Participants in research: Routine extrapolation of randomised controlled trials is absurd

doi:10.1136/bmj.331.7508.109-b

BMJ 2005;331:109-110 (9 July), doi:10.1136/bmj.331.7508.109-b

Letter

Participants in research

Routine extrapolation of randomised controlled trials is absurd

The first 150 words of the full text of this article appear below.

EDITOR—For more than a decade it has been an article offaith in evidence based medicine that randomised controlledtrials are "best evidence" and their findings can routinelybe extrapolated to clinical situations.¹ In his editorial Sackett,the founder of evidence based medicine, seeks retrospectivelyto reassure clinicians that this practice was justifiable, butthe accompanying study by Vist et al fails to tackle the question.²

Most randomised controlled trials are not hypothesis testingexperiments but epidemiological comparisons of rival treatmentsthat measure the effect size of the treatment in a sample drawnfrom a population.³ For the measurement to be precise, a sampleneeds to be sufficiently large; and to be extrapolated to thepopulation, a sample needs to be representative (a census ofall cases or a random sample).⁴ Problems of small samples canbe overcome by megatrials, but ever larger randomised controlledtrials and meta-analyses . . . [Full text of this article]