In two recent cases, the Human Fertilisation and Embryology Authority (HFEA) was called upon to decide whether to allow the deliberate creation of 'saviour siblings': children created from embryos selected following Human Leukocyte Antigen (HLA) tissue typing to ensure compatibility with an existing sick child in need of tissue donation. We have written at length elsewhere about the subsequent judicial review of the HFEA's permission of HLA tissue typing on the first occasion, and the ethical merits of its decision in that and the second case (1). Here we confine ourselves to scrutinising the ethical and legal basis for the HFEA's contrasting decisions in these two cases. We argue that the two decisions are irreconcilable and that the Authority's reasoning is defective.

The Two Cases: Hashmi and Whitaker

Raj and Shahana Hashmi have a three-year old son, Zain, who suffers from the blood disorder, Beta Thalassaemia (BT). Zain has to undergo regular blood transfusions and may die without a bone marrow transplant. BT is hereditary and both of the Hashmis are carriers, which means that any child they produce carries a one in four chance of having BT. Zain was diagnosed as suffering from BT at four months. Two months later, Mrs Hashmi conceived naturally in the hope that they might be able to create a match for him. The resulting child, Haris, though free of the disease, was not a tissue match for Zain. His parents then launched a worldwide search for a donor and, when that failed, began to consider alternative options.

The fertility clinic which was treating the Hashmis applied to the Human Fertilisation and Embryology Authority (HFEA) for permission both to carry out pre-implantation genetic diagnosis (PGD) on embryos to ensure that the Hashmis have a child born free of the disease, and also to conduct tissue typing. This would enable them to identify which, if any, of the embryos created by 'test tube' fertility techniques is a perfect blood match for Zain, so that umbilical cord blood might be used to save his life. Permission was granted.

Michelle and Jayson Whitaker have a three-year old son, Charlie, who suffers from Diamond Blackfan Anaemia (DBA), a rare form of anaemia where the bone marrow produces few, or no, red blood cells. Symptoms are similar to other forms of anaemia and include paleness, an irregular heartbeat and heart murmurs because of the increased work the heart needs to do to keep oxygen moving round the body. The disorder can lead to irritability, tiredness and fainting and requires intensive therapy including painful daylong blood transfusions and daily injections of life-saving drugs. DBA has no cure although bone marrow transplants can help. If the Whitakers were able to have another child who would be a matching tissue type donor, then cells created by him/her could help Charlie's body to create red blood cells, giving him a 90% chance of recovery. The Whitakers requested that their doctor, Mohammed Tarranissi, be allowed to test embryos taken from Mrs Whitaker using PGD. Their case was urgent: a transplant needed to be carried out in the next 18 months to have a good chance of success. Like the Hashmis, the Whitakers claimed they wanted another baby anyway and would not view a new child purely as a donor infant.

While very similar to the case of Zain Hashmi, Charlie's case differs in one important respect: the DBA from which he suffers is 'sporadic' rather than hereditary. This means that the chances of his parents having another baby with the disease are no greater than those present in the general population: five to seven per million live births. As such, there is no reason to believe that the Whitakers' embryo would have the same defect. On this basis, the HFEA rejected the Whitakers' application. The justification given was that the 'tissue typing' procedure would be performed solely to find a match for Charlie, and not in order to check whether the embryos themselves carried a genetic disorder. For the HFEA, the Whitakers' case therefore was relevantly different from the Hashmis' since, for the Hashmis, the procedure was in the interests of the new child as well as the interests of Zain. Whereas in the case of the Whitakers, only Charlie would directly benefit, and at some point in the future, his new brother or sister might suffer from the knowledge that she or he had been chosen, and other embryos discarded, primarily to save Charlie's life.

Distinguishing Hashmi and Whitaker

In the course of the deliberations leading to its decision to grant the Hashmis the permission they sought, the HFEA released a list of criteria which would guide its future decision-making:

a) the condition of the affected child should be severe or life threatening, of a sufficient seriousness to justify the use of PGD;

b) the embryos conceived in the course of this treatment should themselves be at risk from the condition by which the existing child is affected;

c) all other possibilities of treatment and sources of tissue for the affected child should have been explored;

d) the techniques should not be available where the intended recipient is a parent;

e) the intention should be to take only cord blood for purposes of the treatment, and not other tissues or organs;

f) appropriate implications counselling should be a requirement for couples undergoing this type of treatment;

g) families should be encouraged to participate in follow-up studies and, as with PGD, clinics should provide detailed information about treatment cycles and their outcomes;

h) embryos should not be genetically modified to provide a tissue match. (2)

These criteria largely map those recommended in a prior report of the HFEA's Ethics Committee, with the notable exception of the second criterion: the Ethics Committee had advised that HLA typing should not be limited to cases where the embryo itself might be at risk of a particular genetic disorder (3). This criterion did provide a clear distinction between the Hashmis and the Whitakers, as Charlie Whitaker's illness was sporadic. Any embryos produced by his parents were at no more risk of developing into a child suffering from Diamond Blackfan Anaemia than would be any other randomly selected embryo. What is at issue, then, is the ethical basis for this second criterion. What motivated the HFEA to introduce it, and to rely on it (against the recommendation of its own ethics committee)?

