A DARPA-like culture at NIH can drive biomedical and health advances

By: White House Office of Science and Technology Policy (OSTP) Director Dr. Eric Lander, National Institutes of Health Director (NIH) Dr. Francis Collins, OSTP Assistant Director for Biomedical Science Initiatives Dr. Tara Schwetz, and NIH Principal Deputy Director Dr. Lawrence A. Tabak

The biomedical research ecosystem has delivered advances that not long ago would have been inconceivable, exemplified by highly effective COVID-19 vaccines developed by global partners and approved in less than a year. The United States stands at a moment of unprecedented scientific promise and is challenged to ask: What more can we do to accelerate the pace of break-throughs to transform medicine and health? Toward that end, President Biden recently proposed to create a new entity, the Advanced Research Projects Agency for Health (ARPA-H), within the National Institutes of Health (NIH) “to develop breakthroughs—to prevent, detect, and treat diseases like Alzheimer’s, diabetes, and cancer,” requesting $6.5 billion in the fiscal year 2022 budget (1). The idea is inspired by the Defense Advanced Research Projects Agency (DARPA), which follows a flexible and nimble strategy, undeterred by the possibility of failure, and has driven breakthrough advances for the Department of Defense (DOD) for more than 60 years. To design ARPA-H, it is critical to understand what is working well within the biomedical ecosystem, where there are crucial gaps, and the key principles of DARPA’s success.


Progress in medicine and health in recent decades has been driven by two powerful forces: pathbreaking fundamental re- search and a vibrant commercial biotechnology sector. Fundamental research is typically performed in university, nonprofit, and government labs. In the United States, it is mostly funded by the federal government, largely through the NIH. By steadily pursuing important fundamental questions in biology and medicine, scientists have made great progress in discovering the molecular and cellular mechanisms underlying health and disease—often suggesting new ideas for clinical treatment. Such fundamental research is what economists term a public good, in that it produces knowledge available to everyone and thus requires public investment. Some have estimated that every dollar of federal investment yields at least $8 in economic growth, and suggested that every new therapeutic approved by the US Food and Drug Administration (FDA) can be traced, in part, to fundamental discoveries supported by NIH (2, 3). Given its outsized impact, robust federal investment in fundamental research remains crucial to health and to the economy.

The commercial sector is largely focused on research, development, and marketing of specific products, to bring sophisticated therapies and devices to patients. Biotechnology companies have access to abundant capital to develop products—provided they can protect their intellectual property and recoup the costs by generating sufficient profit in a short enough period of time. Currently, more than 8000 medicines are in development, including 1300 for cancer (4, 5).

In many cases, these two components are all that’s needed to drive progress toward clinical benefit—though subsequent regulatory approvals, reimbursement, and adoption in health care systems can also be optimized. It’s becoming clear, though, that some of the most innovative project ideas, which could yield breakthroughs, don’t always fit existing support mechanisms: NIH support for science traditionally favors incremental, hypothesis-driven research, whereas business plans require an expected return on investment in a reasonable time frame that is sufficient to attract investors. As a result, some of the most promising ideas may never mature, representing substantial lost opportunity.

Bold ideas may not fit existing mechanisms because (i) the risk is too high; (ii) the cost is too large; (iii) the time frame is too long; (iv) the focus is too applied for academia; (v) there is a need for complex coordination among multiple parties; (vi) the near-term market opportunity is too small to justify commercial investment, given the expected market size or challenges in adoption by the health care system; or (vii) the scope is so broad that no company can realize the full economic benefit, resulting in underinvestment relative to the potential impact. Evaluations by companies also may not consider the impact of projects on inequities that persist in our health ecosystem. In short, projects with a potentially transformative impact on the ecosystem may not yet be economically compelling or sufficiently feasible for a company to move forward. At the same time, there are no public mechanisms to propel these public goods at rapid speed.

Many such bold ideas involve creating platforms, capabilities, and resources that could be applicable across many diseases. Whereas most NIH proposals are “curiosity-driven,” these ideas are largely “use-driven” research—that is, research directed at solving a practical problem.


