Estimation of total mortality due to COVID-19 • Hacker News Board

Re: Estimation of total mortality due to COVID-19

#371

Post by zhaksbs » Sat, May 08, 2021, 2:45 AM UTC

vkou wrote: post #274

zhaksbs wrote: post #226

Earlier quoted context omitted.

China literally welded people in their homes. They probably lied about their numbers. New Zeland, and Australia had few cases when they did lock down, and are islands. Korea has a very disciplined population compared to the West. Really you should be comparing the US to Europe. The EU didnt do much better, but was far more draconian. Or within the states. But anyway, i never know if people are serious when they propo…

> But anyway, i never know if people are serious when they propose the China or Australian model. Those were brutal and Im glad to be voting for a constitutional amendment next week to limit the governor's lockdown authority. I never know if people are serious when they suggest that a few hundred thousand dead is a worthwhile footgun-sacrifice to escape the draconian rule to which Australians and New Zealanders were…

Death is not the worst things in life, an unreasonably large fear of death is much worse. A death cowering in fear, or full of regret is pathetic.

Who do you think you are picking who dies like a tyrant doling favors? Do you not think I have elderly loved ones I worry about? Do you think I would enjoy getting COVID myself, or potentially orphaning my daughters? No! I just refuse to blow my kids mind with endless lock downs.

My obligation is to take COVID seriously, and do my utmost not to get sick myself. If I do, ill lock myself up in the basement for as long as it takes.

But with my obligations I also have rights, and Ill be damned if someone too scared to have ever pondered their own mortality will ruin life for me without protest.

Re: Estimation of total mortality due to COVID-19

#372

Post by zhaksbs » Sat, May 08, 2021, 2:49 AM UTC

worik wrote: post #339

neither_color wrote: post #254

Earlier quoted context omitted.

I agree with the last point and I don't think it's a bad thing. I'm against lockdowns period. I'm not against vaccines, not against masks, but I don't want this administration, the last administration, or any other to have the power to put us on indefinite house arrest. This IS a hill I wont budge from. With China, we don't have enough information to make the claim that they under-reported deaths. If you CTRL-F China…

Golly. If that is a hill you will not budge from it may be a hill you die on. All the cherry picking data in the world will not help you if you get sick. I am really sorry that some abstract principal becomes a hill to die on.

Ya, i think thats what he meant. Hes willing to do all the reasonable things and all the unreasonable theatrics but he refuses to be at homw-arrest.

I cant speak for the OP, but for those of us who have given life and death some thought, lockdowns are worse than death.

Re: Estimation of total mortality due to COVID-19

#373

Post by dang » Sat, May 08, 2021, 3:55 AM UTC

harambae wrote: post #369

dang wrote: post #364

We changed the title from "True U.S. death toll from Covid is more than 900k, study finds". That was editorialized, which breaks the site guidelines. It was also arguably both misleading and baity, which means that it broke the site guidelines on all three points: " Please use the original title, unless it is misleading or linkbait; don't editorialize. " Please don't do that! https://news.ycombinator.com/newsguidelin…

Original source was this: https://www.mediaite.com/news/true-u-s-death-toll-from-covid... Which is where the title came from. At the last minute I remembered there was an actual scientific paper so I changed the URL to that.

That makes sense. We sometimes leave a title up after changing the URL too.

Re: Estimation of total mortality due to COVID-19

#374

Post by Uberphallus » Sat, May 08, 2021, 4:33 AM UTC

username90 wrote: post #314

Uberphallus wrote: post #308

Earlier quoted context omitted.

Social distancing was never as much about saving people directly as about allowing the healthcare system to handle the mess. In computer terms, it's about launching processes in a controlled manner vs let them run freely and make the system paginate. You don't let critical systems paginate.

Right, I agree with that. But it was only valid for the first few weeks. The healthcare systems weren't critically overrun anywhere in the world. Yes, from time to time not every patient they wanted to put on a ventilator got one, but ventilators doesn't dramatically improve the odds of survival so it wasn't a big deal anyway. So why did California continue with their measures for a year? They still did among the wor…

> But it was only valid for the first few weeks. The healthcare systems weren't critically overrun anywhere in the world.

