If I were in the position I don't think I'd put a number on it. I would hope that my priority would be to mitigate the amount of lives lost, and if closing large segments of the economy would be in furtherance of that goal then that's what would happen until such point that it wasn't. If I had to pick a number though it would be 69
Had a coworker send me a video of Bill Bennett saying that projections now show only 60000 will die which is the same as the flu. So we shut everything down for nothing of course. How the fuck can people be this stupid? I sent a reply back showing actual numbers that would have died if we had done nothing. I'm sure it will be hand waived away. I give up, it's like people want to see a bunch people die.
I can’t speak for your friend’s rationale, but why are mortality rates not factoring in the asymptomatic? Granted it’s a large guesstimate, but still- Covid asymptomatic is a rough 50%. The flu is 14%. Provide a fair estimate instead of a scenario where no one is asymptomatic. That’s plain wrong. They’re not being factored in at all from any data I’ve seen. So we’re automatically inflating the mortality rate.
I think people are stupid (and by that I mean they don't understand the complexities involved other than simple numbers presented by the media), and combine that with the fear of what could be, and that is enough to have them not care about what could be, especially if they are not personally hit or affected by this. "It's not happening to me, so what's the big deal?" If a relative of their was on an incubator I think that'd be a different story... but a lot of people have very little empathy for others. Especially when the alternative is for them to lose their job, or have their lives ruined because they can't work or don't have the finances to weather a good storm.
There is the case fatality rate, which is fatalities/identified cases. There is also the infection fatality rate, which is fatalities/total infections. The case fatality rate for any disease is pretty much always higher than the infection fatality rate because not all infections are identified. This is normal, it's just that people are sloppy about definitions, and refer to "death rate" or "fatality rate" without specifying which of the two they are talking about. Infection fatality rate is the more useful number, but case fatality rate is known, whereas infection fatality rate can only be estimated. Case fatality rate in the United States right now is about 4%; it has been higher elsewhere and lower elsewhere (10-12% in Italy/Spain/UK/France, 2-3% in Germany/Canada). As mentioned it depends heavily on how thoroughly infections are identified and become cases, as well as where you are in the progression of the disease (if you're still in the exponential growth part of the infection curve, the case fatality rate will be artificially low because a disproportionately high number of your total cases haven't run their course to recovery/death yet). Most of what I've seen estimates the infection fatality rate somewhere near 1%, taking all factors into account (under-testing, under-counting deaths, etc). 1% is about 10x worse than the flu.
The Diamond Princess numbers are as good as we can get because everyone got tested. The death rate is 1.6% with around 46% of the infected being asymptomatic. Odds are it will rise farther as some are still in the hospital. And that's with excellent medical care. You let it run wild and the death rate looks like Italy or Spain. https://www.statista.com/statistics/1099517/japan-coronavirus-patients-diamond-princess/
https://fivethirtyeight.com/features/a-comic-strip-tour-of-the-wild-world-of-pandemic-modeling This is the best explanation of modelling I've seen so far and it's in a Goddamned comic strip, because that's the world we live in now. This is another possibility that scares me. Bottom line: we are confident at least 600,000 people have this virus, and 25,000 have died from it since March 1. That was SIX WEEKS AGO. 4,000 per week over a full year puts this at #3 on the death list, behind cancer and heart disease https://www.cdc.gov/nchs/fastats/deaths.htm. Also, incidentally 4k * 52=208,000, so that might be where Trump got this number, assuming that the spread magically stops somehow. Except, it's not 4k/week evenly. It's 1000-2000 people dying from this every day now. https://covidtracking.com/data/us-daily We'll be at 4k THIS week by Wednesday. Saturday to Saturday, over 10,000 people died. We have two more weeks of April. At 30k a day being confirmed, that's another 480,000 cases and at 1.5%, that's 7200 dead. It's likely we'll see far more than 7200 people die from this before May. What we don't really know yet is the percentage of recovered that have permanent damage to their hearts, brains and lungs. Some of the people with this virus are now forever disabled by it, and we've not even begun to untangle that. Sure, the dead may cost us nothing, but thousands of disabled Americans is a nightmare none of us want to be a part of. It's not simply binary: recover or die is a false dichotomy. Some people will never return to normal. So, for these people suggesting we open the economy back up: as long as it's your ass out there running the risk like the rest of us, sure. No panic room, no private nurses, no masks, no PPE. If the leaders, CEO's and government officials clamoring for people to get back to work aren't sharing the risk that they are asking their employees to face, then they can go fuck themselves. It's callous to ask someone to risk death or permanent disability for an hourly wage, when they themselves are too cowardly to do it. I'm not much of a gambler, but I'm not fucking around with a 46% chance of not having symptoms, a 1.5% chance of death, and a not-insignificant chance of permanent lung damage. Sorry, but baseball games and my daily commute aren't worth it. I know math is hard, and biostats are even harder. It's why we have experts, pay them a lot of money and should be fucking listening to them.