DARPA was launched in the wake of Sputnik with a singular mission: to make pivotal investments in breakthrough technologies for national security. DARPA has played a key role in generating bold advances that have shaped the world—such as the internet, Global Positioning Systems, and self-driving cars—and has contributed to the development of many others, including messenger RNA vaccines. However, failure, especially failing early, and learning from that failure are also hallmarks of DARPA.

DARPA has a distinctive organization and culture that contrasts with traditional approaches in biomedical research. It is a flat and nimble organization whose work is driven by approximately 100 program managers (PMs) and office directors. The PMs are often recruited from industry or top research universities, and they come for limited terms of 3 to 5 years. They typically bring bold, risky ideas, and they are given the independence and sufficient resources to pursue them, mitigating risk through metric-driven accountability and by pursuing multiple approaches to achieve a quantifiable goal.

DARPA can support research at three stages (basic research, applied research, and advanced technology development); can fund efforts in multiple sectors (industry, university, national labs, and consortia across these sectors); can provide the critical mass of funding needed to tackle bold goals; and is empowered to pro mote collaboration and integration across performers. DARPA does not perform its own internal research. Although proposals are reviewed on a competitive basis, PMs have authority to select a portfolio of projects intended to achieve a particular program goal.

DARPA has long encouraged a culture that values a relentless drive for transformative technical results and a willingness to take risks. Notably, it does not focus on merely accelerating ordinary products to the market or making incremental progress, but on creating true breakthroughs. To act in this way, DARPA makes broad use of flexible hiring, procurement, and contracting authorities, provided by law.

Although DARPA is an excellent inspiration for ARPA-H, it is not a perfect model for biomedical and health research. It serves the needs of a single customer, the DOD, and its mission is focused on national security. Its projects typically involve engineered systems. By contrast, health breakthroughs (i) interact with biological systems that are much more complex and more poorly understood than engineered systems, requiring close coupling to a vast body of biomedical knowledge and experience; (ii) interact with a complex world of many customers and users—including patients, hospitals, physicians, biopharma companies, and payers; (iii) interact in complex ways with human behavior and social factors; and (iv) require navigating a complex regulatory landscape. ARPA-H can learn from DARPA, but will need to pioneer new approaches.


NIH has some experience with running large, complex programs using DARPA-like approaches to drive highly managed, use-inspired, breakthrough research. A classic example was the Human Genome Project, aimed at reading out the complete 3 billion–nucleotide human genetic code. When the project began in 1990, the technology to accomplish the goal hadn’t been invented. By driving innovation, it was completed ahead of schedule and ultimately decreased the cost of sequencing a human genome from $3 billion at the outset to $500 today (6). Though estimates vary, it is clear that the overall economic return on investment has been enormous, with notable analyses estimating a nearly 180- fold return (7, 8).

A very recent example is the NIH’s response to the COVID-19 pandemic. Within weeks, NIH created two programs. The Accelerating COVID-19 Therapeutic Interventions and Vaccines (ACTIV) program is an unprecedented partnership with government, industry, nonprofits, and academia to drive preclinical and clinical therapeutics, developing master protocols for testing prioritized compounds in rigorous randomized clinical trials. These efforts accelerated the development and testing of several of the vaccines that are now being widely used. The Rapid Acceleration of Diagnostics (RADx) program used an “innovation funnel” approach to identify promising ideas for COVID-19 tests and support 32 new technology platforms that collectively are contributing 2 million tests per day, mostly at point of care (9).

Although these programs have been successful, they required bespoke solutions and herculean efforts to get them off the ground. Because NIH lacks a regular framework for such projects, many bold ideas are hard to realize. That’s where ARPA-H can help.


ARPA-H should have a clear mission. Building on DARPA’s mission statement, an initial mission could be: “To make pivotal investments in breakthrough technologies and broadly applicable plat- forms, capabilities, resources, and solutions that have the potential to transform important areas of medicine and health for the benefit of all patients and that cannot readily be accomplished through traditional research or commercial activity.”

