Can Public Health Become Precise?

In 2003, scientists completed one of the most ambitious research projects in human history: the Human Genome Project. After more than a decade of work and billions of dollars in investment, researchers had produced the first complete map of the human genetic code. The achievement promised nothing less than a revolution in medicine.

The vision was compelling. Rather than treating every patient with the same drug, the same dosage, and the same assumptions, physicians could someday tailor care to each individual’s unique biology. Cancer treatments could be matched to the mutations driving a specific tumor. Medications could be selected based on a patient’s genetic profile. Diseases might even be predicted and prevented before symptoms appeared.

Over the next two decades, that vision evolved into what we now call precision medicine. While many of its promises remain works in progress, precision medicine has already transformed fields such as oncology, pharmacogenomics, and rare disease diagnosis. Millions of patients have benefited from therapies designed not for the “average” patient, but for people with specific biological characteristics.

Yet this success raises an intriguing question.

If medicine can become precise, can public health?

At first glance, the idea seems contradictory and paradoxical. Public health has traditionally focused on populations rather than individuals. Vaccination campaigns, tobacco control policies, clean water systems, injury prevention programs, and disease surveillance efforts are designed to improve health at scale. The goal has never been to determine which intervention is best for a particular person, but rather which intervention will produce the greatest benefit for a community, city, state, or nation.

But the world that gave rise to traditional public health is changing rapidly.

Today, researchers have access to enormous streams of data that would have been unimaginable even a decade ago. Electronic health records capture patterns of disease in near real time. Satellites monitor environmental exposures from space. Smartphones generate detailed information about movement and behavior. Genomic sequencing can identify emerging pathogens within days. Artificial intelligence systems can sift through millions of records to uncover patterns that humans might never detect.

Together, these advances are creating the possibility of something that once seemed impossible: delivering the right public health intervention to the right population, at the right time.

This emerging field, often called precision public health, seeks to combine the population focus of public health with the data-driven targeting capabilities that transformed precision medicine. Rather than asking how to personalize treatment for a single patient, precision public health asks how to better identify which communities face the greatest risks, which interventions are most likely to work in specific contexts, and how to allocate scarce public health resources more effectively.

The idea has generated both excitement and skepticism. Supporters argue that precision public health could dramatically improve disease prevention, outbreak response, environmental health monitoring, and health equity. Critics worry that the concept may overpromise, divert attention from the social determinants of health, or simply rebrand tools that public health practitioners have used for decades.

So is precision public health the next great frontier in population health or merely a buzzword in search of a definition?

The answer lies at the intersection of genomics, data science, epidemiology, artificial intelligence, and one of the oldest questions in public health: how can we use limited resources to improve health for the greatest number of people?

How precision medicine set the stage

Precision public health did not appear out of nowhere. It inherits its vocabulary, optimism, and tools from precision medicine, whose modern institutional history runs through the Human Genome Project, the postgenomic data revolution, and the U.S. Precision Medicine Initiative. The Human Genome Project officially began in October 1990 and was completed in April 2003; along the way, it produced a draft of the human genome in June 2000, accelerated sequencing technology, and helped normalize rapid data sharing through the Bermuda Principles. NHGRI estimates the Human Genome Project’s original projected cost at about $3 billion, while its scientific payoff extended far beyond the reference sequence itself. 

A second milestone came in 2011, when the U.S. National Research Council’s report Toward Precision Medicine helped formalize the term and argued for linking molecular data with clinical information to refine disease classification. Then, in January 2015, President Obama announced the Precision Medicine Initiative, described by the White House as a “bold new research effort” launched with a proposed $215 million investment; NIH later developed the PMI Cohort Program, which became All of Us. Today All of Us describes itself as a nationwide precision-medicine program, and in recent data releases it has combined surveys, physical measures, EHRs, genomic analyses, and wearable data at large scale. 

Another quiet but important milestone came in 2022, when the Telomere-to-Telomere consortium announced the first truly complete human genome sequence, filling gaps that the Human Genome Project could not resolve with earlier technologies. That matters for public health because “precision” depends on the completeness, representativeness, and analytic usefulness of the reference systems on which downstream tests and risk models rely. 

What precision public health means

Precision medicine and precision public health overlap, but they are not the same enterprise. CDC’s Office of Public Health Genomics has framed the distinction neatly: if precision medicine is about the individual, precision public health is about populations delivering “the right intervention at the right time, every time to the right population.” CDC’s later discussion refined the point further: precision public health includes both using public health to realize the population-level benefits of precision medicine and using new tools to sharpen the classic public health functions of assessment, policy development, and assurance. 

That definition matters because the term is still contested. Reviews of the literature show that “precision public health” has been used to describe several related ideas: more granular risk prediction, sharp geographic targeting, integration of genomics with social and environmental data, and population-level delivery of interventions informed by data science.

Some authors welcome the term; others warn that a technology-heavy framing can distract from upstream determinants of health if it privileges better measurement over better policy. 

What now makes it plausible

What changed is not one technology but a stack.