We live in a different planet? I live across the border with North Italy, and it was only 2 weeks what it took for the situation to go from "oops, looks like there's a problem" to "let's mobilize doctors out of retirement and potentially kill them to save people". There have been different levels of restrictions since. Same in France, Spain, Germany, pretty much anywhere in Europe.

> Yes, from time to time not every patient they wanted to put on a ventilator got one, but ventilators doesn't dramatically improve the odds of survival so it wasn't a big deal anyway.

Ventilators were abused at the beginning out of ignorance about the disease, but they're still the best last-resort treatment for critical cases. In my region, during the second wave, most deaths happened to people waiting for a ventilator, or during transfer from oxygen to a ventilator, and those on ventilators generally survived.

That said, now there's a more defined and effective protocol before reaching a ventilator that requires hospitalization nonetheless, and makes you require one if not followed.

> So why did California continue with their measures for a year? They still did among the worst in the world so those measures seems to have had little to no effect at all.

California had terrible numbers despite the fact that measures worked. It's been studied ad nauseam and social restrictions is the among most effective measures to be taken in every study, with a solid interval of confidence. One of my favourite studies is [0] because it shows that some things were done right (lockdowns, closing schools, remote work) but some had little to no effect (surface disinfection).

[0] https://www.nature.com/articles/s41562-020-01009-0

Re: Estimation of total mortality due to COVID-19

#375

Post by lazide » Sat, May 08, 2021, 4:35 AM UTC

spekcular wrote: post #370

timr wrote: post #367

Earlier quoted context omitted.

A prospective test is blind. You don't know what the data looks like, so you can't cheat. > Also, if I understand correctly, none of the modeling done in standard texts such as Gelman et al.'s Bayesian Data Analysis involves prospective validation under your definition. Should we then classify the examples in that book as "mathematical fairy tales"? This seems like a fairly strict standard! It's surprisingly common f…

> A prospective test is blind. You don't know what the data looks like, so you can't cheat. Suppose I blind myself to the holdout set. What's the difference? > It's surprisingly common for models to pass all of the cross-validation you want to throw at them, and fail in the real world. I don't care what statistical techniques you've applied, if you don't conduct blind tests, you don't know how your model performs. Wh…

Not the poster, but happy to answer since I have a decent amount of experience in similar areas.

Their point is that while in theory you could blind yourself to the holdout set - that rarely REALLY happens. At least in a truly good way. Almost everyone is going to either peek at the data, massage the data, or try the validation and when it fails tweak the model and try again, or any number of shenanigans that no one admits happen - but happen all the time.

The failure mode that a prospective test would guard against in this case is - you can't peek or massage the future (and if you can, you can retire a billionaire pretty easy and stop worrying about this) - so you can't accidentally fool yourself or others by peeking, massaging, etc. And if you tweak your model, then do another prospective test - then you still can't peek at the future, and you'll have a different set of data, so you can't be overfitting to the old data/p-hacking.

You can do it over and over again, and as long as you don't just say 'all good' on a failed test without doing another different prospective test, you either have a predictive model or you clearly do not.

That is something the traditional validation models cannot do, since they aren't looking for truly never before seen data.

It still doesn't tell you if once in production you will stop fitting and fail over, but it does at least tell you that it fit a never been seen, new set of data at least once. Which is better than the other tests can do by their very nature.

Re: Estimation of total mortality due to COVID-19

#376

Post by seppin » Sat, May 08, 2021, 5:07 AM UTC

chmod600 wrote: post #13

It looks like Florida is two shades "better" than California in the diagram, which would put the total Florida death rate as better than California. That would be interesting because Florida did almost no lockdown. How could that be?

> How could that be?

