Landmark Studies That Changed the World

Statistics is not only a set of techniques; it is a record of arguments that changed how millions of people live. When doctors began quitting cigarettes in the 1950s, they were reacting to numbers, not to anything they could see or feel. The idea that a cigarette today could seed a cancer decades later is invisible to any single person. Only patterns traced across thousands of lives make such a risk visible at all.

The studies in this lesson mattered because they turned invisible risks into decisions: to quit smoking, to vaccinate a child, to check blood pressure before symptoms appear. Understanding why these results were believed teaches you what strong evidence actually looks like, so you can tell a genuine landmark from a passing headline.

Cohort and case-control designs

A case-control study starts from the outcome. It gathers people who already have a disease, finds similar people without it, and looks backward at how their histories differ. A cohort study runs the other way: it starts with healthy people, records their habits, and follows them forward for years to see who falls ill. Cohorts are slower and costlier, but they do not depend on faulty memory, so when both designs point the same way the conclusion is hard to dismiss.

The Bradford Hill viewpoints

In 1965 Austin Bradford Hill set out considerations for moving from association to causation: the strength of the link, its consistency across studies, a dose-response gradient, biological plausibility, and temporality, meaning the cause comes before the effect. No single item proves a cause. Taken together, they build a case a reasonable person can accept.

Scale and replication

Large numbers shrink the role of chance, and independent teams finding the same pattern rule out one laboratory's quirks. A result that survives across countries, decades, and study designs is what we mean by convincing.

Imagine following three groups over many years: non-smokers, light smokers, and heavy smokers. Suppose lung cancer deaths rise steadily from the first group to the last. That ordered pattern, more exposure bringing more disease, is a dose-response gradient, and it is one of Hill's strongest signals. Random noise rarely lines itself up so neatly across several groups. The British Doctors Study reported exactly this shape: death rates that climbed with the number of cigarettes smoked per day.

Not every dramatic association is causal. Around 2000, observational studies suggested that hormone replacement therapy protected women's hearts, because women taking it had fewer heart attacks. But those women were also, on average, wealthier and more health-conscious, both confounders. When the randomized Women's Health Initiative trial tested the therapy directly, the apparent heart benefit disappeared and some harms surfaced. Even large, consistent observational findings can mislead when a hidden factor sorts people into groups.

In 1950 Richard Doll and Austin Bradford Hill published a case-control study in the British Medical Journal linking smoking to lung cancer. They then launched the British Doctors Study in 1951, a prospective cohort that followed tens of thousands of British male doctors for decades and confirmed the link with a clear dose-response gradient.

In 1954 a field trial of Jonas Salk's polio vaccine enrolled roughly 1.8 million American schoolchildren. A rigorous portion was randomized, double-blind, and placebo-controlled, and it was evaluated by Thomas Francis Jr. On April 12, 1955 the results were announced: the vaccine worked, and mass immunization began.

Begun in 1948 in Framingham, Massachusetts with about 5,200 residents, the Framingham Heart Study followed people for years and helped popularize the term risk factor, identifying high blood pressure, high cholesterol, and smoking as drivers of heart disease.

• "A single study proved it." These landmark findings were confirmed by repeated, independent replication.
• "Observational studies can never show causation." They can build a strong case when many Hill viewpoints align.
• "A bigger sample guarantees a correct answer." Size beats chance, not bias; a huge biased study is still biased.
• "These conclusions were obvious at the time." Each was fiercely contested before the evidence accumulated.

Pick a current health claim you have seen in the news. Score it against four Bradford Hill viewpoints: Is the association strong? Is it consistent across studies? Is there a dose-response gradient? Does the cause clearly precede the effect? Count how many it satisfies, then decide whether you are looking at a landmark or a headline.

Think Like a Maester: A result becomes convincing not when one study is dramatic, but when many designs, done by many hands, keep pointing the same way.

1. What is the key difference between a cohort study and a case-control study?
2. Why is a dose-response gradient considered strong evidence for a causal link?
3. Which feature of the 1954 Salk polio field trial made its results especially credible?
4. What did the Framingham Heart Study contribute to how we talk about heart disease?
5. Why does replication across countries and decades strengthen a conclusion more than one large study alone?

A handful of statistical studies reshaped modern life: Doll and Hill on smoking, the Salk polio field trial, and the Framingham Heart Study. They convinced the world not through a single striking number but through sound designs, large scale, dose-response patterns, and repeated replication. The Bradford Hill viewpoints give you a checklist for weighing such evidence, and the hormone-therapy reversal reminds you that even strong associations must survive tougher tests before they earn belief.

• Statistics in Medicine and Health
• Causal Inference: From Correlation to Cause