As someone who is essential infrastructure and going to work everyday in Florida, a state that was laughed at for being late to the party to shut things down and where it was "any minute now" that we would be the next hot spot, I'm not as worried as I apparently should be. I have read most of the comic strip as time allows; the model is a best guess based on a series of assumptions but it doesn't cover the part where the expected outcome from the model doesn't match reality and then the PDCA (plan, do, check, adjust) cycle on the model occurs in a timely manner and there is a public health reassessment of where we really stand. The reason it doesn't cover that is because it's not occurring. At this point, I am willing to make a prediction. The AAR on this is going to come out and say this was an overreaction, while exposing that had this been something more lethal, we were woefully unprepared because the fucksticks at the CDC are busy fighting obesity and racism instead of being prepared to fight a pandemic. It will likely come out that this shouldn't have been a nation-wide lockdown, it should have been regional or maybe even municipal level. The damage done to the economy was never worth it. Also, I am starting to sense the quarantine narrative is going to start to change quickly (I think I am reading it already), this is going to change from this was done to flatten out the curve - everyone will eventually get it, it's just timing to this quarantine was done to limit infection to the greatest extent possible because look at the damage at the individual level. If we are really lucky, we will look at some of the red tape regulations that were, in essence, road blocks from allowing quick reaction by industry to address the issues and ensure that never happens again.
So you didn't read the entire comic that explains why a model doesn't exist, then say that the model that doesn't exist is putting out results that don't match up with reality. Then you make predictions based on nothing more than how not worried you are. You do you, man.
It would be a lot different if everyone had the same body type, same set of illnesses, same age, same physical health, same diet, same medical history, same genetic history, same financial/geographical/demographical resources and same ability to recover from illness.
Are you not paying attention? The model exists. That says 538 hasn't created a model. The model, which has been revised multiple times due to accuracy level that amount to total useless garbage, can be found here: https://covid19.healthdata.org/united-states-of-america You didn't look at these links, did you? Based on assumed infection rates, the model shows you assumed hospital usage accounting for social distancing. You live in Washington State, correct? The location of the first known outbreak in the U.S. at an assisted care home? Where one could reasonably expect a higher level of infection, especially where the hysteria said the case load is doubling every 3 or so days if I recall correctly? Look at the data your state is reporting here: https://www.doh.wa.gov/Emergencies/Coronavirus ; on 4/12 387 patients hospitalized. The model is here for the expected hospital usage: https://covid19.healthdata.org/united-states-of-america/washington
Welp, I'm out of work. Woke up with symptoms and made the call. I can't get a test because I'm not experiencing severe chest pains (yet). No idea as to when I can go back. I have a follow up virtual appointment on Thursday, same as today. My place of employment has no known cases yet, so they think it's likeliest I caught it from the grocery store. I'll know 90-95% by Thursday if I have Covid, but at this point it's just as likely to be the flu or something similar. Basically I'm in 100% self isolation for the next 2 weeks regardless of diagnosis on Thursday, and then I'll be 'released' following increased availability of testing, or if there's no symptoms for at least 14 days. Not sure if my employer will accept that, or I'll be out for the next 3 months or so. My neighbor has been hacking like crazy and I've heard them a couple times while INSIDE my home. /life in the city circa 2020.
