While some exploitative transactions are harmful to the exploitee, as is obvious in Nazi Trial and is likely in Psychotherapy, we often describe as exploitative cases in which both parties gain, as might be true in Student Athletes, Surrogacy, Kidneys, Lumber, Lecherous Millionaire, and Rescue. It will, therefore, be useful to make two sets of distinctions. First, we can distinguish between harmful exploitation and mutually advantageous exploitation. By harmful exploitation, I refer to those cases in which the exploiter gains by harming the exploitee. By mutually advantageous exploitation, I refer to those cases in which both parties—including the exploitee—reasonably expect to gain from the transaction as contrasted with the pre-transaction status quo or more accurately, as contrasted with their pre-transaction rights or entitlements. In any case, it is the advantageousness that is mutual, not the exploitation.

We can similarly distinguish between nonconsensual exploitation and consensual exploitation. By nonconsensual exploitation, I refer to cases in which the exploitee does not token consent or does not give morally transformative consent. Setting aside the possibility that exploitation itself may render consent non-transformative in certain ways (see Chapter 3), consensual exploitation refers to cases where the exploited party has given consent that would otherwise be regarded as morally transformative; that is, the consent is voluntary, competent, informed, and so forth.

The two distinctions overlap but are not equivalent. There can be cases of mutually advantageous but nonconsensual exploitation. If, in a version of (p. 202 ) Rescue, A threatened to hurt B if B did not agree to pay A $200, B has not given morally transformative consent because he was coerced, but the transaction might still be mutually advantageous. There can also be cases of harmful but consensual exploitation, as when a self-loathing B allows A to benefit by harming her. Nonetheless, because these distinctions tend to converge, and because it will simplify matters, I shall rely on the distinction between harmful non-consensual exploitation (HNE) and mutually advantageous and consensual exploitation (MACE).

HNE is typically morally worse than MACE, but it is of less theoretical interest. It is not difficult to explain why it is wrong to harm another or to act without another’s consent and it is relatively easy to justify prohibiting such actions. By contrast, it is more difficult to explain why and when MACE can be wrong and whether and why we might be justified in prohibiting or refusing to enforce such transactions. And so it is on those latter cases on which we will focus.

The elements of exploitation

Let us start with the claim that A exploits B when A takes unfair advantage of B. Taking unfair advantage could be understood in two ways. First, it may refer to some dimension of the outcome of the exploitative act or transaction, that is, the transaction is substantively unfair. And this, it seems, has two elements: (1) the benefit to A, and (2) the effect on B. We may say that the benefit to A is unfair because it is wrong for A to benefit by harming B in a case of HNE (in addition to the wrong of the harm itself) or because A’s benefit is excessive relative to B’s benefit in a case of MACE. Second, to say that A takes unfair advantage of B may imply that there is some sort of defect in the process by which the unfair outcome comes about; for example, that A coerces B into giving (apparent) consent or that A deceives B or fails to provide B with relevant information or manipulates B. If mutually advantageous and consensual transactions can be exploitative, then a defect in consent is not a necessary condition of exploitation. Rather, a moral defect in the outcome is both necessary and sufficient to constitute exploitation.

The Outcome of a Transaction

When Is A Transaction Consensual?

Although I have argued that one can consent to be exploited, the claim that A exploits B is often associated with the claim that B does not give morally transformative consent to the transaction. At first glance, it seems plausible to argue that A does not exploit if B voluntarily agrees to what might otherwise be an unequal or unfair distribution of advantages and burdens. Motive is important. I do not think that a hospital exploits its volunteer workers even if others are paid for comparable work if the volunteers prefer to work altruistically for no pay than not to work. Intentional altruism aside, although a transaction can be exploitative even if B’s decision is voluntary, well informed, and rational as in Lumber and Rescue, the moral force of exploitation may turn on whether the exploitation is consensual. In particular, there may be good reason to prohibit nonconsensual exploitation, but to permit consensual exploitation. We have at least prima facie reason not to interfere with a person’s autonomous choice to allow herself to be exploited.

Because I considered consent in some detail in Chapters 3 and 4, I will make only a few brief remarks here.

Voluntariness

We generally say that consent is voluntary when it is not coerced. I have argued that A coerces B to do X in a way that renders B’s consent non-transformative only if A proposes (threatens) to violate B’s rights if B chooses not do X. On this view, A does not coerce B in cases such as Lumber, Lecherous Millionaire, Surgery, Kidneys, or Surrogacy. For A does not propose to violate B’s rights if B rejects A’s proposal in any of these cases. Whatever we think about A’s proposal in Lecherous Millionaire, A has made an offer not a threat, and offers do not coerce. The mere fact that B may have no reasonable alternative to accepting A’s proposal does not entail that B’s decision is coerced. We would not say that A commits a battery in Surgery on the grounds that B has no reasonable alternative but to agree to surgery. Moreover, even if B’s background situation is unjust, it does not follow that A’s proposal is coercive. Even if a just society would provide for B’s daughter’s (p. 212 ) surgery in Lecherous Millionaire, B has no right that A pay for her surgery. And so whatever else we think about A, his proposal does not coerce B into having sexual relations with him.

Information

B’s token of consent may not be morally transformative if there are defects in B’s information such that B’s decision is not likely to advance B’s interests considered from an ex ante and all things considered point of view. In some cases, B’s informational deficiencies may be due to fraud, as when A deliberately sets back the odometer in a car that he is selling. In other cases, A may withhold or fail to disclose information that A has an obligation to provide. And so B’s agreement to purchase a home from A is not binding if A fails to tell B that the roof leaks. And in the paradigmatic medical context, B’s consent to surgery is not valid if A fails to inform B of the risks that are involved.

Competence

Even if A provides B with all the relevant information, B may not have the competence to process that information. We often assume that minors and those with mental impairments do not have the cognitive or emotional capacities to give morally transformative consent. Those who lack adequate education may be unable to make reasonable evaluations of the alternatives. Moreover, even if B is an otherwise competent person, her competence can be temporarily disturbed as in Psychotherapy, where B may be in the grips of transference. Furthermore, B’s judgment can also be distorted when A makes what I call a seductive offer, where the lure of short term benefits causes B to excessively discount the long-term costs and to make a decision that does not serve her long-term interests, as may be true of cases such as Kidneys and Surrogacy.

Consensual Exploitation

All that said, there are numerous cases of alleged exploitation where B would not have agreed to A’s proposal under better or perhaps more just background conditions, but where A has no special obligation to repair those conditions or to transact with B, and where B is fully informed as to the (p. 213 ) consequences of various choices and fully capable of making such choices. Without prejudging whether these transactions are actually exploitative or unfair, I suggest that such conditions may obtain in Kidneys, Surrogacy, Lecherous Millionaire, Lumber, and Rescue. Precisely because B’s objective situation is what it is, it may be reasonable for B to agree to proposals to which those who are better situated would not agree.

Once again, I do not deny that A’s transaction with B may be wrongfully exploitative even if B gives morally transformative consent. Moreover, I do not preclude the possibility that a transaction should be prohibited even if it is mutually advantageous and consensual. The present point is that we should not treat B’s consent as non-transformative on grounds of coercion, information, competence, or rationality, just because the transaction is exploitative.

The moral force of harmful and nonconsensual exploitation is relatively unproblematic. If a transaction is a case HNE, there is at least a prima facie case for prohibiting it. But what could justify not allowing the parties to enter into a mutually advantageous consensual transaction or refusing to enforce its terms given that such refusal is effectively equivalent to prohibiting the transaction? Consider a version of Rescue.

Cash-Poor. B is prepared to pay A $200 to tow his car, but has only $50 in cash. He has no checks. He is prepared to sign an IOU for $150.

If it is known that the state will not enforce such agreements on the grounds that it constitutes an “unconscionable contract,” A may be unwilling to rescue B. Can such a policy be justified? Perhaps.

Let us assume that the following (Table 5.1) represents the utility gains of A and B in Rescue as contrasted with the baseline in which A does not rescue B.

I intentionally stipulate that B gains more utility than A in the unfair transaction (2) to emphasize the earlier point that the exploitee often gains more utility in an exploitative transaction than the exploiter. It is entirely possible, for example, that B would have paid $500 to be rescued under the circumstances, but that we would still regard $200 as an excessive and unfair price. (p. 214 )

In any case, the present question is not whether A’s behavior in (2) could be wrong given that A’s behavior in (1) is permissible, but whether it is possible to justify interfering in exploitative transactions such as (2) given that we do not require A to aid B in (1). After all, those who claim that a practice is exploitative typically claim or assume that exploitation provides a reason for intervention. Those who argue that commercial surrogacy exploits the birth mothers also maintain that surrogacy contracts should be unenforceable or entirely prohibited.⁴⁶ Those who argue that a market in kidneys exploits the poor typically argue that such sales should be banned. Those who argue that studies such as SCAT or ST are exploitative believe that such studies should not be allowed. Those who support intervention may assume or argue that the practices are nonconsensual or harmful, but we need to ask whether intervention can be justified if the transactions are consensual and mutually advantageous.

That line of argument is more difficult to sustain. As a starting point, it seems reasonable to assume a moral default position in which we should not interfere with or prohibit transactions which are beneficial to both parties and to which both parties consent. Consider, once again, the payoff structure of Rescue as described in Table 5.1. Suppose A proposes (2) (where B pays $200). B counters that they should agree on (3) (where B pays $50). A rejects (3) and tells B that it is either (1) (no transaction) or (2). Given these options, B is prepared to accept (2). At first glance it seems that society is not justified in preventing B from accepting (2), its unfairness notwithstanding and that it should also be prepared to enforce its terms, as in Cash-Poor, if doing so is necessary to facilitate such transactions.

(p. 215 ) Reasons for Intervention

In this section I consider seven arguments for intervention with mutually advantageous consensual exploitation. Although I do not think that most of these arguments are successful, at least one argument is quite plausible.

Paternalism

As we saw in Chapter 2, we may be justified in interfering with transactions on paternalistic grounds in order to protect B from making a decision that does not advance her own interests. But a paternalistic argument for interference simply does not apply in the present context if the exploitative transaction is advantageous to the exploitee. We can hardly forbid B from agreeing to pay the exorbitant price in Rescue on paternalistic grounds if A would otherwise leave her stranded. Similarly, we could not justify preventing parents from allowing their infants to participate in ST on paternalistic grounds given that the infants can expect to be better off if they do participate. So we can set paternalism aside as a justification for interfering with MACE.

