Minichan

No people who had those things died of covid-19

@1,119,740 (dreamworks)

It's the decrease in pollution mang.

Well homicides and car accidents being down makes sense

I like this graph for comparing COVID-19 with flu, traffic accidents, heart disease, and other stuff. It's good we're keeping the meat packing industry forcibly open though. We wouldn't want the numbers for heart disease falling behind at a time like this.

@previous (F)
That's a bit confusing. I was about to say it seemed weird car crashes were basically unchanged (would have expected a visible decrease from stay at home/reduced traffic) then I saw the little "1/1/18" on the car crash line indicating this is 2018 car crashes which were pandemic-free.

Also why is flu broken down to with and without pneumonia? If the pneumonia is caused by the flu it doesn't make sense to exclude it. I mean could you imagine separate "Covid-19 with/without pneumonia" graphs? No one would die from Covid without pneumonia ?

Also I think the 1957 Asian flu line should be adjusted for the much smaller population of the United States 60 years ago.

(Edited 1 minute later.)

@previous (Meta !Sober//iZs)
They're updating it! Here's the latest, and it's headed downward. So that's good. It's almost less relevant than cancer and heart disease.

> Also why is flu broken down to with and without pneumonia? If the pneumonia is caused by the flu it doesn't make sense to exclude it. I mean could you imagine separate "Covid-19 with/without pneumonia" graphs? No one would die from Covid without pneumonia ?
They actually go into some detail about the uncertainty involved in those numbers on the page they posted the graph and links to the data from the CDC. In their words: "This was the deadliest recent flu season. The chart shows one line for deaths attributed directly to flu, and another for deaths attributed to either flu or pneumonia. The smaller line is an undercount of flu-caused deaths, the larger is an overcount, with the real number lying somewhere in between." This way you can see what both overcounted and undercounted flu deaths would look like in this context.

> Also I think the 1957 Asian flu line should be adjusted for the much smaller population of the United States 60 years ago.
The Y-axis is measuring deaths per million, so that's already accounted for in the units. You're seeing the adjusted figure.

> I was about to say it seemed weird car crashes were basically unchanged (would have expected a visible decrease from stay at home/reduced traffic) then I saw the little "1/1/18" on the car crash line indicating this is 2018 car crashes which were pandemic-free.
Again, it's in deaths per million so monthly variations of a few hundred deaths month-to-month nationwide doesn't show up dramatically. Nor does it really spike too much. It rises from ~2300 in February to ~3000 nearer the holidays with a bit of an increase in the summer months. If you look at the figures they are working with from the National Highway Traffic Safety Administration, you do see the kind of variation you would expect for seasonal and holiday travel. It just doesn't vary too much or show up at this scale.

The real take-home message is awareness. There are people out there claiming that COVID-19 is less of a worry than X, be that the flu or traffic deaths or heart disease. The dramatic spike in the graph makes it painfully obvious that COVID-19 has done more in about a month than any of those things over the same time frame. Through March and April it's almost doubled the fatality rate of traffic accidents for all of 2018. Going by the numbers, the threat of COVID-19 isn't being over-represented or hyped up. It is, right now, the most statistically relevant threat out there.

@previous (F)
> There are people out there claiming that COVID-19 is less of a worry than X, be that the flu or traffic deaths or heart disease.
Oh yeah I'm quite familiar with this. The one I heard most often is "Flu killed X people last year and COVID-19 has only killed 0.5X people and we don't panic about flu so it's stupid". COVID has now passed the 2018 flu death toll so I don't hear this anymore but they were right: until two weeks ago COVID had indeed not killed as many people as flu. The fallacy was taking the entire year's flu deaths and comparing it to the first 3 months of COVID.

What I'd like to see (and there's no fuckin way this is going to happen so this is pure fantasy) is some way to calculate how many years of life COVID has actually taken. It seems most of the deaths were people who were already on the edge - the very elderly, people with serious underlying conditions. People who likely wouldn't live 10 more years in any case, or could have also succumbed to regular flu. So for each COVID death, they calculate a score and say "Person X would have lived Y more years without COVID".

You can actually see this in the graph in the OP where it appears non-COVID deaths "decreased" by about half. It would seem a lot of the heart attack, stroke, etc, deaths got turned into COVID deaths but also that a lot of the COVID deaths would have happened in this year, anyway, without COVID.

The other thing I'd like to see is how many deaths per day (all causes) in 2019 vs how many deaths per day (all causes) so far in 2020. I think this would be easier for most people to understand if you can say "The Grim Reaper is running at 109% normal speed". Because if you just report corona deaths in a vacuum, it's hard to grasp how big a problem is. You can say "Okay 64,000 people in the US died, that's 0.02% which means 99.98% of people have nothing to worry about".

(Edited 5 minutes later.)

@previous (Meta !Sober//iZs)
Do NOT question the narrative.

@1,120,556 (Meta !Sober//iZs)
> You can actually see this in the graph in the OP where it appears non-COVID deaths "decreased" by about half.
Unfortunately, the graph in the OP is made of fiction and bullshit.

That graph doesn't appear in the NYT article linked in OP's image and the CDC mortality stats don't back up anything remotely like those numbers. Here are the datasets for the CDC COVID19 weekly death counts and CDC weekly deaths by cause over region.

I've tried to recreate OP's graph working from the actual numbers for New York from the CDC. The CDC divides New York into regions for the New York City metro area and the rest of the state. OP's graph is unclear about which region(s) are included. OP's totals almost match up for the state minus city deaths, but COVID deaths for this region are in the 4000s rather than 10000+ as OP's graph implies. As far I can tell, none of OP's number match with anything remotely close to reality.

The CDC breaks death figures down by week. I'm not sure which chunk of time OP's graph decided was "mid-march to mid-april". The weeks don't quite line up from one year to the other, but I've included the dates on mine to indicate that these are equal spans of time during the same period from each year.

The graph I've made includes all areas (NYC+NYSTATE) within the state of New York.

> The other thing I'd like to see is how many deaths per day (all causes) in 2019 vs how many deaths per day (all causes) so far in 2020. I think this would be easier for most people to understand if you can say "The Grim Reaper is running at 109% normal speed". Because if you just report corona deaths in a vacuum, it's hard to grasp how big a problem is. You can say "Okay 64,000 people in the US died, that's 0.02% which means 99.98% of people have nothing to worry about".
Well, now that I have the datasets here maybe I'll try to do something like that later when I have some time.

There are analyses of "excess deaths" out there, that is, the death rate that is above what would be predicted by previous years' figures after you subtract out the COVID deaths. All of these analyses seem to indicate that COVID deaths are being undercounted or increased stresses related to the situation are contributing to more deaths. For instance, this article from The Economist points out that cardiac arrests in NYC are soaring. The "why" of that is probably a thorny question, but the raw numbers on 911 calls and deaths indicate a definite rise.

@1,120,556 (Meta !Sober//iZs)

> You can say "Okay 64,000 people in the US died, that's 0.02% which means 99.98% of people have nothing to worry about".

That disguises the scale of the tragedy by disguising it with statistics

@1,120,870 (F)
I find this one even more puzzling. I'm quite surprised to see a spike in non-COVID mortality.

If you had asked me (without showing me either graph or any numbers), I would guess the non-COVID death rate to be less than usual from people staying home. I would think social distancing works about equally well for all contagious illnesses so if it's reducing COVID, it should also be reducing everything else transmissible through aerosols and you would see that in the death rates.

@previous (Meta !Sober//iZs)
Stress actually is really bad for you. It can cause your immune system to fuck up and exacerbate existing conditions. And people are stressed about the situation.