Just what proportion of trials are kept under wraps is unclear - one of Donald Rumsfeld's known unknowns - but a 2010 study from the National Institute of Health Research suggests that half are never published, and those with positive results twice as likely to get into print.