Publication bias... again

Publication bias seems like a problem that just won't go away. PLoS Medicine published a paper (doi: 10.1371/journal.pmed.0050217) late last year that indicated that this practice is alive and well in what is probably the most concerning area of all, clinical trials. The paper, by Lisa Bero and colleagues looked at trials submitted to the Food and Drug Administration, and found that a quarter of trials submitted in support of new drugs applications remain unpublished a year later. The study also found that of the published trials, discrepancies between the FDA submission and the published paper tended to lead to presentation of the drug in a more favourable light. What is heartening is that now such analyses do at least tend to be widely discussed — as my colleague Andrew Hyde noted in his round up of coverage of this paper.