Testing whole genome sequencing in the paediatric clinic

Appropriately implementing novel technologies involves critically considering how and when to use them. The technology of next-generation DNA sequencing and its application in whole genome sequencing (WGS) is a key example of where this 'how and when' problem arises. Genomics challenges presumptions, such as whether possible applications of a technology should lead its implementation, or whether it should instead be guided by a core clinical question. And when the individual being tested is a child, these considerations become more significant.

While the idea of using a particular test to help obtain a diagnosis is certainly not new, some features of WGS arguably set it apart from more ‘traditional’ clinical genetics.1 It can give rise to significant volumes of information. Not all of this is yet understood, some has been previously overinterpreted and its meaning will almost certainly change with time.2 WGS will not always lead to certainty, and may introduce new uncertainties.3

This is not to say that WGS is inherently problematic. It has huge potential to positively impact clinical practice, mitigating the effects of illness and improving quality of life for children and their families. We are already seeing families benefiting from obtaining a diagnosis following genomic testing.4 But at the same time, we should be mindful not to bestow more to genomics than it is capable of providing.

Anderson et al 5 evaluated one such implementation of paediatric clinical WGS, the Genome Clinic study. The study involved WGS in children with an underlying condition to identify: (i) primary variants to explain the child's clinical presentation, (ii) other medically actionable variants for conditions that present in childhood; (iii) pharmacogenetic variants relevant to the child's illness and (iv) medically actionable variants for conditions that typically present in adulthood (secondary variants (SVs)). Parents considering their child's potential participation in the study were required to accept information relating to (i) to (iii), but were able to choose whether to receive information about (iv).6

It is laudable that the Genome Clinic is first implementing paediatric WGS in a research context, accompanied by psychosocial and ethical research. This allows for protocols and theories to be tested to inform (and perhaps to change) subsequent clinical practice.ⁱ It is also reassuring that parents showed a positive overall attitude to WGS. However, it should also be noted that returning SVs in a paediatric context remains a contested topic.7

In this commentary, I scrutinise three issues highlighted by this study: (1) the non-separation of the choice over agreeing to diagnostic WGS and whether to receive adult-onset SVs; (2) the value of engaged and critical deliberation regarding paediatric WGS and (3) the place of best interests in debates over returning adult-onset SVs.

(1) Contemporaneous choices

In the Genome Clinic approach, parents of children recruited to this study needed to make a decision about adult-onset SVs (ie, (iv) above) at the same time as agreeing to diagnostic WGS. It was not possible to delay a decision about adult-onset SVs.

These parents were recruited to the WGS study because their child has a condition. Such families may already be living with diagnostic or prognostic uncertainty and are likely to be invested in obtaining meaningful results. Their willingness to obtain information beyond their child's core condition may be because they are already adept at handling health data. They may see adult-onset SVs as a way to gain back control in the face of uncertainty. Or, they may have adopted the kind of positive, maximising approach to ‘information’ often observed in genomics studies.

The parents did not, however, hold a uniformly positive view of the choice to receive adult-onset SVs. Despite this, they felt a moral obligation to learn about SVs; that they would be ‘remiss… to not know what is knowable’.5 Anderson et al discuss the ‘“weight” of inflicted insight’—something that occurs when a ‘subject is given insight into his [sic] flaws, although such insight is painful to him and although he has not bargained for such insight’.8 Anderson et al reframe this as ‘inflicted ought’—some parents were given insight into genomic knowledge that they did not necessarily want to know, but felt they should nonetheless come to learn.

While settling the debate over reporting adult-onset SVs in childhood is beyond the scope of this commentary, it is important to note that the choice to receive adult-onset SVs does not have to be contemporaneous with the primary decision to participate in a genomics study. Other approaches, including first addressing the core clinical question informing testing to then come back to SVs at a later point in time, could also be used.

It is also important to question whether this inflicted ought reflects a defensible obligation. My intuition is that such an obligation does not (yet) exist. The current social construction of genomic testing is one where individual control is emphasised, information per se is valorised, often while overlooking its lack of complete penetrance or potential for overdiagnosis or overtreatment. The meaning and value of genomic information remains too unsettled to assume that it should be given to all.

(2) Safeguarding deliberation?

