Skip to Content

Here’s what we have to do to show a coronavirus vaccine works

Everyone is trying to avoid covid-19. But finding clear proof for a vaccine will mean hoping some people get it.
Ms Tech | Getty, Pixabay

The moonshot program to come up with a vaccine against covid-19 is advancing faster than anyone could have hoped. At least four experimental vaccines have been shown to protect monkeys, and three of those are already being given to brave human volunteers.

The aim is a vaccine by January, and money is no object. On May 21, the US said it would throw $1.2 billion behind a vaccine from Oxford University and AstraZeneca, part of what President Donald Trump called “Operation Warp Speed.”

It’s all good news, and many scientists believe a vaccine is likely. But the next hurdle is the biggest one: proving that a vaccine candidate actually works.

By early summer, expect to see researchers try to launch several huge efficacy studies involving thousands of volunteers. This is going to be the most costly part of testing a vaccine, and also the hardest to speed up. That’s because researchers will have to wait for people in the study to get accidentally infected with the virus, and then check how often those given the vaccine get sick.

Trump, Fauci and Azar
President Trump along with Anthony Fauci, director of the National Institute of Allergies and Infectious Diseases and Secretary of Health and Human Services Alex Azar in March after returning from the NIH's vaccine research center.

“It’s invariably the most expensive part of vaccine development and the longest of the three phases,” says Stanley Plotkin, former head of vaccines at Sanofi Pasteur. More than anything, he says, the time frame “depends on the incidence of the disease.”

The global effort to create a vaccine on such short notice is unprecedented. New technology has led drug makers to move quickly, and regulators have never pushed paper so fast. But with covid-19 getting beaten back in cities like New York, where new cases have fallen from more than 6,000 a day to less than 600, successful efforts to suppress the disease could perversely make it harder to test a vaccine.

Vaccine makers say it’s a concern. Speaking to Wall Street analysts on a conference call this week, Tal Zaks, chief medical officer at Moderna, whose mRNA vaccine was first to enter human studies in March, said: “The challenge is, how do I ensure I have enough cases? If I go and vaccinate a lot of people, it doesn’t matter how many if there is no circulating virus.”

The irony is that the process will go faster if the covid-19 outbreak keeps flaring up. Vaccine researchers are also expected to pick nurses, doctors, and other at-risk groups to join their studies, so there’s a bigger chance of subjects catching it. They will advise people to stay safe, even while hoping some get sick.

“You still have to tell people to try not to get infected. You don’t say ‘Stop wearing a mask, or ‘Why don’t you meet people in a closed space.’ How is that for a weird dilemma?” says Arthur Caplan, a bioethicist at NYU Langone Health in New York. “The world is trying to get this under control, which I admire, but it does undermine the ability to study a vaccine.”

Trial design

While companies are the ones to decide what a trial looks like, in the US they need approval from the Food and Drug Administration, which has already said it expects to see randomized double-blind trials, the gold standard for proving that a treatment really works.

That means in the big studies starting this summer, some people will get the real vaccine and others will likely get a placebo shot. Then researchers will look to see how many in each group get infected or develop covid-19.

According to Clinton Hermes, an attorney at Bass, Berry & Sims and an outside counsel to the Coalition for Epidemic Preparedness Innovation, a “placebo arm” is necessary to keep both scientists and subjects in the dark. “Knowing who got the vaccine can bias the researcher on a subconscious level in the way they collect data, and bias the behavior of the subject too,” says Hermes. “What you find is that if people are told they were given a vaccine, they are more likely to expose themselves to a virus than if they are uncertain. It’s just human nature to engage in riskier behavior if they think they have been vaccinated.”

Going faster

To run their studies, vaccine companies are likely to look for volunteers among higher-risk populations in big, dense cities, according to Cynthia Dukes, vice president for drug development services at Icon Clinical Research, which is coordinating several covid vaccine and drug trials.

Most people to be vaccinated will likely be first responders, health care workers, or members of the National Guard, she says. “It’s a good population to go into and get an answer. We do want them to have a risk of exposure. If they are sitting at home, it’s not going to work.”

Dukes says she’s seen draft plans for vaccine trials involving 6,000 to 10,000 volunteers, in which researchers estimated that as many as 3.7% of volunteers would be exposed to the virus. That means they’re expecting about 100 to 150 infections in the placebo arm and fewer, or ideally none, among those who get vaccinated. That would be statistical proof the shot works.

During the trials, people will regularly be checked to see if they are infected as well. “In a clinical study, at the first sign of a sniffle you get tested,” says Dukes.

According to Dukes, it is possible to speed up the studies by filling them more quickly. She expects that won’t be a problem; volunteers have been calling in by the thousands. Another way to get them done faster is simply to make them larger. That is what the Department of Health and Human Services signaled it planned to do with the AstraZeneca trial, which it said would involve 30,000 people in the US.

Like a horse race, Operation Warp Speed could end up backing trials of several vaccine contenders. That means the total number of volunteers needed could soar to more than 150,000, Reuters reported. But it's far from clear there will be enough covid-19 around. Adrian Hill, the Oxford University scientist behind the AstraZeneca shot said he thought there was a 50% chance a trial would lead to "no result at all."

"It's a race against the virus disappearing," Hill told The Telegraph.

