Let's use a simple example why one needs to be careful reporting mortality rates (and again I'm using simplifying assumptions but it works generally for our purposes and we can add necessary complications). In the US in 2017-2018 flu season 61,000 people are estimated 1/n
To have died from the flu according to the CDC (link below). So what is the mortality rate? Well there were roughly 1m flu tests with (let's assume they are all positive) positive diagnoses. Is that the mortality rate? No. What about the 810k 2/n cdc.gov/flu/about/seas…
Hospitalizations due to the flu? No. What about the 20.7m doctors visits about the flu? No. The CDC produces an estimate of 44.8m SYMPTOMATIC cases of the flu. Let's extend this thinking to corona. (Again simplifying). Right now everyone getting tested in many places 3/n
We read about are say the hospitalization phase to make the CDC comparison. So all your sampling is already taken from moderate to severe cases. If you will note, as testing broadens the picture rapidly and dramatically improves. To use the CDC case, you are moving from 4/n
Hospitalization to doctor to symptomatic with no medical attention. Even this is under counting the number. In other words, the headline mortality rates are taking moderate to severe cases and then being shocked there are moderate to severe outcomes. There are to final notes 5/n
Mortality rates of the elderly from the flu according to the CDC are...(drum roll) similar as a share of deaths compared to corona. Furthermore, CDC mortality rates are similar when compared hospitalized or accounting for medical testing rates. That is exactly what we would 6/n
expect based upon where observations are being recorded and the type of patient. In other words, while my own read of the data is still that this will come in worse than the flu, expect it to continue to really trend down when set against population numbers.
