Should all babies have their genomes sequenced at birth? The question has been hotly debated for the past 15 years. Unlocking the genome’s 3.2-billion-letter code promises understanding of both health and disease. But practical and ethical challenges loom large.
We are now at a critical juncture. Multiple genomic newborn screening studies are launching worldwide—with cohort sizes ranging from 1,000 to 100,000 babies. These studies must not only show real health benefits for newborns, families and health care systems, but resolve the ethical, legal and implementation issues raised by the application of genomic sequencing for public health benefit.
Sequencing the first human genome was a 10-year, $3 billion, transcontinental effort. In the 20 years since, advances in technology have made it possible to sequence millions of individuals worldwide, at ever-decreasing cost and increasing speed. The results are providing fundamental insights into disease biology. Genomic sequencing now delivers direct health benefits by transforming rare disease diagnosis, guiding cancer therapy and improving surveillance of global pandemics. The benefits are arguably clearest in critically ill babies with rare diseases. Rapid genomic diagnoses already guide precision therapies and other major treatment decisions, such as organ transplants, in real time and at scale.
But what if we could identify newborns at risk of serious, but treatable, rare diseases before they became unwell? This idea is not new. The first newborn screening test was developed by Robert Guthrie in the 1960s. Blood obtained by pricking a baby’s heel was collected on filter paper and tested for phenylketonuria, a rare metabolic condition that, if untreated, causes intellectual disability. Since then, newborn screening programs have grown to incorporate multiple biochemical markers. They have rightly been heralded as exemplary successes of public health screening, delivering life-saving diagnoses with quick turnaround times and at low cost.
However, adding new conditions to newborn screening programs is slow. In the U.S. it takes an average of 9.5 years to add a single condition to the recommended uniform screening panel (RUSP). Wide variability also exists among, and sometimes even within, countries. Meanwhile, more and more rare conditions are becoming treatable. Incorporating genomic sequencing into newborn screening programs offers the possibility of suddenly expanding them from detecting tens to detecting hundreds of rare conditions, including those without other readily available biochemical markers, such as cardiac and neurological diseases. But policy and implementation ought to be informed by empirical evidence, not hype.
The challenges of incorporating genomics into newborn screening are formidable. Even if we can demonstrate the technical ability to generate and accurately interpret genomic data from thousands of newborns within clinically meaningful time frames, important questions remain. First, how do we select which conditions to screen for? Should we maintain focus on severe but treatable childhood-onset disorders as most argue, or use this as an opportunity for broader population screening such as for breast cancer risk and other adult-onset conditions? What exactly constitutes treatability? Using narrow definitions of treatability will focus programs on conditions where treatment such as bone marrow transplant or enzyme replacement therapy cures or prevents the onset of disease. However, broader definitions are also possible, for example to include early intervention therapies in intellectual disability and improve learning outcomes. Second, how do we ensure appropriate consent? Currently, most newborn screening programs operate on implicit or minimal explicit consent. Generating genomic data introduces complexities, including privacy, data usage and insurance implications. How and when is this information best presented to parents to enable carefully considered, informed choices?
It is important to remember that all screening programs inevitably cause harms. These can be at the individual level, such as disruption of parent-child bonding, or it can be at the system level, by diverting resources from diagnostic services. Perhaps most importantly, introducing genomics into newborn screening requires careful consideration of equity. This includes equity of access, which may be limited by the ability to engage with the digital consent tools likely to be central to population-scale program rollouts. Historic underrepresentation of diverse communities in genomic data sets could lead to inequitable screening outcomes unless addressed. And, particularly outside of publicly funded universal health care systems, access to precision treatments may be limited by the parents’ ability to pay, further entrenching inequality.
The best way to overcome such challenges and inform policy is with high-quality empirical evidence generated in large cohorts and in a variety of health care systems. For example, we already know from surveys, interviews, focus groups and formal public dialogues that members of the public and prospective parents generally hold positive attitudes toward genomic newborn screening and would like to see a broad range of conditions included. However, we have also learned that high interest does not necessarily translate into uptake if the offer of genomic newborn screening is made in the few days after birth, when most new parents are feeling overwhelmed. This highlights the need to develop and test different models of consent. Exploring these simultaneously in different settings will allow us to compare outcomes and gather evidence faster. Similarly, the independent generation of gene lists by multiple expert groups will increase the rigor of condition selection processes by identifying areas of international consensus and opportunities for harmonization. Trying different sequencing and analytical approaches concurrently will enable real-world comparison of performance as well as cost effectiveness.
With the technological barriers now largely resolved, we must generate high-quality evidence to inform public policy. There is no substitute for learning from doing, provided it is done in an ethically sound, considered and transparent manner. Otherwise, we risk fragmented and commercially driven implementation of genomic newborn screening, which would only exacerbate concerns about consent, data usage and equity.
It is often tempting to express strong opinions on topics like genomic newborn screening. What’s harder, and can require more courage, is to work with experts and the public to design and run studies at scale that will generate the evidence to move the debate forward. We believe now is the time.
This is an opinion and analysis article, and the views expressed by the author or authors are not necessarily those of Scientific American.