raw-al, you are forgetting the massive other side of the coin. That being that we don't have an account for all the mass shootings that might have been prevented because of the efficacy of the psychotropic drugs. I'm actually surprised by the incredibly low prevalence of mass shootings (defined as 4+ or more at one time) in the US being somewhere around 2-3 per year for the last 30 years.
This huge blind spot that is unmeasurable makes it very difficult to determine whether these drugs are causing any harm or actually helping to an enormous degree. I can go further and say that perhaps the shooters you mention above that were these pyschotropic dugs killed less people than they would have where it not for the drugs. Maybe they would have gone into a football stadium or what not? This is obviously just pure speculation but the argument could be made.
As an extreme example, say in the future we invent a drug that will 100% stop mass shootings from occurring, only that it would take 5 days for the drug to kick in and that we could somehow identify every single person that would fit the profile of a person who would be a mass shooter before they became one. The only piece of information we would not have would be the time until the person committed the mass shooting. So, all the people we would identify as mass shooters would, with this drug, be prevented from becoming a mass shooter except for the people that were identified too late (< 5 days). So, inevitably some mass shootings would occur and all of these shooters would be on the drug. There would be a 100% correlation here and because of the absence of evidence of the mass shootings not happening, from an initial look at the data it would appear that the drug caused the mass shootings, but the exact opposite is happening.