Understanding Risks and Benefits in Research on Reproductive Genetic Technologies

Abstract

Research protocols must have a reasonable balance of risks and anticipated benefits to be ethically and legally acceptable. This article explores three characteristics of research on reproductive genetic technologies that complicate the assessment of the risk-benefit ratio for such research. First, a number of different people may be affected by a research protocol, raising the question of who should be considered to be the subject of reproductive genetic research. Second, such research could involve a wide range of possible harms and benefits, making the evaluation and comparison of those harms and benefits a challenging task. Finally, the risk-benefit ratio for this type of research is difficult to estimate because such research can have unpredictable, long-term implications. The article aims to facilitate the assessment of risk-benefit ratios in research on reproductive genetic technologies by proposing and defending some guidelines for dealing with each of these complicating factors.

Keywords:

• benefits
• genetic technology
• reproductive technology
• research
• risks

I. THE REQUIREMENT FOR A REASONABLE RISK-BENEFIT RATIO

The Belmont Report identifies beneficence–the avoidance of harm and the promotion of good–as a fundamental principle in ethically acceptable research with human subjects (1979). Further, the Belmont Report requires that “research be justified on the basis of a favorable risk/benefit assessment” (1979, Part C 2). A general consensus exists that research with human subjects must follow this ethical guideline.

An acceptable risk-benefit ratio is also a legal imperative for most research done in the United States. Federal regulation requires institutional review boards (IRBs) to determine that, “risks to subjects are reasonable in relation to anticipated benefits, if any, to subjects, and the importance of the knowledge that may reasonably be expected to result” in order to approve a research protocol (45 CFR 46.111(a)(2)). This requirement is stated more clearly in the Office of Human Research Protections' IRB Guidebook, a document intended to clarify the federal regulations: “Risks to research subjects posed by participation in research should be justified by the anticipated benefits to the subjects or society” (IRB Guidebook, Chapter IIIA). ¹ Although the requirement takes slightly different forms in these various documents, it is clear that the balance of risks and possible benefits should be a central consideration in the evaluation of research.

The IRB Guidebook defines “risk” as “the probability of harm or injury (physical, psychological, social or economic) occurring as a result of participation in a research study” and a benefit as “a valued or desired outcome; an advantage” (IIIA). The same document notes an important asymmetry in these definitions: that the definition of risk incorporates an element of likelihood (by referring to the probability of harm or injury) whereas the definition of benefit does not include any recognition of the fact that possible benefits are not guaranteed. The Office of Human Research Protections (OHRP) therefore recommends that IRBs should evaluate the ratio between risks and anticipated benefits (IRB Guidebook, Chapter IIIA). ²

As noted, the requirement for research to hold out a reasonable risk-benefit ratio is both an ethical guideline and a regulatory requirement for most research in the United States. It is therefore vital that a research protocol's risks and anticipated benefits be accurately and completely assessed. However, this is not an easy task.

Even in paradigm cases of research with human subjects, determining whether a research protocol has an acceptable risk-benefit ratio is challenging because this concept has significant room for interpretation. A primary point of ambiguity is the term “reasonable.” How should ‘reasonable’ be defined? Is a reasonable risk-benefit ratio simply less than one (where the benefits outweigh the risks)? A second difficulty is the lack of guidance about how risks and anticipated benefits should be quantified and compared against each other. How much weight should be put on different types of risks and anticipated benefits and how much should those risks and anticipated benefits be discounted based upon their likelihood of occurring? A third area of ambiguity exists around the question of which kind of risks should be taken into account when assessing the risk-benefit ratio. That is, should all risks and anticipated benefits be taken into account? Or should there be limits on the types of risks and anticipated benefits that are assessed? These questions also must be answered before the risk-benefit ratio for any given research protocol can be specified.

For research on reproductive genetic technologies, the assessment of the risk-benefit ratio is even more complex. Unlike most clinical research, in which identifiable subjects are directly, physically, and immediately affected by participation in a protocol, research on reproductive genetic technologies can affect numerous individuals indirectly and in a wide variety of ways, sometimes many years down the road. The challenge of figuring out whether research has a reasonable ratio of risks and anticipated benefits is therefore even greater for protocols involving reproductive genetic technologies than for paradigmatic clinical research.

In this article, I attempt to provide some insight into the assessment of risks and anticipated benefits for research on reproductive genetic technologies. After reviewing the nature of reproductive genetic technologies, I discuss three characteristics that complicate the evaluation of their risks and anticipated benefits and suggest some guidelines for dealing with these complexities.

