Early disease fighters, such as Edward Jenner, Louis Pasteur, and William Farr, suspected if enough people were vaccinated, it could eradicate a disease. At the dawn of the 20th century, veterinarians more interested in livestock than people seized on the idea and coined the term “herd immunity.” By the 1920s, clever studies with hundreds of thousands of mice vaulted the idea into the mainstream, stirring optimism that making a fraction of a population immune could forestall a devastating outbreak.
But even the trailblazers researching herd immunity were mystified by how to deploy it in practice. This conundrum has featured in battles against many modern plagues—such as smallpox, polio, and measles. And now it is part of the debate as the COVID-19 pandemic continues to flourish around much of the world.
Some prominent leaders wonder if herd immunity created as people are naturally infected with SARS-CoV-2 coronavirus would be enough to restore society to working order. For evidence, they point to hard-hit epicenters such as New York City, where approximately 20 percent of the residents have been infected and the caseload has been low and steady for months. This sustained recovery must be due to herd protection, they argue.
But based on simple math, past experiences with outbreaks, and emerging evidence from the ongoing pandemic, this claim is a fantasy.
“If we had reached sufficient herd immunity in New York, you would expect incidents to continue going down, not to be holding steady,” says Virginia Pitzer, an epidemiologist at the Yale School of Public Health who specializes in the mathematical modeling of how diseases spread.
A vaccine with 50-percent efficacy could spare hundreds of thousands from hospitalizations, chronic health issues, and death—but it couldn’t hit the herd immunity threshold on its own even if everyone took it. The safest bet is a vaccine above 75 percent efficacy.
% of population has been infected
A vaccine 50% effective
% has been infected
A vaccine 80% effective
% has been infected
NOTE: Each SEIR scenario includes 1,000 dots. Each dot equals 1,000 people. The model assumes an infection-fatality ratio of 0.75% and a basic reproduction number of 3.0. It also assumes that immunity develops after infection, so does not allow for reinfection, and that a vaccine is the sole intervention among the masses.
The reality is that most of the world—including 90 percent of the United States—remains susceptible to infection by the coronavirus virus, despite the global toll so far. Banking on natural infection to control the outbreak would lead to months, if not years, of a dismaying cycle in which cases subside and then surge. Even if such community-mediated protection were established, it would be constantly eroded by the birth of children and the real possibility that immunity in those previously infected would wane.
Only two infectious diseases have ever been eradicated: the human scourge of smallpox and the cattle-borne germ rinderpest. All other known afflictions—including such Old World pestilences as rabies, leprosy, and bubonic plague—have either been managed through human intervention or remain uncontrolled.
“It's very unlikely that we're going to see elimination of COVID-19 altogether from the population simply through the buildup of natural immunity,” says Pitzer. But if we add a highly effective vaccine on top of that, Pitzer says, “then it is theoretically possible that we could eliminate the virus” or at least control it.
A 237-page report from the National Academy of Medicine, published October 2, lays out how to distribute such a vaccine in an equitable manner—while also showing how hard this process will be. A crucial step will be communicating how good the vaccine needs to be to stop transmission. While major health agencies, including the U.S. Food and Drug Administration and the World Health Organization, say a COVID-19 vaccine should be at least 50-percent effective to be approved, this benchmark would actually be too low to establish protective herd immunity.
“It doesn't mean that a vaccine that's below this certain threshold will not be useful,” says Bruce Y. Lee, professor and executive director of Public Health Computational and Operations Research (PHICOR) at the City University of New York School of Public Health. “But if you want to be in a situation where you don't have to do social distancing and these other things anymore, then the vaccine really needs to be over 80 percent efficacy.”
What we mean when we talk about herd immunity
Herd immunity’s prominence in fighting epidemics can trace its origins to the 1920s and the University of Manchester in England. Inside a lab there, about 15,000 mice per year scurried through what looked like moon bases in miniature. Intricate residential pods—each about a foot wide—were connected by cylindrical tunnels, allowing the rodents to move freely around the Lilliputian cities.
But occasionally, the mouse cities would experience epidemics—ones started intentionally by the project’s leaders, William Whiteman Carlton Topley and Graham Selby Wilson. Members of one city would be exposed to lethal bacteria, while those in a separate city would receive doses of a vaccine along with the dangerous germ. The duo’s findings—published in 1923—demonstrated that immunity in a portion of a population could slow an outbreak and protect otherwise susceptible individuals.
“They called it experimental epidemiology,” says Paul Fine, a professor of communicable disease epidemiology at the London School of Hygiene & Tropical Medicine, who has written extensively about the origins of herd immunity. Topley and Wilson—along with some help from their contemporaries—helped popularize the idea, particularly through a textbook that’s still used by students to this day.
