Severe methodological problems in a recent review and analysis of the association between lithium concentration in potable water and suicide rates.