Review in trials on reporting bias/witholding data in (industry) medical research - with lots of examples
Well worth reading. http://www.trialsjournal.com/content/pdf/1745-6215-11-37.pdf
The overall message and the recommendations aren't new, but thus is a succinct and informative review by the German health technology asessement body that's worth citing and using in talks. It includes and goes well beyond the usual suspects (rofecoxib, gabapentin, SSRIs, rosiglitazone, oseltamivir) and gives a really nice overview of all the recent attempts to make people behave better eg through trial registration.