Herd immunity
A Socratic walk-through of herd immunity — reasoned out one step at a time, not lectured.
The question we started with
THE QUESTION #Why can vaccinating most of a population protect the people who were never vaccinated?
A newborn too young for the measles vaccine, and a child on chemotherapy who cannot safely receive it, both walk through the same crowded places as everyone else. Neither carries any protection of their own. Yet in a well-vaccinated town they very rarely catch measles — not because they are shielded, but because the disease does not arrive. How can a person's safety be produced entirely by other people's immune systems?
Reasoning it through
REASONING #Start by asking what an epidemic actually needs. Not victims — chains. One infected person must, on average, hand the infection to more than one other person, or the outbreak shrinks. Call the average number of new infections one case produces in a wholly susceptible population R0. If R0 is above 1, cases multiply; if it falls below 1, each generation of infection is smaller than the last, and the chain dies out.
Now notice what vaccination does to that number. It does not make a susceptible person invisible; it removes them from the pool of people the infection can pass through. If a proportion p of the population is immune, then of the R0 contacts an infected person would have infected, a fraction p are now dead ends. The effective reproduction number becomes R0 x (1 - p).
Set that below 1 and solve, and the whole idea falls out in one line: transmission collapses when p exceeds 1 - 1/R0. That is the herd immunity threshold, and it is worth pausing on what it does not say. It does not say the unvaccinated are protected by proximity to the vaccinated. It says the pathogen cannot sustain a chain long enough to reach them.
Put numbers to it. For a pathogen with R0 = 2 the threshold is 1 - 1/2, or 50 per cent; at R0 = 4 it is 75 per cent. Measles, with an R0 usually estimated between 12 and 18, gives 1 - 1/12 = 92 per cent at the low end and 1 - 1/18 = 94 per cent at the high end. That is why measles returns first when coverage slips: its threshold sits so close to the ceiling that a few percentage points of decline crosses it, while a less transmissible disease still has slack.
Ask a sharper question, though: is the threshold a coverage figure or an immunity figure? Immunity. No vaccine is perfectly effective, so if a vaccine protects a fraction e of those who receive it, coverage must satisfy coverage x e above the threshold. Two doses of measles vaccine are roughly 97 per cent effective, so a 92 per cent immunity requirement needs about 95 per cent coverage. The arithmetic is unforgiving in exactly the region where it matters most.
The analogy
THE ANALOGY #Think of a forest fire and a firebreak cut through it. The cleared strip is not protecting any particular tree by standing over it; it protects the far side of the forest by ensuring the fire runs out of fuel before it gets there. Cut enough of the trees and a spark that lands anywhere burns a small patch and stops.
Fire spreads to whatever is adjacent, whereas people mix by choice and habit — so a forest can be uniformly thinned, while a population's unvaccinated trees cluster together in dense stands, and it is precisely there that the fire keeps going.
Clarifying the model
THE MODEL #That last point is the most important caveat. The 1 - 1/R0 formula assumes random mixing: everyone equally likely to meet everyone. Real populations do not mix that way. Unvaccinated people cluster — by school, by neighbourhood, by conviction — and inside a cluster the local immunity level can be far below the national figure. A country reporting 93 per cent measles coverage can contain a school at 60 per cent, and an outbreak in that school is entirely consistent with the national number. Herd immunity is a local property wearing a national label.
Second, R0 is not a constant of the pathogen. It is a property of a pathogen in a population, and it moves with contact patterns, density, season, and behaviour. The threshold derived from it inherits that softness, which is why any single figure quoted for a new disease should be read as an estimate under stated conditions rather than a physical constant.
Third — and this one undermines the concept rather than refining it — the argument assumes vaccination blocks transmission. Some vaccines prevent severe disease very well while reducing onward transmission much less, and some immunity wanes. Where that is true, a vaccinated person is still a link in the chain, and the coverage-to-threshold arithmetic no longer describes anything real: the vaccine remains valuable for the person receiving it, but the protection of the unvaccinated bystander weakens or disappears. This was much of the argument about COVID-19 vaccines, and it is the reason "herd immunity" transferred badly from measles to that setting.
Finally, a misconception worth correcting: reaching the threshold does not mean the disease vanishes, only that sustained chains cannot be maintained. Imported cases still occur and still infect people. What ends is the outbreak, not the risk.
A picture of it
THE PICTURE #How to readEach bar is one value of R0 with its threshold computed as 1 - 1/R0, rounded to the nearest per cent. Read left to right as the pathogen gets more transmissible and watch the curve flatten as it climbs: the first bars rise steeply, but past R0 = 6 the requirement creeps toward a ceiling it never reaches. That flattening is the trap — doubling transmissibility from 6 to 12 adds only nine points to the requirement, yet those are the nine hardest points in the whole range to achieve.
What became clearer
WHAT CLEARED #Herd immunity is not a shield anyone stands behind. It is a statement about chains: once enough contacts are dead ends, each infection produces fewer than one successor and the outbreak shrinks itself out of existence before it reaches the people who have no protection. The threshold follows directly from how transmissible the pathogen is, which is why measles demands so much more than most diseases — and why the concept fails quietly wherever the unvaccinated cluster together, or the vaccine stops disease without stopping spread.
Where to go next
ONWARD #- Why ring vaccination worked for smallpox eradication without ever reaching national thresholds.
- How network structure, rather than average coverage, determines where an outbreak actually goes.
Key terms
TERMS #| Term | What it means |
|---|---|
| R0 (basic reproduction number) | the average number of secondary infections one case causes in an entirely susceptible population. |
| Effective reproduction number | the same quantity in the real population, reduced by existing immunity and by behaviour. |
| Herd immunity threshold | the immune fraction, 1 - 1/R0, above which sustained transmission cannot be maintained. |
| Sterilizing immunity | immunity that prevents infection and onward transmission, not merely severe disease. |
Every term the collection defines is gathered in the glossary.