[Pond, James Pond]

Obviously -- okay maybe not so obviously, but nevertheless I promise -- I'm not that stupid.

I, in no way, meant to imply you are stupid.

My initial point was in response to the expressions of shock at some of the reported adverse events in drug leaflets, and that the presence of that warning may have been due to one single incident that could not be totally excluded as a possible factor at some point in the clinical evaluation process. Some asserted that with this as a warning, then they (drugs) must be to blame for shootings.

At what point do you say that, although other factors clearly are involved, certain drugs increase the risk of suicide, murder, or mass murder, and they increase that risk in a way that the medical establishment can't predict or prevent? Do you never say that, because there are other factors?

Companies, despite licencing and sales, still continue to collect data from patients and institutions to add to their analyses. The issue I see with your scenario is that if there were an affect by a drug that could cause these behaviours, or make people more prone to them, they would manifest in other ways too. Aberrant behaviour manifests in many ways. They would all be pooled in. That would increase the statistical power.

I am not a statistician, by any stretch though: checking my shopping bill gives me a headache, so stats are not my strong suit!