Epidemiology
The study of how disease is distributed across a population, used to find its cause by counting and comparison — plotting who falls ill, where, and what they share, then removing the common exposure to see the outbreak stop.
✦ Wait, really?
In 1854 John Snow ended London's deadliest local cholera outbreak not with any medicine but by unscrewing a pump. He plotted some 500 nearby deaths on a street map, saw them cluster around a single water pump on Broad Street, and on 8 September persuaded the parish to remove its handle so no one could draw water. Honest to the end, he later admitted the epidemic was already fading — because so many residents had fled — so even he would not claim the handle alone had stopped it.
What it is
Epidemiology is the science of disease in populations rather than in single patients. Instead of asking "what is wrong with this person," it asks "who is getting sick, where, when, and what do they have in common" — and uses the pattern to find the cause. Its core tools are counting and comparison: tally the cases, map them, and compare the rate of illness between people who shared some exposure and people who did not. Where the rates diverge sharply, you have a suspect.
Why it mattered
For most of history, the cause of an epidemic was guessed from theory — bad air, imbalance, divine judgment — and the guesses could not be settled because no one measured. Epidemiology made cause a matter of evidence: you can find and stop a killer without a microscope, without knowing the pathogen, purely from the geometry of who falls ill. That is how cholera, a disease whose bacterium had not yet been isolated, was traced to contaminated water and defeated by plumbing. It remains the discipline that detects outbreaks, tests whether a treatment or a risk factor is real, and tells public health where to spend its effort.
What it unlocked
Epidemiology is the engine of public health. The same logic — count, map, compare exposed to unexposed, remove the cause — traced smoking to lung cancer, contaminated food to poisoning outbreaks, and new pathogens to their sources long before anyone could culture them. The controlled comparison at its heart grew into the clinical trial, which is how every modern drug and vaccine is judged. It is proof that populations, counted carefully, will tell you things no individual body can.
Minimum viable version
List every case with where and when it struck, mark them on a map, compare how often the sick shared a suspected exposure versus the well, then cut off the leading suspect and watch whether new cases fall.
Bootstrap recipe
You need
- · A defined outbreak: a disease appearing in more people, in one place and time, than expected
- · A way to find and list cases — who is sick, where they live, and when they fell ill
- · A map of the affected area
- · A list of the plausible shared exposures: water sources, food, wells, contacts, workplaces
- · Something to count and tally with — paper, a pen, patience
Steps
- 01Build a line list: one row per case, recording name or household, location, date of onset, and known exposures. Chase down every case you can, including the dead — missing cases bias everything that follows.
- 02Plot the cases on the map as marks stacked at each address (a 'spot map'). Look for where they cluster.
- 03Mark the candidate sources on the same map — every pump, well, market stall, or standpipe.
- 04For each suspect source, compare the sick and the well: did those who fell ill share an exposure that the healthy nearby did not? Compute attack rates — cases per hundred exposed versus per hundred unexposed.
- 05Hunt hard for the exceptions that would break your hypothesis. Snow's proof included cases far from Broad Street who had gone out of their way to drink from that pump, and a nearby brewery whose workers drank only beer and stayed well.
- 06When one exposure stands out — a far higher attack rate among those who shared it — interrupt it: close the well, remove the handle, recall the food, isolate the source.
- 07Keep counting after you act. A genuine cause should be followed by a fall in new cases beyond what the outbreak's natural decline would give.
How you know it worked
The exposed group's attack rate is dramatically higher than the unexposed's, and it holds up against the exceptions — sick people who were exposed despite living far away, well people who avoided the exposure despite living close. Snow's strongest evidence was a natural experiment: across South London, houses supplied by a company that drew sewage-tainted water had roughly eight to nine times the cholera death rate of houses supplied by a company drawing cleaner water upstream — same city, same streets, differing only in the water at the tap.
What goes wrong
- ⚠ Confusing nearness with cause — a source can sit in the middle of a cluster and be innocent; only comparing rates between exposed and unexposed separates the two.
- ⚠ Incomplete case-finding — if you miss cases (or the dead), the map lies and the ratios are wrong.
- ⚠ Acting after the peak — if the outbreak is already declining, removing the source may look effective when it changed nothing, or look useless when it worked. Measure against the expected natural course.
- ⚠ Long or variable incubation — cases appear days after exposure, so the cluster in time can smear and mislead if you assume illness follows exposure immediately.
This entry is awaiting its full account — the cartographers are at work. Its place in the graph is already verified.
Unlocked
Frontier — nothing charted yet.
Sources
- — John Snow, *On the Mode of Communication of Cholera*, 2nd ed. (London, 1855)
- — Steven Johnson, *The Ghost Map* (2006)
Something wrong on this page? Every claim here is meant to survive challenge. Suggest a correction →