Ethics leads to research bias

[injuryupdate]

Good study in the BMJ. Ethics commmittees, which generally insist on very hard line protocols for research projects (perhaps to justify their existence) end up getting patient populations which are biased. Rather than a random selection of a population, once you have read their rights to patients and made them sign a 20 page consent form, a much smaller and less representative patient population remains. Not surprising but good to see it proven:

http://bmj.bmjjournals.com/cgi/content/full/331/7522/940