Source Count: 14 | Weighted Score: 35 | Source Confidence: [4/5] | Primary Tier: 1 | Last Updated: April 2, 2026
Keywords: evidence-based-medicine, randomized-controlled-trial, cochrane, meta-analysis, systematic-review, blinding, placebo, clinical-trial-design, p-value, intention-to-treat
Category Tags: evidence-based-medicine, clinical-trials, research-methodology, medical-epistemology
Cross-References: X_2_15 — Mind Body Medicine · ZE_1_01 — Ethics Overview · P_1_01 — Philosophy Overview

QUICK SUMMARY

Evidence-based medicine (EBM) — the systematic integration of the best available research evidence with clinical expertise and patient values to guide medical decision-making — was formalized as a paradigm by Gordon Guyatt and David Sackett (McMaster University, 1992) but has roots extending to the first controlled clinical trial: James Lind's scurvy experiment (1747, HMS Salisbury), in which he compared six treatments on 12 sailors sized into pairs, demonstrating that citrus fruits cured scurvy. KEY FINDING The randomized controlled trial (RCT) — the gold standard of clinical evidence — was pioneered by Sir Austin Bradford Hill with the 1948 Medical Research Council Streptomycin Trial for pulmonary tuberculosis, which introduced random allocation of patients to treatment and control groups using sealed envelopes, blinding of outcome assessors, and statistical analysis of results. The RCT revolutionized medicine by replacing "eminence-based" authority with "evidence-based" empiricism. The hierarchy of evidence places systematic reviews and meta-analyses of RCTs at the top, followed by individual RCTs, cohort studies, case-control studies, case series, and expert opinion. The Cochrane Collaboration (founded 1993, named after Archie Cochrane; now "Cochrane") coordinates the production of systematic reviews across all medical specialties — by 2024, the Cochrane Library contained >8,900 systematic reviews. Modern EBM faces challenges including the replication crisis (only ~50% of high-profile clinical research reproduces), publication bias (positive results are ~2× more likely to be published), the limitations of p-value-based reasoning, and tensions between population-level evidence and individual patient needs.

1. VERIFIED CLAIMS (Tier 1 — Peer-Reviewed / Established)

• KEY FINDING The MRC Streptomycin Trial (1948, British Medical Journal): 107 patients with pulmonary tuberculosis were randomly allocated (by sealed envelopes based on random sampling numbers) to streptomycin + bed rest (55 patients) or bed rest alone (52 patients). Result: 4/55 (7%) deaths in the streptomycin group vs. 14/52 (27%) in the control group at 6 months — a statistically significant difference. This was the first RCT reported in accordance with modern standards of randomization, concealment, and statistical analysis.
• James Lind (1747): a Royal Navy surgeon on HMS Salisbury, conducted the first controlled clinical experiment by selecting 12 sailors with scurvy and dividing them into pairs receiving six different treatments (cider, elixir of vitriol, vinegar, seawater, oranges and lemons, or nutmeg/barley). Only the citrus fruit pair recovered rapidly. Published in A Treatise of the Scurvy (1753). Despite this evidence, the Royal Navy did not mandate lemon juice until 1795.
• The evidence hierarchy (formalized by the Oxford Centre for Evidence-Based Medicine, now CEBM): Level 1a — systematic reviews of RCTs with homogeneity; Level 1b — individual RCTs; Level 2 — cohort studies; Level 3 — case-control studies; Level 4 — case series; Level 5 — expert opinion. This hierarchy guides clinical practice guideline development worldwide.
• Publication bias: Easterbrook et al. (1991, Lancet) demonstrated that studies with statistically significant results were ~2.5× more likely to be published than those with non-significant results. The AllTrials campaign (2013, coordinated by Ben Goldacre) advocates for registration and reporting of all clinical trials. The U.S. ClinicalTrials.gov registry (established 2000) now contains >480,000 registered studies.
• Intention-to-treat (ITT) analysis — analyzing patients according to the group they were originally randomized to, regardless of whether they completed treatment — was established as the standard for RCTs by Fisher et al. (1990) to prevent bias from selective dropout. ITT preserves the benefits of randomization and provides conservative estimates of treatment effects.

