Excess mortality is not an easy thing to measure, but some of these numbers are truly startling. An extra 200 people dying for every 100,000 is a lot.

That said, I found this article to be confusing. For one thing, it's not just bitter middle-aged undereducated people who drink and pop Vicodin: apparently ALL white people aged 22 - 64 are dying at greater than expected rates. The gap starts getting really big around age 45, which is about the age when everyone starts getting a lot more heart disease and cancer. But I couldn't tell whether the article was saying somehow that all those extra white people are dying because they don't have Obamacare, or they are getting sick slightly earlier due to class-based lifestyle choices, or they aren't married therefore no spouses to nag them about going to the doctor, or what.

My takeaway from the article is that all of those are contributing factors. 

There's no single reason. There are many reasons. 

Another contributing factor: Living in the South. 

Not only are middle-aged white people drinking more, using more opioids, and killing themselves at higher rates, more of them are getting sick with the diseases that usually kill older people. And when they do get sick, they don’t get better.

This trend was especially concentrated in the South, they found. “In seven southern states—West Virginia, Mississippi, Oklahoma, Tennessee, Kentucky, Alabama, and Arkansas—the gap between actual and expected mortality in 2014 exceeded 200 deaths per 100,000 people. In West Virginia, mortality rates were higher than at any time since 1980,” they write. The report was not published in a peer-reviewed journal, and its raw data was not made available.

Some statisticians have been digging into this since the NYT Upshot blog did an initial post back in November. Here's the latest comment I found:  http://andrewgelman.com/2016/01/19/death-trends-update-its-all-about-women-in-the-south/

Fascinating. I had not considered this fact:

When performing your reverse causal inference, remember that people move, and, as we’ve discussed before, the cohorts are changing. 45-54-year-olds in 1999 aren’t the same people as 45-54-year-olds in 2013. We adjust for changing age distributions (ya gotta do that) but we’re still talking about different cohorts.