Whilst the HFEA's press release giving its decision in Whitaker is silent on its reasons for drawing this distinction, subsequent interviews and comments reveal three arguments in support of it, which we here address in turn.

i) It is wrong to create a child for a particular purpose (as a 'means to an end')

The first argument is that a saviour sibling would be 'a commodity rather than a person' and would be wrongfully treated as a means rather than as an end in itself. As Boyle and Savalescu point out, this worry has its philosophical roots in Kant's famous dictum, 'never use people as a means but always treat them as an end.' (4) However, as these authors go on to point out, this does not work as a general argument against saviour siblings. Firstly it relies on a misreading of Kant who counselled not against treating people as means, but rather against treating them merely or solely as a means. Secondly, this argument fails to say what is wrong about creating a child as a saviour sibling, when creating a child for a number of other 'instrumental' purposes is widely accepted. Given that (for example) attempting to conceive a child in order to provide a playmate for an existing child is seen as reasonable, how would we distinguish this from the reasons advanced by the Hashmis or Whitakers? And why should we use the opportunity presented by their need for assistance and an HFEA licence as an excuse to scrutinise their reasoning?

Thirdly, even if the concerns expressed above were convincing reasons for not allowing the deliberate creation of saviour siblings (and we believe they are not), they surely provide no basis for distinguishing between the Whitakers and the Hashmis. Both couples want a child capable of saving the life of a sick sibling. The fact that in the Hashmis' case, they would need simultaneously to screen for disability can have no conceivable relevance in terms of the arguments set out above.

ii) It is wrong to create a child with which to perform this particular procedure, because the child would grow up knowing that it had been created for this particular function and this knowledge would be psychologically damaging;

The claims made about the possibility of psychological harm to the child are speculative and unconvincing as a way of distinguishing the Hashmis from the Whitakers. For we would need to believe not just that a child is likely to be adversely affected by the knowledge that she had been selected as a saviour sibling, but that the adverse psychological impact is likely to be greater in the case where she knows that cells taken from the embryo had not first undergone a separate screening process for genetic disorders. This is deeply implausible. And in any case, it is not supported by the kind of weighty empirical evidence which would be necessary to overturn the presumption in favour of attempting to save a child's life.

iii) PGD may generate as yet unknown health risks for the child created and so should only be used when there is a known benefit to this particular child (i.e. not merely to third parties such as siblings).

The argument which comes closest to an intellectually robust defence of the HFEA's position is this third one. Ann Furedi, at that time the HFEA's Director of Communications, explains it as follows:

'... if you are carrying out a procedure to prevent a child being born with a serious illness then … one could say that the benefits outweigh the risks for that particular child that would be born. But when you get to PGD solely for the purpose of tissue typing … strictly from the point of view of the physical wellbeing of that child, you cannot say that it is for its benefit, you may even be doing something which is [harmful].' (5)

Does this provide a justification for treating the two cases differently? The underlying principle here is that an embryo should be exposed to the risks of PGD only if it (or the person it becomes) is likely to derive enough benefit to outweigh those risks. On this view, the potential child is thought to be like an existing patient and as such doctors should expose it to danger only if, on the balance of probabilities, it will be a net beneficiary.

If this 'net beneficiary principle' is accepted then there does seem to be an important difference between selecting a saviour sibling and screening for a serious genetic disorder, since (it is argued) only the latter procedure benefits the child created and so only the latter is acceptable. But the 'net beneficiary principle' relies on some confused thinking about what it means to 'benefit an embryo'. It appears to depend on something like the following model. In Hashmi, an embryo (H) is subjected to an intervention (T) which has the following effects: (6)

(a) T prevents H from having a serious genetic disorder;
(b) T involves as yet unknown long-term health risks for H.

So in Hashmi, subjecting H to T can (on this model) be justified solely by reference to H's interests because the benefit of (a) (to H) outweighs the harm or risk involved in (b).

In Whitaker however things seem importantly different. For an embryo (W) is subjected to an intervention (T*) with only the following effect:

(b*) T* involves as yet unknown long-term health risks for W.