Externalities

We might interfere with mutually advantageous and consensual transactions because they impose harm or negative externalities on others or because interference provides benefits or positive externalities to others. For example, even if commercial surrogacy is beneficial to the participants directly involved, it might have negative effects on the way in which society perceives and treats women or children.⁴⁷ And even if Lecherous Millionaire were beneficial to B, a system which allows A to take advantage of society’s failure to provide health care may postpone the day in which society fulfills its obligation.

Three points about this line of argument. First, if a transaction between A and B has negative effects on others, there is a prima facie case for intervention, although there are some contexts in which we do and probably should ignore such harms. If A offers B a job for which C has applied, the transaction between A and B has a negative effect on C. But this hardly constitutes a reason to interfere with the transaction because C has no right to be protected from competitors. Second, it is an empirical question—and often an extraordinarily complex empirical question—as to whether interfering with a (p. 216 ) practice would actually reduce negative externalities or promote positive externalities. The assertion that commercial surrogacy has harmful effects on the way in which society regards children and women does not show that commercial surrogacy actually has those effects.⁴⁸ Third, even if the externalities argument provides a rationale for intervention, it has nothing to do with the interests of the exploitee or exploitation as such, and so we can set it aside.

Strategic Intervention

What I call “strategic intervention” is the most plausible justification for interfering with MACE. It is, in essence, a collective action argument. The core idea is best explained by an example. Thomas Schelling once argued that professional hockey players preferred to wear helmets for safety but that few players would do so voluntarily because they believed that wearing helmets compromised performance, putting them at a competitive disadvantage, and also detracted from their “macho” reputation.⁴⁹ Schelling argued that virtually all players would prefer that the NHL require them to wear helmets, in which case they can gain the safety benefits of helmets without bearing the competitive or reputational costs. Wearing hockey helmets constituted a collective action problem among the players that could be solved by strategic intervention. Virtually all players would prefer (1) helmets to (2) no helmets, but will end up choosing (2) if left to their own devices, and so requiring helmets shifted the equilibrium from (2) to (1). Minimum wage laws can also be understood as a strategic intervention. Left to fend for themselves, it may be perfectly rational for many workers to accept a sub-minimum wage rather than remain unemployed. Yet it would be better for most (perhaps not all) if they were prevented from agreeing to sub-minimum wage proposals—if employers would then offer higher wages rather than not hiring them at all.

Strategic intervention bears some resemblance to paternalistic intervention because it interferes with the agent for her own good. It is, however, importantly different. Whereas paternalistic intervention overrides the individual’s preferences, strategic intervention serves to help or enable a person get the result that she already prefers but cannot obtain without such intervention.

The strategic argument may apply to interference with exploitative transactions. Recall the payoff structure of Rescue, where A can make a credible threat to opt for (1) (no transaction) rather than (3) (fair transaction) and so (p. 217 ) get B to agree to (2) (unfair transaction). The strategic argument maintains that if the state prohibits A from entering into transaction (2) or announces that it will refuse to enforce its terms in Cash Poor on the grounds that it is an “unconscionable contract,” then A will propose (3) rather than (1). After all, given that A benefits from (3), A has no reason to opt for (1) over (3) if the state effectively takes (2) off the table. In short, the strategic argument says that we may be justified in not allowing A to exploit B if preventing such exploitation will cause A to offer a fairer deal to B rather than refusing to transact with B.

The relevance of the strategic argument to a given case depends on the facts. It is an empirical question as to how potential contractors will react to such intervention. The strategic argument might justify interfering in cases such as Rescue where A is the only seller and B is the only buyer, but might not justify interfering in cases such as Lumber where there may be many potential buyers if B did not want to contract with A at A’s proposed price. Whether barring “unconscionable contracts” in a case such as Cash Poor helps the exploitee, similarly depends on the likely outcomes of such intervention. My tentative conclusion is that the strategic argument provides an important and plausible rationale for interfering with mutually advantageous and consensual exploitation in clinical research. The empirical question is whether interference with such research actually works to the advantage of prospective subjects.

Epistemological Arguments

There may be situations where it is difficult to determine whether a transaction is a case of HNE or MACE. Given these epistemological difficulties, it may make sense to prohibit all transactions of a certain type because the expected harms associated with cases of HNE may be greater than the expected benefits associated with cases of MACE. Interfering with genuine cases of MACE is simply the price we have to pay for prohibiting HNE. As I argued in Chapter 2 with respect to group soft-paternalism, something like this may be true of situations such as Psychotherapy. Suppose that most cases of psychotherapist/patient sexual relations are harmful to the patient and not based on morally transformative consent, but that some cases have no harmful (or even beneficial) effects and are robustly consensual. It does not follow that it should be permissible for psychotherapists and patients to engage in such relations when they believe it is benign and consensual. For given that (p. 218 ) neither psychotherapists nor patients are well positioned to have confidence that their relationship is mutually advantageous or consensual (and perhaps not exploitative), it may be better to prohibit all psychotherapist/patient sexual relations.

As with the strategic argument, the force of the epistemological argument depends on facts. For example, although it is clear that some cases of commercial surrogacy work out badly for the surrogate mother (as in the case of Mary Beth Whitehead), it is possible that most cases work out reasonably well.⁵⁰ If so, we would not be justified in prohibiting all commercial surrogacy transactions by appeal to the epistemological argument. If the independent review of research protocols is working reasonably well, we do not need to trust investigators to determine whether a protocol is beneficial or consensual, so there is no reason to think that the epistemological argument would justify interfering with mutually advantageous and consensual research on the grounds that there is a high probability that we are mistaken about its advantageousness and consensuality.

A Brief Interlude: The Permitted Exploitation Principle (PEP)

Before considering three additional arguments for interfering with mutually advantageous and consensual transactions, it is worth noting that the previous four arguments are person affecting arguments. They attempt to justify interfering with transactions on the grounds that interference is better for someone. Paternalistic and strategic arguments claim that interference advances the interests of the exploitee. The externalities argument claims that interference may be good for others. The epistemological arguments claim that interference will be bad for those who are not harmed by the transaction, but will be good for those who are harmed.

But could we justify interfering with mutually advantageous and consensual transactions when interference is better for no one? To get a grip on the problem, let us make the following assumptions which might be true of a case such as Rescue or Lecherous Millionaire, and may also be true of some cases of clinical research.

1. 1. A is under no obligation to transact with B on any terms.

2. 2. A proposes to transact with B on unfair terms.

(p. 219 )

3. 3. A’s transaction with B on unfair terms is to A’s benefit (ex ante) and also serves B’s interests (ex ante), all things considered (setting aside paternalistic arguments).

4. 4. A’s transacting with B on unfair terms will not harm other persons (setting aside externalities).

5. 5. B makes a voluntary, informed, and rational decision to transact with A on unfair terms and we can reliably determine that this is so (setting aside epistemological arguments).

6. 6. If A is not allowed to transact with B on unfair terms, A will not choose to interact with B on fair terms that are more favorable to B (setting aside strategic arguments).

If such conditions 1-6 obtain, then we must ask whether we could justify preventing a transaction that is good for the parties involved, to which the parties consent, that is worse for no one else, and in which interference will not improve the outcome for the disadvantaged party.

Let us refer to the claim that it would be wrong to interfere with such transactions as The Permitted Exploitation Principle (PEP). The PEP does not claim that it is morally permissible for A to exploit B. Rather, PEP claims that it is wrong to prevent Pareto superior or win-win transactions on the grounds that the terms of such transactions are unfair. The PEP refers to the morality of regulating transactions, and not the morality of transactions themselves.

The distinction between the morality of an act and the morality of regulation or interference is, of course, at the heart of liberalism. Consider freedom of speech. Although it is wrong to make a speech in which one denies that the Holocaust occurred, it may also be wrong to interfere with such a speech. Some cases are more difficult.

The Bigot’s Will. After many years, B’s son has declared that he is gay. B asks his lawyer, A, to re-write his will such that B is disinherited.

If B has a right to leave his money to whomever he wishes, then it would be wrong to prevent him from doing so. In addition, if B needs a lawyer’s assistance to accomplish his aim, there is not much distinction between preventing B from rewriting his will and refusing to assist him in doing so. On that somewhat uncomfortable view, it is not only sometimes right not to interfere with another’s wrongdoing, it is sometimes right to assist others in doing wrong if they have a right to do wrong.

(p. 220 ) Justice

With this interlude behind us, there are at least three (arguably) non-person affecting arguments for rejecting PEP. A deontological or egalitarian argument might maintain that we are prima facie justified in prohibiting an exploitative transaction simply because it is unfair or unjust.⁵¹ Because I consider this argument in more detail in Chapter 6, I shall be brief here. I do not think it incoherent to claim that there might be deontological reasons to reject PEP, but such a view is difficult to defend. Clearly, one cannot reject PEP by appealing to some of the traditional arguments against consequentialism. The PEP does not claim that we should allow transactions just in case they have better aggregate consequences, where the advantages to some outweigh the harm to others. The permitted exploitation principle says that we should allow transactions when it would be better for all parties to the transaction and worse for no one else. Moreover, because PEP precludes intervention only when the parties give valid consent to the transaction, one cannot easily reject PEP on the grounds that it allows A to use B as a mere means to his own ends. Although we can disagree as to the force of the “mere means” principle and whether B’s valid consent completely blocks that charge, B’s valid consent surely goes at least some (if not) all of the way towards defusing that charge.

A Symbolic Argument

A symbolic argument for intervention maintains that we are justified in prohibiting unfair but mutually advantageous transactions in order to symbolize the view that the exploiters are behaving wrongly. The symbolic argument concedes that interference will prevent some win-win transactions to the detriment of the parties involved, but maintains that doing so will encourage fairer transactions in the long run. As Seana Shiffrin puts it, “One should not underestimate the benefits of having an environment in which exploitation does not occur, or at least, in which the state forbears from facilitating clearly unfair treatment.”⁵²

Now to the extent that a policy of interfering with or refusing to uphold MACE transactions motivates people to transact more fairly, the symbolic argument turns out to be a person-affecting (externalities) argument. A pure symbolic argument would argue that it is justifiable to interfere even if the symbolization has no effect on anyone’s behavior. I find this very difficult to believe.

