Visualization of data is great. And sometimes it tells us something…though we don’t always know what. Slate has an interactive feature showing the rise of diabetes in America by county. Nothing too surprising.
But follow the gradient from El Paso to the Illinois-Missouri border. The differences are small across state lines, but the consistent differences along the borders really don’t make. Are there state-level policies or regulations causing this? Or, are there state-level differences in measurement? This weird pattern shows up in other CDC data I’ve seen.
Update: I think the mystery is solved in the comments:
Very interesting. I suspect the answer has to do with the manner in which the county estimates are produced. I went to the original data source, the CDC, and then to the relevant FAQ:
There they say that the diabetes prevalence estimates come from the “CDC’s Behavioral Risk Factor Surveillance System (BRFSS) and data from the U.S. Census Bureau’s Population Estimates Program. The BRFSS is an ongoing, monthly, state-based telephone survey of the adult population. The survey provides state-specific information”
So the CDC then uses a complicated statistical procedure (“indirect model-dependent estimates” using Bayesian techniques and multilevel Poisson regression models) to go from state to county prevalence estimates. My hunch is that the state level averages thereby affect the county estimates. The FAQ in fact says “State is included as a county-level covariate. ”
This is just a guess, but I think it is quite possibly the answer. (I should note that I looked very briefly at maps of unemployment by county and did not see the same pattern; county unemployment rates have to be estimated because there is not enough data at the county level in a survey of 60000 households, but perhaps the BLS does not use state covariates.)
Quite a lot of national data is collected by local agencies, which normally use standards (and budgets!) set by their parent, state-level agencies. Thus I would lean towards differences in state-level data-gathering or data-reporting policies as the likely explanation.
I’m going to bookmark this as something to come back to later. I’d like to see how stark the differences across state boundaries are with a more continuous scale.
Chris
Oklahoma’s boarder sticks right out also
http://rxnm.wordpress.com/ miko
Different data sources, collection methods, manipulation in state reporting of county stats?
Very interesting. I suspect the answer has to do with the manner in which the county estimates are produced. I went to the original data source, the CDC, and then to the relevant FAQ:
There they say that the diabetes prevalence estimates come from the “CDC’s Behavioral Risk Factor Surveillance System (BRFSS) and data from the U.S. Census Bureau’s Population Estimates Program. The BRFSS is an ongoing, monthly, state-based telephone survey of the adult population. The survey provides state-specific information”
So the CDC then uses a complicated statistical procedure (“indirect model-dependent estimates” using Bayesian techniques and multilevel Poisson regression models) to go from state to county prevalence estimates. My hunch is that the state level averages thereby affect the county estimates. The FAQ in fact says “State is included as a county-level covariate. ”
This is just a guess, but I think it is quite possibly the answer. (I should note that I looked very briefly at maps of unemployment by county and did not see the same pattern; county unemployment rates have to be estimated because there is not enough data at the county level in a survey of 60000 households, but perhaps the BLS does not use state covariates.)
Heidi
I agree with Anthony. It’s important to remember that different states have different surveillance systems. States with strong surveillance systems and strict reporting requirements will “catch” more cases of diabetes than those with weaker systems.
Clark
So that weird blip in Arizonan and New Mexico is due to the Navajo having a disproportionate amount of diabetes? Anyone know why its incidence is so high in the four corners area?
isamu
Injuns are known to have high rates of diabetes. All the blips out west and in the eastern part of OK are due to reservations. The southeast has high rates due to blacks (note the orange swath of Appalachia). The only area I can’t figure is West Virginia/East Kentucky.
http://blogs.discovermagazine.com/gnxp Razib Khan
The southeast has high rates due to blacks (note the orange swath of Appalachia). The only area I can’t figure is West Virginia/East Kentucky.
are you trying to pretend to be stupid? or are you genuinely that ignorant? there’s lots of intra-white variance in rates of morbidity in this country:
are you trying to pretend to be stupid? or are you genuinely that ignorant? there’s lots of intra-white variance in rates of morbidity in this country:
Well, probably it’s a stupid remark but if you compare the racial map and this map, the areas with a lot of blacks overlap the high rate of diabetes.
http://blogs.discovermagazine.com/gnxp Razib Khan
Well, probably it’s a stupid remark but if you compare the racial map and this map, the areas with a lot of blacks overlap the high rate of diabetes.
i know that, don’t assume i’m a retard. the original facts about discrepancies in diabetes rates by race are on the mark. but confusion as to patterns easily explained by differences across regions irrespective of region. if one doesn’t immediately consider that, one is either stupid or ignorant. there is no sin in stupidity, but there is also no palliation. so i hope that the person was just ignorant 😉
no more discussion about this topic. i already offered my judgement as to the original comment.
I am amazed that diabetes levels are that high. No wonder they called it the silent epidemic.
Katharine
As usual, the South is the fount of nastiness in the country.
http://blogs.discovermagazine.com/gnxp Razib Khan
watch it katherine.
a) the south is the fount of nastiness
b) black people in the USA have roots in the south
c) so you’re racist?
though the real south (not athens, austin, etc.) is culturally interesting, though i have the same biases as most people from northeastern snowland.
http://washparkprophet.blogspot.com ohwilleke
Most likely the surveys break out states by metro areas and use intrastate regions for non-metro areas, except for places outside the county reporting system (like Indian Reservations), at least as a first order approximation.
http://ironrailsironweights.wordpress.com/ Peter
Injuns are known to have high rates of diabetes. All the blips out west and in the eastern part of OK are due to reservations.
Alaskan Natives don’t seem to be affected in the same manner. The heavily Native parts of northern and western Alaska are only modestly higher than the rest of the state.
http://blogs.discovermagazine.com/gnxp Razib Khan
wut do they eat? have they shifted to processed carbs?
http://rturpin.wordpress.com Russell
Some of the state borders that are most obvious are those with very low population densities, such as the boundary between Texas and New Mexico. Oldham county, Texas, has 2,185 residents, per the 2000 census. Just slightly more than 1 per square mile. Yoakum county has 7,322 residents. These low numbers will amplify a variety of anomalies caused by differences in data collection.
Paul Ó Duḃṫaiġ
Folks I think we are missing the real elephant in the room here. It’s not due to race it’s down to levels of Poverty/low income. They often point to a correlation between low-income and an increase in diabetes/obesity due to low quality food etc. The areas with rates above 10% are probably among the poorest/most deprived in the US (well I’m assuming that, as an Irish person I could be talking out of my arse in this regards).
If any of ye have seen the South Park KFC episdoe you’ll understand 😉
Eric
It interests me that there’s a sharp line between northern WI and the UP of Michigan. I’ve spent significant amounts of time in both areas and demographically and culturally they’re very similar (norhter WI may have a few more tourists from IL, but I’d guess they’re not factored in). The accent is slightly more pronounced in da yoop. But that shouldn’t account for an uptick in diabetes.
Sign up to get the latest science news delivered weekly right to your inbox!
Gene Expression
This blog is about evolution, genetics, genomics and their interstices. Please beware that comments are aggressively moderated. Uncivil or churlish comments will likely get you banned immediately, so make any contribution count!
About Razib Khan
I have degrees in biology and biochemistry, a passion for genetics, history, and philosophy, and shrimp is my favorite food. In relation to nationality I'm a American Northwesterner, in politics I'm a reactionary, and as for religion I have none (I'm an atheist). If you want to know more, see the links at http://www.razib.com
MENU
Pingback: Tweets that mention Diabetes stops at the state line! | Gene Expression | Discover Magazine -- Topsy.com()
Pingback: Diabetes in America « Anti-Kryptos – Against the Hidden()