"But, like CFR, we don’t know how high the rates are if we have no idea how many people have or had it."

While it's true that the hospitalization / ICU / mortality rates may be lower because we don't know the full number of cases, it still seems to be an immediate and acute problem.

Regardless of the rates, the absolute number of people being hospitalized in Italy was enough over overrun their healthcare system. The same situation looks to be on track in many European countries (Spain, France, UK).

So, the worst-case scenario may be less bad than predicted (let's certainly hope this is the case!), the number of case's we're identifying and having to deal with is still dangerously high.

> These folks aren’t even that young and they describe it as being quite mild...

Take that with a grain of salt. The people for whom it's not mild aren't going to be giving interviews.

It’s literally survivorship bias.

Whatever they are, they’re high enough to be causing major problems for hospital systems in places like Italy. They seem high enough that many countries desperately need to slow the spread of the virus or they will face dire consequences.

Anecdotes are not sufficient data

They are when there's little quantitative data to work with.

A lot of people's biggest questions are what the hell "COVID-19 symptoms" and apparently it means just about anything.

I'm suspicious of the 2 week long illness my family has just gotten over. Hit each of us differently. Predominantly nasal mucus, but the kids had some chest congestion, wife had a fever, and I had what felt like a sinus infection and a bit of a dry cough.

None of it particularly severe. Likely not COVID-19, but given the scattershot presentation of the disease in non-critical cases, I'm not ruling it out.

No, the symptoms are pretty clear:

Fever, dry cough, fatigue, sputum production, shortness of breath (dyspnoe), muscle/joint pain.

https://ourworldindata.org/coronavirus#the-symptoms-of-covid...

We don’t have sufficient data because we are, in our infinite wisdom, not testing every person with a pulse. Or at least testing from as random of a population subset as we can manage. At least in the US. Other countries mileage may vary.

But even then it is my understanding that there currently is no way to test if you already had it.