Statistics from Altmetric.com
Editor,—We read with interest the recent editorial on evidence-based child health.1 We were struck by how easy it is to overlook a randomised controlled trial (RCT) in theArchives (or any other journal) if this is not indicated in the title or abstract. We wondered if this was the case in commonly used bibliographic databases, such as Medline.
With the help of the Cochrane Collaboration Cystic Fibrosis Group, we retrieved the results of a handsearch for RCTs from theArchives from 1986–90. Each RCT (a trial in which participants were allocated to groups by random allocation, excluding pseudorandom methods such as date of birth or case number allocation) was assessed as to whether or not it contained the word ‘random’ or some variant in the title or abstract/summary section.
These results were compared with the simple (and easily remembered) Medline search strategy of ‘random*’—where * is the truncation character.
The simple search strategy found two thirds (59/90) of the handsearch defined RCTs, of which 78% (46/59) had the word ‘random’ or variants in the title or abstract. However, of the 31 RCTs which failed to be retrieved by this strategy, none (0/31) had ‘random’ or variants in the title or abstract (p<0.001).
We propose that the Archives adopts some of the CONSORT recommendations for the reporting of RCTs,2 by encouraging authors to include the word ‘randomised’ in the title and abstract of all these papers, enabling easier retrieval of useful clinical studies.
The editors comment: We agree and we are currently adopting this proposal.