Genetics and Public Health: Toward an Efficient Division of Labor

The question of how genetics should figure in public health resource allocation has been muddied by two competing distortions. Genomic enthusiasts oversell the clinical actionability of common-disease genetic markers and underspecify what changing risk estimates actually changes about clinical advice. Environmental determinists dismiss genetic evidence as a distraction from the structural levers that drive population health and underestimate the genuine heterogeneity in vulnerability that genetic data can reveal. Both positions serve coalition interests. Neither serves patients efficiently.
The honest account begins with a distinction that has been resisted because it threatens established research programs and funding streams. For rare, severe conditions caused by identifiable mutations, genetic testing is among the most cost-effective tools in medicine. Identifying BRCA1/2 mutations, Lynch syndrome, familial hypercholesterolemia, or Huntington’s disease changes clinical decisions sharply and saves lives. The case for genetic screening in these domains is strong and the resource investment is justified. For common, polygenic conditions, including most cases of type 2 diabetes, coronary artery disease, and major depressive disorder, the situation is fundamentally different. The genetic variants associated with these conditions have small individual effects. Knowing that someone’s polygenic risk score places them in the upper quintile for depression does not change the clinical recommendation they receive. It tells them what careful family history collection already suggested. The information adds cost without altering advice.
This does not mean polygenic risk scores have no role in public health. It means their role is specific and its limits require honest acknowledgment. The emerging evidence on gene-environment interaction offers a more defensible case for genetic stratification than the standard PRS argument does. If genetic variants that influence stress reactivity interact with specific environmental exposures to produce disorder in some individuals but not in others, then identifying high-vulnerability individuals before those exposures occur allows targeted prevention rather than universal intervention. This uses genetic information to make environmental interventions more precise. A person with high polygenic loading for stress reactivity who is about to enter a high-adversity environment represents a different prevention opportunity than a low-loading individual in the same environment. Directing intensive support toward the first person while allowing lighter-touch approaches for the second uses the information where it has genuine decision-relevance.
The conditions for this to work are demanding and worth stating plainly rather than burying in methodological fine print. Polygenic risk scores must be validated across ancestries. The current literature is heavily Eurocentric, and scores calibrated on European ancestry populations perform substantially worse in other groups. Applying poorly calibrated scores in clinical or public health settings would direct resources away from high-risk individuals in underrepresented populations, compounding existing health disparities rather than reducing them. Until ancestry-adjusted scores with validated cross-population performance are available and independently audited, broad public health deployment creates harm that the efficiency argument does not override.
Even well-calibrated scores face the operationalization problem that Horwitz’s work makes visible from a different angle. Risk stratification is only efficient if the interventions directed at high-risk individuals are themselves effective. For many common conditions, the most effective interventions are not genetic or pharmacological. They are environmental, behavioral, and structural. A high-PRS individual for type 2 diabetes benefits most from access to affordable whole food, safe walking infrastructure, and reduced occupational stress, not from earlier statin prescriptions or genome sequencing. The genetic information identifies who needs the environmental intervention most urgently. It does not change what the intervention is. If the infrastructure for effective environmental intervention is absent, the efficiency gain from genetic stratification is illusory. The high-risk individual gets identified and then offered nothing that changes their trajectory.
This points directly to the resource allocation question. Public health spending on broad genetic screening for common diseases, in the current environment where the infrastructure for effective environmental intervention is underfunded, misallocates resources in a way that is empirically demonstrable rather than ideologically charged. The billions directed toward genome-wide sequencing programs, common-disease biobank expansion, and PRS development for conditions where the genetic information does not change clinical advice represent an opportunity cost that falls primarily on the environmental interventions with the largest documented returns.
Air quality improvement stands above other candidates on the evidence available. Fine particulate matter and nitrogen dioxide drive cardiovascular events, stroke, respiratory morbidity, and emerging evidence links chronic pollution exposure to neuroinflammatory pathways relevant to depression and cognitive decline. These are population-level exposures affecting everyone, with the largest burden falling on low-income urban communities with the least political capacity to demand remediation. The cost-effectiveness evidence is strong across multiple independent analyses. The U.S. Clean Air Act has returned roughly twenty-five to thirty dollars in health and economic benefits for every dollar spent on enforcement. Benefit-to-cost ratios at this magnitude are rare in public health and essentially absent in genomic medicine for common diseases.
The efficiency argument for redirecting resources toward air quality improvement does not rest on dismissing genetics. It rests on applying the same cost-effectiveness logic to both interventions and accepting what the evidence shows. Genetic screening for common diseases, in the absence of ancestry-adjusted scores, validated GxE models, and effective downstream interventions, adds cost and creates risk of iatrogenic labeling. Air quality improvement reduces disease burden at the population level without requiring individual screening, without creating new diagnostic identities, and without the equity risks of poorly calibrated genetic tools.
Food systems offer a near equivalent case. Sugar-sweetened beverage taxes with revenue directed toward fresh produce access generate health gains through a direct causal pathway, reduce healthcare costs, and have documented effects on obesity and diabetes incidence in natural experiments across multiple countries. The return on investment is strong.
Income transfers and housing quality interventions work through all downstream determinants simultaneously and produce the most durable long-term improvements in population health. Their weakness as candidates for redirected screening budgets is that their effects operate over decades rather than years, making the case for reallocating research and screening funds to social policy harder to prosecute in any specific budget cycle, even if the lifetime returns are superior.
The honest conclusion for policymakers is a division of labor that reflects what each tool actually does. Genetic screening serves patients with rare, high-penetrance conditions and potentially serves high-risk stratification for prevention when ancestry-adjusted scores and effective downstream interventions are both available. Environmental intervention, led by air quality and food systems, serves population health for common diseases with documented efficiency that genetic medicine for those conditions cannot match. Allocating resources accordingly is not ideological. It is the application of the same cost-effectiveness logic that public health uses to justify any other resource decision, applied without the institutional protections that genomic research programs have accumulated by attaching themselves to the prestige of precision medicine.
What clarity requires is a willingness to end programs whose cost-effectiveness case has not materialized, and a funding structure that rewards honest evaluation of returns rather than the continuation of programs whose survival depends on the belief that the clinical actionability of common-disease genetics is just around the corner. That belief has been just around the corner for twenty years. Clarity means naming that.