{
  "id": "herd-immunity",
  "title": "Herd Immunity",
  "lang": "en",
  "source_lang": "en",
  "summary": "Herd immunity is the indirect protection against infection that arises when population immunity reduces transmission of a pathogen.",
  "category": "health",
  "subcategory": "public-health",
  "type": "concept",
  "aliases": [
    "Community immunity",
    "population immunity",
    "herd protection"
  ],
  "generated_at": "2026-11-03",
  "model": "pro/gpt-6.1-sol",
  "url": "https://aiwiki.page/en/herd-immunity/",
  "translations": {
    "zh-Hans": {
      "url": "https://aiwiki.page/zh/herd-immunity/",
      "json": "https://aiwiki.page/zh/herd-immunity.json"
    }
  },
  "sections": [
    {
      "heading": "",
      "text": "Herd immunity is the indirect protection against an infectious disease that occurs when immunity in a population reduces the opportunities for its pathogen to spread. People who are immune are less likely to become infected, transmit infection, or both, interrupting chains of transmission and lowering exposure among other people. This protection can extend to individuals who remain susceptible, including infants, people with certain forms of immunodeficiency, and those who do not develop an adequate response to vaccination.\n\nThe concept is central to epidemiology, vaccination policy, and the control of communicable diseases. It explains why the effects of a vaccine can extend beyond its recipients and why the distribution of immunity matters alongside its average level. Herd immunity is not an absolute shield, a property shared equally across all infections, or necessarily a permanent state. Its strength depends on the pathogen, the kind and duration of immunity, patterns of contact, and the conditions under which transmission occurs.\n\nA distinction is necessary between herd protection, meaning any reduction in risk produced by other people's immunity, and the herd-immunity threshold, a model-dependent level of immunity above which sustained epidemic growth becomes impossible under specified conditions. Substantial indirect protection can exist below that threshold, while outbreaks can occur in susceptible groups even when a population-wide average appears high."
    },
    {
      "heading": "Historical development and scope",
      "text": "The origins of herd immunity lie in observations that an individual's risk of infection depends partly on the susceptibility of surrounding people. In 1917, the American veterinarian George Potter discussed contagious abortion in cattle in terms of herd-level resistance. In 1923, the British researchers W. W. C. Topley and G. S. Wilson explicitly examined “herd-immunity” through experimental studies of bacterial infection in groups of mice. Their work distinguished the resistance of individual animals from the behavior of infection in a population continually affected by the addition of susceptible animals.\n\nDuring the following decades, the concept became increasingly important in human infectious-disease research. Studies of naturally recurring epidemics showed that transmission often declined after susceptible people had been depleted, then returned as births and population movement replenished susceptibility. In the 1930s, A. W. Hedrich analyzed measles in Baltimore and related epidemic behavior to the estimated proportion of children who remained susceptible. These observations helped establish that epidemic cycles could arise from changes in population susceptibility rather than changes in the pathogen alone.\n\nMathematical work supplied a more explicit framework. The epidemic models developed by William Kermack and Anderson McKendrick in 1927 demonstrated how an outbreak could decline while some susceptible people remained. Infection did not have to reach every susceptible individual: transmission could become unsustainable because infectious people encountered susceptible contacts too infrequently. Later vaccination programs transformed this theoretical insight into a practical question—how much immunity, distributed in which groups, would prevent continued spread?\n\nThe term has subsequently acquired several overlapping meanings. “Population immunity” can describe the overall distribution of immune responses, even without demonstrated indirect protection. “Herd protection” emphasizes the observed reduction in exposure among people who are not themselves protected. “Herd immunity” is often used for both the underlying population effect and the condition in which transmission cannot sustain itself. These meanings are related but not interchangeable; a population can have substantial immunity without having enough transmission-blocking protection to prevent recurrent epidemics.\n\nThe mechanism also has limits. It is most straightforward for infections transmitted between people when immunity reduces the probability of onward spread. Tetanus, acquired mainly when environmental spores enter wounds, does not produce conventional person-to-person herd protection: one person's vaccination does not substantially reduce another person's exposure to spores. For vector-borne infections such as malaria, transmission depends on interactions between humans and mosquitoes, so population effects require models that include vector abundance, biting patterns, and immunity affecting infection or infectiousness. Herd immunity is therefore a relationship among hosts, pathogens, and transmission systems, rather than a universal percentage applicable to every disease."
