Yeah, and that made me wonder whether survivorship bias could be at play. That is, populations at lower elevations see less opportunities for suicide due to individuals dying of things like heart disease, etc., before they have a chance to off themselves at a later age.
But they do mention that they "[controlled] for percent of age >50 yr." Does that imply that, per county, they excluded high-elevation suicides that wouldn't have happened at lower elevations, using known mortality rates for common causes-of-death for age > 50 individuals?
I'm sure the answer is in the paper somewhere, but I need to get back to work. :-)
HN is an echo chamber.