Getty / The Atlantic

The United States and many other nations are launching a life-threatening experiment. They are rapidly and perhaps prematurely easing restrictions on businesses and social activity—even as the novel coronavirus, SARS-CoV-2, is still prevalent and much of the population remains susceptible to the disease it causes, COVID-19. The understandable desire to restore normal life quickly has raced ahead of the scientific knowledge necessary to do so safely. The result could be mortality rates that exceed even those occurring now, in the first wave of the pandemic.

Since late last month, I have been meeting frequently online with a group of nine colleagues: David Baltimore, Mike Brown, Don Ganem, Peggy Hamburg, Richard Lifton, Marc Lipsitch, Dan Littman, Shirley Tilghman, and Bruce Walker. All are well known for their work in areas such as virology, immunology, genetics, and epidemiology. All have served in one or more leadership roles: as presidents of universities or other academic institutions, as heads of government agencies, as advisers to drug or biotechnology companies, or simply as pioneers and mentors in their field. All have sought solutions to the great medical problems of our time. None of us can recall a crisis as stark as COVID-19.

Everyone in our group agrees that most governments were unprepared for a pandemic and underestimated this one, and that health-care workers were imperiled by a shortage of protective gear and diagnostic-test kits. Without effective treatments to reduce frightening death rates or vaccines to protect against infection, societies have had to resort to simple measures to blunt the pandemic. Seven centuries ago, in The Decameron, Giovanni Boccaccio wrote about 10 fictional citizens of Florence who hunkered down in a forest to tell stories, while escaping a world largely shut down by a plague. Today’s distancing and sheltering are even more severe, but at least information technology is allowing people of varied expertise—including the 10 scientists in our group—to meet and think about how to put an end to the present dilemma.

After a very shaky start, the United States may finally have developed enough tests to diagnose most patients who come to hospitals for treatment of COVID-19 symptoms. Nevertheless, we believe that expanding current testing capacity remains a matter of extreme urgency—one that justifies a level of intense, coordinated work at a national, even international, scale that resembles the campaigns we associate with world wars. This means, at a minimum, marshaling the enormous physical and intellectual capacity of biomedical-research labs across the United States. The shortfall in testing isn’t just a problem for individual patients and their doctors. It is also holding back large-scale surveys of seemingly healthy populations, in workplaces and elsewhere, and scientific research into fundamental properties of the virus and the disease it causes. The ability of states and communities to reopen without risking calamity depends on a rapid acceleration of scientific discovery.

With the number of new cases declining in many places, the purposes and required scale of testing have shifted. While doctors in hospitals still need to make diagnoses of individual patients, there is an escalating need to test much larger groups repeatedly—to track the spread of the virus as restrictions ease—and to carry out population-based studies that will reveal more about how this virus behaves.

Testing for the virus is based on detection of the coronavirus’s unique RNA sequences in swabs of the nose or throat or in saliva. Most tests in the United States are currently performed in large reference laboratories using automated commercial instruments that can each process only 1,000 to 3,000 tests a day, yielding the current national total of about 150,000 tests a day. While that number could double over the next few months, the methods now in use are unlikely to yield the level of testing—in the range of 2 million tests or more a day—necessary to monitor large populations and conduct ambitious experiments.

Could alternative technologies meet these new testing demands? Considerable evidence indicates that they could. An especially compelling argument can be made for harnessing powerful DNA-sequencing machines, each of which could test tens or hundreds of thousands of samples a day for viral RNA. The needed sequencing machines are already in wide use—for medical research and detection of mutations in patients with cancers or other diseases—in numerous academic and commercial laboratories around the country. Several teams of investigators are devising and testing new protocols that could achieve the new aims.

In one proposal that our group considered to reach this kind of scale, samples from each person would be given a unique identifier—in essence, a bar code. Thousands of samples would be mixed together and sequenced, allowing the determination of which samples contained viral RNA and which did not. Scientists now have extensive experience in using this approach—for instance, to measure amounts of thousands of different cellular RNAs in millions of individual cells pooled in a single tube—building confidence that the method should work. Sequencing has not been used previously at this scale for clinical tests, primarily because its enormous capacity has not been needed, but the FDA has used the method to survey the food supply for contaminating pathogenic organisms.

The use of such sequencing methods has a notable shortcoming: the need to transport samples to places where the sequencing machines reside, creating expenses and delays. That approach might be acceptable for some of the studies we propose below, but it is less helpful in determining whether an individual is safe to enter a workplace or school on a given morning. Ideally, for the later purposes, tests would be conducted swiftly and at high volume at the places where samples are taken. Numerous ideas for such tests have been proposed, but none is yet validated to work at the scale required. One such approach, still in development, would exploit the ability of the well-known bacterial gene-editing system known as CRISPR to recognize coronavirus RNA.

While the need for greatly expanded testing in the next phase of this pandemic is widely acknowledged, the United States has no coordinated plan for how to achieve it. The technical building blocks are in hand, but how to put them together is not yet clear. Moreover, major regulatory hurdles limit the use of the results from novel tests in patient care, especially in certain states such as New York. And the logistics of deploying enough personnel to track samples and deliver results are daunting. Because of the complexity and importance of such testing, a centralized program, run by a strong scientific leader and paid for with federal dollars, may be the only solution.

Congress has recently provided $25 billion to be used for various facets of the testing conundrum. We propose that at least some of these funds—and much more if needed—should be placed under the direction of a single office housed within a large government agency, with the authority to synchronize the currently fragmented efforts to develop better tests; to offer financial incentives for biotechnology firms with promising ideas; to reduce and harmonize regulatory requirements; to implement efficient new testing methods throughout the country; and to help assemble the teams needed to collect and process these tests. This is an enormous undertaking. But we are at war with the invading coronavirus, and wars require no less.