    },
    {
      "heading": "Biological mechanisms of indirect protection",
      "text": "Herd protection begins with changes in the probability that exposure leads to another infectious person. The immune system can interfere at several points: preventing a pathogen from establishing infection, limiting its multiplication after infection, shortening the infectious period, or reducing the amount of infectious material released. Immunity that completely prevents productive infection is often called sterilizing immunity. Such immunity is not required for every population benefit, but protection that acts only against severe illness, without changing infection or infectiousness, does not by itself interrupt transmission.\n\nDifferent immune components contribute differently. Antibodies can neutralize pathogens before they enter susceptible cells, while T cells can recognize and eliminate infected cells. Immune responses at the respiratory or intestinal mucosa may be especially important for blocking acquisition and shedding at those sites. Circulating antibodies and cellular immune memory may remain effective against serious disease even when protection at an entry surface is incomplete. Consequently, strong protection against hospitalization cannot automatically be interpreted as equally strong protection against transmission.\n\nTransmission can be understood as a sequence involving contact, exposure, infection, and infectiousness. If an immune person is less likely to become infected, that person is less likely to serve as a link in a transmission chain. If infection nevertheless occurs but infectiousness is reduced, fewer secondary cases may result. These mechanisms can operate together. Their relative importance differs by pathogen, vaccine, time since immunization, and the characteristics of the exposed person.\n\nSome vaccines also reduce asymptomatic carriage. This is particularly important for bacteria that commonly inhabit the upper respiratory tract without immediately causing disease. A conjugate vaccine can reduce carriage of targeted bacterial types, lowering their circulation and protecting unvaccinated age groups. The resulting reduction in invasive disease among nonrecipients is evidence of an indirect effect, even though those individuals have not acquired vaccine-induced immunity themselves.\n\nBiological protection is neither perfectly uniform nor always binary. Some people develop stronger responses than others; protection may also decline gradually rather than disappear at a single moment. Previous infection, vaccination, age, and immunosuppression can produce different immune profiles within the same population. “Immune” and “susceptible” are useful modeling categories, but real populations often contain a spectrum of susceptibility and infectiousness.\n\nIndirect protection should also be distinguished from passive immunity. Antibodies transferred during pregnancy, or administered as an antibody product, act within the recipient. They may directly protect that person for a limited period. Herd protection instead arises because other people's immunity changes the recipient's probability of encountering an infectious source. Both mechanisms can protect infants, but they operate through different biological and population pathways."
