Precision Medicine Needs Better Infrastructure — And We Already Have The Models For It
By June Cha, Ph.D., MPH
Precision medicine aims to match patients to the treatments most likely to work for them. Over the past decade, that promise has begun to deliver. Targeted oncology therapies, pharmacogenomic screening, and biomarker-driven trial designs have transformed how we develop and deploy medical products. But as the science of treatment grows more precise, the science of measuring whether clinical trials represent their intended populations has not kept pace.
In a study my colleagues and I recently published in PLOS One, we examined how standard methods for evaluating clinical trial representativeness perform when the population a trial is targeting differs, even modestly, from the broader disease population recorded in a patient registry. We found that these standard methods perform poorly. The implications go beyond measurement and into policy, resource allocation, and the infrastructure that connects patients to both healthcare and research.
The Measurement Problem
The standard approach to evaluating clinical trial representativeness compares the demographic makeup of trial participants to the demographics of the corresponding disease population in a national registry. This works when the registry population matches the population the trial is targeting. But as precision medicine moves toward therapies designed for specific molecular subtypes, defined by genetic mutations, receptor profiles, or biomarker signatures, the trial’s target population increasingly differs from the broader disease population in the registry. These subpopulations may have different demographic compositions, meaning the comparator used to judge representativeness may itself be wrong.
Using U.S. cancer clinical trials registered between 2017 and 2023, we simulated the effects on representativeness assessments when the target population differs from the registry population by fixed amounts. At a difference of just 5 percentage points, roughly a quarter of studies were misclassified by sex, about a third by ethnicity, and 40 percent by race. At 15 percentage points, two-thirds to three-quarters of studies were misclassified. These margins are well within the range of documented differences between cancer subtypes and their parent disease populations, and they will only grow as precision medicine targets narrower subgroups.
The Gap Between Healthcare Access And Research Visibility
The tools we currently rely on to assess whether clinical trials reflect their intended populations are not built for the era of precision medicine. They were designed for a world in which clinical trials targeted broad disease categories like breast cancer, prostate cancer, and colorectal cancer, and where national registries could serve as a reasonable proxy for the relevant patient population. That assumption is less valid today.
If we care about whether a therapy works for the patients it is designed to treat, and whether that therapy will perform in the real world the way it performed in the trial, then we need representativeness metrics that are anchored to the right populations. Building those metrics requires better data on who the relevant patients actually are, and no statistical technique alone can provide it.
The federal government has made significant commitments to expanding healthcare access in underserved areas, particularly rural communities. Investments in rural health infrastructure, telehealth expansion, and modernized care delivery are meaningful steps that address real gaps. But healthcare access and research capacity are not the same thing. U.S. cancer registries, for instance, are quite complete at the diagnostic level because state law mandates reporting. The patients are recorded. The question is whether they can participate in precision medicine research for which their recorded diagnoses should make them eligible.
In most cases, they cannot. When community oncology providers in the rural Southeast were surveyed about why their patients do not enroll in clinical trials, 70 percent cited distance from the trial center, 55 percent cited lack of transportation, and 50 percent cited the absence of local trials. About half of community oncology centers cite limited staffing as a top barrier to running trials. Rural practices are far less likely to have the genomic testing resources, genetic counselors, or molecular tumor boards that precision trials require. A patient can appear in a cancer registry and still have no practical path into a clinical trial for the molecular subtype of disease they carry. The registry knows they exist, but the research system cannot reach them.
Cancer has the most mature research infrastructure of any disease in the United States, backed by decades of federal investment. If research capacity gaps are pronounced in oncology, the challenge facing other therapeutic areas as they enter the precision medicine era is considerably steeper. Cardiology, neurology, immunology, and rare diseases are all moving toward biomarker-driven treatments, but most lack anything close to oncology’s network of community research sites, established trial pipelines, or population-level surveillance. The infrastructure deficit that limits representativeness in cancer trials will only widen as precision medicine expands.
Why Knowing Better Hasn't Led To Doing Better
The case for investing in research infrastructure is not new, and it extends well beyond oncology. As precision medicine expands across therapeutic areas, each will face its own version of the same problem. Targeted therapies need to be tested in populations that reflect the patients who will use them, and the research capacity to do that does not exist in most communities. Policymakers, regulators, industry, and patient advocates broadly agree on this point. Federal investment in rural health and community-based care is expanding access in areas that have long been underserved. The regulatory landscape increasingly supports decentralized trial designs that can bring research participation closer to patients. And yet, investment in the link between healthcare delivery and clinical research remains inadequate. Why?