II. THE NATURE OF REPRODUCTIVE GENETIC TECHNOLOGIES

Reproductive genetic technologies (RGTs) are those techniques that are used with the aim of creating a human being with (or without) particular genetic characteristics. They combine genetic testing or gene therapy with various assisted reproductive technologies (ARTs) to determine at least some of the genetic makeup of the individuals they are used to bring into existence.

There are several technologies in use today that can be categorized as RGTs. Sperm sorting is a process in which sperm containing the XX (female) genotype are separated from those containing the XY (male) genotype using a centrifuge. This technique, combined with artificial insemination using only the sperm with the desired genotype, significantly increases the likelihood of conceiving a child of the desired sex.

Another RGT currently available is preconception genetic diagnosis, which involves the testing of an oocyte's first polar body (which is produced during oogenesis). Using this technique, oocytes can be tested for the presence or absence of certain genes before fertilization without destroying the cell. Only those oocytes that have not received the undesired genes (or those that have received the desired genes) from the mother are fertilized and then implanted.

Preimplantation genetic diagnosis, an RGT that can be used after conception to screen for a number of genetic diseases, is also currently possible. This technique involves the genetic testing of embryos fertilized in vitro by removing a single cell from each embryo at the eight-cell stage and using it to screen for specific genetic sequences. Only those embryos with (or without) those sequences are transferred to a woman's uterus to continue development. A similar but more powerful technology, preimplantation genetic haplotyping, has recently been developed that allows embryos to be screened not only for known genetic sequences but also for genetic patterns present in family members who are affected by a genetic condition (Renwick et al., 2006).

Several other RGTs are currently under development. Some of these are now being tested on animals while others remain only theoretical possibilities. Human germline genome modification, the alteration of an individual's genes in a way that will change the genotype of any offspring that individual may have, is not currently accepted for use on humans, but has been tried on animal models (DHHS, 2002). Embryonic gene therapy, in which the genetic makeup of embryos could be altered using vectors containing desired genes, followed by the implantation of only successfully altered embryos, may also someday be a possibility.

Finally, an RGT that has been used successfully on animals but not yet on humans (at least to public knowledge) is reproductive cloning. ³ In reproductive cloning (in contrast to therapeutic cloning) an individual's entire genome is duplicated to create a genetically identical individual. Certainly there are other RGTs that have not yet been conceived of, but will one day become realities.

As noted above, evaluating the risk-benefit ratio for research on these types of RGTs is often more complex than evaluating this ratio for most other research with human subjects. This is due, at least in part, to three features of this type of research:

1. Because many different people may be affected by RGT research, it may be unclear who should be considered to be a research subject in these types of protocols,

2. The variety of possible harms and benefits may be greater for RGT research than for most clinical research, and

3. RGT research has the potential to cause harms or benefits many years in the future.

These three features make the assessment of a research project's risk-benefit ratio a difficult task. In the following sections I discuss the complications that these characteristics generate and propose some guidelines for dealing with them.

III. WHO IS THE SUBJECT?

In contrast to most clinical research, in which an identifiable individual consents to participate in research and bears the risks associated with that research, research on RGTs has the potential to affect a number of different individuals, none of whom is obviously the single subject of the research. Children who are brought into existence as a result of the research protocol ⁴ are central figures in the research. The parents of those children may also be participants in RGT research. In many cases, both the mother and the father are involved in the research, although in different ways.

Third, RGT research has implications for the group of people who share significant parts of the future child's genotype, such as her immediate and extended family and possibly those sharing her genetic heritage (particularly in cases in which the subject is from a tight-knit ethnic group that is relatively genetically homogenous). A final group that could be affected by RGT research is the set of future people who will be descendants of the children created as a result of the research. Because genetic selections and alterations made as part of a research protocol are likely to be passed down to future generations, those people who receive the altered genes could be considered to be subjects of the research.

The fact that so many different parties could be affected by research on RGTs creates ambiguity about which of them should be considered to be a research subject. Clarification of this ambiguity is necessary to accurately evaluate the risk-benefit ratio for a research protocol. The federal regulations stipulate that risks to the subject should be weighed against benefits of the research. They define ‘human subject’ as, “a living individual about whom an investigator (whether professional or student) conducting research obtains

1. data through intervention or interaction with the individual, or

2. identifiable private information.” (45 CFR 46.102 (f)).

The regulations specify that,

Intervention includes both physical procedures by which data are gathered (for example, venipuncture) and manipulations of the subject or the subject's environment that are performed for research purposes. Interaction includes communication or interpersonal contact between investigator and subject. Private information includes information about behavior that occurs in a context in which an individual can reasonably expect that no observation or recording is taking place, and information which has been provided for specific purposes by an individual and which the individual can reasonably expect will not be made public (for example, a medical record). Private information must be individually identifiable (i.e., the identity of the subject is or may readily be ascertained by the investigator or associated with the information) in order for obtaining the information to constitute research involving human subjects.