(p. 221 ) Complicity

A final argument for interfering with exploitative agreements focuses on the “enablers.” It may be thought that even if the parties themselves have a right to enter into an exploitative transaction such as Rescue, “others may reasonably feel that they may or even must not assist” such transactions, say by being prepared to enforce its terms.⁵³ It may be wrong for me to try to prevent a friend from smoking, but it does not follow that I have to buy her cigarettes (even with her money). On this view, we may have “a permission and perhaps a deontological commitment not to assist grossly unfair treatment” even if the exploited party prefers that we do so.

The complicity argument is of obvious relevance to clinical research given that IRBs must specifically approve a study if it is to go forward. If the argument can be sustained, then it might justify IRBs from refusing to approve exploitative research not because it is contrary to the interests of research subjects, but because it’s wrong for IRBs to assist or enable immoral transactions.

I am not convinced that the complicity argument can be sustained. First, I believe that the complicity argument puts too much pressure on the distinction between the “prohibition” of a transaction and “non-enabling” or “non-enforcement” of a transaction. For example, it suggests that the state is not complicit with the unfair transaction in Rescue if it chooses not to prohibit it, but that the state is complicit in a case such as Cash-Poor if it chooses to enforce the terms of the transaction. Given that the transaction will not occur unless the state is prepared to enforce its terms, it is arguable that if the state should not lend its hand in Cash-Poor, then perhaps it should also prohibit the transaction in Rescue. Similarly, it is arguable that if the state should not prohibit in Rescue, then it should lend its enforcement powers in Cash-Poor.

Second, the complicity argument also seems to involve a paradoxical relation to the wrongness of a transaction. If an exploitative transaction is wrongful because it does not do enough for B, then its wrongness is a function of its effects on B. Yet the complicity argument is strangely indifferent to the interests of those who are treated wrongfully. To the extent that the “enablers” should be concerned with advancing the interests of those who are treated wrongfully and should also be concerned to respect their autonomy, then they should be prepared to hold their noses and assist A’s wrongful exploitation of B if, ex hypothesi, the transaction will not occur without their assistance.

(p. 222 ) The Moral Force of MACE: A Problem of Non-Ideal Theory

Where does this leave us? I have argued that exploitation is wrong even if the transaction or arrangement is mutually advantageous and consensual. I have also argued that there are often good reasons to adopt what I have called the Permitted Exploitation Principle (PEP). Although there is a plausible strategic argument for prohibiting or not facilitating such transactions (depending upon the facts), there are also moral reasons to permit and even to enable such exploitative transactions. As a general principle, the moral default position is that we should not prevent and we should positively enable transactions that are beneficial to the disadvantaged and to which they consent.

To put the point in different terms, PEP is a plausible principle of non-ideal moral theory. Roughly speaking, ideal moral theory aims to provide the principles for a just society and a just world. Ideal theory defines our moral aims. By contrast, non-ideal moral theory concerns what a person should do “in circumstances where at least some others are not doing what they are required to do.”⁵⁴ Non-ideal theory seeks to provide the principles by which individuals should act under unjust or non-ideal moral conditions and the principles that should underlie social policies for regulating actions by individuals and organizations under non-ideal moral conditions. Whereas Rawls’s A Theory of Justice is a project in ideal moral theory, Rawls himself acknowledges that the problems of non-ideal theory “are the pressing and urgent matters… that we are faced with in everyday life.”⁵⁵ Ideal theory is logically prior to non-ideal theory because we need to know toward what we should aim. But we also need moral principles for the world in which we find ourselves, imperfect though it is.

It is, of course, a matter of some dispute as to what is the best ideal theory of a just society. It is even more debatable as to what people and societies owe each other. Rawls famously does not extend his principles of a just society to the world as a whole.⁵⁶ Be that as it may, the principles that define a just global distribution of wealth and income as a matter of ideal theory will not tell us what non-ideal principles apply to the unjust world in which we find ourselves.

The principles that define when we should allow or enable transactions are important elements in both ideal and non-ideal theory. The present point is this: just because a transaction would not occur under ideal or just conditions, it does not follow that it is wrong for it to occur under non-ideal conditions or (p. 223 ) that it would be better to disallow it or not enable it to happen. Given the non-ideal background conditions under which people find themselves, there should be a very strong presumption in favor of principles that would allow people to improve their situations if they give appropriately robust consent and if doing so has no negative effects on others. And this is so even if the transaction occurs only because the background conditions are unjust.

Let us now return to the issue of clinical research in developing countries by focusing on the possible use of PCTs such as ST. As I noted at the outset, it is often claimed that clinical research must not be exploitative if it is to pass ethical muster. Commentators have implicitly advanced a version of what I called The Exploitation Argument. That argument maintains something like this:

1. (1) If a practice is exploitative, it should not be permitted.

2. (2) Placebo controlled trials (PCTs) that withhold proven effective treatment are exploitative.

3. (3) Therefore, IRBs or other relevant decision-makers should not permit PCTs such as ST.

This argument also applies, of course, to other ways in which clinical research is thought to be exploitative, such as the failure to provide post-trial treatment to those who received an intervention in the course of a trial. On closer inspection, and as evidenced by the quotations with which we began, those who claim that PCTs such as ST are exploitative actually make several different but related arguments and offer several different prescriptions for nullifying or reducing the alleged exploitation. In this section, I seek to apply the previous analysis of the concept of exploitation to these issues.

Three preliminary observations. First, it is not clear whether those who accept the exploitation argument regard trials such as ST as cases of HNE or MACE. They do not explicitly recognize the distinction or appreciate its potential significance. Second, if we assume that participation in such research is advantageous to the participant and consensual, I hope to have shown that (2) is more difficult to support than is often supposed because we lack a non-problematic theory of the fairness of transactions. Third, even if we assume that (2) is true, I hope to have shown that it is not at all obvious that we (p. 224 ) should accept (1) and (3). Indeed, it is arguable that (1) should be rejected unless one of the arguments for intervention can be sustained and is applicable to this context.

Is The Surfaxin Trial Exploitative?

As we have seen, some commentators have argued that a trial such as ST is exploitative because it would deliberately withhold standard treatment that is available in the home country of a sponsor such as DL. The Declaration of Helsinki states that “The benefits, risks, burdens and effectiveness of a new intervention must be tested against those of the best current proven intervention.”⁵⁷ Call this The Standard Care Principle. ST certainly seems to be incompatible with this principle unless we follow Robert Levine and interpret “best proven current intervention” as “the de facto standard of care” that “would otherwise be available” to the subjects in their locale.⁵⁸ That would be a fudge. In any case, the issue here is not how best to understand the Declaration of Helsinki. The issue is whether a straightforward reading of the Standard Care Principle is a defensible principle for the non-ideal conditions where such care is not available and where investigators have no independent obligation to provide it. If DL is not required to provide any care at all to Bolivian infants, is it unfair to offer them a 50% chance of receiving surfactant therapy? If it is unfair, why is it unfair?

It is worth noting here that analogues to the standard care principle also arise in non-medical contexts.⁵⁹ Just as we can ask what constitutes fair treatment of research participants, we can ask what constitutes a fair wage to workers. Is Nike treating its workers in Thailand unfairly if it offers what are low wages by American standards but higher than the wages earned in comparable work in Thailand? The wages may well be too low, but few believe that fairness requires that Thai workers be paid the same as American workers.

It may be argued that the standard care principle does not reflect a general principle of fairness, but the view that physicians have special moral obligations to their patients that are not replicated in other contexts such as manufacturing. As Sonia Shah puts it, “If researchers know that one option is better than the other, they are ethically obligated to simply administer it: failing to do so puts subjects—who are also their patients—in harm’s way”⁶⁰(emphasis added). Lurie and Wolfe argue that researchers are “obligated to provide the best scientifically proven intervention—independent of (p. 225 ) the economic status of the volunteers and regardless of where the study is conducted.” Not to provide such interventions “is completely inconsistent with the Hippocratic oath…”⁶¹ On this view, investigators exploit subjects when they fail to treat subjects as patients—however consensual their participation as subjects might be.

There are two major difficulties with this line of argument. First, if clinical researchers should regard their subjects as patients, then one needn’t appeal to exploitation to criticize such studies. The claim that researchers exploit their patients would not be false, but it would be entirely otiose. Second, this argument for the standard care principle assumes what has to be shown, namely, that researchers are obligated to treat participants as if they were patients. The issue is not whether investigators have M.D. degrees. Even if the investigators are not physicians, it is possible that the principles of medical practice should govern the relationship of investigators with subjects. But assuming that the researchers are physicians, we cannot assume that the ethical principles that govern clinical research are identical with or entailed by the principles that apply to relations between physicians and patients. Surely, simply referring to participants as patients does not show that investigators must or should treat participants as patients.

One move won’t work. If researchers should treat subjects as patients (or something close to that), they may not be able to avoid such obligations by seeking and receiving consent. Suppose a physician employed by DL says this to the parent of a prospective subject:

Although I am trained as a doctor, it is important that you understand that I do not regard your child as my patient. I am conducting a trial. If you choose to enter the study, your baby will be placed on a ventilator in the hospital. Your baby may or may not receive what we hope will be shown to be a beneficial treatment. In effect, we will flip a coin. If it comes up heads, the air that is pumped into your baby’s lungs will contain Surfaxin. If it comes up tails, the air that is pumped into your baby’s lungs will have nothing else. If you agree to enroll your baby in the trial, your baby’s chances of survival will be better than and certainly no worse than if you did not participate in the study. But my goal is to determine whether and how well Surfaxin works, and not to provide the best possible treatment for your baby. It is unfortunate that your baby is in this situation and that medical care is not (p. 226 ) otherwise available to you, but we are not responsible for that. If you don’t want to participate with us on those terms, that’s your choice.

In my view, uttering this statement does not suffice to get investigators off the moral hook. It is at least possible (although not obviously true) that some of the ethical obligations of researchers are inalienable, that is, they cannot avoid these obligations by announcing that they do not intend to fulfill them and receiving the subjects’ consent to interact with them on those terms.