First, sequencing costs collapsed. NHGRI’s cost tracking shows an extraordinary fall from roughly $95 million per human genome in 2001 to about $562 in 2021, and NHGRI notes that by the end of 2015 draft whole genomes had already dropped below roughly $1,500 in its funded centers. Falling costs turned genomics from a prestige project into an operational input. 

Second, the COVID-19 era significantly accelerated the field of public health bioinformatics. CDC’s SPHERES consortium coordinated SARS-CoV-2 sequencing across public health labs, academia, and the private sector, while CDC’s genomic surveillance program integrated contracted sequencing, public repositories, and tagged baseline surveillance samples to produce weekly jurisdiction-level estimates. Meanwhile, NIH’s N3C showed that more than 90 institutions could harmonize EHR data in a central enclave and link those data to other sets for research and risk modeling. That is the kind of backbone precision public health needs: not just smarter algorithms, but routinized pipelines. 

Third, public health gained richer nonclinical data. Digital epidemiology, in a classic formulation, uses data generated outside the health system to answer epidemiologic questions. During and after COVID-19, that logic expanded to include mobile phone mobility data, syndromic signals, online behavior, and wearable sensors. Wearable reviews suggest such devices can detect physiologic changes before or around symptom onset for some infections, and All of Us now explicitly includes wearable data in its research environment. 

Fourth, place-based analytics matured. CDC/ATSDR’s Social Vulnerability Index quantifies community vulnerability using census-based variables (or it used to before the shenanigans of 2025), while the Environmental Justice Index and climate-burden modules add indicators of environmental load, heat, wildfire, and extreme events. These tools are not glamorous, but they are exactly what make population targeting concrete. In parallel, wastewater surveillance became a normalized population sensor: CDC says NWSS can help detect outbreaks early, direct prevention where needed, and complement clinical surveillance. 

Where it already works

The strongest case study is SARS-CoV-2 genomic surveillance. CDC reported that genomic surveillance was critical for tracking emerging variants; during 2021 it integrated new data sources and analytic methods to produce more timely, geographically representative estimates of variant proportions. That system documented both Delta’s rapid rise in mid-2021 and Omicron’s even faster displacement of Delta in late 2021, thereby informing public health planning and communication. This is precision public health at its clearest: not individualized care, but faster, more precise situational awareness of populations. 

A second success is wastewater surveillance. CDC launched NWSS in September 2020; by December 2022, it had expanded from 209 to more than 1,500 sampling sites, covering about 47% of the U.S. population. CDC’s public-facing materials describe it as an early-warning tool that can guide decisions such as where to deploy mobile testing and vaccination sites. That is a classic precision-public-health move: a lower-cost, community-level biomarker that helps public agencies act before clinical case data fully catches up. 

A third example is tuberculosis sequencing. CDC explains that conventional TB genotyping examines less than 1% of the genome, whereas whole-genome sequencing examines more than 90%. CDC began retrospective WGS in 2012 and moved to prospective sequencing of all new M. tuberculosis isolates through its National TB Molecular Surveillance Center in 2018. WGS supports cluster detection and possible drug-resistance detection, and WHO’s 2023–2024 guidance on next-generation sequencing for drug-resistant TB shows that this is becoming part of global surveillance and diagnostic strategy, not merely an academic exercise. 

Some of the oldest examples are still among the best. Newborn screening is, in effect, a mature form of precision public health: universal population screening for rare conditions, followed by highly targeted intervention for the few infants identified. CDC reports that newborn screening in the United States began in the 1960s, now reaches more than 98% of roughly 4 million newborns annually, and reduces morbidity and mortality through early identification and treatment. It is population-wide in reach, but precise in action. 

There are also strong place-based examples. During COVID-19, CDC/ATSDR’s SVI helped place 240 additional testing sites in 33 states, with 69% of those sites located in communities with moderate to high social vulnerability; CDC also describes its use in Massachusetts to identify testing disparities and in broader efforts to prioritize vaccine allocation and other resources. These examples show the basic promise of precision public health: using geography, vulnerability data, and real-time need to move resources closer to the people most likely to be missed.

Our own analyses point in the same direction and also show why single indices are not enough (which we’ll be getting out in the next few months). Measures such as SVI and CDI identify overlapping but distinct communities: SVI is especially useful for surfacing urban communities facing housing, language, transportation, and demographic vulnerability, while CDI more directly captures concentrated socioeconomic deprivation, persistent poverty, and financial hardship. That distinction matters because where agencies draw the line can determine whether millions of people are included or left out.

For overdose prevention, CDC’s Montana case study shows how linking law-enforcement ODMAP data with EMS data gave near-real-time visibility into spikes, allowing epidemiologists to warn local harm-reduction groups and EMS teams. And for environmental health, CDC’s Tracking Network and ATSDR’s Environmental Justice Index give agencies a way to map exposure burdens and health vulnerability together, including heat and wildfire risks. Across these examples, the lesson is not simply that more data are better. It is that public health targeting improves when agencies combine multiple forms of place-based information social vulnerability, deprivation, persistent poverty, medical debt, uninsured burden, environmental exposure, and near-real-time service data to identify communities with overlapping need and respond with resources proportional to the severity of that need. 

