Keith Humphreys poses a real-life question:
Here is my colleague’s not-as-simple-as-all-that dilemma. He has pilot data showing that a cheap, generic drug has a potentially life saving medical application. He now has a grant from the NIH (i.e., not industry) to study it in a major trial. A for-profit company has a copycat version of the generic drug that is virtually the same but is still on patent and is thus highly expensive. The question is whether he should study the generic version or the proprietary one.
If the experiment proves that the generic version of the drug is effective, he will get a very nice paper in a medical journal. And that will be the end of it. We have a million studies showing that clinical practice is minimally influenced by journal articles alone. In contrast, if he studies the proprietary drug, a profit making company will send out their armies of marketers to physicians and get many of them to use it, making a great deal of money for themselves in the process…but this will also save lives, because aggressive marketing and promotion such as the industry does has been proven to affect clinical practice dramatically.
Italics mine. So what should he do?