On the other hand, the statement reflects the plausible view that there is an important distinction between the obligations of physicians to patients in personalized medical care and the obligations of investigators to subjects in medical research.⁶² Defending the precise contours of that distinction is an issue on which considerable ink has been spilt. But there must be some such distinction. Although the researcher may be constrained to avoid certain sorts of harms to subjects, the aim of research is to seek generalizable knowledge. By contrast, the aim of personalized medical care is to advance the health of the patient and setting aside certain complications (such as duties to warn third parties), the treating physician has an obligation to pursue that goal. Surely a researcher could not ethically adhere to a study protocol or administer study procedures not required for treatment if he were under an obligation to deliver personalized medical care.⁶³ In any case, the present point is that whereas one cannot claim that ST is not exploitative just because it involves medical research, one can’t assume that it is exploitative simply by appeal to the ethics of personalized medical care.

And so we are back to the question as to whether ST is exploitative. I argued above that a transaction is exploitative only if B receives less from (or pays more for) the interaction than is required by a plausible principle of fairness. This raises several issues. First, I argued that A exploits B only when A gains or expects to gain from the transaction. In the present context, it is ambiguous as to precisely who are the beneficiaries on the side of the alleged exploiters. It is clear that DL is thought to be the principal agent of exploitation in ST, but it is worth noting that such research might be regarded as exploitative because citizens in the developed world, who risk nothing, stand to gain from trials conducted on citizens of underdeveloped societies. In other cases of alleged exploitation, such as SCAT, the study may be conducted by a not for profit institution, such as the National Institutes of Health. Who is gaining from the alleged exploitation? We might say that the (p. 227 ) scientists gain, but, in this case, the intended indirect beneficiaries are the infants born to HIV-positive mothers in developing countries.

Second, if a transaction is exploitative when the gains from the transaction are unfairly distributed, we must determine whether we should evaluate those gains from an ex ante or an ex post perspective, and whether we should adopt the same perspective for both parties. Angela Ballantyne argues that “both the benefits and burdens associated with trial sponsors are primarily guaranteed.”⁶⁴ This is false. Although the costs to trial sponsors may be roughly guaranteed, the scientific and financial benefits are not. That is why the trial is done. Consider this case.

Oil. A proposes to buy land from B in the hopes that he will find oil. On the basis of a geologist’s analysis, A’s expected profit from drilling will be $10 million. A offers to pay B $100,000. B accepts. A finds no oil and loses $1 million on this project.

I believe that A may have exploited B in Oil because A’s expected gain may have been excessive relative to B’s actual gain, even though A’s actual gain was negative whereas B’s actual gain was positive. Similarly, I believe that investigators can exploit trial participants even if, ex post, the participants gain whereas the investigators do not.

Ballantyne argues that an ex ante perspective is insufficient. She thinks that non-exploitative research should provide participants with a positive increase in expected utility and a fair ex post distribution of benefits.⁶⁵ It would seem that sauce for the goose should be sauce for the gander. If, as I have argued, we should evaluate the gains to “A” from an ex ante perspective, I do not see why we should evaluate the gains to “B” from both an ex ante perspective and an ex post perspective. The crucial question is whether investigators have offered a fair ex ante package of benefits to research participants. After all, even if A is offering personalized medical care to B, we do not evaluate the ethics of A’s behavior by reference to its ex post effectiveness.

From an ex ante perspective, then, it is arguable that a PCT such as ST is actually beneficial to all participants, since all participants have a 50% chance of receiving beneficial treatment and, in the actual case of ST, even those not getting the drug would have profited from intubation and better all around care. This is not a conceptual trick. Consider standard medical treatments. If B has an illness for which the standard therapy works 50% of the time, A can expect to be paid even if the treatment does not work for B. I am not here claiming that it is unproblematic to conduct a PCT when standard treatment (p. 228 ) is available. I am claiming only that in asking whether ST is exploitative, we should evaluate the benefits to participants from an ex ante perspective.

Third, it is not clear how to measure the parties’ relative benefits in a case such as ST. The critics of ST tend to emphasize that the participants might not receive surfactant therapy, but say little as to what they do receive. If we evaluate the participants’ benefits in something like “utility gain” relative to the status quo in which they receive no medical care at all, the expected utility gain of participants in Bolivia—an increased probability of survival—is enormous.⁶⁶ Moreover, assuming that there is an adequate supply of potential participants, DL does not gain that much from the participation of any individual participant. I am not claiming that comparative utility gain is the best way to assess the fairness of a distribution of benefits. To the contrary. I suggested above that A can exploit B even if B gains much more from the transaction than A (as in Unfair Surgery). At the same time, the claim that a distribution of benefits is unfair or exploitative must also be defended. And setting aside an appeal to the “standard care” principle, it is not at all clear why the ex ante distribution of benefits between DL and the participants in ST is unfair. And if it is not unfair, then the claim that the subjects are exploited does not get going.

Fourth, some critics claim that trials such as ST are exploitative because they “serve the interests of those who live in wealthy nations.”⁶⁷ We do well to remember that the indirect beneficiaries of research have genuine medical needs. More importantly, and as to whether the subjects are exploited, the issue is not the legitimacy of the interests that are being served, but whether those who serve those interests are being treated fairly. If those who knit oriental rugs are treated fairly—and I do say if—then they are not being exploited just because they are providing goods for affluent westerners.

A fifth point deserves more extensive discussion. Call it the off-shoring objection. Off-shoring and out-sourcing are often conflated. We outsource labor when a person or firm hires others to provide some service. And so a university might “outsource” its custodial work or landscaping to a private firm. We “off-shore” labor when we use persons from other nations—typically separated by oceans—to provide some service. Objections to off-shoring labor and services are often raised in non-medical contexts such as the manufacture of clothing in Asian sweatshops or the use of customer service agents in India. Still, it may be thought that there is something particularly obnoxious about a practice in which the affluent directly or indirectly entice poor people from LDCs to provide a medical service that will primarily (p. 229 ) benefit comparatively affluent people in developed societies, and it may seem particularly worrisome when the inducement to participate is the prospect of receiving life-saving medical treatment as in ST. And while the off-shoring of clinical research is a growth industry, it has been the target of considerable scholarly criticism. More to the point, I have not encountered any sustained attempt to defend the practice in the academic literature.

But what, precisely, is the ethical problem here? At a basic level, outsourcing or off-shoring is simply an extension of the division of labor and trade. We “outsource” labor whenever one party has a comparative advantage in producing a good or providing a service, and a party can have a comparative advantage over another even if one party is superior at both. The principle works within families. My wife may be a better cook and a better cleaner than I, but if she is much better at cooking and only a little better at cleaning then I have a comparative advantage in cleaning. It makes sense for her to outsource the cleaning to me and for me to outsource the cooking to her. Similarly so within firms. A lawyer may be able to type faster than her secretary, but her secretary has a comparative advantage with respect to typing nonetheless.⁶⁸ It is more efficient for her to “outsource” her typing to a secretary who is paid $20 per hour while she engages in legal work for which she is paid $200 per hour. The affluent hire nannies, gardeners, house cleaners, painters, or volunteer (professional) soldiers to perform tasks that they could but would prefer not to perform. This is the stuff of everyday life and as a general principle it is a desirable practice so long as such transactions are consensual, mutually advantageous, and fair. Moreover, we may think that we should not interfere with such transactions if they are consensual and mutually advantageous, even when they are not fair.

In principle, the off-shoring of labor across national boundaries is no different. If Britain has a comparative advantage over France as a producer of cotton, whereas France has a comparative advantage as a producer of wine (even if France is a more efficient producer of both cotton and wine), then it makes sense for France to produce wine for Britain and for Britain to produce cotton for France. Americans (and that includes low income Americans) benefit from the fact that goods are produced at lower cost in China than if they were produced in the United States.⁶⁹ As Adam Smith puts it,

What is prudence in the conduct of every private family, can scarce be folly in a great kingdom. If a foreign country can supply us with a commodity cheaper than we ourselves can make it, better buy it (p. 230 ) of them with some part of the prudence of our own industry, employed in a way in which we have some advantage.⁷⁰

As with products, so too with services. American customers benefit from having telephone inquiries answered by persons in India rather than in Indiana, given that the Indians are paid substantially less than Indianans. We may object to and work to change the circumstances that render it reasonable for people to consent to work for low wages but, as a general principle, we do the people of LDCs no favor if we deny them the opportunity to produce goods and provide services that will mostly be consumed in the developed world.

It is, of course, an empirical question as to whether and how much people—and which people—in LDCs benefit from trade with the developed world. Even if there are aggregate gains from trade, there will certainly be individual losers from trade within every society. If the United States off-shores the production of steel, the American users of steel may benefit, but steel workers will not. If kleptocratic rulers in LDCs capture much of the surplus of trade with the developed world, the majority of people in LDCs may not gain. All that said, it is certainly plausible to think that most workers in LDCs actually do benefit from trade. A moderate left-wing economist such as Paul Krugman writes that “while fat-cat capitalists might benefit from globalization, the biggest beneficiaries are, yes, Third World Workers.”⁷¹ Krugman adds that wherever “new export industries have grown, there has been measurable improvement in the lives of ordinary people.”⁷² Such industries typically offer higher wages than workers could receive elsewhere and developing the export sector typically has a positive “ripple effect throughout the economy.” And, as Martin Wolf points out, this is particularly so for women, “whose status can be transformed in patriarchal societies by opportunities to earn incomes for themselves.”⁷³

I think it fair to say that philosophers and ethicists have shown little interest in the benefits of trade or globalization. As Teson and Klick have noted, “None of the major works on global justice draws on the relevant economic literature, the general consensus of which recommends free trade as a way to enhance global and national wealth and thus benefit the poor.”⁷⁴ It is not clear why this is so. Teson and Klick speculate that philosophers may lose interest because the defense of free markets relies on empirical economic arguments that are not well understood and because the benefits of free trade, such as they are, flow from egoistic behavior rather than a (p. 231 ) commitment to global justice and the amelioration of deprivation—“Promoting trade… is not a sufficiently lofty way to discharge our duties based on justice.”⁷⁵ There may also be a signaling issue. Given that the benefits of free trade are often touted by persons on a side of the political spectrum with which many do not wish to be associated, commentators might be concerned that sympathy for this approach may not convey the appropriate signals about their own political and moral sympathies.

Assuming that the general case for the benefits of free trade is roughly correct, the off-shoring of clinical research has the potential to be a win-win transaction. To put the point in admittedly crude terms, it could be argued that people—including infants—in LDCs have a “comparative advantage” in serving as research participants and that if we are concerned for their welfare, there should be no objection to off-shoring medical research in much the same way that we off-shore the manufacture of apparel or customer service support for computers. Distasteful as it might seem, it may be advantageous for the citizens of LDCs to be used as research subjects, and, if so, we should encourage its development.

