Evidence‐based medicine

The concept of evidence-based medicine was first developed and promoted in the 1980s and 1990s by David Sackett, Gordon Guyatt and colleagues at McMaster University. I will use the abbreviation EBM, conscious that to paediatricians, this is just as likely to mean expressed breast milk. Like many major innovations, initial enthusiasm (Fig. 1) was followed soon by a backlash by detractors who blame EBM for many evils. The human condition seems to favour polarisation. Sackett described EBM as ‘the conscientious, explicit and judicious use of current best evidence in making decisions about the care of individual patients.’ EBM stressed the importance of randomised controlled trials. The first randomised controlled trial was probably that performed in 1747 by the Scottish naval physician James Lind, who was later to sail with Captain Cook. In those days, more sailors died from scurvy than in battle. Lind randomised 12 sailors with scurvy into pairs who received one of six different dietary supplements. The pair given citrus fruits recovered. Lind’s original diary describing this experiment is in the library of the Royal College of Surgeons in London. The Cochrane Collaboration formalised the use of systematic reviews of the evidence and particularly of meta-analysis. Metaanalysis in this context refers to a mathematical pooling of the data from multiple randomised controlled trials and is arguably even more ancient than randomised controlled trials. Philosophical meta-analysis has been dated back to 12th century China; meta-analysis was used in astronomy in the 17th century, and medical meta-analysis started with a 1904 paper which combined studies on typhoid immunisation. Some clinicians have argued that EBM stands in opposition to clinical acumen, which was once described cynically as ‘making the same mistakes with increasing confidence over an impressive number of years.’ I disagree. Good clinicians have always used both the best evidence and their clinical experience. EBM has simply formalised the use of evidence and made the evidence more accessible. Databases of medical studies, such as Medline, and search engines to search the database rapidly, such as ‘PubMed Clinical Queries’, allow the modern physician to find and examine the available evidence in less than 10 min. EBM is open to satire. The seven alternatives to EBM, including vehemence-based medicine and eminence-based medicine, were based on colleagues. Every time we showed an early draft of the paper to a colleague, that doctor said they liked them all except one. Invariably, the one they did not like was the one based on them. Another Christmas, BMJ paper invited advocates of EBM to participate in a double-blind, cross-over trial of wearing or not wearing parachutes when jumping from an aeroplane. This paper makes the valid point that a randomised controlled trial may not always be possible, necessary or sensible. No one will ever test penicillin against placebo to treat meningococcal infection. Like all human activities, EBM is fallible. There can be a selection bias: positive studies are more likely to be published than negative ones. Pharmaceutical companies are inherently conflicted. Quite reasonably, they will fund and promote research that favours their product. Unfortunately, however, some pharmaceutical companies have suppressed trial data unfavourable in terms of safety or efficacy to their product. A meta-analysis of flawed or biased data will itself be flawed: rubbish in, rubbish out. On the other hand, antenatal preterm maternal steroids to prevent neonatal respiratory distress syndrome and the supine sleeping position to prevent sudden infant death are prime examples where a timely meta-analysis of available data would have saved lives. Good evidence is also the basis for good ethical decision-making, when the balance of benefits and harms is often critical. EBM advocates do not say that you must always do a randomised controlled trial, only that you should use the best available evidence to assist in clinical decision-making. If there is no high-level evidence, then the sensible clinician bases management on experience and, if in doubt, consults colleagues judiciously. The available evidence needs to be used cautiously and with a healthy scepticism regarding its validity. However, summary dismissal of EBM risks losing current and future gains.