Doll and Hill study

Doll and Hill study
Title	Doll and Hill study
Authors	Richard Doll; Austin Bradford Hill
Year	1950
Type	Case–control study; cohort follow-up
Location	United Kingdom
Subjects	British physicians; general population cohorts
Key findings	Association between cigarette smoking and lung cancer; dose–response relationship

Doll and Hill study

The Doll and Hill study was a landmark epidemiological investigation that first quantified the association between cigarette smoking and lung cancer, linking risk to amount smoked and prompting major changes in public health policy. Conducted by Richard Doll and Austin Bradford Hill in the late 1940s and early 1950s, the work combined observational methods with statistical inference to argue for a causal relationship and influenced debates involving institutions such as the Royal College of Physicians, British Medical Journal, Medical Research Council, University of Oxford, and University of London.

Background and context

Doll and Hill worked in a period shaped by World War II, the aftermath of the Second World War, and evolving roles for institutions like the Ministry of Health and the National Health Service; contemporaneous figures included Winston Churchill, Clement Attlee, Aneurin Bevan, and researchers at the London School of Hygiene & Tropical Medicine. Interest in occupational and environmental causes of cancer had been growing through reports by investigators such as Percival Pott and organizations like the League of Nations and World Health Organization. Public attention involved companies such as Imperial Tobacco and events like high-profile legal cases in the House of Commons and discussions in media outlets including the Times of London and the Manchester Guardian.

Study design and methodology

Doll and Hill initially used a hospital-based case–control design, comparing patients with lung cancer to patients with other conditions, then extended analyses with prospective follow-up of British physicians enrolled via the British Medical Association and questionnaires distributed through medical journals like the BMJ (British Medical Journal). Methods drew on statistical principles developed by scholars such as Ronald Fisher, Jerzy Neyman, and Karl Pearson and incorporated procedures influenced by institutions including the Royal Statistical Society and the Medical Research Council. Data collection, coding, and classification used contemporary nosology from clinics at St Bartholomew's Hospital, Guy's Hospital, and cancer registries coordinated with the Office for National Statistics. Exposure assessment relied on self-reported cigarettes per day, and design elements included matching and stratification informed by practices at the Nuffield Unit and discussions with contemporaries at King's College London.

Findings and statistical analysis

Doll and Hill reported a strong positive association between cigarette smoking and lung cancer, with greater risk among heavy smokers and evidence of a dose–response gradient; results were summarized with contingency tables, relative risks, and trend tests influenced by methods used by Austin Bradford Hill in later causal criteria papers. Statistical interpretation referenced frequentist approaches advocated by Fisher and hypothesis-testing conventions in journals such as the Lancet and the BMJ (British Medical Journal). Follow-up cohort analyses of physicians provided mortality comparisons and standardized mortality ratios that echoed analytic techniques employed at the MRC Statistical Unit and reinforced earlier case–control estimates.

Interpretation and implications

Doll and Hill argued the observed associations, coherence across study designs, temporality, and dose–response were consistent with a causal link, anticipating later articulation of causal criteria associated with Austin Bradford Hill and debates involving public bodies like the Royal College of Physicians and the World Health Organization. Their conclusions influenced policy discussions in national legislatures such as the House of Commons and international advisory panels including committees of the United Nations. The study catalyzed shifts in clinical guidance disseminated through organizations such as the British Medical Association and professional bodies at universities like Cambridge University and University of Edinburgh.

Criticisms and subsequent evaluations

Contemporaneous criticism came from tobacco industry interests including Imperial Tobacco and commentators in media outlets like the Daily Mail and Daily Express, who challenged causal inference and pointed to potential confounding and selection bias. Academic critiques referenced methodological debates with statisticians tied to institutions like the Royal Statistical Society and commentators influenced by the work of Ronald Fisher and others skeptical of observational inference. Subsequent large-scale prospective studies, meta-analyses in journals such as the Lancet and syntheses by agencies including the World Health Organization and the US Surgeon General corroborated and extended Doll and Hill’s findings, while legal and policy reviews in bodies like the House of Lords and commissions connected to the European Commission evaluated regulatory responses.

Legacy and influence on public health policy

The Doll and Hill study became foundational for tobacco control initiatives implemented by governments and agencies including the National Health Service, World Health Organization, US Surgeon General's office, and national legislatures across Europe and North America. It informed interventions such as warning labels debated in the House of Commons, advertising restrictions litigated in courts like the European Court of Human Rights, and taxation policy discussed by bodies such as the Treasury (United Kingdom). The study’s blend of epidemiology and advocacy influenced generations of researchers at institutions including the London School of Hygiene & Tropical Medicine, Harvard School of Public Health, Johns Hopkins Bloomberg School of Public Health, and spurred legal and social movements involving organizations like Action on Smoking and Health.

Category:Epidemiology Category:Tobacco control