In Whitaker, T* cannot be justified by reference to W's interests since there is some risk but no benefit for W. So, if subjecting W to T* is to be justified at all, it must be solely by reference to the interests of a third party (the existing child). Perhaps third party interests are sometimes sufficient to justify subjecting an embryo to harm or risk but, on the HFEA view, third party interests are not a sufficient reason, at least in cases of this type (a position which is entailed by the 'net beneficiary principle').

What is wrong with this model? The main difficulty is that in Hashmi it is not the case that T (PGD) prevents H from having a serious genetic disorder. Rather, H is selected because it does not have the genetic disorder in question (and so had H been 'naturally' implanted, rather than implanted as a result of T, H still would not have had the disorder). So we cannot think of T as benefiting H in a straightforwardly causal way, because T does not cure H or eliminate a disorder. Rather, T involves choosing H on the grounds that H is already a 'healthy' embryo

Given this, what can it mean to say that H has been benefited by T? The only way to make sense of this claim is to say that H derives benefit because T causes H to be implanted, and being implanted is better for H than not being implanted (assuming that, if implanted, H will go on to have a 'life worth living' and that the alternative to implantation is destruction). So, if there is any benefit at all for H, it is not 'being healthy rather than having a genetic disorder'. Rather, the benefit is 'existing rather than not existing'. Hence, the HFEA argument must rest on the view that, for H, the benefit of existence somehow outweighs the as yet unknown long-term health risks.

This style of argument raises a number of very thorny philosophical problems which we cannot explore in any depth here.(7) But there are also more practical and telling objections to the HFEA view. The most relevant of these for our purposes is that the argument just outlined applies equally to both Hashmi and Whitaker. For if the relevant benefit is being caused to exist (rather than being cured of a genetic disorder) then clearly both H and W stand to gain more or less equally in this respect - since both are caused to exist by the selection process and probably would not have existed without it. And what is more, this will apply (again, more or less equally) to all selected embryos, except in those few cases where the life in question is so bad that it is 'not worth living'. This then is a decisive objection to the HFEA's attempt to justify treating Hashmi and Whitaker differently on the basis of potential physical harm caused by PGD.

A further practical argument against the HFEA position, and in particular the 'net beneficiary principle', is that it would have far more restrictive implications for infertility treatment services than the HFEA is likely to want to allow. The HFEA's position relies on the idea that being implanted benefits the embryo and, what is more, this seems to be the only kind of benefit available from the selection process (at least from the perspective of the selected embryo itself) since, as we have seen, PGD does not cure disorders, it merely selects those embryos that are already 'healthy'. But what about those embryos which are discarded rather than selected, including those which do have disorders? These embryos stand to gain nothing from the PGD process, which is hardly surprising since the alternative to implantation is use in research and/or destruction. But this does not sit happily with the 'net beneficiary principle', according to which an embryo should be exposed to the risks of PGD only if it (or the person it becomes) is likely to derive enough benefit to outweigh those risks. Discarded embryos obviously will not derive any benefit and so, if the HFEA is really committed to the net benefit principle, it should ban all forms of PGD - and probably all forms of IVF (since this inevitably involves discarding embryos). It follows that either the HFEA is not really committed to the 'net beneficiary principle' (in which case the attempt to use it to distinguish Hashmi from Whitaker is disingenuous) or that the HFEA is inadvertently being inconsistent, not having realised that it cannot endorse the 'net beneficiary principle' in this case without also ruling out even very mainstream IVF practices. Significantly, and more generally, this means that the HFEA's general position that PGD should only be permitted where the embryo is at significant risk of a serious genetic disorder remains similarly morally unjustified.

It is also worth noting that in relevant respects, the child to be born is not in the same position as the adult patient who should only receive treatments which, on balance, are for her benefit. The legal context of the provision of infertility treatment services recognises this, specifically allowing for consideration to be given to the interests of third parties - notably, those who are seeking to make use of infertility treatment services to conceive and, very relevantly for our purposes, the interests of existing children. Indeed, what seems to have escaped many commentators who champion the legal paramountcy of child welfare in this context, is that the 1990 Act actually provides that:

S. 13(5): A woman shall not be provided with treatment services unless account has been taken of the welfare of any child who may be born as a result of the treatment (including the need of that child for a father), and of any other child who may be affected by the birth' (our italics).

So when considering saviour sibling cases, the legislation explicitly requires us to take account of the welfare of the existing sick child. How should the courts seek to balance the unknown effects of PGD on the new addition to the Hashmi family, against the known tremendous health benefits to Zain? It seems hard to resist the conclusion that the tangible short-term benefits to Zain outweigh the risks to the saviour sibling which are (at most) speculative.

