Risks from the PGD process

One of Taylor-Sands’ most interesting suggestions is that there might be a difference between selecting a (future) child because it is congenitally deaf and some other forms of selective reproduction. Taylor-Sands accepts thatAn embryo selected because it is deaf could only ever become a deaf child. There are no other options for that child apart from becoming a deaf child, apart from non-existence (p. 18).1

But what goes for the deafness case does not, she claims, apply to the whole of selective reproduction… additional risks of harm associated with the PGD process (which involves ART and embryo biopsy) … could be avoided if the child is born naturally. The non-identity problem is not relevant to these risks of harm because there is another option for the child—to be born as a result of natural conception (p. 18).1

It is difficult, however, to see much difference between the deafness case and the saviour sibling case in this respect. In the choosing deafness case, the parents have a choice between several embryos, one of which will (if implanted) become a deaf child. The Non-Identity Problem engages because, were the parents to have selected a hearing child instead, the deaf child would not have existed. Therefore, it is difficult for the deaf child to complain of harm caused by selection-for-deafness, because, were it not for the selection process, he or she would not exist.

But can't the same be said of saviour sibling selection? If the saviour sibling were to complain of damage caused by the selection process, couldn't the parents answer as follows? We faced a choice between having our next child through ‘natural conception’ or using PGD and tissue typing, and we chose the latter. As a result of that decision you were born in a damaged state but, had we chosen ‘natural conception’ instead, we would have ended up with a different child, not you.

Taylor-Sands suggests that there is another option for such children: being born ‘as a result of natural conception’. However, although this is a theoretical possibility, it is not what would in fact have happened if the parents chose ‘natural conception’. This is because the probability of the very same gametes coming together at conception in the two different scenarios (sexual intercourse vs in vitro fertilisation) is vanishingly small (not least because of the enormous number of possible different sperm that could be involved). And, on a fairly standard view of the Non-Identity Problem as applied to human reproduction, ‘different gamete’ entails ‘different person’.2

It seems therefore that, in saviour sibling cases, the Non-Identity Problem does engage even when we consider the risks of PGD. The ‘natural conception’ option is rarely relevant because it is, at best, a theoretically possible one and not one that would have led to the same child being created in the most relevant alternative possible worlds.

Quality-of-life thresholds

Taylor-Sands cites a view, attributed to Feinberg, that in wrongful life cases the child is harmed only if ‘his/her life is so bad that non-existence is rationally preferable to existence’ (p. 18).1 She then argues that it would be a mistake to apply this quality-of-life threshold to selective reproduction.

Her most important argument for this is that, even if children are not harmed overall by being created (even if their aggregate lifetime welfare is more than ‘zero’) such children still suffer from harmful events during their lifetime. They both suffer harm and live with states of harm. Say, for example, that parents deliberately choose to have a child with a painful condition, X. That child's life may be ‘worth living’ overall but, nonetheless, we can still legitimately say that the child will suffer harmful experiences of the symptoms of X, and that having X is a harmed state.

This seems plausible, but what follows from it? One thing that does not follow from it is that the harm threshold is different depending on whether we are considering wrongful life cases for existing people, or contemplating selective reproduction for possible future people. For the very same things can be said both. In both cases, we can (at least in the abstract) ask whether, all things considered, a life contains more positive than negative elements and perhaps conclude that, even if it is really quite bad, the positive just outweighs the negative. And in both cases, we might still say that, even if it is a life worth living overall, it is nonetheless one which contains harmful episodes and one in which the person lives with harmful states. All of us experience harmful events from time to time (ones that we would be better off without), but that does not mean that we were harmed by having been brought into existence, or that our parents (or other originators) did something morally questionable when they brought us into being.3

Morally objectionable consequences

Taylor-Sands objects that ‘there seems to be something morally repugnant about justifying the use of PGD to select a child who will experience considerable pain and suffering on the basis that his/her life may, overall, be a life worth living’ (p. 18).1 Most people, even those who use this argument, would agree. However, merely pointing this out neither solves the Non-Identity Problem nor shows it to be inapplicable. Rather, it is a restatement of why the Non-Identity Problem is interesting and troubling; intuitively, we want to say that such cases are ‘repugnant’, but it is difficult to support this ‘gut feeling’ because there seems to be no harm.

There are, however, ways of accommodating this intuition which do not require a radically ‘relational’ approach. One is to appeal to impersonal moral principles such as Procreative Beneficence. According to this, when choosing between possible future children, we are morally obliged to choose the one who ‘is expected to have the best life, or at least as good a life as the others’.4 Such principles often give a reason not to select ‘a child who will experience considerable pain and suffering’, that reason being the availability of a different possible future child who would have a better life.

In the case of saviour siblings, Procreative Beneficence with respect to the as-yet non-existent child needs to be weighed against benefits to the parents and to the existing child who needs a donor. Nevertheless, these issues are, in many respects, the same ones that any family, or any doctor, must face when contemplating intrafamilial living donation. As Taylor-Sands herself puts it,It is … unsatisfactory to focus solely on the individual interests of the child to be born in assisted reproduction (p. 71).1

This seems entirely right, and it would be hard to imagine a justification of saviour sibling selection that did not appeal to wider familial interests. Whether however accommodating this requires a fundamentally different ‘relational’ approach to bioethics or to the welfare of the child is far from clear.