It is worrying that in this study even those parents who expressed reservation about receiving adult-onset SVs still obtained them. While ‘inflicted ought’ explains some of this, it is also interesting to consider the relationship between information, consent and autonomy and problems with how these concepts may be operationalised in the clinic.

There is a tendency in some bioethics discourse to construe autonomy superficially, such as presenting it as a property of decisions and inextricably tying it to informed consent.ⁱⁱ If a decision is supported by information and is made voluntarily with appropriate understanding, then it is said to be autonomous. However, this places too much emphasis on information (and its transfer) at the expense of the process of the decision and the psychological properties of the person involved. Deliberation about complex decisions, such as that faced by the parents in the Genome Clinic, also needs to draw on interpretations of autonomy that recognise that it is not just about decisions but about persons too. There needs to be space for critical reflection consistent with one's values, recognition that a maximising approach to information could impact autonomy and better embedding of social and relational contexts and constraints.

A case for the need for structured deliberation about testing, with opportunity for critical reflection, is gaining traction in genomics. Gornick et al,9 for example, tested how a deliberative democracy process influenced public attitudes towards returning SVs. They found that (in the shorter term at least) participants favoured policies that did not support return of results in contexts such as predicting adult-onset disease in children. A majority of participants also disagreed with a ‘no choice’ approach to policies over returning results. The deliberative process appeared to increase appreciation of the nuances involved in such testing decisions.

Anderson et al arguably recognise this in their paper, when they say that: “seemingly clear choices may obscure the complex weighing of costs and benefits behind [them]”.5 Their suggestion for building in advance consideration of ambivalence as part of the deliberative process leading to testing is laudable.

It is also worth noting that the mere offer of a test may not be neutral. If a particular suite of information is being offered, that offer may be interpreted as implicit encouragement to accept it. This too could contribute to inflicted ought.

(3) Broad best interests

The ‘best interests’ debate in paediatric genomics has two dimensions: (a) controversy over the concept of ‘best interests’ and (b) whether providing adult-onset SVs meets a best interests test.

Regarding how ‘best interests’ should be conceptualised, Anderson et al 10 have previously rightly claimed that much of the debate over ‘best interests’ with regard to returning adult-onset SVs in children arises because differing conceptualisations of this concept are being used implicitly, obscuring meaning. While settling debates over concepts of best interests is beyond the scope of this commentary, the Genome Clinic study adopted a so-called ‘broad’ conception.5 Under this, a child's best interests incorporate relational and future-oriented components. This can justify the return of adult-onset SVs, as a benefit to a child's parents can be seen to also benefit the child. Thus, the answer to (b), on this broad conception at least, is ‘yes’.

Yet despite its wide acceptance in bioethics, clinical practice and policy, the legitimacy of using ‘best interests’ at all remains contested. Its requirement for maximising ‘the best’, controversy on when it enables intervention in parental decision-making and lack of consensus as to its interpretation have led some to call for it to be abandoned altogether.11 Additionally, as Anderson et al 5 observe, the justification for adopting approaches such as broad best interests can fall over when parents appear to adopt it when agreeing to receive adult-onset SVs in their children, but then decline to receive the same information for themselves, as happened in this study.ⁱⁱⁱ Alternatives are also available: basic interests,12 the harm principle,13 the zone of parental discretion14 and a child-focused approach1 are some that could be examined more closely in the context of paediatric WGS. Other non-interests-based concepts, such as filial obligation or familial well-being, are also relevant.

Ambivalent obligations

The study by Anderson et al demonstrates that debates over the place of WGS in paediatric clinical genomics are far from settled. As genomic testing could also soon extend to healthy populations, the need for normative ethical reflection is ongoing. In this study, some parents reported ambivalence about obtaining information pursuant to adult-onset SVs. Others chose to receive them but then declined similar information for themselves, contradicting broad justifications of children's interests in this testing.

The history of genetic testing tells us that advance enthusiasm to receive predictive health information does not always translate to subsequent health behaviour. Both the ambivalence and the inconsistency in applying broad best interests observed in this study suggests that parents may be uncomfortable or unable to weigh the complexity of considering future health information in the shoes of their child. Ultimately, if we are to remove what has traditionally been a presumptive right for competent individuals to decide, we need to be sure that the information will have immediate purpose within the family.