I agree, and I wasn't quite going for a point to point comparison per se, just made up a story where the obvious, logical answer was ignored. The second logical answer in the case of meds, would be that the meds didn't work, not that they caused an incident. Lastly you would look at the meds based on the percentages of violentmurderrampages/to not (of course a real study would have many, many more variables, and I'm being purposefully reductive for your reading pleasure), I do agree that meds can have adverse effects on individuals, which is why there are so many different types of drugs used for similar problems. Also we can't really keep track of people that are prescribed meds, and lie about taking them, and don't take the right dosages, or skip days here and there, or feed them to their dogs, or sell them to their friends.
Meds are definitely over-prescribed IMO, but I'm not a doctor so what do I know. Someone ask me about computers. Quick.