Statistics from Altmetric.com
Editor,—Craft1 is right to point out that steps to control selection bias by alternation or randomisation to treatment groups did not begin with the Dutch trials of paludrine in malaria and the Medical Research Council’s (MRC) first trial of streptomycin in pulmonary tuberculosis. However, neither were these important design features introduced with the trial of patulin conducted in the early 1940s, as he suggests. Readers interested in earlier examples of efforts to control biases in treatment comparisons should enjoy visiting the “Controlled trials from history” web site, which has been established by the Royal College of Physicians of Edinburgh and the UK Cochrane Centre. The web site currently contains over a dozen examples of reports of controlled trials published before 1940 in which investigators used randomisation or alternation to create comparison groups (www.rcpe.ac.uk/cochrane).
The design feature of the MRC’s streptomycin trial that makes it a methodological milestone was the care taken by the investigators to prevent knowledge of the allocation schedule among those responsible for recruiting patients to the trial. Empirical evidence has shown just how important this precaution against biased allocation is, and how difficult it is to achieve if the allocation schedule is based on alternation.2
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.