RichardGibbons

So, that's the testing half of the numbers. What about the tracing part? Is there some way to quantify the tracing? I've been thinking about it, and have found nothing but anecdotes. One of the challenges is that effectively it's a system with feedback. Like, if you do good tracing, then the virus doesn't spread, which means that there are fewer "high risk" people to test, which means your per capita test rate can be low. This is an interesting discussion, because pretty well everyone accepts after the first explosion of cases, South Korea got the pandemic under control with test and track. So it would be interesting to have something quantitative that supports or contradicts the "track" part of the thesis. I guess one of the other things to keep in mind--which is obvious but tends to be ignored because people are too busy creating things like the 10 Commandments--is that multiple different strategies might work. Like, maybe "masks + handwashing + a culture that does these things when told to" is as effective at stopping the virus as "test + track + lock up people who were exposed".

It's far worse than a tax. It's a third party capriciously changing contracts after signing.

Can you please show data to prove that? I was mainly talking about Taiwan, Germany, and South Korea. (Full disclosure: a couple of weeks ago, Singapore would've been on my list of test & trace success stories, but it seems to be breaking down.)

Have either of you heard a plausible explanations for the increase in insurance rates? That sort of jump seems very strange to me. Like, were these things priced to make a huge loss five years ago, or are the new high prices just to make a massive profit? I would've guessed that the insurance markets are close enough to efficient that a 100-300% increase in premiums without an extreme event would never happen. But I'm clearly wrong, so what's the deal here?

Yeah, I think the increase is a result of insurance rates skyrocketing.

This warning from a reddit post is so true....

Nah, your assumption is wrong. As far as I can tell, cobafdek is thoughtful and smart, and has a high degree of integrity.

This is a really great point, but for some reason I'm still hopeful. I think a few good leaders could reverse this trend. Trump and most Republicans have abandoned good faith efforts to rule, but I still believe that people in aggregate are intrinsically good, and most people who get into politics really do want to do what they believe is best for the country. The only question is whether this phase in USA politics is a short-term thing, or if the system has been broken to the point where it's impossible for good leaders to be elected. If the latter, then I'd expect to see American power to continue to plummet in the world, as it drives away its allies and its anti-science/anti-education mentality dramatically weakens the country's competitive position. Just as happened with the Soviet Union, I think it's very hard to sustain (for much more than a century) a system whose values don't match reality. But I still think it's more likely that USA will pull out of this descent and remain a superpower. (Or that the blue states separate and become the new superpower, producing 2/3rds of the USA GDP--effectively the same thing. Secession isn't likely today, but after a few elections where the majority is repeatedly thwarted through tricks like voter suppression and gerrymandering, I think it becomes more likely.)

Sure. It's pretty simple. You were saying early on that hundreds of thousands or millions of people in the USA probably had it, and that testing was pointless. Then, hundreds of thousands of people in NYC actually got it and more than 10K died, which wouldn't have happened if all those people were infected already. And the countries that have done the best have been the ones who were good at testing and tracing. That said, I have come down more harshly on you than others simply because you're dishonest, and because--if you are actually a doctor--I think that dishonesty makes you dangerous. (I would have no problem whatsoever with you if you were simply wrong and adjusted your beliefs as the evidence changed. In fact, I'd have a whole pile of respect for you, if that were the case.) I also think that anyone who makes the argument that the worst case didn't arise so the worst case scenario couldn't possibly be true--after there was a massive, month-long lockdown to prevent the worse case scenario--is either disingenuous or stupid or both.

No, they were appallingly bad, to the extent that I'd lose a some degree of respect for all doctors if it weren't for Dalal's sensible posts on this thread to pull me back off that ledge. Frankly, it terrifies me that doctors exist who ignore evidence for gut feel, make wild speculations not even supported by common sense let alone facts, and then when the evidence proves them wrong, continue to insist that they were right. (What the heck is one to do if one seriously needs a doctor, and this is the doctor one gets? Just roll over and die?)

Yeah, kudos to you for saying early on that you thought it was unlikely to work.

This is unfortunate. https://www.scmp.com/news/china/science/article/3080055/anti-malarial-treatment-hailed-trump-has-no-benefit-coronavirus

Yeah, any solution won't be perfect. It's nice, however, that everyone involved recognizes that the perfect is the enemy of the good. I imagine that the epidemiologists' understanding of exponential growth will allow them to see that a solution that reduces the R0 by 70% is worthwhile even if it doesn't reduce it by 100%.

Yeah, obviously it requires test and trace. But it doesn't make much sense to say, "we haven't done this well in the past, therefore this article sucks because it suggests we do it in the future." Like, no kidding that the USA screwed up early and that it isn't capable of doing test and trace today. While you claim he's saying to open the country today and simultaneously scale up test and trace, the author doesn't say that. In fact, he implies the opposite. Test and trace are two of the Ten Commandments. Commandments because those two things are super important to do, because they're not just irrelevant details that we can ignore if we don't have the capability today. The reason the article's interesting and compelling is because it lays out what needs to be in place to start opening things up, including a effective test and trace strategy. So if you're looking to see if we're ready to open up, one criteria might be whether you've got the resources to do test and trace on the expected number of new infections. I agree with everyone else though--saying that it's OK to open schools because kids don't die doesn't make much sense, and doesn't fit with his "avoid events where people mix with a lot of different people" reasoning.