2. CREDIBLE CLAIMS (Tier 2 — Academic / Debated but Supported)

• Archie Cochrane (Effectiveness and Efficiency, 1972) argued that medical practices should be evaluated by rigorous evidence rather than tradition or authority. His advocacy for systematic evaluation became the intellectual foundation for the Cochrane Collaboration (established by Sir Iain Chalmers, 1993).
• The CONSORT Statement (Consolidated Standards of Reporting Trials, first published 1996, updated 2010): a 25-item checklist and flow diagram for reporting RCTs, adopted by >600 medical journals. CONSORT has significantly improved the quality and transparency of trial reporting.
• Number needed to treat (NNT, introduced by Laupacis et al., 1988): the number of patients who must be treated for one additional patient to benefit (NNT = 1/absolute risk reduction). NNT provides an intuitive measure of clinical significance alongside statistical significance — e.g., statins for primary prevention of cardiovascular disease have NNTs of ~60–100 over 5 years, meaning 60–100 people must take statins for 5 years to prevent one cardiovascular event.
• The replication crisis in clinical medicine: Ioannidis (2005, PLoS Medicine, "Why Most Published Research Findings Are False") argued that the majority of published research findings are likely false due to small sample sizes, small effect sizes, researcher flexibility in analysis, financial interests, and the hot-topic bias of medical journals. This paper (~8,000 citations by 2024) triggered widespread reforms in trial registration, reporting standards, and statistical practice.
• Bayesian approaches to clinical trial design (adaptive trials, Bayesian meta-analysis) are increasingly advocated as supplements or alternatives to frequentist p-value-based analysis, allowing continuous updating of evidence and more flexible trial designs. The RECOVERY trial (COVID-19 treatment evaluation, UK, 2020) demonstrated that adaptive platform trials can evaluate multiple treatments simultaneously and produce results rapidly.

3. SPECULATIVE CLAIMS (Tier 3 — Possible but Unverified)

• Whether EBM can be fully applied to complex, individualized conditions (chronic pain, mental health, multimorbidity) where standardized trial populations poorly represent typical patients is an ongoing debate about "external validity" vs. "internal validity."
• Whether AI-assisted analysis of electronic health records can replace or supplement RCTs for some clinical questions (real-world evidence, emulated trials) is a promising but unproven approach with significant confounding risks.

4. DUBIOUS CLAIMS (Tier 4 — No Credible Source / Contradicted by Evidence)

• Claims that EBM means "cookbook medicine" (mechanically following guidelines without clinical judgment). Sackett (1996) explicitly stated that EBM integrates three components: best evidence, clinical expertise, and patient values — it is not a rigid algorithm.
• Claims that alternative medicine treatments are valid despite repeatedly failing RCTs, on the grounds that RCTs are "inappropriate" for evaluating holistic therapies. While trial design may need adaptation, the fundamental requirement for controlled comparison cannot be abandoned without sacrificing the ability to distinguish effective treatments from ineffective ones.

Counter-Arguments & Criticisms

Against EBM: Greenhalgh et al. (2014, BMJ, "Evidence Based Medicine: A Movement in Crisis?") identified five threats to EBM: the drug and device industry distorting evidence, overwhelming volumes of evidence, statistically significant but clinically marginal benefits, inflexible guidelines, and under-powered evidence for multimorbidity. Some argue that EBM privileges a narrow definition of "evidence" that marginalizes clinical experience and patient narrative.

For EBM: Before EBM, medical practice was dominated by eminence, tradition, and unsystematic observation — leading to widespread adoption of ineffective or harmful treatments (radical mastectomy for breast cancer, routine tonsillectomy, prolonged bed rest after heart attack). EBM's insistence on rigorous evaluation has demonstrably improved patient outcomes.

IMAGES

No images assigned yet.