    },
    {
      "heading": "Reproduction numbers and the threshold principle",
      "text": "The simplest quantitative account uses the basic reproduction number, \\(R_0\\): the average number of secondary infections generated by a typical infectious person in a wholly susceptible population under specified conditions. It is not an immutable property of a pathogen. Contact patterns, population density, behavior, and environmental conditions all influence its value. The effective reproduction number, often written \\(R_{\\mathrm e}\\) or \\(R_t\\), describes transmission under the immunity and other conditions currently present.\n\nIn a homogeneous model with random mixing and complete, uniformly distributed immunity, let \\(p\\) be the immune proportion. The susceptible proportion is \\(1-p\\), giving\n\n\\[R_{\\mathrm e}=R_0(1-p).\\]\n\nThe critical immune proportion is therefore\n\n\\[p_c=1-\\frac{1}{R_0}.\\]\n\nAbove this value, \\(R_{\\mathrm e}<1\\), and successive generations of infection decline on average. If \\(R_0=3\\), the threshold is approximately \\(66.7\\%\\); if \\(R_0=15\\), it is approximately \\(93.3\\%\\). These figures illustrate the model, not universal targets for particular communities.\n\nVaccine coverage and immunity are not identical. If a vaccine prevents infection with effectiveness \\(E\\), and vaccination is randomly distributed, a simple approximation is\n\n\\[R_{\\mathrm e}=R_0(1-cE),\\]\n\nwhere \\(c\\) is coverage. The corresponding critical coverage is\n\n\\[c_c=\\frac{1-1/R_0}{E}.\\]\n\nFor \\(R_0=5\\) and \\(E=0.90\\), this gives coverage of approximately \\(88.9\\%\\). If the calculated requirement exceeds \\(100\\%\\), vaccination with that effectiveness cannot, by itself, bring transmission below the threshold under the assumed conditions. Models in which vaccination separately affects acquisition and infectiousness require a more detailed formulation.\n\nThe threshold concerns epidemic growth, not the immediate disappearance of cases. When immunity is accumulated during an uncontrolled epidemic, many people may already be infectious when the threshold is crossed. Transmission continues while those infections resolve, producing additional cases. This phenomenon, called epidemic overshoot, means that the final infected proportion can be substantially larger than the threshold. In the simplest closed susceptible–infectious–recovered model, an initially almost wholly susceptible population has an approximate final infected fraction \\(z\\) satisfying\n\n\\[1-z=e^{-R_0z}.\\]\n\nCrossing the threshold therefore does not mean that the epidemic stops at that exact population fraction.\n\nNor does \\(R_{\\mathrm e}<1\\) imply that every infectious person transmits to fewer than one other person. It is an average across transmission events. Chance clusters and imported cases can still occur, particularly in small populations or settings with highly variable transmission. A threshold is best understood as a boundary in a specified model of population dynamics, rather than a guarantee of individual safety or an observable switch that can always be identified from a single coverage figure."
    },
    {
      "heading": "Contact networks and uneven immunity",
      "text": "Real populations do not mix randomly. People have repeated contacts within households, schools, workplaces, religious communities, and social networks. Children often have disproportionately frequent contact with other children, while adults' connections vary with occupation and mobility. These patterns affect both the spread of infection and the value of immunity in particular groups. Immunizing a person who would otherwise connect many transmission chains can have a larger indirect effect than immunizing a person with relatively few exposure opportunities.\n\nIn age-structured or otherwise stratified models, transmission is represented by a next-generation matrix. An entry describes the expected number of infections in one group caused by an infectious person in another. The dominant eigenvalue of this matrix determines whether infection grows. Changing immunity in one group alters the matrix and potentially its dominant eigenvalue. There is therefore no single immune percentage sufficient to describe every possible allocation of immunity across the population.\n\nClustering of susceptibility is especially consequential. A country may have high average vaccination coverage while particular neighborhoods or communities have much lower coverage. If susceptible people preferentially interact with one another, infection can spread within these groups even when the national average exceeds a simple homogeneous-model threshold. Geographic averages can also conceal underimmunized birth cohorts or groups that have missed later doses. The relevant unit for transmission may be a school, settlement, or connected social network rather than an administrative region.\n\nHeterogeneity can also lower model-estimated thresholds under particular assumptions. During infection-driven epidemics, people with more contacts or higher susceptibility may be infected earlier, leaving a remaining population less able to sustain transmission than a randomly immunized population of the same size. This selective depletion can produce a lower calculated threshold. However, the result depends on whether differences are stable, how contacts change over time, and whether immunity persists. It does not establish a fixed, universally low threshold or show that allowing infections is a low-cost route to protection.\n\nPopulation movement links otherwise separate settings. Infection imported into a highly immune population may produce only a short chain; the same introduction into a susceptible cluster may cause a substantial outbreak. Repeated introductions can generate recurring cases even where local transmission is usually subcritical. Schools, transport routes, and cross-border movement can therefore shape the practical meaning of “community.”\n\nThese relationships connect herd immunity with social determinants of health. Barriers to healthcare, interrupted vaccine supply, displacement, and unequal access to services can create linked pockets of susceptibility. The resulting exposure is not confined to individuals who remain unvaccinated: it also affects vaccinated people whose protection is incomplete. Analyses based only on average coverage can consequently miss both transmission risk and its unequal distribution."
