What Is Herd Immunity? Protecting People Who Are Not Immune

The basic reproduction number, written R nought, is the average number of people an infected person would infect in a fully susceptible population with no interventions. If it exceeds one, an outbreak grows; below one, it dies out. Herd immunity works by reducing the effective reproduction number below one, and the fraction of the population that must be immune to achieve that is one minus the reciprocal of R nought. The consequences are arithmetic. A disease with an R nought of two requires half the population immune; one of four requires seventy-five percent; measles, among the most transmissible diseases known with estimates from twelve to eighteen, requires somewhere between ninety-two and ninety-five percent, which is why measles returns first whenever coverage slips and why it functions as an early warning indicator for a vaccination programme. The number is not a property of the pathogen alone, since it depends on contact rates, population density and behaviour, which is why it differs between settings and over time.

The simple threshold formula rests on conditions that are never fully met, and each failure matters:

• •Homogeneous mixing, meaning everyone is equally likely to contact everyone else, which is false in every real population and means clustering of unvaccinated people can sustain outbreaks even where national coverage exceeds the threshold
• •Immunity that blocks transmission rather than only preventing severe illness, which some vaccines do well and others do not, and a vaccine that prevents disease without preventing infection contributes far less
• •Durable immunity, since protection that wanes returns people to the susceptible pool and can make a threshold unreachable
• •A stable pathogen, since a virus that evolves to escape existing immunity resets the calculation
• •No animal reservoir, since a pathogen circulating in another species can reinfect regardless of human immunity, which is why eradication has succeeded for smallpox and is far harder for diseases with wildlife hosts

The evidence for the mechanism is strong and specific. Smallpox was eradicated worldwide by 1980, using ring vaccination around cases rather than universal coverage, which exploits the same principle locally. Polio has been eliminated from all but a small number of areas. Rubella vaccination protects pregnant women and their foetuses primarily through population immunity, since the direct benefit to the vaccinated child is modest and the serious harm is congenital. Introducing pneumococcal and meningococcal conjugate vaccines in children produced sharp declines in disease among unvaccinated adults, an effect measured repeatedly and attributable to reduced carriage in the group doing most of the transmitting. The converse is equally documented: measles outbreaks follow reliably wherever coverage in a community falls below the threshold, and the return of measles to countries that had eliminated it tracks coverage declines closely.

The term entered general discussion during the pandemic and was frequently used to mean something the epidemiology does not support, namely a strategy of allowing widespread infection to build immunity. The objections are specific rather than ideological. Reaching a threshold through infection means the deaths and long-term illness associated with that infection rate occurring first, which is the cost the whole approach exists to avoid. It assumes that infection produces durable transmission-blocking immunity, which for coronaviruses it does not. It assumes the pathogen will not evolve to evade that immunity, which it did. And it assumes vulnerable groups can be shielded while infection runs through everyone else, which requires a degree of separation that no society has achieved. The historical usage is worth noting: the term was coined in the 1920s in veterinary science and has always described a threshold produced by immunity however obtained, with the public health application built entirely around vaccination because that is the route that does not require the disease.

Population immunity is a public good in the technical sense, since it is produced collectively and its benefit cannot be withheld from anyone, which creates a predictable incentive problem. An individual who declines vaccination while everyone around them accepts it receives most of the protection and bears none of the small risk or inconvenience, which is individually rational and collectively unsustainable if enough people reason the same way. The policy responses are the standard ones for public goods: making the behaviour easy and free, which removes most of the friction that accounts for a large share of under-vaccination; requirements attached to school entry or employment, which raise coverage substantially and provoke resistance; incentives; and communication addressing specific concerns rather than treating all hesitancy as a single phenomenon, since the research consistently finds that the majority of unvaccinated people are not ideologically opposed but face access barriers, confusion about schedules, or particular worries that are answerable.

The threshold for herd immunity is one minus the reciprocal of the reproduction number, which requires half the population immune for a disease that spreads to two people per case and ninety-five percent for measles. The formula assumes even mixing, transmission-blocking and durable immunity, a stable pathogen and no animal reservoir, and each failed assumption weakens it, which is why clustering of unvaccinated people sustains outbreaks in highly vaccinated countries. It has eradicated smallpox and protects those who cannot be vaccinated, and its use to describe deliberate mass infection was not what the concept means.