I hear the groans. It might be thought that it is one thing to employ Indians to answer phones, and quite another to employ them as subjects in a trial of a new drug for Type-2 diabetes, whether they do it for money or for access to medical treatment that they would otherwise not receive. Shamoo and Resnik say that

it is especially important to prevent pharmaceutical companies from using people in developing countries as cheap labor to test drugs that will only be used in the developed world, because this would constitute an egregious form of exploitation.⁷⁶

They do not say why this practice is exploitative much less “egregious” exploitation. Moreover, if the egregious exploitation that they seek to prevent is a form of mutually advantageous and consensual exploitation, it is not clear why we should want to prevent it. Would they seek to prevent running shoe manufacturers from “using people in developing countries as cheap labor to manufacture shoes that will only be used in the developed world” if what is cheap labor to us are among the better jobs available in such societies? And if Shamoo and Resnik believe that we should not seek to prevent such transactions in the non-medical sphere, why should we seek to prevent such transactions in the medical sphere?

(p. 232 ) We have been assuming that the transaction between investigators and participants in ST is mutually advantageous. The question is whether investigators are exploiting subjects nonetheless when they conduct a placebo controlled trial if an efficacious intervention is available. I have argued that if we set aside direct appeals to the “standard of care” principle, the critics have advanced no independent arguments as to why the distribution of benefits between the parties is unfair—perhaps because they think it so obvious that ST is exploitative that an argument is not necessary. After all, the investigators would be “right there” and would be withholding potentially life-saving treatment from infants they are using to develop generalizable knowledge. If that’s not exploitative, what is? Although I think the issue is more complicated than that, I want to assume, for the sake of argument, that ST is exploitative and that DL must do more for its subjects (or perhaps others) if it is to avoid the charge of exploitation.

As should now be clear, even if we assume that ST is exploitative, the question remains as to whether DL should be permitted to go forward with the trial if participation is both advantageous (ex ante) to the exploitee and consensual. Lurie and Wolfe say that “An unethical research study is an unethical research study even with informed consent.”⁷⁷ This is clearly too quick, for the question is whether morally transformative consent is sufficient to transform a study that would otherwise be unethical into a study that is rendered ethical. But even if a subject’s transformative consent does not render the study ethical, such consent might justify not interfering with the study. It is to the issue of consent that I now turn.

Consent

There are several questions that we can raise here: (1) do the participants (or their surrogates) in studies such as ST typically give morally transformative consent? (2) If not, is it possible that such participants can give morally transformative consent? (3) Is “community consent” necessary or sufficient to satisfy this dimension of ethical research?

Clearly, the answer to questions (1) and (2) will turn on our criteria for morally transformative consent (which I considered in Chapter 3) and the relevant empirical data. Although we cannot resolve that here, I am sure that the answer to (1) is at least sometimes no. There are no doubt numerous and egregious cases in which investigators have not met the requirements of any plausible view of morally transformative consent. They may have deceived (p. 233 ) or failed to give relevant information to prospective participants. There may be cases in which the participants are not competent to give morally transformative consent. Moreover, it is entirely possible that the parents will not fully understand that the researchers are prepared to withhold life-saving therapy from their infants.

It may or may not be feasible to resolve these sorts of cognitive mistakes. What then? As I argued in Chapter 2, a major function of the independent review process is to assure that it would be reasonable for the participant to give consent when she is not in a position to make such judgments. If prospective subjects (or surrogates) are not themselves capable of giving morally transformative consent because they lack the ability to comprehend what the research involves, there are two (compatible) options. First, an independent review process can determine that it is reasonable for the subject to consent and, in effect, ask the subject to give the functional equivalent of a child’s “assent.” Second, we can require that the participant’s (or parent’s) expressed consent be supplemented by the consent of a surrogate such as a representative of the government or a health official who can reliably determine that the participant’s decision is consistent with her long term interests.

We should, however, distinguish between the sorts of cognitive deficiencies in a subject’s consent that can be remedied by IRBs or surrogates and worries about the voluntariness of the subject’s consent. It is sometimes argued that research such as ST is not based on morally transformative consent because the participants have no reasonable alternative but to consent to participate. On this view, impoverished parents who would ordinarily have no medical care available to treat their infants with respiratory distress syndrome do not give morally transformative consent to participate if that is their only chance of obtaining medical care for them. They are coerced by their circumstances as it were.

As I argued in Chapters 3 and 4, the fact that prospective subjects do not have viable alternatives to participation does not preclude their giving morally transformative consent. We do not say that a patient’s consent to surgery is involuntary, coerced, or invalid just because her only alternative is to die, and if we do not think that the prospect of death undermines valid consent to treatment, there is no reason to think that the prospect of receiving no medical care undermines valid consent to participate in research. I have argued that A’s proposal is coercive in a way that renders B’s consent non-transformative only if A proposes to violate B’s rights if B rejects A’s proposal. And the central fact is that DL does not propose to violate a (p. 234 ) potential subject’s rights in ST should a potential subject decide not to participate.

Annas and Grodin disagree. They argue that

in the absence of health care, virtually any offer of medical assistance (even in the guise of research) will be accepted as ‘better than nothing’ and research will almost inevitably be confused with treatment, making informed consent difficult⁷⁸ (emphasis added).

But parents would not be confusing treatment with research in this case if participation represents their only access to treatment.

There may be other worries about consent. As I have argued, a proposal might constitute a seductive offer or undue inducement even if it is not coercive and, if so, it would compromise consent in a different way. CIOMS worries that the offer of medical care may “induce prospective subjects to consent to participate in the research against their better judgment.”⁷⁹ But as I argue in Chapter 4, it is important to distinguish between two claims:

1. (1) an inducement constitutes a seductive offer that distorts the judgment of the subjects and motivates them to consent to participate when doing so does not advance their interests;

2. (2) given the subject’s objective circumstances, the inducements make it rational for the subjects to participate.

When (1) is true, there is good reason to regard consent as not transformative, but there is no reason to regard consent as not transformative in (2).

The point, disturbing as it might be, is that although extreme poverty can give rise to the distorted judgments in (1), it can also lead to judgments that are crystal clear. As Shamoo and Resnik put it, “We should not equate economic vulnerability with lack of sound judgment.”⁸⁰ And if (2) is true, we do not respect the subject’s humanity or rationality by denying their capacity to make their lives less miserable than they already are. David Rothman writes that “abject poverty is harsh enough without people having to bear the additional burdens of serving as research subjects.”⁸¹ But if serving as a research subject is a benefit, all things considered, we might say that abject poverty is harsh enough without denying people the opportunity to make their lives somewhat less miserable by participating in biomedical research and receiving benefits that they would not otherwise receive.

(p. 235 ) Let us assume that the participants (or their surrogates) do or can give morally transformative consent to participate in ST. Some commentators have argued that medical research such as ST not only requires the consent of the participants (or their surrogates), but also requires the consent of the community. This view need not claim that a community’s consent could override the absence of consent from an individual. After all, a major purpose of insisting on informed consent is so that an individual’s interests cannot be involuntarily sacrificed for the interests of the community. It does say that the community’s consent might also be necessary.

Broadly speaking, there seem to be at least four arguments for requiring community consent in LDCs. First, some commentators reject what they regard as a “highly individualistic” conception of rights and duties that underlies the significance of informed consent, one that allegedly assumes an “atomistic view of the person” that they find wanting or, perhaps, that is not accepted in some developing societies even if it is accepted here. This is a large topic that we cannot here pursue in depth. But suppose that a society believes that a woman should secure her husband’s permission to enroll in a trial and that her own consent is not sufficient. It might make sense as a political or practical matter to accommodate such beliefs, but I am not prepared to grant that refusing to do so assumes an unjustifiable “atomistic view” of the person. Here I would prefer the company of the Universal Declaration of Human Rights, which also appears to regard individual rights as a universal value.

Second, it may be argued that a community’s consent can be required because its present and future members have interests that are not identical with the interests of the individual subjects. A study might have genuine negative externalities for the community, and, if so, it is perfectly sensible for the community to demand that its consent is also necessary to protect its interests. For example, research might draw a community’s medical personnel away from their traditional work to the detriment of the community (although, perhaps, to the advantage of the medical personnel). It is another question, however, as to what sorts of alleged or perceived interests justify denying the opportunity for individuals to participate in research if they consent. Charles Weijer observes that “Providing information on disease treatment may negatively affect beliefs regarding traditional healing.”⁸² No doubt true. But it is doubtful that a community’s interest in preserving false beliefs about healing justifies preventing its citizens from receiving efficacious medical care to which they consent.

(p. 236 ) Third, a community may require that its consent be obtained so that it can leverage the individual’s participation to promote the interests of other members of the community. I will not here assess the merits of this argument save to note that it has absolutely nothing to do with protecting the subject from exploitation.

Fourth, the community or its government may be well positioned to act as a proxy for the subject, given that the subjects may exhibit the sorts of cognitive deficiencies I noted above. As I argued in Chapter 2, there is nothing untoward in this form of “soft paternalism” if the community can and will make sensible decisions on behalf of persons whose competence is suspect. In this case, however, the community’s consent is required not because it has its own interests to protect but because it is better positioned to protect the interests of the subject than the subjects themselves.

But this raises a difficulty with some of the arguments advanced by the critics of studies such as ST. If we take seriously the community’s authority to reject ST on behalf of its members, then it is difficult to justify ignoring its acceptance of ST. There may be a strategic argument that justifies prohibiting a community from accepting such a trial if doing so will prevent a “race to the bottom” among the societies that might compete for being used for such studies and if so doing will result in studies that provide better care for the participants. But bracketing such reasons, it will be difficult to argue that a trial should not be permitted if both the prospective subjects and the community’s representatives consent to it.

Reasonable Availability and Fair Benefits

As I noted at the outset of this chapter, those who claim that much multi-national research is exploitative focus not solely on the trial design itself, but on the availability of benefits when the trial is completed. Annas and Grodin say

Unless the interventions being tested will actually be made available to the impoverished populations that are being used as research subjects, developed countries are simply exploiting them in order to quickly use the knowledge gained from the clinical trials for the developed countries’ own benefit⁸³ (emphasis added).

