Review of publication bias in studies on publication bias: Mandatory publication of data may helpBMJ 2005; 331 doi: https://doi.org/10.1136/bmj.331.7517.638-b (Published 15 September 2005) Cite this as: BMJ 2005;331:638
Editor—Glymour and Kawachi suggest a modest proposal to reduce publication bias, reviewed by Dubben and Beck-Bornholdt.1 Their suggestion is indeed modest because many statistically insignificant or negative studies are not necessarily submitted to a medical journal for publication.
The only effective solution would be mandatory registration of all trials in a centralised website before the trial starts, with mandatory publication of the raw data on that centralised website irrespective of whether the trialists submit their results to a medical journal for publication. That solution would allow meta-analysts to collect data from all relevant studies and not only studies published in medical journals.
Competing interests None declared.