How can medical journals help prevent poor medical research? Some opportunities presented by electronic publishing

Dr Iain Chalmers, MSc ,Douglas G Altman, DSc

Medical journals share with other elements of the clinical  research community certain responsibilities to the public  for the quality of medical research. Traditionally, they have  tried to fulfil these responsibilities by assessing the reports  submitted to them for publication and then subjecting  some of these to “treatment” of various kinds before a  final decision is taken about what should be published.  The persisting inadequacies of these “therapeutic”  efforts of journals have been made more explicit with  recent improvments in the scientific quality of reviews of  research.1,2 For example, there have been over 500 reports  of controlled trials of a wide variety of drugs used to  prevent or reduce the unpleasant movement disorders  suffered by people taking antipsychotic drugs. These have  been published over several decades in journals that claim  to use peer review, but recent systematic reviews have  shown that this vast body of research yields no  information upon which practice can be based with any  confidence.3 There is depressing evidence that the  therapeutic efforts of journals remain inadequate; in only  one of the 122 reports of randomised controlled trials of  selective serotonin reuptake inhibitors were adequate  details given about the method of randomisation,4 and  these reporting deficiencies are present in some of the  world’s most prestigious medical journals.5  Many journals are now trying to deal