Florida SWAT raided the home of the person trying to publish the real numbers https://www.bbc.com/news/technology-55230764

Re: Estimation of total mortality due to COVID-19

#377

Post by des1nderlase » Sat, May 08, 2021, 8:14 AM UTC

grumpyautist wrote: post #348

worik wrote: post #346

Earlier quoted context omitted.

Conspiracy BS because it cherry picks ideas to support a conclusion. The implied conclusion is that Covid has been exaggerated to sell vaccines Cherry picking exapmple: "criticized GAVI for focussing on funding expensive new vaccines and neglecting to give children cheap old ones" Two minutes reading the article on Wiki tells me that the purpose of the organisation is to facilitate access to new vaccines, expensive p…

So, the study is funded by groups that admit to existing to sell new vaccines? And you don't see a conflict of interest?

There could be conflict of interest, but someone still needs to show that these numbers are substantially wrong.

Re: Estimation of total mortality due to COVID-19

#378

Post by raarts » Sat, May 08, 2021, 8:22 AM UTC

I think calculating weekly expected deaths and drawing conclusions from that is wrong given the seasonality of covid-19 and diseases in general.

Excess deaths should be looked at over a long period of time.

Seems similar to the difference between weather and climate.

Re: Estimation of total mortality due to COVID-19

#379

Post by timr » Sat, May 08, 2021, 3:52 PM UTC

teachingassist wrote: post #205

timr wrote: post #177

Earlier quoted context omitted.

> [Edited to note: the linked article discusses this with further detail and evidence] Here's the sum total of evidence provided: > An analysis by the Netherlands statistical agency suggested that all excess deaths in the Netherlands were directly due to COVID-19 So they use a single study out of Europe to extrapolate here. > The second driver of excess mortality is reduced health care utilization for many causes;3 h…

On the other hand: "in many European countries there was a spike of excess deaths in weeks 31–35 during a period when COVID-19 reported deaths were extremely low. This period coincided with a heat wave and national reports of deaths due to the heat wave. We excluded these weeks of data from subsequent analyses." I personally wouldn't have excluded these weeks - unless you're going to exclude historical (near-annual)…

> No estimation is perfect, nor expected to be, and so your criticism is unreasonable. You're doing exactly the things that you are criticising the study for, and more so. You provide zero evidence to counter their sourced evidence, and ignore and dismiss points that don't match what you are trying to say.

They have provided no "evidence" -- they made a model. They have provided no validation of that model, so it could be completely incorrect. Nobody knows.

Furthermore, I just showed you that of the sourced evidence they did consider, they dismissed nearly everything that reduced the death rate attributable to Covid, and said it was "too dificult" to quantify.

Mathematical models without prospective validation are not evidence. They are not science. They are bedtime stories.

Re: Estimation of total mortality due to COVID-19

#380

Post by timr » Sat, May 08, 2021, 4:00 PM UTC

lazide wrote: post #375

spekcular wrote: post #370

Earlier quoted context omitted.

> A prospective test is blind. You don't know what the data looks like, so you can't cheat. Suppose I blind myself to the holdout set. What's the difference? > It's surprisingly common for models to pass all of the cross-validation you want to throw at them, and fail in the real world. I don't care what statistical techniques you've applied, if you don't conduct blind tests, you don't know how your model performs. Wh…

Not the poster, but happy to answer since I have a decent amount of experience in similar areas. Their point is that while in theory you could blind yourself to the holdout set - that rarely REALLY happens. At least in a truly good way. Almost everyone is going to either peek at the data, massage the data, or try the validation and when it fails tweak the model and try again, or any number of shenanigans that no one…

Yes, exactly. But I'd extend this one part:

> Almost everyone is going to either peek at the data, massage the data, or try the validation and when it fails tweak the model and try again, or any number of shenanigans that no one admits happen - but happen all the time.

While "peeking" at the data is certainly something that happens all the time, usually what happens for even the most rigorous of model-makers is that your model fails "out of domain" -- the past doesn't accurately reflect the future.

Even if you get everything else right, it's incredibly common to fool yourself into believing you're capturing "reality", when in fact you're just making a good fit to past random fluctuations.