Essay

The Birth of Evidence Based Medicine

The term evidence based medicine first appeared in print in 1991 and was popularized by a 1992 JAMA article, but the approach grew from a much longer argument about how physicians should know whether a treatment works. Its history runs through bedside observation, numerical comparison, hospital records, controlled trials, medical statistics, and the effort to make clinical authority answerable to explicit evidence.

Its historical significance lies in the shift from medicine grounded mainly in expert training and customary judgment toward medicine that asked clinicians to combine experience with critically appraised research, patient circumstances, and transparent standards of proof.

Period: Eighteenth-century therapeutic comparison to the evidence based medicine movement of the 1980s and 1990s
Key settings: Naval medicine, Paris hospitals, epidemiology, medical statistics, randomized trials, academic medicine, journal clubs, and clinical guideline work
Historical weight: Evidence based medicine changed the language of medical authority by making research quality, uncertainty, and reproducible comparison part of everyday clinical reasoning.

Read the essay Back to essays

Historical Setting

Before the name, medicine already argued about evidence

The phrase "evidence based medicine" is recent, but the problem it named is old. Physicians had long relied on a mixture of learned authority, personal experience, patient testimony, anatomical theory, and memorable cases. The difficulty was not that earlier medicine lacked observation. It was that observation could be selective, records could be uneven, and improvement after treatment did not prove that treatment had caused recovery.

This problem was especially visible in therapies that seemed plausible but produced uncertain results. Bleeding, purging, proprietary remedies, diet, rest, and surgery could all be defended from experience because some patients improved after receiving them. Yet patients also recovered naturally, declined despite treatment, or differed in ways that made comparison difficult.

The birth of evidence based medicine therefore belongs to a wider history of controlled comparison. It is closely connected to the history of clinical trials, the history of medical statistics, and the growth of medical records as tools for comparing groups rather than merely preserving individual case stories.

Early Comparisons

Therapeutic testing began before modern trial design

Long before randomization and systematic review, some practitioners tried to discipline experience by arranging comparisons. These early efforts were limited, but they showed why medical judgment needed more than memory and professional confidence.

James Lind made comparison memorable

In 1747, naval surgeon James Lind compared six treatments for scurvy among sailors aboard HMS Salisbury, assigning two sailors to each remedy, including citrus fruit, vinegar, and sea water. The experiment was tiny by modern standards and did not follow modern trial rules, but it became a landmark because it set treatments beside one another under broadly similar conditions. Lind published the comparison in 1753 in A Treatise of the Scurvy, and it made therapeutic comparison part of the story of medical reform.

Paris physicians counted outcomes

In 1835, Pierre Charles Alexandre Louis, a physician of the Paris hospitals and a student of the physician Jean-Baptiste Broussais, published Recherches cliniques sur la saignée considérée comme moyen curatif, in which he compared outcomes across groups of hospital cases to ask whether phlebotomy actually cured the illnesses it was prescribed for. The "numerical method" he and others associated with the Paris clinical school did not end therapeutic controversy, but it challenged the authority of dramatic individual cases by asking whether counted results supported customary treatment.

Epidemiology connected care to population evidence

Nineteenth-century public-health investigation also mattered. In 1854, the physician John Snow mapped the cholera deaths around London's Broad Street pump and showed that the outbreak clustered around a shared water source, an argument that carried force even though the microbial mechanism was not yet established. Careful mapping, comparison, and inference could overturn accepted miasma explanations before all mechanisms were settled. That tradition connected clinical evidence to the broader history of epidemiology.

Trial Culture

Randomized trials changed what counted as reliable proof

The twentieth century gave therapeutic comparison new institutional force. Statistical methods, state-supported research, wartime and postwar public health, and stronger drug regulation all encouraged more formal testing. The crucial change was not simply that doctors observed more carefully. It was that investigators increasingly designed studies to reduce bias before treatment began.

Random allocation became central because it helped prevent investigators from placing sicker or healthier patients into one treatment group. The Medical Research Council's streptomycin trial for pulmonary tuberculosis, reported in the British Medical Journal in 1948, became a famous example of this new culture of planned comparison. It randomized 107 patients: 55 to streptomycin with bed rest and 52 to bed rest alone. It was designed with the advice of the MRC's statistician Austin Bradford Hill. It did not invent every element of modern trial design, but it helped make randomization, defined eligibility, follow-up, and statistical interpretation models for later clinical research. Later critics also noted its ethical tension: the control group received no streptomycin even as the drug's benefit became apparent, a problem that shaped later rules about when a no-treatment control is acceptable.

The 1954 field trial of the inactivated polio vaccine, organized in the United States by the National Foundation for Infantile Paralysis, enlarged the scale of what a trial could be. About 1.8 million children were enrolled, and the results, announced on April 12, 1955, showed the vaccine to be safe and effective. The trial demonstrated that randomized comparison could be organized at a national scale and that its results could immediately change public health policy.

Bradford Hill also shaped the standards by which non-experimental evidence was judged. His 1965 lecture "The environment and disease: association or causation?" argued that causation in medicine had to be inferred from a body of consistent evidence rather than from any single study, a framework that remains central to epidemiology and to the appraisal of evidence that trials cannot provide.

This mattered for medicine because it altered the moral and practical status of uncertainty. If physicians genuinely did not know which treatment was better, then a carefully organized trial could be presented as a responsible way to learn rather than as a departure from care. That claim would later become central to research ethics, trial regulation, and the acceptance of controlled studies in ordinary medical practice.

