THREAD Covid-19 is spreading from America's cities to its exurbs & rural heartland. Which areas are likely to face the worst of it?

We modeled the disease's fatality rate & find the South & Appalachia are particularly vulnerable. Policymakers, take note. https://t.co/a84Z4eRPdL

— G. Elliott Morris (@gelliottmorris) April 23, 2020

Who we think is vulnerable to covid-19 is based on many factors, including their age, underlying health conditions, access to health care and recent exposure to potentially infected people. We acquired data on these indicators as well as case and death counts for every US county.

— G. Elliott Morris (@gelliottmorris) April 23, 2020

We're primarily interested in who is susceptible to dying from the virus, rather than just contracting it. This is called the fatality rate, & although we only have data on detected (vs all real) cases, we think the CFR is a good approximation for actual infection fatality rate.

— G. Elliott Morris (@gelliottmorris) April 23, 2020

One big barrier to a proper covid model is that testing is not constant across America. Because states that have higher rates of testing officially detect more cases of the virus, they have on average lower case fatality rates (CFRs). But within states, testing rates are similar.

— G. Elliott Morris (@gelliottmorris) April 23, 2020

So we built a model to identify the traits shared by counties with CFRs far above or below their own state’s average—and predict which places not yet ravaged by the virus will suffer most once the disease comes to town.

— G. Elliott Morris (@gelliottmorris) April 23, 2020

We found that geography has been an important predictor of covid-19's CFR early on. That's probably because cities were hit first, when the US was underprepared for the virus, and because rates of transmission appear to be higher there, and hospitals are near or at capacity. pic.twitter.com/JZsEC9zvP3

— G. Elliott Morris (@gelliottmorris) April 23, 2020

But demographics have also predicted county-level CFR's. We find that areas with higher rates of poverty, more African Americans, and higher rates of heart disease and diabetes have been hit harder than others. And, as expected, people that isolated earlier fared better.

— G. Elliott Morris (@gelliottmorris) April 23, 2020

So, which US counties score better or worse on the variables that have so far predicted covid-19's case fatality rate? If every county sees an outbreak—which now seems likely—where will coronavirus extract its largest tols?

— G. Elliott Morris (@gelliottmorris) April 23, 2020

Often, we see people wagering that cities will be hit abnormally hard by covid-19. In fact, if it does infect most Americans, the predicted CFR in coastal cities, high based on population density, is offset by young, healthy, well-off populations and good hospitals.

— G. Elliott Morris (@gelliottmorris) April 23, 2020

The South & Appalachia will likely fare the worst. There, people are more likely to live in poverty, have heart disease or diabetes or smoke and are older than in other areas. Parts of the Midwest with older populations that don't have ICU capacity are also predicted hotspots. pic.twitter.com/R45AnQ4NJM

— G. Elliott Morris (@gelliottmorris) April 23, 2020

These findings present some public policy concerns. People in the South are most vulnerable to covid, yet their leaders have taken it the least seriously. They didn't self-isolate early. Georgia, Tennessee and South Carolina all announced plans this week to relax their lockdowns.

— G. Elliott Morris (@gelliottmorris) April 23, 2020

Hopefully, analyses like these can inform policymakers when they are deciding the best courses of action.

And 👏👏 to @_rospearce for putting together beautiful graphics and to @DanRosenheck for a good edit.

— G. Elliott Morris (@gelliottmorris) April 23, 2020