The British pharmaceutical company pledged to have 300 million doses ready by October, but some are skeptical that large trials will have delivered answers that soon. “Experts keep saying we’ll have a vaccine in the fall, but we won’t even have any data by then,” says Caplan, who adds that it would be “ethically impossible” to distribute a vaccine before there’s proof it works safely. “People wouldn’t take it. Also, you might be worried about an adverse event that is one in 25,000,” he says. “That’s rare, but if you are vaccinating a billion people, then relatively rare events are going to be more common.”

Anthony Fauci, head of the US National Institutes of Allergy and Infectious Diseases, projected that it will take at least four or five months for a big trial to generate enough evidence that a vaccine works. “If we are successful, we hope to know that in the late fall and early winter,” he told a Senate hearing in May.

Hermes, the vaccine lawyer, agrees that “you won’t see anything deployed in general population without proof of efficacy,” but he thinks governments will be likely to permit emergency use of a vaccine in frontline workers even before all the data is in—perhaps within a few months. “If a vaccine has a good safety profile and might possibly work, you could see people on the front line getting it,” he says.

Immunity correlates

As clues roll in, the whole world is hanging on any hint a vaccine will end the covid crisis. On May 18, Moderna Pharmaceuticals sent the entire US stock market up nearly 4% after it said volunteers who got its vaccine in a safety trial made antibodies at levels similar to convalescent covid-19 patients, a hint that its vaccine might be effective.

In some volunteers, the antibodies were “neutralizing,” meaning they countered the virus when their plasma was tested in a petri dish.

The company was quickly attacked for doing “science by press release,” but Fauci, whose agency is testing the Moderna vaccine, says there is good reason to believe it could work. “It’s definitely not a long shot,” he told Senator Mitt Romney during the Senate hearing. “This is a virus that induces an immune response, and people recover … the very fact that the body is capable of spontaneously clearing the virus tells me that at least from a conceptual standpoint, we can stimulate the body with a vaccine that would induce a similar response.”

That’s a very different situation from HIV—where (with rare exceptions) people never naturally defeat the virus, something that may explain why vaccines for AIDS have flopped time and again. With covid-19, by contrast, about 99 out of 100 people survive and appear to eventually cleared the virus.

The early data from the Moderna trial is an example of what vaccine makers call “correlates of immunity.” A correlate is something they can measure, like antibody levels, and which they can also cause to increase to a certain level with a vaccine. What they need to prove now, in the efficacy trial, is whether or not these factors can predict—or actually cause—real protection against the virus.

“There are lots of kinds of immune responses. So which of these are you creating, and which is responsible for protection, if there is protection? That is the question,” says Plotkin. “Neutralizing antibodies are not always sufficient, but it’s a good guess. For many diseases it’s the key to protection, but not always.”

What could go wrong

What’s extraordinary now is that covid-19 vaccines are advancing to big tests even as the parallel scientific effort to understand our immune response to the coronavirus remains in its early stages. Zaks, the chief medical officer of Moderna, compared the process to "flying a plane" while building it.

And some voices are rising to warn that Operation Warp Speed could easily crash on takeoff. “Normally, it takes up to 10 years to make a vaccine,” Rick Bright, the former head of the US agency responsible for the vaccine push, told lawmakers on May 14. “A lot of optimism is swirling around a 12-to-18-month time frame, if everything goes perfectly. We’ve never seen everything go perfectly.”

So what can go wrong? Dukes, who runs studies on behalf of companies, says that in her experience it’s usually manufacturing. If there's a problem scaling-up supplies, the process can face a big setback, since everything from animal studies to human studies is supposed to be based on an identical product.

For his part, Plotkin says the usual downfall of a vaccine is that researchers pick the wrong part of the virus to include in it or fail to inject enough of the substance. Other times it’s the biology of the virus that’s the problem—as in the case of HIV, a shape-shifter that attacks the immune system.

It’s also well known that a perfectly effective vaccine can fail to make a difference for a much more mundane reason: people don’t get vaccinated.

That happens more often than you think. Many people don’t bother with flu vaccines, anti-vaccine movements are sowing fear, and there are regions of the world where shots for preventable disease never get delivered.

With covid-19, all those problems remain on the horizon. There’s also the fact that it’s going to be hard to make enough vaccine for everyone. “In the context of a pandemic, we expect demand to far outstrip supply,” Zaks said of his company’s vaccine. During the call with stock analysts, his boss, Stéphane Bancel, the company’s CEO, agreed: “Trust me, it was not part of our business plan to have a billion doses.”

Deep Dive

Biotechnology and health

Scientists are finding signals of long covid in blood. They could lead to new treatments.

Faults in a certain part of the immune system might be at the root of some long covid cases, new research suggests.

Gene editing had a banner year in 2023

This year, gene editing finally started living up to its potential

The first gene-editing treatment: 10 Breakthrough Technologies 2024

Sickle-cell disease is the first illness to be beaten by CRISPR, but the new treatment comes with an expected price tag of $2 to $3 million.

This baby with a head camera helped teach an AI how kids learn language

A neural network trained on the experiences of a single young child managed to learn one of the core components of language: how to match words to the objects they represent.

Stay connected

Illustration by Rose Wong

Get the latest updates from
MIT Technology Review

Discover special offers, top stories, upcoming events, and more.

Thank you for submitting your email!

Explore more newsletters

It looks like something went wrong.

We’re having trouble saving your preferences. Try refreshing this page and updating them one more time. If you continue to get this message, reach out to us at with a list of newsletters you’d like to receive.