This and similar statements are ambiguous as to who has a claim to receive such benefits. If they are claiming that the research subjects themselves have (p. 237 ) a claim to receive benefits when the trial is over, then we must ask whether such benefits are required to render the package of benefits fair from an ex ante perspective. But that seems an unlikely interpretation of their view. For one thing, it is irrelevant to research such as ST. If the infants survive the trial, they will not need surfactant therapy in the post-trial environment.

It is more natural to regard “them” as referring to the community. On that view, if we are to avoid exploitation in multi-national research, the community whose members participate in a trial should somehow benefit from the research. On one view, there must be some reasonable assurance that products derived from the research will be available in that country. As CIOMS puts it, “As a general rule, the sponsoring agency should agree in advance of the research that any product developed through such research will be made reasonably available to the inhabitants of the host community or country at the completion of successful testing.”⁸⁴

Advocates of the “fair benefits” view take a more catholic approach to community benefit.⁸⁵ They argue that the amount and not the type of benefit is crucial to avoiding exploitation. They note that reasonable availability applies only to Phase III research of a successful product, and would have little bearing on Phase I and Phase II research, as well as Phase III research that did not yield a successful product. In addition, they point out that reasonable availability does not consider other potential benefits of research in developing countries, including training of health care or research personnel, and the development of the health care infrastructure. They also suggest that insisting on reasonable availability is condescending, as it “precludes the community’s deciding which benefits it prefers.”⁸⁶ I consider this issue in more detail in Chapter 6. The present question is whether and how the provision of benefits to the larger population would negate or mitigate exploitation of research subjects. Consider these four claims.

1. 1. There is an intrinsic connection between the exploitation of the research subjects and the exploitation of the population, such that research that would otherwise not exploit the research subjects does exploit those subjects if the wider population is not treated fairly.

2. 2. There is an intrinsic connection between the exploitation of research subjects and the exploitation of the population such that research that would otherwise exploit the research subjects (p. 238 ) would not exploit the research subjects if the wider population is treated fairly.

3. 3. Although the research may exploit the subjects, the moral weight of this exploitation would be offset if the community receives sufficient benefits such as to render the research permissible all things considered.

4. 4. Although the provision of benefits to the larger population is entirely distinct from the exploitation of research subjects, ethical research requires (1) that the research subjects not be exploited and (2) that the larger population from which the subjects are drawn also benefit from the research.

I am inclined to reject (1) on conceptual grounds. I simply do not see why a participant is exploited because the community does not benefit. But perhaps that is too quick. Winston Chiong argues that

Subjecting a patient to the risks of clinical research, without making good faith efforts to ensure that the benefits provided by this research will be available in the patient’s own community, suggests a lack of respect and concern for a patient’s broad social circumstances that we would not reasonably want physicians to bear towards any of their patients.⁸⁷

Chiong’s claim rests on two questionable assumptions, namely that research subjects should be understood as “patients” and that participation is risky rather than (ex ante) beneficial. But setting that aside, Chiong is not actually arguing that there are independent reasons for providing benefits to the community in addition to the subject. He seems to be arguing that not providing those benefits to the community shows a lack of respect for the subjects themselves. It would seem, however, that the relevance of the investigator’s respect for the “broad social circumstances” of the subject to showing respect for the subject would be for the subject to decide. If the mother of a child with respiratory distress syndrome is prepared to have her child participate in ST even if DL will not make Surfaxin available to other Bolivian infants, we can’t bar her participation by reference to respect for her or her child.

I would similarly reject (2) on conceptual grounds. I do not see why the most generous provision of benefits to the community would negate or mitigate the exploitation of the subjects themselves. Suppose, for example, that (p. 239 ) football players at University X should receive a fair share of any surplus that X receives from their efforts, but that X uses that surplus to support money-losing athletic programs for women. Although the use of the surplus for those ends might justify the exploitation all things considered, it does not cancel or negate the exploitation itself.

Claim (3) is more plausible. From a consequentialist perspective, it is possible that the provision of benefits to the community might justify or override the wrongness inherent in the exploitation of subjects even if it does not remove or negate the wrongness of the exploitation. But to the extent that the principles of ethical research are primarily concerned with the interests and autonomy of research subjects, we should be loathe to tolerate such exploitation as a means by which to generate benefits to the community. So (3) is on shaky grounds.

On (4), research is ethically justifiable only if it is justifiable with respect to the subjects themselves and with respect to the larger populations from whom the subjects are drawn. As Ezekiel Emanuel puts it, “it is an ethical imperative that populations receive sufficient benefit from research in which they participate”⁸⁸ (emphasis added). Although this is a plausible view, I do not think it is persuasive. If “they” refers to the populations, it does not seem as if “they” participate in research at all. It is the subjects who participate, not populations. And if ST is beneficial to the participants and if they or their proxies give morally transformative consent, it seems prima facie wrong for the community to hold them hostage to DL’s willingness to provide benefits to others. (Of course, communities may provide resources of various kinds—personnel, logistical support, infrastructure, etc.—that could have been used for other purposes. If so, communities can reasonably ask to be compensated for their investment. I consider that claim in more detail in Chapter 6.)⁸⁹

Now the view that investigators should provide benefits to the community might not be justified by the goal of avoiding exploitation but, rather, by principles of beneficence or distributive justice that seek to provide more health care resources to the poorest areas of the world. That is plausible. And a more fully developed non-ideal theory than I have offered would provide principles for moving from our current non-ideal world to a world that is more just. Still, we must be careful not to conflate valid concerns about the distribution of medical resources in the world with concerns about the relationship among the participants or the nations involved in particular studies and with the exploitation of those participants.

(p. 240 ) In this connection, it is also important to distinguish between the claim that the distribution of resources in the world is unjust and the claim that the sponsor or the home country of the sponsor is causally responsible for the injustice. Crouch and Arras argue that the misery of people in underdeveloped societies “must be due in no small measure to the flagrantly unjust behavior of the former colonial powers, which plundered their natural resources and subjugated their peoples…”⁹⁰ Along similar lines, Thomas Pogge has argued that the affluent nations of the world have harmed—not merely failed to help—the worse off by upholding an institutional regime that, among other things, allows the rulers of many nations to enrich themselves by selling natural resources to wealthy nations while refusing to share the benefits with the population.⁹¹ I am sure that this causal view is at least part of the truth, but I suspect that it is only partially correct. Some of the poorest societies of the world have had little economic contact with the highly industrialized nations. They have suffered not because imperialism has made them worse off, but because the affluent nations have “found too little there to be exploited.”⁹² Of course, even if the industrialized nations have not caused the poverty of the underdeveloped society, they may still have obligations of justice or beneficence to ameliorate that poverty. Moreover, it is arguable that given the special competencies of medical investigators and the resources available to their sponsors, they should do their part in this general effort through medical research. But the force of those arguments has little to do with exploitation.

The Permitted Exploitation Principle Revisited

I have argued that the claim that ST is wrongfully exploitative of the research subjects or the community is more difficult to sustain than is often supposed. But let us assume, for the sake of argument, that ST or some comparable trial is exploitative. If ST is harmful or non-consensual, then these are good reasons not to allow it to go forward, although we would not then need to invoke its exploitativeness as the rationale for interference. But if ST is a case of MACE, I have argued that there is good reason to endorse the Permitted Exploitation Principle as a default position. I considered several rationales for resisting PEP. The question is whether they apply in the present context and to what effect.

Remember that PEP is a claim about the ethics of interference and not a claim about the ethics of a transaction itself. In a commentary on ST, (p. 241 ) Robert Temple confuses these two issues: “One must ask… why, if everyone in a trial is better off because of participation, and no one is denied anything otherwise available to them, the trial is not ethically acceptable.”⁹³ I argued above that it is not ethically acceptable for A to exploit B in Rescue just because B is better off and is not denied anything otherwise available. And the same is true for clinical research. Temple may be right in thinking that ST should be permitted, but wrong to think that ST is ethically acceptable because it renders everyone better off (at least ex ante).

Alex John London makes a similar mistake. London refers to the essence of PEP as the “minimalist view” because it claims that we should permit a transaction so long as it does not worsen anyone’s position and is consensual. Setting aside the derisive language (who would want to support a position described as “minimalist”?), London is mistaken as to what PEP claims.

The minimalist view accepts the status quo in the host community as the appropriate “normative baseline” against which proposed research initiatives are evaluated – meaning that the status quo is treated as the threshold of a person’s moral entitlements in this particular sphere.⁹⁴

Not so. PEP does not “accept” the status quo in any deep moral sense. PEP need not deny that all Bolivian infants with RDS are morally entitled to surfactant therapy. But it does not follow that DL, in particular, has an obligation to provide surfactant therapy to any or all Bolivian infants with RDS. It is important to note that London does not claim that DL has an obligation to conduct its research in Bolivia just so that it can make surfactant therapy available to those to whom it would not otherwise be available. London might claim that DL has an obligation to provide surfactant therapy to its subjects because the moral entitlement of research subjects shifts as soon as they become research subjects. But that is to presuppose precisely what is at issue.

London also maintains that PEP assumes that transactions are fair simply “because they provide mutually beneficial terms of cooperation that each party can freely accept.”⁹⁵ Not so. PEP does claim that transactions are fair just because they are mutually advantageous. It claims that there is reason to permit such transactions even if they are unfair.

In addition to making questionable claims about the content of PEP, London also makes questionable claims about the consequences of accepting PEP.

(p. 242 )

… the minimalist position prevents one from assessing the extent to which individuals in the host community have a legitimate claim against their own government to better conditions. It also keeps one from assessing the influence of third parties, such as foreign governmental and corporate entities, on the community’s basic social structure.⁹⁶

Not quite. To accept the view that we should permit ST because it is advantageous to all concerned hardly “prevents” anyone from also asking whether the Bolivian government is doing all that it could or should for the children or from asking whether the decisions of third parties have harmed the basic social structure of Bolivia, and it is hard to see why anyone would think that it would.

Of course, even if the “minimalist” view does not “prevent” us from addressing other or larger issues, it is arguable that it “does little to bring attention to the root causes of the developing world populations’ most pressing health needs” and that its acceptance would perpetuate “an ad hoc and piecemeal approach to the health needs of populations that already bear the greatest burden of disease and deprivation.”⁹⁷ This is not a claim about the truth of the minimalist view, but an empirical claim about the effects of adopting that view. It is possible that the acceptance of a view that encourages mutually advantageous and consensual transactions within a context of injustice and deprivation does less to address those background conditions than a view that focuses on the background conditions themselves even if the latter view might lead to avoidable harm to those who would benefit from such transactions. On the other hand, it is also possible that precisely because acceptance of the “minimalist view” will permit studies such as ST, it will draw attention to the desperate background conditions that make it rational for parents to enroll their infants. It’s hard to tell.