This seems to be a plausible way to think about research subjects for both legal and ethical purposes. Under this broad definition, however, it is possible that more than one of the parties identified above could qualify as subjects of research. Which of these parties should be considered to be research subjects and therefore play a central role in the risk-benefit analysis?

Because the children who will be created as a result of RGT research protocols play a central role in research, they are obvious candidates to be considered to be subjects of that research. As the entities who are being created and manipulated, future children are the targets of the investigation. These possible future children are the individuals who would have the most to gain or lose from the research protocol. Depending upon the RGT used, they could owe their existence (or non-existence) to that research. Further, it is possible that at least some of their genetic makeup would be determined by the RGT being investigated. As a result, these future children could live with the effects (both the harms and the benefits) of the research their entire lives.

One could object, however, that the future child cannot be a subject of research on RGTs because he does not exist when the research is begun. Instead, he is brought into existence through the use of those technologies. This fact raises the question of whether a future child can be harmed or benefited from this type of research. If the child's parents had not decided to conceive a child as a part of the research protocol, it is almost certain that a different combination of egg and sperm would have been used to produce a child, and, as a result, a different child would have come into being. A child created through RGT research, therefore, had only two possibilities: to have been a research subject or not to have existed at all. ⁵

The consequence of this fact, one could argue, is that even if the child experiences birth defects or other problems because of the method of her conception, she is actually no worse off than she otherwise would have been because otherwise she would not have existed. Similarly, it makes no sense, one could claim, to suggest that the child was benefited by her parents' choice to conceive her through RGT research because her only other possibility was not to exist (rather than to exist with the genetic condition avoided by the use of RGTs). In light of these observations, the assumption that the subject of RGT research can be harmed or benefited by participating in the research is brought into question. This argument applies slightly differently depending upon the RGT under discussion, but is applicable to all of them in some way.

While theoretically fascinating, these arguments are unconvincing because they draw their conclusions from the perspective of hindsight. That is, the moral claims they are based upon view the situation from some time after the child has been brought into being and therefore, significantly, after the genetic identity of the child is established. Choices about participation in research, in contrast, are made from the perspective of foresight — they are made before the precise genetic identify of the child is known. As a result, the decision-making process concerns some undetermined “future child” as the subject rather than a particular (genetically identifiable) future child. The appropriate way to think about the subject of the research, from a moral perspective, is therefore as a placeholder — a genetically undetermined future child, who could in fact turn out to be a wide variety of actual genetically determined children. If the subject is conceptualized in this way, the arguments above are circumvented and the future child can, in fact, be made better or worse off by participating in a research protocol. It seems, therefore, that it is coherent to discuss the risks and benefits of research participation for children created through the use of RGTs.

The central role of the future child in RGT research protocols indicates that those future children must be considered to be research subjects. It is clear that future children are manipulated by the research and that private information is gathered about them, so that they meet both of the individually sufficient conditions for being a research subject.

Another set of candidates who could be considered to be research subjects in RGT protocols are the women who provide the ova or gestate the fetus as part of or following the research. These women may be involved with the research in a number of different ways and their participation is necessary for carrying out the protocols. The harvesting of ova clearly counts as a manipulation of an individual for research purposes. The genetic testing that is generally necessary as part of RGT research is a way of gathering these women's private information. Risks to women are also involved with the implantation of an embryo and with carrying and delivering a child. In some cases, more than one woman could be involved in a given research protocol if, for example, one woman provides the ova and another woman carries the child. Participation in a research protocol in any of these ways, however, would generally fall under the above definition of subject, meaning that these women should be considered to be research subjects.

Men who provide sperm for RGT research could also be considered subjects of such research in some cases. The collection of sperm, despite posing minimal risk, constitutes an intervention with these men. More importantly, private information is often collected in the course of the genetic testing required by most RGTs. Like the women participating in this research, then, at least in some cases the men involved in RGT research should be considered to be research subjects.

Genetic selections and alterations made in the course of RGT research will be passed down to the descendants of the children created through that research, suggesting that those descendants could be thought of as subjects of RGT research. Genetic manipulations could have unpredictable effects on future generations, as they are passed down and interact with other genes during the process of reproduction. Further, the number of individuals affected in this way by RGT research is likely to multiply over time as the number of the subject's descendants grows.

It may seem that because these future people do not yet (and may never) exist, it would be unreasonable to consider them to be subjects of RGT research. It could be argued, however, that they are the entities targeted for intervention in the course of the research and that identifiable private information would be gathered about them. They would receive genes from the child created through the research and so their genetic makeup would be manipulated by that research. In some sense, then, the research constitutes an intervention with those future individuals.

