Intended for healthcare professionals


Many reports of RCTs give insufficient data for Cochrane reviewers

BMJ 1999; 319 doi: (Published 24 July 1999) Cite this as: BMJ 1999;319:257
  1. E Haydn Walters, Professor and director (Haydn.Walters{at},
  2. J A E Walters, Lecturer (Julia.Walters{at}
  1. Department of Respiratory Medicine, The Alfred and Monash Medical School, Prahran, Melbourne, Victoria 3181, Australia
  2. Department of Respiratory Medicine, Monash Medical School, Prahran, Melbourne, Victoria 3181

    EDITOR—We endorse the call in Editor's choice in the issue of 31 October for improvement in the performance of clinical trials. As part of the Cochrane Airways Group we are carrying out a systematic review of trials on the use of β agonists in asthma. In many of our selected publications we have found it difficult to tell whether allocation of treatment was adequately concealed.

    The reviews undertaken for the Cochrane Collaboration aim to include only those studies in which selection bias has been minimised. For our review of regular short acting β agonists versus use as required we identified 30 studies that were randomised controlled trials Only five gave sufficient information for us to say with certainty that there had been adequate concealment. Many of the others, however, had been carried out in major centres of high quality research where they almost certainly did use accepted randomisation methods, details being held by a third party. Editors of journals should insist that, rather than giving the general statement that the design was randomised and double blind, reports should give a short description of the randomisation method used.

    A second difficulty we have faced has been authors' tendency not to publish their results in a form that is directly usable by reviewers. In our series we have been able to extract fully all the data on reported outcomes in only six of the 30 papers; 15 yielded none, because what was presented was derivative (such as the change from baseline) or merely the P value for some statistical comparison. It would seem sensible for editors to ask for the basic data to be published, made available on request, or made available on the internet. When we or other readers write and request the basic data later they can be difficult to access or the people who were originally responsible for them may have moved. The lack of response to such requests has been a problem for some Cochrane reviewers.

    Although perfection may be impossible, editors of journals could make a positive impact by insisting on more appropriate criteria for the presentation or availability of data, with a specific view to future systematic reviews.

    View Abstract

    Log in

    Log in through your institution


    * For online subscription