In any case, the present question is not whether accepting PEP represents the most effective mechanism of change, but whether PEP should be accepted on its own terms. The case for accepting PEP is simple and straightforward, although not less meritorious for that: if a transaction is better for the exploitee and if the exploitee consents to be exploited, then we should allow the transaction to go forward. I have argued, however, that even if we start with PEP as the default position, there is a plausible strategic argument for prohibiting exploitative transactions such as ST. If refusing to allow ST will result in a less exploitative or non-exploitative trial, then we can justify (p. 243 ) the prohibition with respect to the interests of the participants. Whether the strategic argument actually justifies a given prohibition depends on the counterfactual: what will happen if we do not allow the transaction? If refusing to allow the transaction will cause the exploiter to offer less exploitative terms, then such refusal makes sense. If refusing to allow the transaction will cause the exploiter to go elsewhere and not transact with these prospective participants, then the refusal does not advance the interests of these exploitees. It is an empirical question as to what the effects will actually be.

Consider two scenarios. In Scenario One, the costs of doing research in Bolivia are greater than the costs of doing research in the United States, say because it is necessary to provide the infrastructure that makes such research possible. In that case, DL would prefer to conduct its trial in the United States unless differences in the trial design made it less costly to do the research in Bolivia. The advantage of a PCT is that sample size can be smaller, the results can be obtained more quickly, and costs are lower because there is no need to purchase drugs produced by rival companies. On this scenario, DL might rank the options in this way:

1. 1. PCT in United States

2. 2. PCT in Bolivia

3. 3. ACT in United States

4. 4. ACT in Bolivia

Option (1) is a complete non-starter since such a trial would not be approved for nations where infants would normally receive surfactant therapy, and, even if it were approved, parents would never consent to participate. If (2) is also disallowed as exploitative, then DL will prefer to conduct an ACT in the United States (3) than to conduct an ACT in Bolivia (4). In this case, the strategic argument does not apply; intervention does not advance the interests of Bolivian infants.

In Scenario Two, however, the costs of doing research in Bolivia are lower than the costs of doing comparable research in the United States, perhaps because salaries are lower or because the prevalence of the condition under study is greater and it is possible to recruit a sufficient number of subjects at fewer sites, etc. In Scenario Two, DL’s ranking is this:

1. 2. PCT in Bolivia

2. 1. PCT in U.S.

3. 4. ACT in Bolivia

4. 3. ACT in U.S.

(p. 244 ) As with Scenario One, a PCT in the U.S. (1) is off the table. But if DL is not allowed to choose option (2), on the grounds that it is exploitative or because, say, it is inconsistent with the principles incorporated in the Declaration of Helsinki, then given the choice between (3) and (4), DL will prefer (3). In this case, the strategic argument for prohibiting (2) serves to advance the interests of Bolivian infants.

It is an empirical question as to which scenario is closer to the truth in a given case. In the present context, it appears that Scenario One is closer to the truth. Robert Temple of the U.S. Food and Drug Administration conjectured that if DL were required to conduct an active controlled trial, it “almost surely would be conducted in the United States or another developed country” and there would be “no trial of any kind of a surfactant in Bolivia.”⁹⁸ And this is precisely what happened. When DL’s proposed protocol was widely criticized as exploitative, it abandoned the proposal for a PCT in less developed countries and conducted an ACT in the United States and other developed societies. This does not show conclusively that the standard care principle is wrong. It does suggest that advocacy of the principle can be self-defeating if its aim is to advance the welfare of prospective subjects.

Similar issues arise with respect to whether investigators should be required to provide post-trial treatment to participants in clinical trials, an issue that has been particularly salient with respect to the provision of anti-retroviral drugs to persons who become HIV positive in the course of an HIV-vaccine trial or who respond positively to anti-retroviral therapy (ART) in the course of a study of the ART. Using rather neutral language, The National Bioethics Advisory Committee says “It has been pointed out that expecting industrial sponsors to provide expensive drugs free of charge after a trial is over might curtail interest among companies in developing interventions specifically for diseases prevalent in developing countries.”⁹⁹ But it’s not just that this requirement might work to the detriment of the populations in developing countries; it might work to the detriment of the trial participants themselves. The NBAC does not say whether it thinks that this argument has merit or takes a position on its force. It says simply that this concern “has been pointed out” and that the unwelcome consequence “might” occur. But the point is arguably quite crucial. We must ask whether we want to endorse an ethical view that, if it is followed, motivates investigators to abandon research from which prospective participants can expect to benefit and to which they would consent.

(p. 245 ) Pogge’s Challenge

In an important and insightful analysis of exploitation in international research, Thomas Pogge seems prepared to bite precisely that bullet.¹⁰⁰ He argues that to adopt the Pareto principle or what I have called PEP is, in effect, to encourage investigators to scour the earth in search of infants with RDS whom they can intubate but refuse to infuse their air with a surfactant. And, he adds, “It is hard to have confidence in a morality that encourages and rewards such predatorial behavior.”¹⁰¹ When put that way, it seems hard to disagree.

In developing his analysis, Pogge makes use of three standard moral categories: moral freedom, moral goodness and moral constraint (or requirement). Moral freedom refers to that which one is permitted to do. Moral goodness refers to the moral quality of the state of affairs one’s actions bring about. Moral constraint refers to that which one is not permitted to do or that which one is required to do. Pogge accepts that DL has the moral freedom not to conduct its study in a poor country rather than the United States (or not at all). He also acknowledges that “It is morally bad… that many human beings… should die as a result of a drug company’s decision to test a new treatment in its rich home country rather than among the poor abroad.”¹⁰² Nonetheless, Pogge argues that testing Surfaxin against a placebo when effective live-saving treatments are available is a violation of the operative moral constraint or requirement, namely, that physician-investigators should not deliberately withhold life saving treatment from those they are using as subjects when they could provide treatment without incurring extraordinary additional costs.

Now in arguing that ST is unethical, Pogge does not appeal to a general duty to rescue the infants that DL could save, in which case DL might lack the moral freedom to conduct its research in the United States. His point is that it is wrong for DL to use these infants as research subjects by not saving them.

In developing his argument, Pogge asks us to consider a provocative example (which I paraphrase)

The Filmmaker. A filmmaker wishes to produce a documentary about behavior of fishermen whose boats are in distress. He believes that the film will help others avoid counterproductive panic-induced behavior. A successful film requires that the fishermen not be rescued until it is too late. The film crew flies to a location off the coast of a poor country and waits for a (p. 246 ) radio signal of a ship in distress. When it receives a distress signal, the film maker radios back with the following proposal. If the fishermen agree, the filmmaker will flip a coin. If it comes up heads, the filmmaker’s crew will fly by helicopter and save them. If it comes up tails, it will fly by helicopter and film them and make no effort to save them, but will remain in the area until another ship is in distress and it will then save those fishermen. Because the fishermen have no better option, they readily agree.

Pogge maintains that it would be wrong for the filmmakers not to save the fishermen given that they are in a position to do so, even though the fishermen consent to the arrangement, even though the fishermen clearly benefit ex ante from the arrangement, even though the filmmakers have a legitimate purpose in making the film, and, most importantly, even though the filmmakers would have absolutely no reason to be in a position to save the fishermen if they would be obligated to save them once they found themselves in a position to do so.

The italicized passage is crucial. One can deny that the filmmakers have an obligation to put themselves in a position where they can rescue the fishermen and yet also accept that one has a general duty to rescue when such occasions occur in the normal course of events (as when passing by Peter Singer’s pond in which a child is drowning) and even when doing so imposes considerable costs.¹⁰³ Consider this example that Liam Murphy borrows from Shelly Kagan.

Astronomers. For years some astronomers have been planning an observation that must take place at sea at a particular date and time; many hundreds of thousands of dollars have been spent in preparation. Just as they are about to take their readings, they receive a Mayday signal. If they rescue the sailors in distress, the project will come to nothing. But, as we intuitively think, rescue they must.¹⁰⁴

There is an important distinction between the astronomers and the filmmakers. The astronomers have reason to be at sea even if there is a risk that they will have to scuttle their project in order to rescue the sailors. Given that they are in a position to rescue the sailors, it is arguable that they must do so. By contrast, the filmmaker would not put himself in a position to rescue the (p. 247 ) fishermen if he would be required to do so. As Pogge puts it, “The filmmaker has no earthly reason to be near the ocean with his radio equipment and helicopter if this can win him no exciting filming opportunity, but can only slap him with a moral rescue burden.”¹⁰⁵ Pogge recognizes that acknowledging that the filmmakers have the relevant moral freedom not to make the film while insisting on compliance with the operative moral constraint not to let the fishermen perish may yield a world that is less morally good than it would have been in the absence of that constraint. No fishermen will be saved. But, as he puts it, “morality cannot plausibly be purged of such counterproductivity entirely.”¹⁰⁶

But even if morality cannot be purged of all counterproductivity, this is not the sort of counterproductivity that we have come to expect of a sensible morality. We expect that respect for one’s “moral freedom” not to be maximally beneficent will result in deaths or harm that one could have prevented if one had been more beneficent. We also expect that acknowledging deontic moral constraints will result in preventable deaths when it prohibits us from sacrificing some for the sake of others. But we do not ordinarily expect deontic constraints (or requirements) to prohibit acts that would prevent harm to the persons on whose behalf the deontic constraint or requirement is imposed. In the case at hand, the moral requirement to save the fishermen does not serve to protect the fishermen from being sacrificed for the benefit of others; it leads to the preventable deaths of the fishermen themselves. If it is hard to have confidence in a morality that allows that allows filmmakers to deliberately refrain from rescuing fishermen in distress, it is also hard to have confidence in a morality that renders it certain that the fishermen will not be rescued.

