Remember kids... always wear a helmet. (The almighty bikeradar helmet thread)

I genuinely can't be fucked with this thread- but I've tried to respond to the thread so many times in the past hour that I feel I must write something:
Will- you're right to pick me apart for the inaccurate use of stats.

Charlie- you've misunderstood the basics of the trials. These are unblindable, rather like the famous Doll and Hill doctors study, however, any medical statistician will tell you that to remove bias (or innacuracy or whatever you wish to term it) in an acceptable way is simple, using randomisation of the sample as one method. I haven't got the time to go read the cochrane study, but i'd be surprised if every attempt to remove bias wasn't made.
Also the most recent meta analysis I could find easily- ie first page of google- was from 2005, also under the 'Cochrane' title......
Anyway this timewaste of a thread is now on ignore.
The nature of the trials is my point. They are 'unblindable' therefore they can never meet the gold standard set up by John Cochrane when he established the review system.
Most of the criticism of the early trials relates to their failure to remove bias, or successfully randomise the cases. The 2005 edition of the review is a re-edit of the 2000 version after they removed the fictional quotes attributed to Mayer Hillman.
The main reason why the predictions from these 20 year old trials keep being repeated is not that they are the best science but they claim the biggest benefit and so are beloved by the helmet promoters and the press.