My statement above is asking how to test whether the treatments are contributing to a social contagion effect. If they are, we can't test treatment vs. no treatment at a snapshot in time because the very existence of the treatment adds to the contagion effect which in turn produces suicides in treated individuals! To make sure social contagion is not the bigger factor we want to look at how these disorders manifested prior to the existence of the treatment and compare that to the effect of the treatment now.