Integrated genomic screening found far more high-risk adults than gene testing alone
An integrated model combining monogenic variants and polygenic risk scores identified substantially more people at elevated risk for Tier-1 conditions than a monogenic-only strategy in a real-world U.S. biobank. The finding lands at an important moment for genomic medicine, where risk stratification has often been framed as a binary choice between rare high-penetrance mutations and common low-effect genetic burden. In this cohort, the broader architecture was more clinically useful than the narrower one.
Among 20,477 participants in the Genomic Health Initiative, the median age was 64 years, 65% were female, and 36% were of non-European ancestry. Tier-1 conditions were common, affecting 5,381 participants, or 26% of the cohort. The conditions under study included breast, colorectal, and prostate cancers, along with coronary artery disease. Those diagnoses are often treated as separate clinical silos, but their shared feature is risk accumulation across biology, age, and inherited susceptibility.
Polygenic burden carried more of the population risk than rare pathogenic variants
Across the conditions examined, polygenic risk scores explained a substantially larger share of population attributable risk than pathogenic variants. The polygenic contribution ranged from 3.44% to 13.96%, while the monogenic contribution ranged from 0.17% to 2.65%. In practical terms, that gap means common inherited variation was doing more of the population-level work than rare single-gene findings. For clinicians accustomed to treating pathogenic variants as the central genomic signal, the result is hard to ignore.
A monogenic-only testing strategy restricted to people with family history identified 166 pathogenic variant carriers. Those carriers represented only 1.5% of all cases. The number is small because family history is an imperfect filter: it misses late-onset disease, small families, uninformative family structures, and relatives whose diagnoses are unknown or undocumented. A gene-first approach limited to obvious hereditary phenotypes therefore leaves much of inherited risk outside the frame.
GenProb captured a broader swath of inherited susceptibility
The integrated model, GenProb, performed better than pathogenic variant testing alone for identifying high-risk people. GenProb flagged 6,563 individuals as high risk, including 1,259 cases, and accounted for 23.4% of all cases. In a screening context, that difference matters because it moves genomic risk assessment closer to the reality of everyday practice, where disease emerges from overlapping genetic effects rather than from a single dramatic mutation.
The biology behind this performance is straightforward. Rare pathogenic variants can confer large risk in a small subset of people, but they do not explain most common disease burden. Polygenic scores capture the cumulative effect of many common variants, each contributing a modest increment that becomes clinically meaningful when aggregated. For disorders such as coronary artery disease and common cancers, that cumulative burden can shape who develops disease, when it appears, and which patients should be prioritized for prevention efforts.
A narrow hereditary model would have missed most of that signal.
Real-world biobank data sharpen the case for combined risk models
This was not a theoretical exercise in a highly selected pedigree clinic. The Genomic Health Initiative is a community-based healthcare biobank, which gives the results immediate relevance to broad clinical populations. The cohort included substantial ancestry diversity, with 36% of participants from non-European backgrounds. That matters because inherited risk tools that perform only in European ancestry populations have limited utility in routine care and can widen disparities if they are adopted without validation.
The study used Cox regression models to estimate population attributable risk percentage, C-index, and net reclassification index. Those are technical measures, but their clinical message is plain: a combined model improved risk assessment beyond pathogenic variants alone. In screening medicine, classification is not an academic detail. It determines who gets a colonoscopy sooner, who enters a breast surveillance pathway, and who may deserve earlier cardiovascular prevention.
What this means for Tier-1 disease screening
Tier-1 conditions already occupy a special place in clinical genomics because they are actionable. Risk identification is not valuable simply because it is precise; it is valuable when it changes prevention, surveillance, or counseling. The current study suggests that if programs rely only on family-history-triggered monogenic testing, they will identify only a small fraction of affected people. Integrated genomic assessment offers a larger net, and the extra catch is not trivial.
That said, the study also exposes the operational challenge ahead. A model that identifies more people as high risk only helps if health systems can interpret the results, communicate them clearly, and route patients into preventive care without overtesting or unnecessary anxiety. Polygenic risk also requires careful calibration across ancestry groups, clinical settings, and disease categories. A score that is statistically strong but operationally opaque will not change outcomes on its own.
Still, the direction is clear from these data. In a community-based cohort, polygenic burden contributed more to population risk than rare pathogenic variants, and the combined GenProb model reached far more cases than a family-history-based monogenic strategy. For Tier-1 conditions, inherited risk is not just a matter of finding the rare mutation. It is also a matter of recognizing the common variation that quietly shapes disease risk across the general population.
The remaining question is not whether integrated genomic assessment can outperform a narrow monogenic approach; this cohort shows that it can. The harder task is building clinical pathways that use that information safely, consistently, and at scale.