I did an analysis like this for a multivariate statistics class in 2009.
Death rate due to accidents correlated with state square area.
Generally large state = low population density = less mass transit and more/longer driving, therefore more accidents per capita.
And as we know from the electoral map distortions -- low population density = more likely red.
I ended up plotting cancer deaths on one axis and accident deaths on another. The states that were low on accidental deaths were predominantly blue; high in accidental death were predominantly red both in 04 and 08, but the ones that turned blue from 04 to 08 were more likely to be those with low-to-moderate cancer death rates than those with high cancer death rates.
Don't know if those cancer death rates mean more smoking, less medical care, or more pollution...