Notably, ARPA-H’s focus should be broad—ranging from molecular to societal—because breakthrough technologies are needed and are possible at many levels (see the box). When Pres ident Biden challenges researchers to “end cancer as we know it,” many basic scientists naturally think about solutions at the laboratory bench: powerful ways to enlist DNA and RNA readouts, genetic regulation, novel chemistry, and the immune system to prevent, detect, and treat cancers. Technologists think about new sensors and artificial intelligence-–assisted medical decision-making. As importantly, though, there are also opportunities for highly impactful breakthroughs at the macro level to ensure equity in health care access and health outcomes for all patients. Equity considerations (including race, ethnicity, gender/gender identity, sexual orientation, disability, and income level) must be woven throughout the ARPA-H mission—with some projects focused directly on addressing equity and all projects considering equity in their design. Breakthroughs aimed at the most vulnerable groups are not only just and necessary, they will likely improve care for all patients.

ARPA-H’s mission will clearly be different from the mission of the existing NIH Institute and Centers (ICs). For example, the name and mission of the National Center for Advancing Translational Sciences (NCATS), an NIH institute created in 2011, might suggest some overlap. However, NCATS’ primary focus is to support a national network of clinical research centers and a drug screening hub. These two programs account for nearly 90% of its resources. A modestly sized component within NCATS, the Cures Acceleration Network, is aligned with the general directions of ARPA-H.

Similarly, the NIH Common Fund, a program created by law in 2007, is aimed at a different goal than ARPA-H’s use-driven objective: It supports programs to explore new areas of foundational research that cut across multiple ICs—for example, the human microbiome effort. ARPA-H would also be distinct from other existing agencies, such as the Biomedical Advanced Research and Development Authority (BARDA), which focuses on medical coun termeasures for public health security threats.


ARPA-H should be housed as a division within NIH, rather than being a stand-alone entity, for two reasons. First, the goals of ARPA-H fall squarely within NIH’s mission (“to seek fundamental knowledge about the nature and behavior of living systems and the application of that knowledge to enhance health, lengthen life, and reduce illness and disability”). Second, ARPA-H will need to draw on the vast range of biomedical and health knowledge, expertise, and activities at NIH. Setting up ARPA-H within NIH will ensure scientific collaboration and productivity and avoid unproductive duplication of scientific and administrative effort.

It is important to acknowledge, however, that a DARPA-like approach is radically different from NIH’s standard mechanisms of operation and will require a new way of thinking. The creation of ARPA-H will benefit from transparency, accountability, and a healthy skepticism to ensure that the entity does not become a typical NIH institute.

Taking many features from the DARPA model, ARPA-H needs to have a distinctive culture, organization, authorities, leadership, and autonomy (10, 11). ARPA-H’s organization should be flat, lean, and nimble. The culture should value bold goals with big potential impact over incremental progress. The organization should lure a diverse cohort of extraordinary PMs from industry or leading universities, for limited terms, with the chance to make a huge impact. They should be empowered to take risks, assemble port- folios of projects, make connections across organizations, help clear roadblocks, establish aggressive milestones, monitor progress closely, and take responsibility for the project’s progress and outcomes. Projects should be bounded in time, typically a few years, with longer periods allowed for efforts that are highly complex. ARPA-H should expect that a sizable fraction of its efforts will fail; if not, the organization is being too risk-averse. The best approach is to fail early in the process, by addressing key risks upfront. To determine which risks should be taken and to evaluate proposed programs and projects, ARPA-H should adopt an approach similar to DARPA’s “Heilmeier Catechism,” a set of principles that assesses the challenge, approach, relevance, risk, duration, and metrics of success (12).

The ARPA-H director should have substantial authority and in- dependence to act. To keep the entity vibrant, the director should typically serve a single term of 5 years, with the possibility of a single extension in rare cases. For ARPA-H to accomplish its goals, it will need to be provided by Congress with certain authorities parallel to those provided to DARPA, including the authority to recruit, attract with competitive pay, and quickly hire for a set term extraordinary PMs.

