%0 Journal Article %T Can you believe what you read in the papers? %A Mike Clarke %J Trials %D 2009 %I BioMed Central %R 10.1186/1745-6215-10-55 %X There are ever increasing numbers of papers available in healthcare journals, and even more articles appearing in newer media such as the Internet. At first sight, the depth and breadth of this material might mean that people making decisions about their own care or that of others have never had it so good. Surely, they will be able to find research in the relevant topic area. They will. But the problem is: some of this research might not be reliable and the decision maker might not be able to find a sufficiently unbiased collection of the research to help her to make the right decision.For decisions about the effects of health care, randomised trials should boost the chances that comparisons are not confounded by factors other than the interventions being compared. They are, therefore, a more reliable guide for estimates of the differences between the actual interventions [1]. However, the problem of publication bias means that trials which have findings that do not favour the experimental intervention are less likely to be published quickly or at all [2,3], making the available literature a potentially biased and unreliable source of knowledge.It is possible that the recent growth in the number of trial reports being published is a sign that publication bias is being overcome. During this first decade of the twenty-first century, at least 25,000 reports of randomised or controlled trials have been published each year [4]. However, we will not know if this is a fair reflection of the volume of research being done until recent initiatives on widespread trial registration provide a means of tracking large cohorts of trials over time, and there may still be some way to go before all trials are registered prospectively. For example, the World Health Organisation's International Clinical Trials Registry Platform shows that nearly 20,000 trials were registered in the constituent registers in 2008, an increase of more than 4000 compared to 2007 [4]. However, it will only %U http://www.trialsjournal.com/content/10/1/55