In 1950 two British researchers, Richard Doll and Austin Bradford Hill, published a study of patients in London hospitals. They compared people with lung cancer to similar patients without it and asked both groups about their habits. The difference was stark: almost all of the lung cancer patients smoked. Doll gave up cigarettes himself during the research. The same year, the American researchers Ernst Wynder and Evarts Graham published similar results. At the time, roughly four in five British men smoked, and doctors smoked as much as anyone. These studies began one of the great public health battles of the century.

The epidemiological transition

As sanitation, vaccines and antibiotics reduced deaths from infection, the leading causes of death changed. In wealthy countries, and increasingly everywhere, people now mostly die from noncommunicable diseases (NCDs): heart disease, stroke, cancer, chronic lung disease and diabetes. Researchers call this shift the epidemiological transition. According to the WHO, NCDs kill about 41 million people a year, roughly three quarters of all deaths worldwide. These diseases develop over decades and have many causes, so studying them demanded new methods.

Studying slow killers

Two designs became central. A case-control study, like Doll and Hill's first one, starts with people who have a disease and compares their past exposures to those of people who do not. A cohort study follows a large group forward in time and watches who gets sick. In 1951 Doll and Hill began following about 40,000 British doctors, a study that eventually ran for 50 years. It showed that smoking caused not only lung cancer but heart disease and many other illnesses, and that lifelong smokers lost about ten years of life on average. In 1948 the Framingham Heart Study began following residents of a Massachusetts town, helping to identify high blood pressure, high cholesterol and smoking as major causes of heart disease. Framingham researchers also popularized the term risk factor.

From correlation to cause

Tobacco companies argued that the link between smoking and cancer was only a correlation. In 1965 Bradford Hill set out a list of considerations for judging when an association is likely to be causal, now known as the Bradford Hill criteria. They include the strength of the association, its consistency across studies, a dose-response pattern in which more exposure brings more disease, the right timing, with exposure coming before illness, and a plausible biological mechanism. Smoking met them all. In 1964 the US Surgeon General, Luther Terry, had already issued a landmark report concluding that cigarette smoking causes lung cancer in men.

Manufacturing doubt

The tobacco industry knew much of this. Internal documents later released through lawsuits showed that companies understood the dangers of smoking and the addictiveness of nicotine while publicly denying both. A 1969 memo from one cigarette maker put the strategy bluntly: "Doubt is our product." The industry funded friendly research, attacked scientists and promoted "light" cigarettes as safer, which they were not. In 1998 major US tobacco companies signed the Master Settlement Agreement, agreeing to pay states more than $200 billion over 25 years and to accept limits on marketing. The same playbook of sowing doubt has since been used in other debates about health and the environment.

What worked

Countries that cut smoking used a combination of measures:

• Taxes, which raise prices and deter young people especially. The World Bank found that a 10 percent price rise reduces cigarette consumption by about 4 percent in high-income countries, and more in poorer ones.
• Smoke-free laws for workplaces, restaurants and public transport, which also protect nonsmokers from secondhand smoke.
• Advertising bans and plain packaging with large graphic warnings, which Australia introduced in 2012.
• Help to quit, from telephone helplines to nicotine replacement and counseling.

In 2003 the WHO adopted its Framework Convention on Tobacco Control, the first global health treaty, now joined by more than 180 countries. In the United States, adult smoking fell from about 42 percent in 1965 to around 11 percent by the early 2020s. Still, the WHO estimates that tobacco kills more than 8 million people a year, including about 1.3 million nonsmokers exposed to secondhand smoke.

Worked example: a city tackles heart disease

A city health department finds that heart disease is its leading cause of death, concentrated in low-income neighborhoods. Applying Geoffrey Rose's prevention paradox from Level 1, it looks for measures that shift risk for everyone rather than only treating the sickest. It might raise local tobacco taxes and fund free quit programs, work with food manufacturers and restaurants to lower salt, offer blood pressure checks at pharmacies and barbershops, and build safe parks and sidewalks so walking is easy. Each change is modest for any one person. Together, over years, they can prevent far more deaths than a new cardiac ward. The city then tracks rates of smoking, high blood pressure and heart attacks to see whether the plan is working.