Dangerous drugs getting by is not the only issue, and as you point out, rarely happens. What does happen often is mediocre or simply ineffective treatments that are lauded due to financial incentive. This can happen even in highly controlled trials by over exaggerating moderate to mild findings to a public that either doesn’t have access to or doesn’t understand the literature. Also, all methods of data analysis have their flaws, and there are plenty of ways to both incidentally and purposefully end up with significant findings where there are none in reality.

And I’m glad you mention how the replication crisis is worse in psychology. I’m currently in graduate training to be a research psychologist, and the first month of our clinical research methods class has had a continual discussion of scientific integrity, how to identify flawed or overblown findings, and the danger (and reality) of financial incentive’s effects on research.