Why the concept can fail

The biggest mistake is to think precision public health is mainly a data problem. It is also a governance, equity, and implementation problem. WHO’s surveillance ethics guidelines emphasize legitimate purpose, proportionality, and careful decisions about collection, sharing, and use of surveillance data, while WHO’s AI governance report warns that AI in health raises concerns around autonomy, safety, accountability, and fairness. Those warnings apply with special force when models label neighborhoods, not just patients. 

Equity is the second hard constraint. Precision systems can easily become more precise for people who are already visible in the data and less precise for everyone else. That is why diversity and representativeness matter so much. The 2024 Nature paper on All of Us described 245,388 clinical-grade genome sequences and emphasized that 77% of participants in that genomic release came from historically underrepresented communities and 46% from underrepresented racial and ethnic minorities—a corrective to older genomic datasets, but also a reminder of how skewed the field has been. Reviews of precision public health repeatedly flag the risk that predictive systems can deepen, rather than reduce, disparities if fairness is not designed in from the start. 

Then there is infrastructure. N3C’s success required harmonization across more than 90 institutions. NWSS’s own surveillance summary stresses the need for long-term capacity and early standardization of methods. CDC’s TB pages implicitly make the same point: sequencing is not just a machine; it is a laboratory workflow, a bioinformatics workflow, and a public-health interpretation workflow. Precision public health is therefore constrained not by imagination but by staffing, interoperability, procurement, maintenance, and trust. 

Finally, there is a conceptual risk: precision can crowd out solidarity.

Public health often works through universal measures with broad social benefits. A model that helps place one testing van or one naloxone team is valuable. But a society cannot algorithm its way out of weak primary care, poor housing, labor precarity, unsafe water, or underfunded health departments. The literature’s most sensible position is therefore not “precision versus public health,” but precision in service of core public health. 

A realistic path forward

So is precision public health feasible? 

Yes, if we define it modestly and build it institutionally. 

The feasible version is not individualized prevention for eight billion people. It is better surveillance, better targeting, and better timing for public-health action. It works best where the intervention is already known, the signal is actionable, and the governance is trusted: variant tracking, wastewater early warning, newborn screening, TB resistance detection, overdose spike response, and vulnerability-informed outreach. That conclusion is an inference from the strongest available case studies, and it is where the evidence is currently most persuasive. 

For researchers, the priority should be to evaluate interventions, not just predictive performance. A hotspot map or risk score matters only if it changes outcomes. That means prospective evaluation, subgroup reporting, calibration checks, external validation, and explicit inclusion of social and environmental variables alongside molecular ones. It also means building privacy-preserving linkage methods and transparent documentation into study design rather than treating them as add-ons. 

For public health agencies, the practical agenda is infrastructure and governance: interoperable data standards, trusted research environments, tiered access, audit logs, community advisory mechanisms, and a workforce that includes laboratorians, epidemiologists, GIS specialists, informaticians, and ethicists. The All of Us and N3C models show that secure cloud or enclave-based access can widen the research community while still centering stewardship and privacy. 

For policymakers, the essential safeguard is equity by design. Fund public-health basics and precision tools together. Require bias audits and impact assessments for AI and predictive systems. Protect lawful public-health data use while setting limits on secondary use and re-identification risk. And insist that place-based targeting must never be used as a pretext to neglect universal access to vaccines, treatment, clean air, safe housing, and preventive care. Precision public health can be real—but only if it remains, at heart, public health. 

Precision Without Losing the Population

The future of public health is unlikely to look exactly like precision medicine. Public health will never be about creating a unique intervention for every individual citizen. Its mission remains fundamentally collective: protecting communities, preventing disease, and improving health at population scale.

But that does not mean public health must remain blunt.

For most of its history, public health has often relied on broad categories, county averages, and one-size-fits-all interventions. Today, advances in genomics, geospatial analytics, artificial intelligence, environmental monitoring, electronic health records, and real-time surveillance are making it possible to see risk with far greater clarity. We can identify emerging outbreaks sooner, pinpoint neighborhoods facing overlapping vulnerabilities, anticipate environmental threats, and direct resources where they are likely to have the greatest impact.

The real promise of precision public health is not that it will personalize public health. It will make public health smarter.

Yet technology alone will not be enough. The greatest determinants of health remain poverty, education, housing, discrimination, environmental conditions, and access to care. Precision public health succeeds only if it helps us address these realities more effectively rather than distracting us from them. Better data should lead to better decisions, not simply more sophisticated dashboards.

Perhaps the most important lesson from the field’s early successes is that precision public health is not about predicting who will get sick. It is about identifying where risks are concentrated, understanding why those risks exist, and delivering the right intervention to the right community at the right time.

If precision medicine seeks to treat the right patient, precision public health seeks to build the right conditions for health. In a world of growing data abundance and persistent health inequities, that may prove to be one of the most important scientific challenges of the twenty-first century.