    },
    {
      "heading": "Evidence from vaccination programs",
      "text": "Measles provides a clear illustration of the demanding conditions for sustained herd protection. It is highly transmissible, with commonly cited \\(R_0\\) estimates around 12–18 in susceptible populations, although estimates vary by setting. The corresponding elementary threshold is roughly 92–94% immune. Because vaccination does not produce protection in every recipient and coverage is uneven, measles programs generally use a target of at least 95% coverage with two doses in every district. High national coverage alone does not prevent outbreaks if susceptible people remain concentrated in connected groups.\n\nThe eradication of smallpox demonstrated the population effects of vaccination combined with surveillance and containment. The last naturally occurring case was recorded in Somalia in 1977, and the World Health Organization declared global eradication in 1980. Smallpox control did not consist simply of attaining a uniform global coverage threshold. Identification of cases, isolation, contact tracing, and vaccination of contacts and surrounding communities were central to the final phase. Its exclusively human reservoir and recognizable clinical illness made these measures particularly effective.\n\nBacterial conjugate vaccines provide strong evidence of indirect protection across age groups. Vaccination against Haemophilus influenzae type b reduced invasive disease in young children and also lowered circulation of the organism. Pneumococcal conjugate vaccines reduced carriage of targeted serotypes in vaccinated children and were followed by reductions in vaccine-type invasive disease among unvaccinated adults. Such effects are not fully captured by counting illness prevented only in vaccine recipients.\n\nThe meningococcal serogroup A conjugate vaccine introduced in the African meningitis belt in 2010 supplies another important example. Mass campaigns were associated with major reductions in serogroup A disease and carriage, including effects beyond directly vaccinated individuals. By contrast, the strength of carriage reduction and indirect protection cannot simply be assumed to be identical across all meningococcal vaccine products. Vaccine composition, target antigen, population, and transmission ecology matter.\n\nPoliomyelitis illustrates differences between vaccine types. The oral polio vaccine, a live attenuated vaccine, induces intestinal immunity that helps reduce fecal–oral transmission. Inactivated polio vaccine provides strong protection against paralytic disease but, especially in people without prior intestinal exposure, generally produces less intestinal immunity than oral vaccination. The live vaccine's transmission advantages must be considered alongside its rare capacity to cause vaccine-associated paralysis or give rise to circulating vaccine-derived polioviruses where population immunity is inadequate.\n\nNot every successful vaccine produces an equally strong herd effect. Tetanus vaccination principally protects recipients because infection is not ordinarily passed between people. For pertussis, protection against symptomatic disease and protection against infection or carriage are not equivalent, and immunity wanes. These contrasts show why the existence, magnitude, and persistence of indirect protection require pathogen- and product-specific evidence rather than inference from a vaccine's effectiveness against severe disease alone."
