Thread by @MidwestHedgie

Link to official Heinsberg coronavirus data summary. This was the area of Germany hardest-hit early. Approximately 2/3 of the families contacted agreed to participate in testing. Preliminary data is from first 500 people tested. 1/ land.nrw/de/pressemitte…

Both PCR (for active infection) and Ab (resolved infection or infection > a few days) were done. 2% pcr positive, 14% Ab positive, total 15%.

If you assume deaths and infections were evenly distributed across the region, total deaths / total infections = 0.37% thus far. 2/

The antibody test is reported as >99% specific, so should be few false positives.

Given roughly 13% (2% of 15%) of observed infections are ongoing, case-fatality rate estimate will probably end up more like 0.4% than 0.37%. 3/

This is somewhat good news. If you avoid a situation where hospitals are overwhelmed, the case fatality rate of COVID-19 is "only" 4x that of flu in these data, meaning even if 70% of Americans were _gradually_ infected, you might expect only 916,000 or so deaths at that CFR. 4/

Of course, case fatality rate varies widely by age, health status, and access to appropriate medical care, so the confidence intervals around that number are wide (especially to the upside). 5/

Some fringe people/fringe media (like the Daily Mail) saying things like "Europe is close to Herd Mortality." No. That's wrong. In a highly-localized epicenter of infection, only 15% of people have been exposed so far. Most of Europe and the U.S. have had much less exposure. 6/

If 15% of Americans had been exposed, the German data suggest we'd see 200,000 deaths. However, we've had only about 20,000.

Ergo, roughly 1.5% of Americans have been exposed to COVID-19 so far.

Whatever policies we take from here need to deal realistically with that.

/fin