1747: Lind compares scurvy treatments aboard HMS Salisbury; the comparison is published in 1753 and later remembered as an early landmark in controlled therapeutic comparison.
1835: Louis's Recherches cliniques sur la saignée challenges bloodletting by comparing outcomes across groups of hospital cases.
1854: John Snow's mapping of the Broad Street cholera outbreak ties a disease cluster to a shared water source.
1948: The MRC streptomycin trial for pulmonary tuberculosis, designed with Austin Bradford Hill, helps establish randomized trial design as a model of therapeutic evidence.
1954-1955: The Salk polio vaccine field trial enrolls about 1.8 million children; results announced April 12, 1955, show the vaccine safe and effective.
1965: Bradford Hill's lecture on association and causation sets out nine viewpoints for considering whether a statistical association may be causal.
1972: Archie Cochrane's Effectiveness and Efficiency criticizes medical practice for failing to organize reliable evidence about what works.
1992: The Evidence-Based Medicine Working Group's JAMA paper popularizes the modern phrase and teaching program.
1993: The Cochrane Collaboration is founded in Oxford to maintain systematic reviews of randomized trials.
1996: Sackett and colleagues define evidence based medicine in the BMJ as the explicit, judicious use of current best evidence for individual patients.

Cochrane and Review

Archie Cochrane made effectiveness a public problem

Evidence based medicine also grew from dissatisfaction with how medical systems used research once studies had been completed. The problem was not only producing trials. It was finding, judging, summarizing, and applying them without letting habit or prestige decide the answer.

Cochrane criticized unsupported practice

British epidemiologist Archie Cochrane argued that health services should pay closer attention to evidence about effectiveness. His 1972 book Effectiveness and Efficiency: Random Reflections on Health Services, published by the Nuffield Provincial Hospitals Trust, became influential because it tied clinical uncertainty to public responsibility: if resources were limited, medicine needed better ways to know which interventions actually helped. Cochrane's own experience as a doctor with scarce resources during the Second World War shaped his insistence that scarce resources should go to treatments proven to work.

Systematic reviews changed the scale of evidence

A single trial could be misleading or too small. Systematic review methods developed to gather studies according to explicit criteria, assess their quality, and synthesize results. In the 1980s, work in perinatal medicine at Oxford, led by Iain Chalmers and collaborators, reviewed controlled trials of interventions in pregnancy, childbirth, and the neonatal period, and showed how organized reviewing could reveal patterns that scattered reports left unclear. The resulting two-volume synthesis, Effective Care in Pregnancy and Childbirth (1989), and the Oxford Database of Perinatal Trials helped demonstrate how systematic review could inform care and reveal gaps in evidence.

The Cochrane Collaboration institutionalized the ideal

Founded in 1993 in Oxford, the Cochrane Collaboration formalized the perinatal review work into an international network that maintains systematic reviews of randomized trials. Its name linked the new review movement to Cochrane's demand that medicine become more accountable about effectiveness.

McMaster Movement

The 1990s turned evidence into a clinical teaching program

The modern phrase "evidence based medicine" is most closely associated with McMaster University in Canada and with clinicians including Gordon Guyatt, David Sackett, and their colleagues. In 1992, the Evidence-Based Medicine Working Group published "Evidence-based medicine. A new approach to teaching the practice of medicine" in JAMA. They presented evidence based medicine as a new approach to clinical learning: ask answerable questions, search for the best available evidence, critically appraise it, and apply it to the care of individual patients.

This teaching program did not say that clinical experience was useless. Its claim was sharper and more unsettling: unsystematic experience and expert opinion should not sit at the top of medical authority simply because they were senior, customary, or confident. Research design, magnitude of effect, bias, and applicability had to be examined explicitly. In 1996, Sackett and colleagues gave the movement its standard definition in the BMJ: evidence based medicine is "the conscientious, explicit, and judicious use of current best evidence in making decisions about the care of individual patients," combined with clinical expertise and the patient's own circumstances and values.

The movement fitted a broader late twentieth-century medical world shaped by expanding journals, computerized databases, cost pressure, medical specialization, patient advocacy, and concern about variation in care. It also depended on changes in medical education, because evidence based medicine had to be taught as a set of habits: framing questions, reading papers, interpreting statistics, and judging whether research findings applied to the person in front of the clinician.

Debate

Evidence based medicine was influential because it was contested

The rise of evidence based medicine did not end argument about medical judgment. It reorganized that argument around hierarchies of evidence, guideline authority, patient individuality, and the place of professional experience.

Trial evidence did not answer every clinical question

Randomized trials are powerful for some questions, especially comparisons of interventions, but they cannot by themselves settle all problems of diagnosis, prognosis, causation, patient preference, or local feasibility. Critics warned that a narrow reading of evidence could make medicine appear more certain than it was.

Population results had to meet individual patients

Evidence based medicine worked with averages, probabilities, and measured outcomes, but clinical care still occurred one patient at a time. Its most careful defenders therefore emphasized integration: research evidence, clinical expertise, and the values and circumstances of the patient had to be considered together.

Guidelines could clarify care or harden authority

Evidence based guidelines promised to reduce ineffective variation and make recommendations more transparent. They also raised concerns about bureaucracy, cost control, industry influence, and the risk that complex judgment might be reduced to compliance with simplified rules.

Legacy

A new language for medical authority

Evidence based medicine did not create modern medicine by itself. It inherited older traditions of observation, experiment, public health, hospital statistics, record keeping, and trial design. Its novelty was to gather those traditions into a clinical ethic: medical decisions should be open to question, supported by the best available evidence, and clear about the strength and limits of that evidence.

Its legacy can be seen in journal clubs, clinical guidelines, systematic reviews, evidence hierarchies, critical appraisal teaching, and the expectation that claims about treatment should be traceable to methods rather than reputation alone. That expectation changed the tone of medical authority. A recommendation now had to answer not only "Who says so?" but "What kind of evidence supports it, and how well does that evidence fit this case?"