In an essay that deserves more attention than it has received, Michael Lockwood says that

If a project is a worthwhile one, harms no one and benefits many, then it may… be morally permissible, in pursuing the project, to let people die or otherwise omit to save them from harm—when one could do so with the greatest ease—if doing so would be damaging to the project itself.¹⁰⁷

Lockwood’s view applies more clearly to SCAT than to ST because the aim of SCAT is arguably more worthwhile and because whereas the scientific question posed in SCAT could not be answered without a placebo control, DL could have answered the scientific question while providing surfactant (p. 248 ) therapy to all participants. But the issue here is the principle. Is Lockwood correct? I find myself torn. It does seem wrong for the filmmakers to let the fishermen die. But it also seems wrong to impose a requirement (that the filmmaker rescue the fishermen) that is so demonstrably self-defeating given that the requirement will cause the fishermen not to be saved. Who is right? Pogge or Lockwood?

Let us assume that it would ordinarily be wrong or even “grotesque” for the filmmakers to offer the coin-flip deal once they are on site. That leaves open the question as to whether it would be wrong for the filmmakers to embark on their plan and, if so, whether the legitimacy of the plan alters the morality of the (otherwise wrongful) proposal. When put that way, Pogge’s filmmaker example raises an issue for moral theory (somewhat) analogous to Gregory Kavka’s famous problem about rationality—The Toxin Puzzle.

[A billionaire] places before me a vial of toxin that, if I drink it, will make me painfully ill for a day, but will not threaten my life or have any lasting effects.… The billionaire will pay me one million dollars tomorrow morning if, at midnight tonight, I intend to drink the toxin tomorrow afternoon. He emphasizes that I need not drink the toxin to receive the money; in fact, the money will already be in my bank account hours before the time for drinking it arrives, if I succeed.… All I have to do is… intend at midnight tonight to drink the stuff tomorrow afternoon. I am perfectly free to change my mind after receiving the money and not drink the toxin.¹⁰⁸

Kavka’s puzzle is framed in terms of intentions: could one (rationally) intend today to drink the toxin tomorrow when one knows that it will be irrational to drink the toxin tomorrow? Is it possible to form an intention to perform a future act, given the foreknowledge that you will have a strong incentive not to perform that act? We could also frame the puzzle in terms of the rationality of forming and carrying out a plan to drink the toxin. One would want to be able to embark on such a plan. The question is whether it could be rational to drink the toxin as part of that plan after one has already received the money. Some believe that, having made the commitment, it would be mad to drink the toxin when the time comes. David Gauthier argues that it is perfectly rational to make such a commitment for prudential reasons, that it would be mad not to be the sort of person who would drink the toxin.¹⁰⁹

(p. 249 ) Now I have no (rational!) intention to add to the considerable literature on this puzzle. My present point is that morality may present us with arguably analogous situations. If it is permissible for the filmmakers to embark on a plan that contains what would otherwise be an impermissible failure to rescue, there are two principal possibilities: (1) although the plan not to rescue is permissible, the filmmakers are morally required to rescue nonetheless; (2) the permissibility of the plan converts what would otherwise be an impermissible failure to rescue into a permissible failure to rescue. I have no original argument to make here, but if the failure to rescue is embedded in a permissible plan, then the failure to rescue may thereby become permissible.

One might try to defend the view that the filmmaker is morally required to rescue the fishermen by appeal to non-person affecting moral principles. On that view, not rescuing is wrong not because it is bad for the fishermen, but because it violates an abstract or impersonal moral principle. But Pogge does not take the non-person affecting tack. In acknowledging the potential untoward consequences of adopting his view, he notes that “There is a real danger… that the very people in whose behalf morality imposes such constraints end up worse off than they would have been without it.”¹¹⁰ And so it would seem that Pogge is stuck. On the one hand, he says that we cannot have confidence in a morality that encourages and rewards exploitative behavior. On the other hand, if we do not regard such behavior as permissible, there is a real danger that the persons we regard as the victims of exploitation will end-up worse off. Until convinced otherwise, I prefer principles under which people are not worse off.

The Consequences of Principles

This heading is slightly misleading. Principles don’t have consequences, but the acceptance of principles does have consequences. I believe that we cannot determine whether we should promulgate certain principles such as the “standard of care principle” or PEP without considering the consequences of the acceptance of those principles, particularly given that the acceptance of or compliance with those principles can have unintended consequences. And that is an empirical matter. I am not claiming that we must be committed to an aggregative consequentialism. It is to claim that we should be reluctant to endorse moral principles that are self-defeating or counter-productive in terms of the values that motivate those principles.

(p. 250 ) Consider this analogy. The Federal Aviation Administration once proposed requiring that all infants have an assigned seat on a plane and that they be placed in a child restraint (car seat) during take-off and landing. The FAA argued that this would reduce injuries and fatalities to infants. On examination, it appeared that the proposal would actually lead to an increase in accidental deaths and injuries to infants. If parents are required to buy a separate ticket for their infants, some parents will choose to drive to their destination rather than fly. Given that flying is much safer than driving, an infant that is being held by an adult on a plane is still safer than an infant who is in a car seat in a car. So if the point of the regulation was to save lives of infants—not just the lives of infants on planes—the regulation was unwise if it would have the alleged “diversion effect.” And it was withdrawn.¹¹¹

There is an important lesson here for those who develop and analyze the principles of research ethics. I suspect that many of the well-established principles of research ethics have not been put to this sort of an empirical test. I doubt that those who draft documents such as The Declaration of Helsinki or The Belmont Report or the Common Rule have thought it necessary to examine the consequences of compliance with the principles that they advance.

In criticizing the “double standard” that would allow research in LDCs that would not be allowed in the United States, Lurie and Wolfe argue that the long term health of people in developing countries will be “better served by standards that protect all patients and set a higher standard.”¹¹² Writing about a study of HIV transmission in Uganda, Dirceu Grego says this:

Unfortunately, neither researchers nor the research they produce will solve the problem of global inequalities in health. But if researchers were committed to treating all people as equals in research involving human subjects, then this demonstration of equity might serve as a spearhead for more widespread change. For this reason, it is crucial that the ethical requirements set by an indisputable international reference such as the Declaration of Helsinki be applied uniformly everywhere.¹¹³

I do not see how Lurie and Wolfe can be quite so confident that the promulgation and acceptance of the “standard care” principle will advance the interests of people in LDCs. They provide no evidence for their claim and do not seem to recognize that it is a claim for which evidence is required. Similarly, although Grego is right that the demonstration of equity “might” (p. 251 ) serve as a spearhead for more widespread change—even if it meant forgoing research—it is not clear that it will. Given that much of the criticism of studies such as ST has been based on appeals to the negative consequences of accepting so-called “double standards” that would regard such studies as permissible, we should also consider the consequences of accepting standards that would regard such studies as impermissible.

Of course principles are not always accurately applied. Although few commentators have put their arguments in these terms, it is possible that those who reject PEP are implicitly appealing to a version of the epistemological argument I considered above. Let us suppose, arguendo, that a reasonably accurate application of PEP will have better consequences than its rejection. It is possible that the application of PEP will be seriously flawed and that the moral gains consequent to the correct application of PEP will be outweighed by the costs consequent to the erroneous application of PEP and to the general weakening of important ethical considerations. This line of argument would exemplify a well-known strategy of adopting relatively “absolute” rules (“never go through a red light”) as a second-best strategy even when first best rules (“go through red lights when it is safe to do so”) that allow for greater flexibility are likely to be misapplied. From that perspective, the framers of documents such as the Declaration of Helsinki might argue that while the first best ethical principles would actually allow for trials that its principles preclude, there is a great danger that those responsible for applying such principles would allow trials that should not be approved, and so, in the long run, it may be better not to permit trials such as ST, even if it would otherwise be better to permit it. Once again, it is an empirical issue as to whether this argument works in the present context, but it is important to note that this argument does not reject PEP on its own terms.

Exploitation and Double Standards

As I just noted, some critics have argued that to allow a PCT in an LDC that would not be permitted in a developed society is to countenance moral relativism or a double ethical standard—“Acceptance of a standard of care that does not conform to the standard in the sponsoring country results in a double standard in research.”¹¹⁴ If ST would not be permitted in the United States, then it should not be permitted in Bolivia; we should not use “the abominable state of health care in [impoverished] countries… to justify studies that could never pass ethical muster in the sponsoring country.”¹¹⁵ (p. 252 ) As Marcia Angell puts it, we must avoid an “ethical relativism” that would allow for research programs in Third World populations that “could not be carried out in the sponsoring countries.”¹¹⁶

Angell’s point has little to do with ethical relativism as it is usually understood. No one is claiming that it would be acceptable to conduct ST in Bolivia if it does not violate Bolivian values to do so. We can agree that the same fundamental ethical principles should apply to research wherever it is conducted. The question is whether the right universal values might have different implications in different contexts. The principle “When in Rome, do as the Romans do” has nothing to do with relativism; it is a corollary of the universal principle that one should show respect for others’ norms when in their midst. The principle “serve food that your guests will like” is a universal principle that will yield different results for different guests.¹¹⁷.

Similarly, the principle of informed consent is a universal principle that will yield different results in different cases, given that some people will consent to interventions whereas others will not. Thus one can administer medically necessary blood transfusions to those who consent, but not to Jehovah’s Witnesses who refuse to consent on religious grounds. There are no double standards here, just a consistent application of the principle of informed consent to medical treatment. In the context of medical research, if it would be irrational for Americans to consent to participate in a placebo controlled trial of Surfaxin, but it would not be irrational for a Bolivian to do so, insisting that only “rational” consent can be taken as morally transformative will have different implications in the two societies. Once again, there are no double standards at play.

Or consider the assessment of risks and benefits in research. As we have seen, The Common Rule states that the risks of participation must be “reasonable in relation to anticipated benefits, if any, to subjects, and the importance of the knowledge that may reasonably be expected to result.” Although this provision implies that IRBs can approve research when the importance of the knowledge outweighs the risks to the subject, consider the implications of a subject-centered version of the principle that requires that risks to the subject are reasonable in relation to anticipated benefits to the subject.

A subject-centered version of the reasonable risk criterion could be understood as a universal principle that has different implications in different social settings. Given that the infants in Bolivia would receive no treatment in the absence of ST, the risks of participation are certainly reasonable in relation to the anticipated benefits of participation. By contrast, if we apply that (p. 253 ) principle in a developed society where subjects could expect to receive the surfactant therapy outside of the trial, then the risks of participation to the subjects are not reasonable in relation to the anticipated benefits. Once again, I am not arguing that the placebo-controlled version of the Surfaxin trial passes ethical muster. I am arguing that hyperbolic polemical appeals to the notion of double standards will not resolve that question.