Medical studies with striking results often prove false A statistical analysis finds that study results showing a 'very large effect' rarely hold up when other researchers try to replicate them.

http://www.latimes.com/news/science/la-sci-medical-studies-uncertain-20121024,0,5207372.story