AllTrials campaign
Ben Goldacre (Bad Science), is running a campaign for all medical trials to be published.
It seems that best current evidence indicates that about half of the trials for treatments currently in use remain unpublished.
That's pretty astounding when you think about it. Given that many of the trials that are published only compare new treatments against placebo, rather than against the best available existing treatment, and often barely show them as being more effective than homeopathy^Wsugar, one has to wonder how dire were the results that are not being published.
If you prefer to have science done in the open, then I'd encourage you to visit www.alltrials.net and sign the petition and/or give them a donation.
The fix is pretty straight-forward: Simply insist that clinical trials have to register before they start in order to be considered valid, and that all registered trials have to report their results. It need not be expensive to do -- a git repo. shared across the industry, containing a file for each trial in a standard format would do the trick probably. I imagine they'll come up with a more expensive solution, but in comparison to the cost of running a decent sized trial, it'll still be pocket change.
(A note from the campaign, sent to existing petition signatories, is what prompted this blog post -- please read it and be inspired to help)