Armed with efficient and accurate tests to detect the virus (indicating active infection) and reliable tests to measure antibodies against it (implying prior exposure and possible immunity), public-health programs could paint an accurate picture of the current pandemic. Small and large businesses, schools, health-care facilities, and other organizations could track the outcomes of their attempts to restore normal activities, and scientists could answer key questions about viral transmission and host immunity. Laboratory experiments in animal models and cell cultures are already providing useful information about some of these questions. But decisive answers will come only from studying human beings who are exposed to the virus under real-life conditions. Such studies may be feasible only under circumstances in which natural transmission is occurring at significant rates, as it currently is. Therefore, if we are to get answers to the following questions, we must act now.

Absent a vaccine, can we deploy enough precautions to protect people in the workplace?

An uncontrolled experiment is currently under way among health-care workers who have direct contact with a large number of COVID-19 patients and cannot always maintain physical distance from them or from one another. These workers use personal protective equipment—N95 masks, gowns, and gloves—to reduce transmission within hospital settings, and they depend on frequent monitoring of symptoms and temperature and testing for viral RNA to detect and confirm cases. These practices—and other evidence that physical distancing and even simple cloth masks are helpful—support the hypothesis that similar measures in a variety of other workplaces, combined with close monitoring for infection, will allow people in many other occupations to return to work safely.

This hypothesis could be challenged in various settings during the coming months to establish its validity. Weekly testing of all workers would allow for early detection of new cases. Combined with rigorous contact tracing and enhanced surveillance, researchers should be able to distinguish between new infections that occurred in the workplace and those that occurred elsewhere in the community.

Does naturally acquired infection generate immunity?

Knowing whether patients who recover develop long-lasting immunity against reinfection has far-reaching implications for allowing people to go back to work or school. It would also shed light on the prospects for a vaccine.

Studies to answer these questions require identifying enough people who have recovered, then testing them repeatedly for the appearance of a new infection. Such people are relatively easy to find. They include doctors and nurses in hospitals in hard-hit metropolitan areas such as New York City; staff and residents at nursing homes with high rates of infection; and crews of U.S. Navy ships that have experienced outbreaks of COVID-19.

Results from weekly tests for the virus in the health-care workers who continue to be exposed should provide answers when compared with results from their co-workers who have not previously been infected. When previously infected sailors are deployed to other vessels, they will be surrounded by many unaffected sailors. If another outbreak occurs on such a ship, weekly testing for the virus should reveal whether the attack rates differ between those who were infected earlier and those who were not.

If significant numbers of people prove to be susceptible to reinfection, it will be important to assess whether prior exposure at least makes the disease less severe the second time around. Such studies could also have a significant bearing on the design of a vaccine against the coronavirus because they would help scientists understand which kinds of immune response lead to full protection against reinfection.

What are the important features of asymptomatic infection?

Growing evidence indicates that a significant fraction of people who are infected by the coronavirus are currently showing no symptoms. Widespread and frequent testing for the virus, especially among those at especially high risk of infection (such as health-care workers and slaughterhouse employees) can identify asymptomatic infections. Following up on those cases will shed light on how many asymptomatic people ultimately develop symptoms; how long it takes for them to do so; whether asymptomatic people who ultimately develop symptoms have higher viral loads than those who don’t get sick; whether symptomatic and asymptomatic people have different immune responses; whether other, simpler procedures (such as tests for some chemical abnormality in the blood) might be used to screen for infection; and how large a contribution asymptomatic people make to the ongoing transmission of the virus.

Similarly, as schools reopen, it will be important to know the contribution that children (who rarely develop the symptoms of COVID-19) make to transmission of the virus, especially to contacts at home. Such studies could also be combined with studies of transmission in households with different numbers of occupants and different amounts of space per occupant. Such information would help health officials and the general public alike plan containment strategies wisely.

Most of the work that normally occurs in the world’s laboratories has been slowed or halted, for the same reasons that so many other regular activities in our society have been constrained. At the same time, a remarkable number of scientists, in biomedicine and other fields as well, have turned their attention to matters related to the pandemic, regardless of their former interests.

As a result, in just a couple of months literally thousands of papers have appeared in peer-reviewed journals or simply been posted online. Such rapidity poses risks to the reliability and rigor of scientific work. But the efforts have also produced an extensive picture of the manifold manifestations of COVID-19; accounts of the outcomes of the still highly imperfect methods for care and prevention; and hundreds of strategies to develop vaccines, test existing drugs, and seek new ones. And all of this has been built on top of a rich and long history of discoveries in many fields that have fostered our understanding of human viruses, allowed the development of sophisticated molecular tests, and provided digital tools to track the epidemiological patterns of disease. Despite so much past and recent activity, major questions about the pandemic remain unanswered, a situation we attribute largely to the complexity of the problems and the inherently slow nature of the scientific process.

Making matters worse, some of what science has taught has been ignored. Despite repeated warnings after prior epidemics about the likelihood of new ones caused by novel microbes, the United States and many other countries failed to respond efficiently to this one. Scientists might have detected the new coronavirus much earlier with the better tools for microbial surveillance that already exist; prevented the pathogen’s worldwide spread by more aggressive testing and contact tracing; and supported better and safer health care with larger stockpiles and pipelines for procurement of medical equipment. Humanity should never be this unprepared again.

Now, as nations around the globe consider when and how to reopen their societies, the need for more extensive testing—and for the scientific knowledge it will yield—is more urgent than ever. The world cannot return to normal unless science can deliver lifesaving information at sufficient speed.

We want to hear what you think about this article. Submit a letter to the editor or write to