They have their hands in just about everything - and they consider anything to be a marketing opportunity.
These companies hire "authors" to write "articles" about their medicines, touting the efficacy/effectiveness of the drugs while minimizing the side effects that might come out. They then publish these "articles" with the backing of the "authors'" name recognition highlighting the positives. The problem is that these "authors" don't write the articles, rather company-paid ghostwriters do. What ends up happening is a 6000 word article written by pharmaceuticals with data-mined information pulling one or two positives from a study and skewing the results in favor of the company paying for the article.
This isn't a new development, though it's arguably a more recent one. Here's a blog post from January 2008 discussing issues with a published article from 2006 highlighting data-mined information. This is but one example, and if you look, you can find several more.
Let's hope Sen. Grassley is serious about his inquiries into this major effort to advertise through "information."