I did understand your point and I maintain my contention. On what are you basing the claim that there is "substantial previous work and knowledge" supporting efficacy of lithium at the low concentrations one would expect to see in those who drink contaminated tap water? The scientific literature generally has not found benefit to bipolar patients having lithium serum concentrations below 0.4 mmol/L - at the very least. (Do a PubMed/MedLine search!) Any claim that serum concentrations orders of magnitude less will prevent suicide would need much further investigation before being accepted as fact, even in the context (and perhaps especially in the context) of past research.

Actually, that's incorrect and reinforces that the correlationisnotcausation tag should still be used for these articles, because obviously even us intelligent Slashdotters can get confused!

While one might be able to say that lithium causes decreased suicides based on other data from other studies, one cannot infer from this study that the low lithium levels in the water (which would presumably not cause lithium blood levels in the range currently targeted for therapy) are what's responsible for the decreased rates. One can only claim correlation.

This is due to the possibility of confounding variables (for example, do wealthier neighborhoods tend to have more pure drinking water?).

Even if you control for all the factors that you know about that could cause such misinterpretation of the data, there is still a very real possibility of an unknown confounder (e.g., could there be another substance in the water that tends to track with lithium levels, and could it be that other substance that is the suicide-protective agent?)

Thanks for the advice. One question, though- is there any reason to choose CVS instead of SVN, other than a case in which a project is already using it?