Unlike DARPA’s focus on a single customer, ARPA-H will need to create breakthrough innovations that serve an entire ecosystem and all populations. ARPA-H should have a senior leader responsible for ensuring that issues of equity are considered in all aspects of ARPA-H’s work—from scientific program development to staff recruitment and hiring.

Within the Department of Health and Human Services, it will be important for ARPA-H to collaborate with other key agencies such as the FDA, the Centers for Disease Control and Prevention, BARDA, and the Centers for Medicare and Medicaid Services—to identify critical needs and opportunities and to partner on complex projects that interact, for example, with public health infrastructure or medical regulation.

DARPA should also play a role in advising ARPA-H on its experiences in driving breakthrough innovation and collaborating on specific projects of shared interest. And, it would be valuable to engage science-based agencies and departments, such as the National Science Foundation, the National Institute of Standards and Technology, and the Department of Energy.

It will be critical for ARPA-H to engage with the broader biomedical community, including patients and their care-givers, researchers, industry, and others, to understand the full range of problems and the practical considerations that need to be ad- dressed for all groups and populations.

The potential opportunity is extraordinary. Through bold, ambitious ideas and approaches, ARPA-H can help shape the future of health and medicine by transforming the seemingly impossible into reality. The time to do this is now.

References and Notes

  1. Remarks by President Biden in Address to a Joint Session of Congress (2021), www.whitehouse.gov/briefing-room/speeches- remarks/2021/04/29/remarks-by-president-biden-in-address-to-a-joint- session-of-congress/.
  2. A. A. Toole, Does Public Scientific Research Complement Private Investment in Research and Development in the Pharmaceutical Industry? J. Law Econ. 50, 81–104 (2007). doi:10.1086/508314
  3. E. Galkina Cleary, J. M. Beierlein, N. S. Khanuja, L. M. McNamee, F. D. Ledley, Contribution of NIH funding to new drug approvals 2010-2016. Proc. Natl. Acad. Sci. U.S.A. 115, 2329–2334 (2018). doi:10.1073/pnas.1715368115 Medline
  4. G. Long, “The Biopharmaceutical Pipeline: Innovative Therapies in Clinical Development” (The Pharmaceutical Research and Manufacturers of America, 2017).
  5. “Medicines in Development for Cancer 2020 Report” (The Pharmaceutical Research and Manufacturers of America, 2020).
  6. DNA Sequencing Costs: Data (2020), www.genome.gov/about-genomics/fact- sheets/DNA-Sequencing-Costs-Data.
  7. S. Tripp, M. Grueber, “The Economic Impact and Functional Applications of Human Genetics and Genomics” (American Society of Human Genetics, 2021).
  8. “The Impact of Genomics on the U.S. Economy” (Batelle Technology Partnership Practice, for United for Medical Research 2013).
  9. “RADx diversifies COVID-19 test portfolio with four new contracts, including one to detect variants” (2021), www.nibib.nih.gov/news-events/newsroom/radx- diversifies-covid-19-test-portfolio-four-new-contracts-including-one-detect-variants.
  10. A. Prabhakar, “How to Unlock the Potential of the Advanced Research Projects Agency Model” (Day One Project 2021).
  11. R. E. Dugan, K. J. Gabriel, in Harvard Business Review (Harvard Business Publishing, 2013).
  12. “The Heilmeier Catechism” (2021), www.darpa.mil/work-with-us/heilmeier- catechism.


The authors thank R. Fleurence and A. Hallett for their helpful input in preparation of the manuscript.

Published online 22 June 2021 10.1126/science.abj8547

Stay Connected

Sign Up

We'll be in touch with the latest information on how President Biden and his administration are working for the American people, as well as ways you can get involved and help our country build back better.

Opt in to send and receive text messages from President Biden.

Scroll to Top Scroll to Top