Seattle Flu Study finds COVID19 likely spreading in Seattle for past 6 weeks | Hacker News Reader

Seattle Flu Study finds COVID19 likely spreading in Seattle for past 6 weeks

twitter.com

36 pointserentz15 comments

This has really hit home for me. I'm pretty sick with a 102 fever, cough, and aches. My guess for various reasons is that it's influenza A, but there are reasons to think it's not crazy to think it's COVID19.

The issue is that there's this kind of positive feedback loop about getting tested. If you have symptoms consistent with COVID19, you're told it's probably the flu and no need to go in and get tested unless you're high-risk. People always say "have there been identified cases of COVID19 near you?" The answer is basically "not that I know of," but if everyone is staying home and not getting tested so as to not panic and overburden the system, you don't know. As a result, it will look like nothing's there until suddenly there's a de novo case that can't be traced to Asia, Europe, or anywhere else that's had large numbers of cases. Then it's too late and is everywhere.

I understand the base rates, and that I probably don't have COVID19, but I don't really have faith in assessments of where things are at if people are rarely getting tested anyway.

> Then it's too late and is everywhere.

So we get the flu. Life goes on.

Later in thread

> Trevor Bedford @trvrb · 14h I'd assess the p-value for this coincidence at 2/59=0.03 and so is statistically significant. Additionally, these two cases are geographically proximal, both residing in Snohomish County. 5/9

That's... really not how p values work. Does everyone have to mention p values now just to try sound like they're authoritative?

I cringed when I read it. At the same time, you can have a meaningful p value from a single observation. If you were assuming novel spread from China and a random selection of virus genotype, you'd expect a coincidence like this to happen about 3% of the time if the null hypothesis held. So it's not completely bullshit.

Big caveat is-- inferring this from a single genetic difference with 59 reference sequences taken at different times is dubious. That 3% number could be drastically different in reality.

Still-- cryptic spread from the earlier case is probably the most reasonable explanation for what we've seen.

This doesn't surprise me at all. Some country (like Japan) is seeing these detected, because they are often seeing a doctor for the common cold. This is very different dynamic in the country like the US where a lot of people don't (or can't afford to) see a doctor for it -- thus I can only imagine there are quite a few infected, but undetected cases behind numbers that are reported.

A good reply that summarizes my reaction:

Brian O'Roak @TheRealDrOLab

Replying to @trvrb

@JShendure and @seattleflustudy Do you really think Twitter is the best place to post this information? Wild speculation on the number of cases without confirmation is just going to panic people.

You don't reduce panic in a Florida hurricane by downplaying the wind strength until after everyone's already feeling it rattle their house. If we care about reducing panic the correct thing is not hiding information, it's having competent people in charge communicating regularly with the public about the facts and telling the public what is going to happen and what they need to do to prevent spread.

You’re correct. And this tweet storm is a great example of someone who is not charge telling the public things that may not be facts (by their own admission).

The point is not that all information must flow via some authority and be approved for release. It’s that there is a competent authority also speaking to the public. We haven’t had that in the USA, and are only now seeing states stepping into that role and getting up to speed.

Guarantee this information is being taken seriously by the WA department of health. The Seattle flu study isn’t some joke, it’s a real scientific research group at Fred Hutch and UW.

I've been interested in what the incentive is to avoid panicking people. What does panic actually look like and why are people so against it? People buying groceries and staying home? If panic looks like getting ready for a snow storm or hurricane, staying at home watching a movie instead of going to the bar, or washing my hands more frequently, I'm not sure what the downside is. It seems like the outcomes are worse if you wait.

Panic looks like cheap hotel rooms and plane tickets :)

Sure you want to get in an airplane?

Airplanes aren’t particularly comfortable, but I don’t see coronavirus making flying any worse.

If I was older or had a family to worry about I might see things differently, but right now this just seems like a good opportunity.

Panic means there's no more supermarket open because first the hordes come and panic-buy, then the supermarket has to close as even the workers don't show up. Panic means people not showing up to work, doctor's offices overcrowded, streets blocked with people going to the countryside. Panic means the system breaks down - electricians don't show up to fix issues, firefighters can't get through as the streets are blocked, local businesses getting looted and going bankrupt as they have to close, etc etc

Mild panic/mild fear/excessive carefulness will inconvenience you. Real panic means people die and things break.

>streets blocked with people going to the countryside

A pandemic isn't a nuclear attack, leaving town won't get you anything.

Fleeing the towns was popular when the miasma theory of disease was in vogue, and would have been useful for avoiding animal vectors, but you wouldn't expect it to happen with COVID-19.

Reply on news.ycombinator.com