CONCLUSION

What then should we conclude with regard to the HFEA's response to the Whitakers? In authorising the HLA typing in Hashmi, the HFEA had already (rightly in our view) recognised that this was a legitimate form of screening for the selection of embryos. This means that the HFEA had already reached a position on the difficult issue of whether tissue typing should be allowed for the benefit of a third party. As such, it seems to us that their refusal to allow the Whitakers to take advantage of the same procedure rests on a distinction without a difference. It was an indefensible decision, reliant on confused thinking, and particularly misguided given that it was liable to result in the death of a child.

While the issue of whether the Whitakers might have succeeded in a judicial review of the HFEA's decision must remain outside of the scope of this paper, it seems clear that there was nothing in the law which prevented the HFEA from deciding their request differently. Further, our review of the ethical issues would suggest that there was nothing which ought to have prevented the HFEA from approving the Whitakers' request.

One more general point may be offered by way of conclusion. It should be noted that the HFEA is guiding the regulation of reproductive technologies in a direction that does not always fit easily with the general principles of health care law as they have developed elsewhere. Paradoxically, it might be argued that the principle of child welfare was intended by the architects of the 1990 Act to enjoy a weaker role in the regulation of human fertilisation and embryology than in other areas of law (8). Yet there is little doubt that the courts would sanction the far more invasive procedure of use of bone marrow to save Charlie Whitaker, were his 'saviour sibling' already born (9). The assertion of child welfare concerns to ground a prohibition of embryo screening to secure a suitable donor for umbilical cord blood might, as such, seem puzzling. One question which is thus clearly raised by its decision in Whitaker is whether the HFEA is too focused on considerations of the welfare of the child to be born, to the unfortunate exclusion of other factors (10). An alternative and, we would suggest, preferable view is that the child to be born is one party amongst others, each having interests worthy of consideration (11), a philosophy entirely consistent with the wording of the 1990 Act.

Notes

(1) Selecting Saviour Siblings' Journal of Medical Ethics (forthcoming), 'Selection of Saviour Siblings' Medical Law Review (forthcoming vol. 12(2)).

(2) Ethics Committee of the HFEA, 'Ethical Issues in the Creation and Selection of Preimplantation Embryos to Produce Tissue Donors' (22 November 2001) and the minutes of the HFEA's meeting on 29 November 2001, where this issue was discussed, available at: http://www.hfea.gov.uk/aboutHFEA/archived_minutes/00028.htm

(3) Ethics Committee report, ibid at para. 3.14.

(4) R. Boyle and J. Savulescu, 'Ethics of using preimplantation genetic diagnosis to select a stem cell donor for an existing person', (2001) 323 B.M.J. 1240, at 1241

(5) Personal interview with Ann Furedi, then HFEA Director of Commuications.

(6) T stands for the process of embryo biopsy, with extracted cells screened for both genetic disorder and tissue compatability. T* is the same as T except that it does not involve screening for a genetic disorder.

(7) One obvious difficulty, for example, is the question of whether it really makes sense to sense to say of an individual that she was benefited by the events that caused her to exist?

(8) While s.1(1)(a) of the Children Act 1989 establishes child welfare as the paramount factor in any decision involving any aspect of a child's upbringing, s.13(5) of the Human Fertilisation and Embryology Act 1990 directs us merely that it is merely a factor of which 'account must be taken'. See also the guidance of s.13(5) offered in the HFEA's Fifth Code of Practice: '3.3 …in deciding whether or not to offer treatment, centres should take account both of the wishes and needs of the people seeking treatment and of the needs of any children who may be involved. Neither consideration is paramount over the others, and the subject should be approached with great care and sensitivity.'

(9) Such donation is held to be in the donor's best interests, notwithstanding the pain and physical risks associated with the procedure, because of the donor's interest in a continued relationship with his/her sibling, see: Strunk v. Strunk (1969) 445 S.W. 2d 145 (Ky. C.A.) for a US authority, and Re Y (adult patient) (transplant: bone marrow) (1996) 35 B.M.L.R. 111 (Fam.) for a UK authority dealing with an incompetent adult donor.

(10) See E. Jackson, 'Conception and the Irrelevance of the Welfare Principle' (2002) 65 Modern Law Review 176 for the argument that the welfare provision in the 1990 Act is unjust, meaningless and inconsistent with existing legal principle.

(11) For this argument made as a criticism of the paramountcy provision of the Children Act 1989, see H. Reece 'The Paramountcy Principle: Consensus or Construct?' (1996) 49 C. L. P. 267.