Yet when most people discuss herd immunity today, they’re really talking about what’s known as the “herd threshold theorem.” It’s what scientists are referencing when they say 75 percent of the population needs to be immune against COVID-19 to stop disease transmission, and it’s surprisingly simple to calculate.
Say a germ lands in foreign world, where an entire population is susceptible. And say it becomes clear that one infected person will transmit it to four others on average—a value known as the germ’s basic reproduction number, represented by an R with a subscript zero and thus called R-naught. To flatten the outbreak’s growth, you want a situation where the afflicted can infect just one person out of four.
“Well, that would be a circumstance where three out of the four were immune. He sneezed in four faces, but three of those individuals were immune,” Fine says. Three out of four is three-quarters, meaning a 75-percent threshold is needed to reach herd immunity.
Different viruses have their own reproduction numbers, so each has its own herd immunity threshold. Try the math again for measles, where one case can infect 18 susceptible people, and you get 94 percent. Polio has an R-naught of seven, so its threshold is 85 percent. These percentages serve as the targets for mass vaccination. Achieve them, and enough people in your community will be protected so that an outsider carrying the germ won’t be able to trigger a sustained outbreak.
While the underpinnings for the threshold theorem arose in the early 20th century, British epidemiologist George Macdonald was the first to incorporate the reproduction number, while studying malaria in Africa in the 1950s. It would be on this continent that a blind spot caused by strictly adhering to the concept would soon be discovered.
Why mass vaccination alone couldn’t beat smallpox
As a 16-year-old volunteer firefighter with the U.S. Forest Service, William Foege learned a key principle that would ultimately save millions of people from the scourge of smallpox: “Separate the fuel from the flames, and the fire stops,” Foege writes in his memoir House on Fire.
This mantra stuck with Foege after he joined the agency now known as the U.S. Centers for Disease Control and Prevention in 1962, and he was eventually stationed in Nigeria as an Epidemic Intelligence Service officer.
Three years earlier, the United Nations, World Health Assembly, and the WHO had launched a global eradication campaign against smallpox. The mass vaccination program quickly squelched the disease in Europe and North America, but nearly a decade later, the disease remained endemic in much of Africa, Asia, and South America, with tens of thousands of cases still reported each year. The virus kept finding hideouts—both in rural areas and high-density cities where it could fester—and ultimately threaten disease-free areas given that the vaccine’s immunity only lasted five years.
The tide turned on December 4, 1966, when a missionary in the southeastern Nigerian region of Ogoja radioed Foege to warn of a new possible outbreak. Trekking 90 miles by motorbike, Foege and his smallpox unit confirmed four cases in one village—but immediately faced a dilemma. Standard protocol called for vaccinating everyone in all the villages within a certain radius, but the team didn’t have enough doses. They would need to improvise.
“If we were smallpox viruses bent on immortality, what would we do to extend our family tree?” Foege writes. “The answer of course was to find the nearest susceptible person in which to continue reproduction.”
They opted to track down and vaccinate the individuals most likely to come in contact with the known cases. Dubbed “ring vaccination” or “surveillance-containment,” this strategy helped clear the final strongholds of smallpox over the next eight years.
It did it by addressing a wrinkle in the herd threshold theorem. That basic equation assumes everyone in a population is equally in contact with one another and spews an infectious virus in the same way.
“The real world violates these assumptions,” says Jeffrey Shaman, an epidemiologist at Columbia University's Mailman School of Public Health. Just look at COVID-19. Young adults drive the bulk of the spread in part because they come into contact with more people. (Millennials and Gen Z are spreading coronavirus—but not because of parties and bars.)
This uneven risk of infection—or heterogeneity—creates hot and cool spots of viral spread. If a public health team can cut off the heavy transmitters, they can control an outbreak with fewer doses of a vaccine. That’s a huge advantage—especially when an epidemic nears elimination and mass vaccination becomes less cost effective.
By 1971, an epidemiologist named John Fox began formulating herd immunity models that would better incorporate heterogeneity, and decades later it is still standard practice for public health researchers. The practice is similar to how firefighters clear trees, shrubs, and other flammable debris to encircle a raging wildfire, and it explains why health care workers, first responders, and people in hot spots such as jails will likely be first to receive an approved COVID-19 vaccine.
“By removing the fuel one step ahead of the virus, we had built a fire line,” writes Foege, who went on to serve as CDC director in 1977, the same year smallpox was eradicated from Africa. He is now the co-chair of the panel behind the National Academies report and a distinguished professor emeritus of international health at Emory University in Atlanta.