    },
    {
      "heading": "Waning immunity, pathogen evolution, and recurrent epidemics",
      "text": "Herd protection changes over time even without a major change in the pathogen. Births add susceptible individuals; migration changes the composition of communities; and immune protection may wane. If these processes raise the effective susceptible fraction, transmission can resume after a period of suppression. Recurrent epidemics may therefore reflect the gradual rebuilding of susceptibility rather than the complete disappearance of immunity. Long-lived protection against one clinical outcome can coexist with declining protection against infection.\n\nInfluenza illustrates the interaction of waning immunity and antigenic change. Alterations in viral surface proteins can reduce recognition by antibodies induced by earlier infections or vaccination. As circulating strains change, a population's immune history remains relevant but does not translate into a fixed fraction that is fully protected against all future strains. Population protection depends on the match between immune responses and currently circulating viruses, as well as age, prior exposures, and the season's contact conditions.\n\nThe COVID-19 pandemic made these distinctions especially visible. Vaccination and previous infection reduced susceptibility and, under many conditions, onward transmission, but these effects were incomplete and varied over time. The emergence of variants with increased transmissibility or immune escape, particularly Omicron, changed the relationship between measured vaccination coverage and protection against infection. Protection against severe disease generally proved more durable than protection against acquiring infection. A single enduring threshold based only on the proportion ever vaccinated or previously infected could not adequately describe these dynamics.\n\n“Hybrid immunity,” generated by both infection and vaccination, is another example of why immune histories cannot be reduced to a simple count. The order, number, and timing of exposures can affect the breadth and magnitude of immune responses. Nevertheless, prior exposure does not imply identical protection in all individuals or against all variants. Repeated infections can occur without making previous immunity biologically irrelevant: immune memory may still change disease severity or the duration of infection.\n\nPathogen evolution can also alter which strains circulate. After pneumococcal vaccination reduces vaccine-targeted serotypes, non-vaccine serotypes may occupy part of the ecological space previously held by those types. This is known as serotype replacement. Replacement does not necessarily eliminate the program's overall reduction in disease, because serotypes differ in their capacity to cause invasive illness. Its importance depends on both carriage changes and the disease burden associated with the replacement types.\n\nSeasonality adds another moving component. Indoor contact, school schedules, humidity, and other environmental factors can change transmission opportunities. A level of immunity sufficient to keep \\(R_{\\mathrm e}<1\\) in one season may be insufficient in another. For infections with appreciable waning or immune escape, population immunity may reduce epidemic size and clinical burden without eliminating circulation. The distinction between suppressing transmission and reducing its consequences is therefore essential when interpreting recurrent outbreaks."
    },
    {
      "heading": "Measuring and attributing herd effects",
      "text": "Herd immunity is not measured directly by a single laboratory result. Investigators combine vaccination records, infection histories, antibody surveys, disease notifications, carriage studies, and transmission models. Each source describes a different component of the problem. Recorded coverage measures vaccine delivery, not necessarily successful immunization; an antibody result measures a particular immune response, not every mechanism of protection; and case counts reflect both transmission and the probability that infections are detected.\n\nSeroprevalence studies estimate the proportion of a sampled population with detectable antibodies to a pathogen or vaccine antigen. They can reveal age groups or locations with immunity gaps, but interpretation depends on assay sensitivity and specificity, sampling design, and the relationship between the measured antibody and protection. Antibodies may fall below a test's detection threshold while immune memory persists. Conversely, detectable antibodies may not provide strong protection against a substantially changed pathogen. Some assays also cannot distinguish vaccination from infection.\n\nA central problem is separating direct from indirect effects. Direct protection compares outcomes associated with an individual's own vaccination. Indirect protection concerns the effect of other people's vaccination on that individual's risk. Total effects compare vaccinated people in a vaccinated population with unvaccinated people in a population without the program; overall effects compare average population outcomes under different vaccination strategies. These contrasts answer different questions and can produce different numerical estimates without contradiction.\n\nA cluster-randomized trial can help identify indirect effects by assigning vaccination strategies to communities, schools, or other groups rather than treating each person as an independent unit. Researchers may then compare outcomes among unvaccinated members of intervention and comparison clusters. Interpretation still requires attention to movement between clusters, differences in uptake, and whether the supposedly unvaccinated groups remain comparable. In infectious-disease studies, one person's treatment can change another person's exposure, violating the independence assumptions used in many conventional analyses.\n\nObservational studies also provide important evidence. Declines in disease among age groups not eligible for a newly introduced vaccine can support an indirect effect, particularly when accompanied by reductions in carriage of targeted organisms. However, changes in diagnosis, reporting, healthcare access, behavior, or other interventions can influence observed trends. Analyses may use comparison regions, age groups, or pathogen types to assess alternative explanations. A decline after vaccine introduction is informative, but timing alone does not establish the entire causal mechanism.\n\nPublic health surveillance supplies the continuing evidence needed to evaluate population protection. Case investigation, laboratory characterization, genomic data, and carriage surveys can identify susceptible clusters and changes in circulating strains. Wastewater surveillance can complement clinical data for some pathogens by tracking community circulation without relying solely on diagnosed cases. Models integrate these data, but their estimates remain conditional on assumptions about contact patterns, reporting, immune protection, and its duration."
