There are a couple of things worth noting here. First of all, I was not aware of this open access medical journal, PLoS. I figure that a few of you might not know of it either, so now you know.
Anyway, I tracked it down from this story on MSNBC.com. The article is on the vast differences in life expectancies across geographical areas of the country. The authors then break down the data into "eight Americas" based on income and race. Hawaii scores well, but the south does very poorly. Of course one can easily criticize this as leaving out many potential other explanatory variables. For example, If industrial pollutants reduce life expectancy of all races but are concentrated in counties of a certain racial makeup, that would be useful to know. (I confess to having no knowledge of whether this would be a significant issue, but an enterprising person with a map of chemical factories could find out in a hurry.) The authors do identify access to health care, even among low-income rural areas of the northern states (e.g. my old stompin' grounds), as being an important positive factor contributing to life expectancy. Food for thought.
For the curious, their life expectancy data is available.
Other articles in this issue of the on-line journal look interesting as well, including a study comparing outcomes in academic medicine (i.e. teaching hospitals) and non-academic medicine.
I commend this journal for making medical studies more accessible to researchers, including those in other fields who will benefit from easy access.
Interesting that I just came across this post, while googling for life expectancy at birth by county. The American Human Development Project reports figures by congressional district, but I dream of relatively recent data/models to the county level - as part of my attempt to craft a Human Security Index for the USA (not national security, but after the 1994 Human Development Report, FDR's 4 freedoms speech, etc.). Fortunately, old posts - and the links from yours in this case, sometimes stay alive.
I recently published a prototype Human Security Index, covering 200 societies.
http://www.unescap.org/publications/detail.asp?id=1345
If you replace the 1345 with 1308 in the link you get another paper, extending the Human Development Index to 230+ economies (~50 more than reported by UNDP).
The stretches (geograhic in the second case, thematic in the first) provided some interesting food for thought - including the question - what might constitute a human security index for the USA, that could be used to crafting improvements (for example in storm warnings and associated advice for response - like evacuating at-risk housing for storm shelters in a manner that would help occupants to take appropriate action).
I haven't yet read deeply enough to see if the cited report and data are actually "data" or "a model" in need of more context - but at least they were locatable.
Thanks for the posting.