“The philosophy of science is to break down the walls of ignorance,” Foege said at a October 2 news conference that unveiled the report. “The philosophy behind medicine is to use that truth for every individual patient, but the philosophy behind public health is to use that truth for everyone.”
But his revelation about fire lines also means fewer people overall need to become immune to tamp down on transmission—relative to what’s predicted by the theorem threshold and mass vaccination goals. Today, this idea has inadvertently propelled a misconception that a lower threshold can be achieved through natural infection to safely thwart COVID-19.
Our future with COVID-19 depends on us
On August 14, Tom Britton, a mathematician at Stockholm University in Sweden, and two other scientists released a model in Science that estimates how social activity might influence the herd immunity threshold. They started with the valid assumption that millennials and Gen Z mix more than older people, and so will more readily spread the virus. Britton’s team landed on a herd threshold of 43 percent—much lower than the 60 to 75 percent you get using the classic equation.
“We don't claim that the number from our model applies in reality,” Britton cautions, adding that the model merely shows the degree to which disease-induced immunity can play a role. “We don't want our paper to have the consequence that people feel relaxed and say, Let's skip restrictions and wait for herd immunity.”
Another limitation of heterogeneity modeling, Columbia University’s Shaman says, is that no one really knows how germs spread among people on the street, so it’s difficult to tell what these reduced thresholds mean for real life.
“[Heterogeneity] is also constantly changing through time because of the measures we put in place. The telecommuting, the closing of schools, the wearing of masks are disrupting all the normal interactions that the virus feeds off,” Shaman says. “That completely changes the landscape.”
Moreover, recent studies of explosive COVID-19 outbreaks in two different regions suggest the classic herd theorem might be valid. In Qatar, the herd immunity threshold appears to have been achieved in about 10 working-class communities.
“So 60 percent of the population of Qatar is migrant workers. Almost all men and South Asian,” says Shaman. “They live in dormitory-style housing. They eat in cafeteria-style settings. They're just about as homogenized, in the sense of their interactions, as you could possibly get.”
In July, researchers began surveying these populations for antibodies, a sign of past infection. They found that 60 to 70 percent of these craft and manual workers—who tend to be young adults—had caught COVID-19 and become immune. Cases in the country have remained low even though officials reopened its borders this summer.
A separate study reported that the Brazilian city of Manaus reached the threshold and dampened its outbreak this summer after coronavirus infected 44 to 66 percent of its population. But a fresh bout of cases raises questions about whether the city truly achieved community protection—or worse, if immunity against the coronavirus wanes.
If the latter, the virus will bounce back even if places reach the herd immunity threshold through natural routes. This vulnerability would be reinforced by children, who are born without immune defenses and thus are susceptible to catching and spreading the disease. Another concern for waning immunity would be frequent reinfections that result in severe symptoms, Shaman says.
“This would suggest we're not going to be done with this any time soon, and that prior exposure doesn't lessen your chance of winding up in the hospital,” he says. Though one severe reinfection has been reported worldwide, there’s no evidence yet this is happening on a broad scale.
If society wants to overcome these bleak possibilities and return to life without social distancing and mask wearing, it needs a vaccine that provides a sufficient amount of what’s known as sterilizing immunity, meaning the drug blocks coronavirus transmission.
“I would say the sweet spot is 80 percent,” says CUNY’s Lee, who co-authored a research paper in July about efficacy goals for the COVID-19 vaccine. The bare minimum standard of 50 percent, set by the FDA and WHO, would only protect half the population if everyone is vaccinated. That falls well below the theorem threshold for COVID-19 of 60 to 75 percent. Such a scenario would be akin to the seasonal influenza vaccine, for which transmission efficacy tends to range between 20 to 60 percent. Mass vaccination doesn’t stop the flu, though it does reduce the disease burden on society.
“We have to make it clear to everyone that the first vaccine to reach the market may not achieve those efficacy levels,” Lee says. “It's not that easy to get an efficacy that high for a respiratory virus.”
That’s because current guidance says vaccine frontrunners can be approved even if they only provide “functional immunity,” which mainly confers protection against the symptoms of the disease.
The ongoing COVID-19 vaccine trials are not designed to estimate the impact the vaccine candidates would have on transmission, write the authors of the National Academy of Medicine report, adding that we may not learn this impact until well after an FDA approval. As they explain, the first priority is to stop the most vulnerable people from dying, especially older people with pre-existing conditions and our limited cohort of frontline health-care specialists and first responders.
“So much of the focus has been on the return to normal,” Lee says, “and we can't have that type of expectation.”
Editor’s note: This story was updated to reflect that polio, not the mumps, has an R-naught of seven and herd immunity threshold of 85 percent.