>Also, the cruise ship cases aren’t all resolved outcomes yet.

No they aren't, but the vast majority have been symptomatic longer than the average time till hospitalization and the number in critical condition hasn't increased in days. It's very unlikely they'll get to anywhere near 20%. And this is in a more geriatric population.

http://www.cidrap.umn.edu/news-perspective/2020/02/study-720...

About 40% of the patients in that study weren't even tested. They were included based on symptoms that include severe pneumonia. The ones who were tested, sought testing because they displayed symptoms severe enough to seek treatment.

This is the very definition of selection bias, and there is absolutely no reason to pretend that a 20% hospitalization rate is a realistic outcome.

Yes, but the reason why the Diamond Princess is interesting is that we know exactly how many of them got infected. We don't know that about any other population.

Also consider my point that many, many of these cases are unresolved. Whereas the China study, which most health experts are relying on currently as our best guess - not the cruise ship — looks at resolved cases.

The number of undiagnosed cases is the great unknown that experts disagree on. It's also the basis for all the wild speculation out there.

Under normal circumstances I would say let's test a random sample of the population in some affected area. But as test kits are scarce doing that would probably be ill advised.

I also agree that there’s a wide range of possibility. But I’m responding to the claim there’s no data with a study from China’s CDC that most experts are citing as the best info we have right now. Is it perfect? No. But the Diamond Princess isn’t a scientific study — it’s a mess.

Edit: for a range of possible outcomes, this article is helpful:

https://www.forbes.com/sites/kenrapoza/2020/02/25/coronaviru...

At the low end are South Korea’s (mortality) numbers, at the extreme high Iran, and Italy aligning roughly with China’s study.

The tl;dr is we have a fairly wide range of possible outcomes, but Italy is lining up with China’s findings.

But with the flu we have large population studies, so we can estimate the number of likely infected based on the number seeking treatment. They study you are citing can't be generalized, and no one but you is claiming it should be used to estimate hospitalization rate.

Public health officials who are trying to produce estimates are using that study along with models of how many un-diagnosed cases might be out there to try to predict hospitalization rates. But they aren't naively throwing out your 20% number.

>But I’m responding to the claim there’s no data with a study

There is no data to support your claim of a possible 20% hospitalization rate. None, it doesn't exist. There is data to support a 20% hospitalization rate among people sick enough to seek treatment, and people sick enough to be clinically diagnosed.

>at the extreme high Iran, and Italy aligning roughly with China’s study.

Italy and Iran also likely have far more actual cases than confirmed.

>No. But the Diamond Princess isn’t a scientific study — it’s a mess.

The study you are presenting makes no attempt to predict the actual hospitalization rate among the general population. And the samples aren't random. It's more a mess than the numbers from the Diamond Princess.

Everyone on board the ship was tested, no sampling bias there. But even then, there is selection bias because the ship's population is older than the general population.