Kristian G. AndersenGeverifieerd account

I think those points are made, but the problem is that then a whole new set of problems are introduced - as many as you're trying to clarify. The NYT articles lays out the issues very well - and Adam has posted a lot of good threads on the issues with CFRs.

I’m puzzled. Can you point to specific errors or issues so we can correct/clarify as needed? The article explicitly states we won’t know CFR for a while bc of under-reporting cases, lag time to deaths, and also that regional variation is likely.

These are same points made in the Times article so I am not sure what the discrepancy is.

Thanks Julia for reaching out to try and clarify. I unfortunately can't point out every mistake as there are many. The main issue is that the rates are calculated incorrectly - see here for how to more accurately calculating CFRs: https://www.worldometers.info/coronavirus/coronavirus-death-rate/ … 1/n

The most important point is that you don't take epidemic growth into consideration (which will increase the CFR). You correctly identify the role of asymptomatics and mild cases (decreasing the rate), but those are not part of CFRs - they're part of infection fatalities. 2/n

Given an exponential increase in the number of cases (which is what we observe in many places), not taking into account epidemic growth is going to dwarf out any effect milder cases will have on rates - especially since the frequency of asymptomatics appear to be low. 3/n

So statements like this "South Korea, for example, where they are testing thousands of people every day, they’ve picked up more than 6,088 people with the virus. Among those, 35 have died. That’s a case fatality rate, for the moment, of around 0.5 percent" 4/n

ends up being wrong - because epidemic growth isn't taken into consideration. The CFR (while much too early to estimate accurately) actually trends at ~3-4% in South Korea when you take epidemic growth into consideration - not 0.5%. 5/n