BIBLIOGRAPHY

1. Medical Research Council | 1948 | "Streptomycin Treatment of Pulmonary Tuberculosis" | British Medical Journal | ∅ | 2.4582::769–782 | ∅ | ∅ | doi:10.1136/bmj.2.4582.769 | ∅ | ∅ | ∅
2. Sackett, David, William Rosenberg, J | 1996 | "Evidence Based Medicine: What It Is and What It Isn't" | British Medical Journal | ∅ | 312.7023::71–72 | A | ∅ | doi:10.1136/bmj.312.7023.71 | ∅ | ∅ | Muir Gray, et al
3. Cochrane, Archie | 1972 | ∅ | Effectiveness and Efficiency: Random Reflections on Health Services | ∅ | ∅ | London: Nuffield Provincial Hospitals Trust | ∅ | isbn:9780900574177 | ∅ | ∅ | ∅
4. Ioannidis, John. e124 | 2005 | "Why Most Published Research Findings Are False" | PLoS Medicine | ∅ | 2.8:: | ∅ | ∅ | doi:10.1371/journal.pmed.0020124 | ∅ | ∅ | ∅
5. Lind, James | 1753 | ∅ | A Treatise of the Scurvy | ∅ | ∅ | Edinburgh: Sands, Murray, and Cochran | ∅ | ∅ | ∅ | ∅ | Reprinted: Edinburgh University Press, 1953
6. Schulz, Kenneth, Douglas Altman; David Moher. c332 | 2010 | "CONSORT 2010 Statement: Updated Guidelines for Reporting Parallel Group Randomised Trials" | British Medical Journal | ∅ | 340:: | ∅ | ∅ | doi:10.1136/bmj.c332 | ∅ | ∅ | ∅
7. Hill, Austin Bradford | 1951 | "The Clinical Trial" | British Medical Bulletin | ∅ | 7.4::278–282 | ∅ | ∅ | doi:10.1093/oxfordjournals.bmb.a073928 | ∅ | ∅ | ∅
8. Greenhalgh, Trisha, Jeremy Howick; Neal Maskrey. g3725 | 2014 | "Evidence Based Medicine: A Movement in Crisis?" | British Medical Journal | ∅ | 348:: | ∅ | ∅ | doi:10.1136/bmj.g3725 | ∅ | ∅ | ∅
9. Easterbrook, Philippa, Ramana Gopalan, Jesse Berlin; David Matthews. . )90201-Y | 1991 | "Publication Bias in Clinical Research" | Lancet | ∅ | 337.8746::867–872 | ∅ | ∅ | doi:10.1016/0140-6736(91 | ∅ | ∅ | ∅
10. Laupacis, Andreas, David Sackett; Robin Roberts | 1988 | "An Assessment of Clinically Useful Measures of the Consequences of Treatment" | New England Journal of Medicine | ∅ | 318.26::1728–1733 | ∅ | ∅ | doi:10.1056/NEJM198806303182605 | ∅ | ∅ | ∅
11. RECOVERY Collaborative Group | 2021 | "Dexamethasone in Hospitalized Patients with Covid-19" | New England Journal of Medicine | ∅ | 384.8::693–704 | ∅ | ∅ | doi:10.1056/NEJMoa2021436 | ∅ | ∅ | ∅
12. Guyatt, Gordon, David Sackett, Jack Sinclair, et al | 1993 | "Users' Guides to the Medical Literature" | JAMA | ∅ | 270.17::2096–2097 | ∅ | ∅ | doi:10.1001/jama.1993.03510170086037 | ∅ | ∅ | ∅
13. Goldacre, Ben | 2012 | ∅ | Bad Pharma: How Medicine is Broken, and How We Can Fix It | ∅ | ∅ | London: Fourth Estate | ∅ | isbn:9780007350742 | ∅ | ∅ | ∅
14. Howick, Jeremy | 2011 | ∅ | The Philosophy of Evidence-Based Medicine | ∅ | ∅ | Chichester: Wiley-Blackwell | ∅ | isbn:9781405196673 | ∅ | ∅ | ∅

CROSS-REFERENCE INDEX

Generated from V4 expansion plan. Last Updated: April 2, 2026

Corrections

• Effectiveness and Efficiency: Random Reflections on Health S — ISBN corrected from 9780900574172 to 9780900574177, verified against Open Library (Effectiveness and efficiency: random reflections on health services, A. L. Cochrane). The previous number failed its check digit.