A new report documents dramatic county-by-county variation in deaths due to alcohol abuse, drug abuse, suicide, and interpersonal violence.

Death due to alcohol, drugs, suicide, and interpersonal violence – sometimes characterized as “deaths of despair” are on the rise in the US, particularly among white males, reversing a centuries-long improvement in life expectancy.

But these deaths are not distributed evenly around the country. In a new analysis appearing in the Journal of the American Medical Association, researchers from the University of Washington used county level data to characterize the extent, and variation, of deaths from these causes.

The study of variation in medicine is more important than it seems at first blush.  High levels of geographic variation in a disease or a treatment strategy suggest that potentially modifiable factors are at play. Put it this way, if I told you that the rate of cholangiocarcinoma was fairly evenly distributed throughout the country, you might conclude that what we’re seeing is more or less a baseline rate, something stochastic, where only really broad changes in diagnosis or treatment will make an impact.

What this study shows us, is that these “deaths of despair” are by no means uniform. In fact, variation is highly pronounced and dramatic.

Let’s look at the rate of deaths associated with drug use nationwide in 2014.

You see a 58-fold range in the death rate from the lowest to the highest counties, with West Virginia and Kentucky particular hot spots.

Looking at the change in death rate from 1980 to 2014 provides a window into the scale of this epidemic.

Deaths due to drug use did not decline over these two and a half decades in a single county in the United States. Not one. And the rate in those hot spot areas has increased fifty-fold.

Moreover, the extent of variation from county to county has increased dramatically. What this tells is that these deaths are not due to some baseline physiology, but rather to local factors. Lead author Dr. Laura Dwyer-Lindgren told me she wasn’t comfortable with the “deaths of despair label”, as they don’t really know what’s driving these changes.

“A lot of the discourse around the just massive increase in deaths from drug use disorders is around the massive increase in deaths specifically from opioids and around how that relates to prescribing and availability”.

We actually have data on the rate of prescribing of opioids on a county level in the US, though not from Dr. Dwyer-Lindgren’s paper.

Here is a map from the CDC looking at opioid prescription rates. Compare this to the map documenting the increase in deaths from drug abuse.

Sometimes a picture is worth 1,000 statistical tests.