Further, information gathered in the course of the research would be able to be tied directly to the child's descendents because of the unique nature of genetic profiles. Even though they are not identifiable at the time the research is begun (because they do not yet exist), private information about those future individuals is known as a result of the research. Considering these descendants to be research subjects therefore might not be as unreasonable as it first appears.

An objector could also make an argument similar to the one described above about harms and benefits to the future children themselves: that it is incoherent to claim that the descendants of a research subject can be harmed or benefited by the subject's participation in RGT research since it is only because of the child's involvement in the research that the descendants exist. If the child had not been brought into existence when and how she was, the descendants would never have come into being. As a result, an objector could claim, they were not made better or worse off by the RGT research.

A response similar to the one given above could be made in rebuttal. Because the exact identities of the descendants of the future child are unknown at the time at which the decision is made, the moral analysis of the effects of the child's involvement uses a placeholder for “the subject's descendants” that can refer to whichever set of descendants happen to come into being. This approach makes it possible to talk sensibly about the risks and benefits to those future individuals.

Finally, RGT research could impact other individuals who are genetically related to the future child. These individuals are unlikely to be directly involved in the research and so are probably not subjects of an intervention related to the research protocol, but they may qualify as research subjects because an investigator could gather identifiable private information about them in the course of the research. When a research protocol requires genetic testing, information about many different individuals is involved. Depending upon how the private information provision included in the federal regulations is interpreted, these individuals could be considered to be research subjects. Extended family members and others genetically related to the future child may be individually identifiable from the information gathered from the future child alone. As a result, an argument can be made that these individuals qualify as research subjects.

As noted above, it is necessary to understand which of these individuals should be considered to be research subjects in order to develop an accurate picture of a research protocol's risk-benefit ratio. The above analysis suggests that any of these parties could be subjects of RGT research if the above definition of human subject is interpreted broadly enough. It seems clear that the children brought into existence as a result of the research are research subjects because the research requires physical intervention with as well as the gathering of private information about those children. A compelling case for considering the women participating in the research to be research subjects can be made in most cases because most RGT research requires women to undergo invasive procedures or genetic testing. When the protocol involves intervention with the men involved with RGT research or requires the gathering of their private information, those men are research subjects.

Future children and the women and men who participate in their creation, therefore, should all be considered to be subjects in RGT research in most cases. RGT protocols generally involve interaction with or manipulation of all of these individuals in the course of the research. Further, these are the people whose participation is necessary for the research to take place. It would be impossible to carry out RGT protocols without each of these parties. These two factors differentiate future children and, generally, the women and men involved in their creation from others who may be affected by the research.

Third parties in research are those individuals who are affected by the research, but who are not researchers or subjects. (Resnik & Sharp, 2006). Descendants of the children created through RGT research are probably more accurately described as third parties than as research subjects. Although descendants' genomes may be affected by RGT research, this intervention does not occur directly, but rather indirectly, that is, through another individual (the future child). Similarly, private information about these individuals is not gathered directly (which is impossible since they do not yet exist), but it is gained by gathering information about another individual.

The same is true of those individuals who are genetically related to the children who are the products of RGT research (but who are not their parents). These individuals' private information is gathered indirectly rather than directly, because it is available only through information about the child.

Because these parties are indirectly rather than directly involved with the research, they do not fall under the above definition of human subject unless that definition is understood in an implausibly broad way. These third parties are not the intended target of the investigator's intervention or manipulation. Secondly, there is generally no interaction between the third party and the investigator. Finally, because the information available about third parties is limited to that information that can be known through other individuals, it may not be truly readily and individually identifiable in the way the definition requires. Additional information about the future child and his family may be necessary to meet the definition as it is described above. As a result, third parties such as the future child's descendants and genetic relatives are more appropriately considered to be third parties affected by the research than actual research subjects. This classification does not entail, however, that the harms and benefits that research might hold for them should be ignored.

If these third parties are not considered to be research subjects, should the harms and benefits the research holds for them be taken into account? Harms to parties not directly involved in research do not seem to be any less morally relevant than harms to those who participate directly. Ethically, then, there is good reason to believe that the effects that research may have on third parties should be included in the assessment of a protocol's risk-benefit ratio. From a policy perspective, however, this issue is more complicated given the practical constraints of research review. The OHRP guidance states that the risks posed to the subject should be weighed against the anticipated benefits for the subject and for society.