    },
    {
      "heading": "Vaccination strategy, equity, and public interpretation",
      "text": "Herd effects change the evaluation of vaccination programs because benefits extend beyond recipients. A program may prevent infections in people who cannot receive a particular vaccine, who have not yet reached the eligible age, or whose immune responses are impaired. Economic evaluations that include these indirect effects can produce substantially different estimates from analyses restricted to direct protection. The size of the difference depends on whether vaccination reduces transmission and how the protected and susceptible groups interact.\n\nProgram design must distinguish several objectives: preventing infection, preventing severe disease, reducing deaths, interrupting transmission, and achieving elimination. These goals can favor different allocations of vaccine. Vaccinating people with high contact rates may strongly affect transmission, while vaccinating people at high risk of severe illness may provide greater immediate clinical benefit per dose. Age structure, vaccine effectiveness, pathogen biology, and available supply determine how these effects interact; a transmission-minimizing strategy is not automatically identical to a mortality-minimizing strategy.\n\nRoutine childhood vaccination, catch-up campaigns, and outbreak-response vaccination address different sources of susceptibility. Routine delivery protects successive birth cohorts, while catch-up programs can close gaps left by missed doses or historical policy changes. For infections in which immunity is durable, sustained high coverage can prevent the accumulation of susceptible people. For others, waning immunity and antigenic change mean that the timing and composition of vaccination also affect population protection. Program evaluation therefore examines continuity and distribution, not only the number of doses delivered.\n\nThe distinction between elimination and eradication is also important. Elimination of transmission refers to interruption in a defined geographic area, with continuing measures often needed to prevent re-establishment after importation. Eradication means the permanent worldwide reduction of naturally occurring infection to zero through deliberate efforts. Herd protection can contribute to both, but biological features such as animal reservoirs, asymptomatic infection, environmental persistence, and available diagnostic tools determine whether eradication is feasible.\n\nEthically, vaccination has both individual and collective dimensions. Indirect protection is relevant to public health because decisions and barriers affecting one group can alter exposure in others. At the same time, observed under-vaccination can arise from supply failures, geographic isolation, conflict, cost, or exclusion from services rather than a single kind of individual choice. Health equity therefore concerns both the distribution of infectious risk and access to the means of reducing it. A population average can conceal substantial differences in who receives protection and who bears continuing disease burden.\n\nPublic interpretation is especially sensitive to the word “threshold.” A model-derived percentage can be mistaken for a promise that risk becomes zero once coverage reaches a particular value. It can also obscure useful reductions in transmission that occur below the threshold. Reporting herd effects accurately requires specifying the pathogen or strain, population, period, immune outcome, and assumptions involved. In practical surveillance, continued cases, changes in local coverage, and the distribution of susceptibility remain informative even when a broad numerical target has been attained."
    }
  ],
  "links": [
    "infectious-disease",
    "immunodeficiency",
    "epidemiology",
    "vaccine",
    "measles",
    "malaria",
    "immune-system",
    "antibody",
    "t-cell",
    "conjugate-vaccine",
    "immunosuppression",
    "pregnancy",
    "matrix",
    "social-determinants-of-health",
    "smallpox",
    "world-health-organization",
    "live-attenuated-vaccine",
    "influenza",
    "covid-19",
    "cluster-randomized-trial",
    "diagnosis",
    "public-health-surveillance",
    "wastewater-surveillance",
    "public-health",
    "health-equity"
  ],
  "sources": [],
  "license": "https://creativecommons.org/licenses/by/4.0/"
}