Should Every Baby's DNA Be Sequenced (Economist)

Summary

A Science & Technology piece on whole-genome screening of newborns. England's Generation Study is sequencing 100,000 babies for around 200 childhood-onset conditions; similar trials run in America (BabySeq2, ~1,000 genes including adult-onset cancers), Australia and parts of Europe. Opening case: Freddie, born April 2025, whose retinoblastoma was caught and treated early because the study flagged the gene. The piece is balanced: more diseases caught earlier versus "patients-in-waiting", harmful follow-up treatment, and a consent problem baked into the design of most programmes. See Newborn Genome Sequencing.

Key claims

• Scale jump versus today's heel-prick screens. Conventional blood-spot screening covers conditions such as sickle-cell anaemia and cystic fibrosis; the broadest national panel (Italy) looks for ~50 diseases. DNA screening lifts that to ~200 (Generation Study) or ~1,000 genes (BabySeq2), and also flags mutations that cause nothing yet.
• Penetrance has been overestimated. Penetrance = share of carriers of a disease-causing variant who actually fall ill. Early studies put rb1 (retinoblastoma) penetrance above 90%; a paper posted online in December (Caroline Wright, University of Exeter) found fewer than a third of adults carrying risky rb1 variants had ever had the cancer. Reason: variants were historically discovered by working backwards from people already ill, inflating estimates. New work suggests the same for ret (medullary thyroid cancer).
• Diagnostic purgatory is a documented harm. Screening hundreds of low- or middling-penetrance genes creates "patients-in-waiting"; in cystic fibrosis it has an acronym, CFSPID (screen positive, inconclusive diagnosis), and parents of such children often question whether the result was worth having (Anneke Lucassen, Oxford).
• Follow-up can cause harm. A risky rb1 result only triggers eye tests; a ret result commonly leads to prophylactic thyroidectomy and lifelong hormone replacement, in people whose true risk may be much lower than assumed.
• Research storage is bundled with screening. Most programmes require consent to both screening and indefinite genome storage for research; "you can't take part in one without taking part in the other" (Lucassen). The research upside is real (Britain's 100,000 Genomes Project let Oxford researchers identify variants behind a childhood neurodevelopmental disorder; future uses include drug side-effect and treatment-response prediction). The downside: genomes are among the most sensitive data there is. The 2023 23andMe breach exposed 6.9m customers; some American insurers can already use genetic-test results in setting premiums (mostly banned in Britain and Australia).

What the article flags as unproven or contested

• Whether population-level benefit outweighs cost once false reassurance, over-treatment and anxiety are counted; that is precisely what the trials are meant to find out.
• Which genes should be reported: one proposed fix is returning results only for treatable, high-penetrance conditions, which would disappoint parents who want everything.
• The ethics of running a screening programme (justified by patient benefit) and a research biobank (justified by uncertain benefit to others) in the same consent, on infants who cannot consent. Jan Friedman (UBC) calls this "mission creep".

Wikilinks

• Newborn Genome Sequencing — concept page: programmes, penetrance problem, consent bundling.
• Polygenic Embryo Screening — the pre-birth version of the same genetic-prediction promise; shares the penetrance / absolute-risk caveat.
• The Economist — Science & Technology.

Captured-by trigger

Source file: raw/processed/Economist 2026-07-04 Science---Techn - Should every babys DNA be sequenced.md.