However, these guidelines do not provide guidance about the extent to which the risks to third parties should be counted in the assessment of the risk-benefit ratio. The Belmont report acknowledges that, “Risks and benefits of research may affect the individual subjects, the families of the individual subjects, and society at large (or special groups of subjects in society).” However, like the OHRP guidelines, the report addresses benefits for the subject and for other parties but limits its discussion of risks to those faced by the subject. Nonetheless, whether consideration of the risks to others was omitted intentionally or accidentally in these documents, this omission is difficult to justify. Risks to third parties should therefore be taken into account in the evaluation of a research protocol's risk-benefit ratio to whatever extent is possible given the practical constraints that limit research review.

This conclusion raises the question of whether risks to subjects and risks to third parties should count equally in the risk-benefit analysis. I argue that they should not. Risks to third parties should be taken into account, but should count for less. There are several reasons that such a “discount” could be applied to the consideration of risks to third parties ⁶ . First, risks to third parties may be more difficult to identify than risks to subjects because they are indirectly rather than directly affected by the research. For a similar reason, the magnitude and likelihood of risks posed to third parties are harder to predict. Finally, risks to third parties are often mediated by actions of other individuals so that the causal connection between the research and the risks is weaker than it is for subjects of the research. The assessment of risks to third parties, then, involves a significantly greater degree of uncertainty and a weaker causal chain than does the evaluation of risks to the research subject. These factors suggest that equal consideration of the risks to third parties would be unwarranted.

IV. WIDE RANGE OF POSSIBLE HARMS AND BENEFITS

Harms and benefits associated with research can be categorized as physical harms and benefits, psychological harms and benefits, social harms and benefits or economic harms and benefits (Levine, 1986, pp. 42–54). In most clinical research, the possible physical harms and benefits are an obvious consideration in the evaluation of a protocol's risk-benefit ratio. The potential for such research to affect the subject and others psychologically, socially, and economically is often limited. In research on RGTs, however, these other types of possible harms and benefits can be of equal or greater importance than the physical ones.

A detailed listing of all of the possible harms and benefits of the many different kinds of RGT research would be too lengthy for the purposes of this article. The following paragraphs, therefore, identify only some of the most significant risks and anticipated benefits that could be taken into account in evaluating a research protocol's risk-benefit ratio.

Physical Harms and Benefits

Although the physical risks of RGTs to the children created as a result of this research have not been thoroughly characterized, there is data to suggest that they can cause physical harm. The known physical risks associated with research on RGTs vary significantly depending upon the technology or technologies being used. The risks associated with assisted reproductive technologies (which usually must be used in conjunction with genetic techniques) are probably best understood and include low birth weight, preterm delivery, perinatal morbidity and mortality, increased likelihood of birth defects and developmental disabilities, and the risks associated with of multiple birth (Jackson et al., 2004; Schieve et al., 2004; Sutcliffe, 2002). Risks associated with preimplantation genetic diagnosis in particular have not been formally studied (Dresser, 2006), so it is unknown whether this technology carries risks in addition to those associated with assisted reproductive technologies.

A sense of the risks associated with germ line gene therapy can be gained from work on somatic cell gene therapy and on animal models. Such research suggests that “it is not now possible to undertake [inheritable genetic modification] safely and responsibly” (Frankel & Chapman, 2000, p. 23). The process of genetic alteration could interfere with the working of normal genes and could create harmful interactions between new and original genes. Finally, the physical risks of reproductive cloning, based upon experiences with animal cloning, are also significant. Only a small minority of cloning attempts develop into live offspring, and those that do have been found to have genetic abnormalities that result in premature aging, overgrowth and other undesirable conditions (Fulka et al., 2004; Rideout et al., 2001).

The physical benefits anticipated for the future children of RGTs are the benefits associated with life without the genetic traits being avoided or selected against (or the benefits associated with a life with the genetic traits being created or selected for). The extent and kind of possible benefits will vary with the technology used and the genetic traits it is used to bring about or avoid, but could include longer life, reduced pain or physical disability, and reduced mental disability. It has further been argued that a variety of benefits of genetic enhancement can come from the use of gene therapy (Frankel, 2003, p. 33).

Women who participate in research on RGTs may also be exposed to physical risks. These risks are for the most part a result of the use of assisted reproductive technologies necessitated by RGTs. The fertility drugs used to gather ova for the research have been shown to cause mild ovarian hyperstimulation syndrome in 25% of women and the severe form of the disease in 0.1–2% of women (Al-Shawaf et al., 2005; Van Voorhis, 2006). Thromboembolic events, renal failure, and myocardial infarction may also be risks of using fertility drugs. ⁷

Finally, the risks associated with multiple gestations, including hypertension, postpartum bleeding, premature labor, cesarean deliveries also face women participating in RGT research (Van Voorhis, 2006). One possible physical benefit of participation for women at risk of having a child with a genetic disease is that the use of RGTs could make the risks associated with prenatal testing or abortion unnecessary (Dresser, 2006).