Several structural barriers help explain the gap. Clinical trial infrastructure functions as a shared resource. Once built, it benefits all sponsors, not just the one who paid for it. This creates a collective action problem where each stakeholder waits for someone else to move first. Current funding generally operates on a study-by-study basis, with sponsors paying for site-specific costs but not for a community's general research readiness. Responsibility is fragmented across agencies. The FDA sets trial requirements, NIH funds research, and CMS determines reimbursement, but none clearly owns the task of building community-level clinical trial capacity. Regulatory signals have been inconsistent, making long-term infrastructure commitments feel risky. And high turnover at existing sites, with trained research staff leaving for better-compensated positions at CROs and sponsors, makes sustained capacity-building difficult even where it has started.
This is a public goods problem. The benefits of clinical research infrastructure are shared across the ecosystem, but the costs fall on whoever builds it first. Agreeing that it matters has not been enough.
Lessons From Across The Life Sciences
What has actually driven investment in health policy is the design of incentive structures in which each party has a reason to participate that serves its own interests. The mechanisms used differed in each case, but they followed a common logic. The HITECH Act used phased financial incentives, payments first and penalties later, to drive hospital EHR adoption from under 10 percent to near-universal within a decade. That infrastructure, built for care coordination, became the foundation for RWE platforms and research networks now used in FDA regulatory submissions. The Orphan Drug Act combined tax credits with seven-year market exclusivity to make rare disease drug development economically viable, producing a roughly 69 percent increase in new clinical trials for established rare diseases. The Cystic Fibrosis Foundation used venture philanthropy, investing $150 million alongside its own patient registry and care network to reduce development risk for Vertex Pharmaceuticals, and later sold the resulting royalty rights for $3.3 billion. Operation Warp Speed used advance purchase commitments to guarantee demand before products existed, compressing a decade of vaccine development into months.
The tools were different. The logic was the same. Each one changed the expected return for the parties being asked to act. And in each case, the returns went well beyond what anyone planned for. EHR infrastructure now supports an entire health data economy. Community sites equipped for research achieved better care quality and patient retention.
Clinical research infrastructure for precision medicine sits in the same territory. The benefits are diffuse, the up-front costs are concentrated, and the externalities are high. The research points toward hybrid solutions that combine demand-side guarantees with governance structures that give each participant a credible reason to invest. No single mechanism will be sufficient, but the toolkit is established. The question is whether policymakers and industry leaders are willing to adopt it.
Getting It Right
Across the ecosystem, there is broad agreement that medical products should work for the patients who will use them, that clinical trials should produce scientifically sound and generalizable results, and that healthcare access, particularly in rural and underserved communities, is worth investing in.
Our research highlights that these goals are connected. Assessing whether a clinical trial represents its target population requires knowing who that population is and whether they can actually participate in research. Right now, for too many patients, the answer to the second question is no. Building research capacity in communities where it currently does not exist requires sustained, coordinated investment across the ecosystem.
The era of precision medicine makes this connection more urgent. As therapies become more targeted, the populations they serve become more specific, and the tolerance for measurement error shrinks. The scientific tools to build better benchmarks exist. The regulatory frameworks to support broader trial participation are in place. Federal investment in healthcare infrastructure is growing. What remains is the work of connecting these pieces deliberately through investment in research-capable community sites and data systems, through investment in the infrastructure that makes trials accessible beyond academic centers, and through a shared commitment to ensuring that the promise of precision medicine is built on evidence drawn from the patients it is meant to serve.
About The Author:
June Cha, Ph.D., MPH, works at the intersection of science, policy, and access to advance biomedical innovation and promote equitable healthcare. Over the past 20 years, she has held leadership roles across the pharmaceutical industry, the U.S. government, and global health organizations — from directing women's health strategy at HHS to advising the Gates Foundation on pharmacovigilance in Africa. Throughout her career, she has connected the dots across the medical product lifecycle — from patient-centered development and regulatory strategy to value assessment and healthcare access — building the conditions for innovation to flourish. She has also shaped prevention-first healthcare strategies, including community-integrated preventive service delivery models. June earned her Ph.D. in biochemistry from the University of Notre Dame and her MPH from New York University.