Yes, which, by itself, would not be a problem (of course, you would need to watch closely for conflicts of interest clouding the issue). No, the real problem is that the drug companies do not have to release all of the data associated with the trials they perform. They don't even have to publish all of the trials they perform.
Given that, I have a very hard time trusting any drug that does not show amazingly stellar clinical trial results. If you have the options of cherry picking what you publish, and your drug only "helps" 50% of the participants, I see that as random noise.
Ben Goldacre[1] is trying to shine a spotlight on this, and is doing a pretty good job. His Ted Talk[2] is well worth watching.
2. http://www.ted.com/talks/ben_goldacre_battling_bad_science