Physical harms and benefits for men participating in RGT research are minimal. Because the research does not generally involve physical intervention that is more than minimal risk with these individuals, it is unlikely that they would be physically affected by that research. Those who are genetically related to the future child are put at no physical risk as a result of the research because they are not physically involved in any way. In contrast, physical risks and anticipated benefits for the descendants of the future child could be significant. These individuals could be exposed to many of the same physical harms and benefits that the children resulting from the research are exposed to because they will share some of their genetic material.

Psychological Harms and Benefits

Because of the novel nature of RGTs, assessing the psychological effects of such research is largely a speculative process. Information about one's genetic makeup that is made available as a result of the genetic testing performed in RGT research is a primary source of psychological risks and anticipated benefits. This information would certainly affect the children created. Further, anyone who has significant genetically similarities to those children, including the children's parents, children's genetic relatives, and descendants, could be affected by knowledge about the subject's genetic makeup because that information also may provide information about the individual's own genetic structure. The knowledge that one has, or is at risk of having, a genetic condition could cause anxiety, angst, or depression. It could, however, also provide one with a sense of control or preparedness. On the other hand, if an individual is aware that she does not carry a certain genetic trait, she could experience relief and happiness, but could also feel guilt or alienation.

Other possible harms and benefits may vary greatly depending upon the success or failure of the research. If research is successful (and if they are later informed of their participation), the children created through RGT research children could feel special and wanted. It is also, possible, however, that these children would feel guilt, a lack of unconditional acceptance, or other negative emotions due to this knowledge. If not successful, future children could feel inadequate or unwanted, or could be resentful of their parents' choice to use RGTs.

The parents of future children could experience psychological effects as well. Men and women could take satisfaction in the fact that they are doing all that they can to promote their future child's well-being. They could be free from the guilt sometimes experienced by parents who pass genetic conditions on to their offspring and from the psychological burdens of prenatal testing or abortion. Parents could also experience the satisfaction of their desires for a child of a particular sex or with a particular characteristic. Psychological harms for parents participating in RGT research are most likely to occur if the desired outcome is not achieved. Such harms could occur if the future child is made worse off by a genetic intervention or if no pregnancy is achieved after repeated attempts.

Social Harms and Benefits

As with psychological harms and benefits, the disclosure of information about individuals' genetic profiles is a possible source of social harms and benefits. If such information is known by others in the community, future children or their parents could be treated differently by those in that community, for better or for worse. In addition, individuals genetically related to the future children, particularly when they form an identifiable ethnic group, could be subjected to risks associated genetic testing even when their individual genomes are not being tested. Genetic information gathered as part of RGT research could enable conclusions to be drawn about these larger groups of people, opening the door for stigmatization of or discrimination against members of those groups. It is also possible that members of these groups would experience some benefits as a result of the availability of this information, such as public awareness or the allocation of public or private resources to help accommodate or ameliorate the conditions.

Additionally, children who are disabled as a result of unsuccessful RGT research could suffer social harms such as stigmatization or discrimination. Parents could also experience social burdens as a result of having a disabled child. On the other hand, successful RGT procedures could benefit both future children and their parents by saving them from having to experience such social burdens.

Economic Harms and Benefits

The final type of risks and anticipated benefits that research subjects and third parties could experience are economic in nature. Again, the disclosure of genetic information is a possible source of such economic harms and benefits. If employers and insurers were to gain access to information about one's own genetic profile or the genetic profile of a family member, the potential for discrimination or exploitation would not be insignificant. Individuals could lose their jobs or insurance coverage on the basis of such information. Similar problems could be faced by members of a genetic community if genetic information about the community as a whole were made publicly available.

The financial burdens of being physically harmed by participation in RGT research could be significant as well. Future children harmed in this way would be likely to experience lost wages or opportunities in addition to increased medical expenses. Women who are harmed by their participation in this research could also experience similar financial hardships. Further, parents raising a child who is disabled as a result of RGT research are likely to incur substantial expenses. On the flipside, analogous financial benefits, that is, the benefits of not incurring these burdens, could be associated with RGT research as well.

One additional benefit that could be considered for inclusion in a RGT protocol's risk-benefit ratio is the benefit of existence itself. However, doing so would be problematic for at least two reasons. First, the federal regulations clearly state that the benefits of therapies subjects would receive even if not participating in the research should not be taken into account in the risk-benefit ratio. Such risks and anticipated benefits should also not be part of the moral analysis because they are not risks and anticipated benefits of the research itself. Therefore, in cases in which a child would be brought into being in some other way, that is, outside of the research protocol, any such benefit should not be considered.

