Is genomic sequencing ready for routine clinical care?

Where does genomic sequencing already work in routine care?

The clearest success is in diagnosing rare diseases in children and adults. In the English National Health Service, genome sequencing as a routine clinical test for children with suspected rare conditions achieved a diagnostic yield of 29% — meaning it found a genetic cause in nearly 1 in 3 children tested [2]. This was significantly higher than the 22% yield from the earlier research-based 100,000 Genomes Project, showing that moving from research to routine care actually improved results [2]. Similarly, a 10-year program at Karolinska in Sweden diagnosed 22.6% of over 15,000 patients with rare diseases, identifying variants in 1,570 different genes [11]. These aren't isolated successes: a large U.S. study of 744 families who had previously gone undiagnosed found that genome sequencing provided a diagnosis for 29.3%, and critically, 8.2% of all families had variants that only genome sequencing could detect — variants missed by standard exome sequencing [4].

In cancer care, the evidence is also strong but more conditional. A landmark Dutch study (the WIDE study) prospectively performed whole genome sequencing on 1,200 consecutive cancer patients and found it was technically feasible in routine practice, with a turnaround time of just 10 working days [7]. Most importantly, it identified additional clinically actionable biomarkers — targets that could guide treatment — in 54% of patients compared to standard molecular testing [7]. A U.S. pilot program at Weill Cornell found that paired germline/somatic whole genome sequencing detected clinically actionable variants in 39% of tumor samples and revealed hereditary cancer risk in 17% of patients [3]. However, this same study noted significant barriers: difficulty identifying which patients benefit most, lengthy consent processes, and limited access to genetic counseling [3].

What's the gap between best-case and typical-case evidence?

The gap is wide. While specialized centers like the Netherlands Cancer Institute and Karolinska have successfully integrated genome sequencing into routine care, most large-scale biobanks and health systems are not returning genomic results to patients. A survey of 24 large biobanks with genomic and electronic health record data found that only 6 (25%) disclosed potentially actionable genomic results to participants [1]. This means that even when the data exists, it's not being used clinically. The same study from Geisinger Health showed that among 175,500 participants who had exome sequencing available, 3.4% (5,934 people) had a pathogenic variant in a disease-risk gene — and nearly 90% of those were unaware of their risk before the study disclosed it [1]. That's a massive missed opportunity.

For population-wide screening, like newborn screening, the evidence is even more cautious. A 2016 survey of Australian parents and health professionals found that while 77% of parents expressed interest in genomic newborn screening, 62% of health professionals felt the technology should not yet be used as an adjunct to current newborn screening [5]. Both groups identified major hurdles: the need for accurate data interpretation, pre- and post-test counseling, and appropriate consent processes [5]. The authors noted that for genomic sequencing to be considered, these ethical, social, and practical implications need much more investigation [5].

Even in areas where genome sequencing works well, it doesn't always outperform existing methods. A study comparing genome sequencing to standard-of-care testing for neurodevelopmental disorders in 150 patient-parent trios found that genome sequencing had a similar diagnostic yield (30% vs. 28.7%) [9]. The advantage wasn't finding more diagnoses, but finding them in a single test rather than integrating results from multiple assays [9]. This efficiency gain is real, but it's not a game-changer in diagnostic power for this specific condition.

What still holds genomic sequencing back from routine clinical care?

Several practical barriers remain. First, sample quality and tumor content are major issues: in the WIDE cancer study, 25% of samples couldn't be sequenced because they had insufficient tumor cells [7]. In the Weill Cornell pilot, 7.3% of cases had insufficient quantity or quality of tissue [3]. Second, the infrastructure for returning results is underdeveloped. The Geisinger study found that most large biobanks don't return results, and even when they do, the process requires significant genetic counseling resources [1]. Third, there's the question of who benefits most. The Parkinson's disease clinic study found that while 82% of patients would share a genetic risk variant with relatives, and 75% wanted to know their polygenic risk score, the actual actionable findings were rare — only 1% had medically actionable variants like BRCA1 [6]. This suggests that for many common conditions, the clinical utility of genome sequencing is still limited.

Cost and turnaround time have improved but remain barriers. The Dutch WIDE study got turnaround time down to 10 working days, which is clinically useful [7]. But the validation study for that same pipeline reported a technical success rate of 95.6% for samples with sufficient tumor content, meaning about 1 in 20 samples that should work still fail [10]. For infection control, routine whole genome sequencing of multi-drug resistant bacteria has proven feasible and useful — a 4-year Australian study sequenced 2,660 isolates and identified 76 transmission clusters, directly informing infection control policies [8]. But this is a very specific, targeted use case, not broad population screening.

The evidence consistently shows that genomic sequencing is ready for specific, high-value clinical scenarios — rare disease diagnosis in children, targeted cancer care, and hospital infection control — but not for routine screening of healthy populations. The gap between what's possible at leading centers and what's actually happening in most health systems remains large, and closing it will require solving infrastructure, counseling, and cost challenges that go beyond the technology itself.

About These Sources

This answer is built on 11 peer-reviewed studies — published from 2021 to 2025, 5 from 2024 or later, 9 in Q1 journals, collectively cited 455 times — selected as the most relevant from 13 studies that passed quality screening, drawn from 74 papers retrieved from a database of over 500 million.

Sources used in this answer

Related research questions