Well, I do have ADHD and I didn't read the post you are referencing until just now. Now that I have, I fail to see how hospital census data correlates to virus infection rates. In your second link from Cuomo, it says that there were nearly 1100 new hospitalizations, but it doesn't say from what. Hospitals and clinics still have people who do get injured and need treatment for new or underlying conditions. Depending on the demographics of the surrounding area, these non-virus related services could vary wildly. Just because social distancing is in effect, it isn't going to stop people from having their cancer get worse, getting into car accidents or having vigorous buttsex to the point of injury. As downndirty noted and 538 explained, models are only as good as their data. Assumptions have to be made because there is no standard for reporting. Speaking as someone who doesn't want anybody, least of all myself, to get sick, I would much rather the model overestimate the potential need for hospital beds and ventilators than not enough. After all, I would rather have something and not need it, than need something and not have it. You are recalling incorrectly. The caseload increase would be exponential based on social distancing not being put into place. Washington put it's shelter in place order on March 23rd, however, the governor put other social distancing policies in place prior to this order. I myself stopped going into work starting March 2nd. The rest of my division was ordered to work from home starting March 15th. Based on Washington social distancing more quickly than most of the country, the infection curve flattened out much more quickly. This graph is based on data that was available to the New York Times a week ago. Model projections are just that. Projections. Assuming a model's projections are iron clad, is a foolish notion. Especially when the model itself is updated, and the data put into it isn't standardized. The whole point is to show what could happen based on available data now. Anybody can cherry pick a data point with hindsight and say how bad the model is. It's like me criticizing the weather person on the news when I get an inch of snow instead of the expected foot.
Of course it correlates. It just doesn't give a full picture. Dude, stop trying to lecture people on this.
If hospital data didn't correlate to infection rates we wouldn't be shitting our pants about hospitals getting overwhelmed. We have one variable causing the surge, and that's Covid 19. Most hospitals are finding that people who need treatments for other ailments are staying away because they think they might catch it there, which is a good thing, even if not entirely rational. We have board members much more knowledgeable on these subjects than myself so I'll leave it to them to better explain if they want, but yeah, the people making these models are using hospital data because they know what they're doing, not because they just have to make one regardless of whether or not it's worth a fuck.
You're still not getting it. You have assumed infection rates. Based on these assumed infection rates you have an assumption on the number of hospitalizations, a subset of that number needs ICU treatment, a subset of that number needs intubation, a subset of that number is a fatality in most instances. You are getting real data on hospitalizatons, real numbers on ICU admissions, real numbers on intubation and real numbers on fatalities. This should be informing the model because the model for this data is overshooting reality in a bad way. Separately, to your point about the NY data - Assume that is 1100 new cases coded as COVID-19 (I think it is) to help you get to the grand total of 15k hospitalized and coded as COVID-19 (I think it is). The model for that day assumed that there would be 65k beds needed for COVID-coded patients. While 15k is bad, it still is far, far less than what was modeled. If those numbers were to include all hospitalizations for other reasons than COVID-19, we would view this virus to be of even less of a concern. We made public health decisions based on this data in the model. We panicked the public based on these models and they are way off. If only you could see one of the three or four earlier revisions to this data from just April alone; the numbers have been revised downward with each revision. Yet the policy isn't moving as fast as these models. At what cost, though? Destroying the economy based on a huge overshoot is not a good idea. I just hope it doesn't turn out that the resulting poverty kills more than the virus. Anyone who understands models knows this. We are in agreement. My point is that the public health decisions aren't updated and matching the revised data.
Ultimately the deaths are a function of infection fatality rate and r0. With a 1% infection fatality rate and an r0 matching the flu, without preventative measures we'd be looking at a couple hundred thousand dead. With an r0 higher than the flu, that number probably hits 1-2 million. We've deliberately driven r0 as far down as we possibly can. Unless you're arguing that the infection fatality rate is significantly below estimates, or the r0 without the measures we've taken is significantly below estimates, the only conclusion is that what you describe as "panic" is what's responsible for the models being off.
You aren't accounting for the nearly 100,000 people left who have tested positive, or the number of people who are symptomatic and haven't been tested. It is extremely difficult to determine with any kind of certainty what percentage of people who test positive will get sicker unless you have parameters in place to account for things like age, underlying health conditions, social distancing, etc. Even if we assume that there aren't any more infections outside of that 113,000 number, and that 20% is symptomatic and require hospitalization, that is 22,600 people. According to the model, there are only 13,010 beds available. The economy was going to go into the toilet as soon as the virus showed up. The resulting economic downturn's length can only be blunted by keeping people home so the infection can run its course quicker. We can recover from an economic downturn a hell of a lot easier with a lot more people than less.