Second, even in cases in which the child would not otherwise have existed, it is not clear that existence itself is, actually, a benefit to the individual brought into being. In contrast, some have argued that it is better to never exist (Benatar, 1997). At the very least, it is clear that it is impossible to compare existence with non-existence, and an assumption that existence is a benefit would be hard to defend or refute. These difficulties suggest that the benefit of existence is not one that should be considered in the risk-benefit ratio for research on RGTs.

All of the above kinds of harms and benefits can be difficult to estimate. Psychological, social, and economic effects may be especially hard to identify and reliably predict because they are often subjective and dependent upon individuals' circumstances. The fact that they can be dependent upon social and cultural factors also makes identifying and accurately estimating them challenging. Certainly some psychological, social, and economic effects can be easily measured. However, in many cases these kinds of harms and benefits are complex, individual, and context-dependent. The central role that these harms and benefits play in RGT research, therefore, complicates the evaluation of the risk-benefit ratio significantly. ⁸

In light of this uncertainty, one could argue that psychological, social and economic harms and benefits should be discounted in the assessment a research protocol's risk-benefit ratio. As noted earlier, risks and anticipated benefits are taken into account based upon two factors: their severity and the likelihood that they will come to pass. Therefore, the current understanding of the risk-benefit ratio already incorporates an element of probability, weighing harms and benefits less when they are less likely to occur. Is the uncertainty associated with psychological, social, and economic risks and anticipated benefits adequately captured by this element of probability? It seems that it is not. The kind of uncertainty associated with the evaluation of these types of risks may be different than the kind of uncertainty associated with the evaluation of many physical risks. It is not simply a matter of probability, but is also highly dependent upon individual differences, contextual factors, and cultural norms. This kind of uncertainty therefore adds another dimension to the accurate assessment of a research protocol's risk-benefit ratio. Just as the notion of the likelihood of a harm or benefit occurring is built in to the assessment of the risk-benefit ratio, this other kind of uncertainty must also be taken into account. The existing notion of probability does not adequately incorporate the difficulties involved with estimating psychological, social, and economic harms.

The extent to which psychological, social, and economic harms should be taken into account should therefore be proportionate to the level of certainty with which these types of risks and anticipated benefits can be estimated and foreseen. However, additional research should be undertaken to enable these risks and anticipated benefits to be better understood and more accurately predicted.

V. LONG-TERM IMPLICATIONS

A third feature of RGT research that makes the assessment of its risk-benefit ratio particularly challenging is the extended length of time required to gain a complete understanding of its risks and anticipated benefits. In most typical examples of clinical research, the effects of the research are experienced and measured reasonably soon after the intervention is carried out. In contrast, effects of the interventions being investigated in RGT research could take the lifetime of the individual they are used to create to become apparent.

Further, because some genetic traits are not revealed until they are put in a specific environment or combined with other types of genetic traits, the full scope of a protocol's effects could take generations to be revealed. A full understanding of a protocol's risk-benefit ratio is also complicated by the fact that both the possible harms and the possible benefits of RGT research can be cumulative. As the prevalence of various genetic traits in the population shifts, communities or societies could be affected in unpredictable ways.

As with psychological, social, and economic effects, the simple idea of probability of an effect coming about does not seem sufficient to account for the uncertainty inherent in estimating the long-term implications of research on reproductive genetic technologies. In addition, with such a long time horizon under consideration, the effects that the research has on individuals far in the future are indirect rather than direct, requiring intervening actions (such as continued procreation) to be realized. These two factors suggest that the long-term implications of research on RGTs should also be discounted to some degree when they are incorporated into the analysis of a protocol's risk-benefit ratio. This is not to say that they should count for nothing, but rather that they should count for less than immediate, direct risks and anticipated benefits.

VI. RESEARCH OR INNOVATIVE THERAPY?

An important objection that could be made to the above conclusions is that most RGTs should not, in fact, be considered to be research but instead are innovative therapies. On this view, these procedures do not need to be written up as research protocols and approved by IRBs. They are instead therapeutic measures used to promote the well-being of each individual patient. The prevalence of this view is reflected in the lack of scholarship available on RGT procedures. Very few studies have been published reporting success rates or side effects of RGTs, even for those technologies that are becoming common practice such as sperm sorting and PGD. If RGTs are used in the context of therapy rather than in the context of research, one could argue, the use of these technologies does not need to be justified by the existence of a reasonable risk-benefit ratio. At least two different replies to this objection can be made.

