Genetic Testing is Messier in Practice than in Theory: Lessons from Neonatology

What is the future of genetic testing during pregnancy likely to look like? Given that the patterns of use of genetic testing in neonatology tends to precede, and thus predict, patterns of prenatal genetic testing in the future, we should expect the increasing use of broad, non-targeted test, such as whole exome sequencing and even genome sequencing, which in neonatology are becoming routine (NICUSeq Study Group et al. 2021).

A detailed look at how genetic testing is currently functioning in clinical practice for neonates raises concerns about the framework proposed here by Bayefsky and Berkman (2021). We agree with the authors premise the increasing and unconstrained access to genetic information poses ethical challenges. We worry, though, that recommendations about which genetic tests physicians should offer will not resolve these issues. The authors’ framework for determining what testing physicians should recommend, offer, or not offer hinges on an assumption that we can predict what information will return from a genetic test and thus, in advance, we can decide which information we do and do not want to discover. Fundamental to the argument are two other assumptions, namely that we have the ability to gather only the genetic information that would be useful in pregnancy (for prenatal testing) or early in life (for neonatal testing) without uncovering other information, and that physicians have the ability to move reliably from genotype to phenotype, which is to say associate a particular genetic result with an expected outcome, including severity of disease, age of onset, and extent of developmental disability. There are firm reasons for doubting all three of these assumptions.

Our own experiences with neonatal testing, as a neonatologist and a clinical ethicist, support what qualitative work has documented for other forms of genetic testing, namely that physicians often send tests with no idea what they will find and that genetic results are frequently uncertain (Char et al. 2018). Additionally, testing results often do not align neatly with the questions they were sent to answer.

Let’s make these issue more concrete. Consider a neonate with congenital diaphragmatic hernia, an anomaly with high associated mortality and morbidity. Suppose that this patient is sicker than most patients with the same anomaly, prompting physicians to order whole genome sequencing. Testing reveals an uncertain but possible pathogenic variant in a gene encoding a surfactant protein, as other mutations in this gene are known to cause respiratory distress of varying severity. Is this patient’s genetic mutation the reason for the more serious illness than other patients with the same anomaly? Should the finding of uncertain meaning change acute medical management, or influence the decision about whether the patient is a candidate for extracorporeal membrane oxygenation?

As genetics progresses toward broader, less targeted testing, “pre-test” restriction of the type of information discovered by carefully choosing which genetic tests to send will be increasingly difficult if not impossible (Kingsmore 2020). If this is the case, then dividing genetic information into the precise buckets that Bayefsky and Berkman outline could only be attempted “post-test”, by filtering results once they have returned. This would pose a number of practical and ethical challenges. Medical societies’ recommendations, which must be broad to address populations of patients, could not successfully accomplish the individualized filtering required. Other forms of guidance would be required. Then would follow practical choices and difficulties. Would we filter information at the level of genetics laboratories, individual physicians, or somewhere else? Who would bear the responsibility for the filtering? With patients’ and parents’ increasing access to their own medical records, how would we conceal certain information (Ross and Lin 2003)? Even if these issues could be addressed, would physicians and the public accept this type of paternalistic restriction? And most fundamentally, even if laboratory scientists, patients, and the public cooperated in such a plan, a great deal of subjectivity and uncertainty remains in assigning information to the various categories proposed in Bayefsky and Berkman’s framework.

Again, let’s illustrate with another hypothetical case. Consider fetal patient with renal anomalies. Whole exome sequencing is sent. The patient is diagnosed with a variant of 22q11.2 deletion syndrome. While other specific deletions causing this syndrome are associated with renal anomalies (Lopez-Rivera et al. 2017), this particular deletion has not been previously associated. Her genetic diagnosis is more certainly associated with an increased risk of developing psychiatric disease in adulthood. Where on the framework would this genetic information fall and who would be responsible to decide? Does this diagnosis meet the proposed threshold for which 70% of women would consider termination? If a woman did opt for termination, would this decision have been based on the renal anomaly or the risk of psychiatric disease?

In light of the limits placed on modern genetic testing practice by various forms of uncertainty, three ethical interests largely dismissed by Bayefsky and Berkman gain traction: a child’s right to an open future, disability rights, and the parent-child relationship. First, the authors suggest that we can reduce effects of genetic information on a child’s future if professional bodies only recommend “information that could be immediately useful for termination or preparation purposes.” The experience in neonatology shows that we cannot easily censor information so as to provide only that which serves a particular purpose. Information revealed in order to decide whether to continue invasive life-sustaining care for a neonate has long lasting effects for patients who then live. We must therefore grapple with the effect of information that does make the child’s future less “open.”

Second, evidence is mounting that genetic testing in neonates is used in ways that do send hurtful messages to the disability community (Deem 2016; Callahan et al.). In part, this stems from difficulties in predicting the nature of disability using genetic tests and in regulating how parents or physicians use genetic results, both explicitly and subconsciously. Bayefsky and Berkman discount these concerns on the basis that they are not backed by empirical evidence. Recent work we have done does, however, provide precisely the support to substantiate such concerns in the neonatal population (Callahan et al.).

Finally, Bayefsky and Berkman also underestimate the effect that a genetic finding can have on a nascent parent-child relationship. Uncertainty is inherent to all parenting, but when increased by medical uncertainty, the cumulative uncertainty can have a formidable effect. Even variants of uncertain significance that are not substantiated with later symptoms nevertheless affect parenting styles through early childhood (Werner-Lin et al. 2019).

If we are correct that the unstoppable trend is toward increasing access to our own genomes, however imperfect and uncertain, and that we cannot limit the effects of genetic information on children’s open futures, perceptions of disability, or parent-child relationships by censoring what we discover, what are we to do? Rather than engaging in rearguard efforts to restrict genetic testing (which we believe are very unlikely to succeed), we should turn our attention to managing the complex information that genetic testing produces. We need research to learn how to enhance our counseling for parents to reveal the breadth of information produced by modern genetic tests so that they can make informed testing decisions (Chen and Wasserman 2017) and handling the ensuing uncertainty. We need research also to improve education for physicians about interpretation and application of genetic findings. We need to remain vigilant for genetic-related biases and additional protections, and enforcement of protections, for children and parents who face genetic diseases and disability of all kinds.