First, even if RGTs can legitimately be considered to be innovative therapies, it could still be the case that they should be the subject of research protocols. According to the Declaration of Helsinki,

“In the treatment of a patient, where proven prophylactic, diagnostic and therapeutic methods do not exist or have been ineffective, the physician, with informed consent from the patient, must be free to use unproven or new prophylactic, diagnostic and therapeutic measures, if in the physician's judgment it offers hope of saving life, re-establishing health or alleviating suffering. Where possible, these measures should be made the object of research, designed to evaluate their safety and efficacy. In all cases, new information should be recorded and, where appropriate, published. (WMA, 2000)

Most RGTs are intended to alleviate suffering in a context in which no proven methods to accomplish that goal are available, suggesting that the physicians should be allowed to use them. However, as their use becomes more prevalent, more and more people will be exposed to the risks as well as the benefits of these techniques. Before the use of these technologies becomes standard practice, then, it would be valuable to have a better understanding of their benefits and their burdens. This perspective is also supported by the Belmont report:

The fact that a procedure is “experimental,” in the sense of new, untested, or different, does not automatically place it in the category of research. Radically new procedures of this description should, however, be made the object of formal research at an early stage in order to determine whether they are safe and effective. Thus, it is the responsibility of medical practice committees, for example, to insist that a major innovation be incorporated into a formal research project. (1979)

Given the paucity of available data on RGTs, it seems particularly vital that data be collected during the use of these technologies, making even innovative treatment a form of research.

Second, the above conclusions about risks and benefits could be useful for decision-making in the context of treatment as well as in the context of research. The predominant ethical paradigm in a treatment context is different from the paradigm that controls in a research context. Risks and benefits to individuals beyond the subject of the therapy are not a primary consideration.

However, this may be in part because few existing therapies have direct consequences for parties other than the patient and so the consideration of those consequences has not taken a significant role in therapeutic decision making. As genetic and reproductive medicine advances, these considerations may begin to play a larger part in therapeutic choices as well as in research approval. As a result, this analysis of risks and benefits for both the subject and for others should be useful in the wide variety of contexts in which RGTs are used.

VII. CONCLUSION

A complete and accurate picture of the risks and anticipated benefits that reproductive genetic technologies hold for subjects and for other parties is difficult to achieve, even though such a picture is necessary for the moral and legal evaluation of such research. It is not always obvious who is the subject of RGT research and therefore whose interests should be considered and to what extent. Further, the variety and temporal distance of the possible harms and benefits a research protocol may hold out complicate the risk-benefit assessment.

I have argued that some of the risks and anticipated benefits associated with RGT research should be discounted in the ethical analysis of that research. In particular, the risks to third parties, uncertain psychological, social, and economic risks and anticipated benefits, and harms and benefits that may take place far in the future must all be taken into account, but should be given less weight in the evaluation process than the direct, immediate, and observable risks and anticipated benefits associated with paradigmatic clinical research. Even with this guidance, however, the complete assessment of risk-benefit ratios in RGT research is an exceedingly complex task, requiring insightful perception, careful consideration, and good judgment.

Notes

1. Additional provisions are included in the regulations for research with fetuses, pregnant women, and embryos. However, these provisions do not apply to most kinds of research on reproductive genetic technologies. Because the embryos that are used in such research are not yet implanted in a woman's uterus, they are not considered fetuses and the women who will receive the embryos are not yet pregnant. Further, existing guidelines on embryo research do not address embryos intended for implantation.

2. This they get from Levine (1986, p. 37).

3. This term covers both germ line gene therapy and germ line gene enhancement. See Resnik (2001, p. 1452).

4. Ooplasm transfer is considered by some to be a form of human cloning because it involves the replication of mitochondrial DNA. For the purposes of this article, however, ooplasm transfer is not considered an RGT because it is done with the purpose of relieving infertility rather than controlling the genetic characteristics of future children.

5. I will call these entities “future children” throughout the article as a matter of convenience. This title is not meant to confer any particular moral status to the gametes or the products of conception that are created throughout the process of RGT research.

6. This is a variation of Derek Parfit's non-identity problem (Parfit, 1984, pp. 351–61). This problem applies to both pre-conception and post-conception RGTs because even in post-conception RGTs (such a preimplantation genetic diagnosis), conception occurs differently than it otherwise would have.

7. The idea of discounting interests comes from Baruch Brody (2001).

8. Fertility drugs have long been suspected to cause an increased risk of ovarian and breast cancers, however, studies have found no evidence to support such a connection (Al-Shawaf et al., 2005).

9. Certainly, there are cases in which physical harms and benefits are subjective and difficult to predict or estimate. The risk-benefit ratio for such research can also be difficult to evaluate. These complications seem to be more prevalent, however, when considering these other types of harms.

World Medical Association.(2000). ‘Declaration of Helsinki’: WMA.

World Medical Association.(2005). ‘Protection of human subjects.” Code of Federal Regulations Title 45 Part 46.