It's an interesting article, making a pretty compelling case for reopening the world except for large gatherings. I'm curious what an epidemiologist would think of it.

OK, now this really starting to feel like a game of Plague, Inc. Add on morbidity risks after half the world is infected....

Yeah, it's clear that they were giving out inaccurate information. The real question I wonder about is whether they were doing that because they're corrupt or because they knew that if they gave out accurate information, China would screw them, potentially making their response even worse. Did they choose the lesser of two evils, or are they just bad? Across all of this, the one definitive lesson to be learned is that the CCP is a big problem and big western nations ought to distance themselves from China, even if it costs money. It's a shame that America isn't in position right now to step up and take international leadership in doing that.

Yeah, I think that epidemiologists who have spent decades of their lives studying diseases are a bit more sophisticated than just copying the HIV strategy. I think they've probably spent a fair amount of time analysing about how our reaction to a disease should be impacted by the way that disease is spread. Like, maybe you're thought about this stuff for three hours a day for the past two months. If that's the case, many of them have probably spent well over 100 times that much time thinking about the issues, and they actually been educated on this topic specifically. At some point, you might want to consider whether the Dunning-Kruger effect might be relevant in this situation.

Further to your point, cherzeca, I actually think the comparable disease isn't HIV, but ebola--I imagine that was more of focus for experts recently than HIV. If ebola were airborne and transmissible by asymptomatic people, you could justify doing almost anything to prevent the spread.

Yeah, I don't think it's possible, and think that this is pretty much to the last hypothesis than anyone should actually test because it makes no sense. But it would be fun to write a science fiction story where this was the case. (Like some sort of scenario where everyone is infected by a disease, and the only way to avoid death is to interact frequently with people who have different genetics--all the people with type A genetics have to interact with types B, C, D, E, and F at least once a week or they die, etc. Figuring that out for the first time would be a bitch.) Empiricism is good, but you are much more efficient and maximize the value if you include some reasonable thought as well. (E.g. why wasn't your theory that infection rates are based on the number of letters in the name of each country's capital city? Or the per-capita number of people in the country who have one leg?) That said, analysis unsupported by data is as pointless as data without reason, so I like your dedication to empiricism.

Hard to know what to read into that article, since Trudeau's hedging all over the place (fair enough--I would be too). IMO, two weeks after each province is passed its peak, he should allow any business with fewer than 200 people in a room to open up, provided that all that businesses' employees and customers are willing to wear a mask for all interactions (except restaurants). (Two weeks is a relatively short time past the peak, but, just as the delay between infection and symptoms is bad on the upswing, the delay is actually good on the downswing since it effectively means four weeks between the peak and when a large number of new cases might come into the hospital as a result of any decrease in social distancing.)

Ok, why don't you walk me through where my proposed theory doesn't match the data (note: I'm assuming your numbers are accurate--I've spent no time looking at India.) My theory is that lockdowns make no measurable difference to diagnosed cases for a minimum of two weeks, and no measurable difference to deaths for at least four weeks, and more realistically five weeks. Why does your "we had exponential growth in deaths for 2.5 weeks after the lockdown" data contract my theory? 2.5 weeks is much less than five weeks, so what am I misunderstanding? Thanks!

The big change in my mental model over the last few days is that the numbers in the USA might not be as accurate as I thought. I was assuming that because USA has a decent healthcare system, it would also have decent data. But, while I thought USA would have more deaths (per capita) than comparable countries with socialized medicine because of people being reluctant to seek medical help, I foolishly didn't clue into the fact that this discrepancy would cause almost all the numbers to be low compared to countries with socialized medicine. (Kind of like China--if we lock someone in their house, they die, and we don't test the body, it doesn't count in the COVID-19 stats!) That New York "at home" deaths comparison is quite eye-opening. It would be useful to get year over year deaths numbers for a variety of countries to get a rough idea of how many COVID-19 deaths are missed in different places.

It actually kind of does look like we were heading for 500,000 deaths in the USA. New York City has 4,111 deaths. And daily "at home" deaths spiked from 20/day to 200 per day when NYC was having 240 recorded COVID-19 deaths. If we assume another 2/3rds of "unattributed deaths at home", then we're talking about 6,500-7000 COVID-19 deaths just in New York while taking precautionary measures. So yeah, when you start extrapolating it to the rest of the country, 500K seems easily achievable if no preventative actions were taken. Low, in fact. This demonstration of how broken your mental model is might humble some people. At least it should.

So you're saying that the shutdown caused exponential growth of deaths? Like, your expectation would be if you isolate people into groups of 4 and never let them have any sort of contact with any other group, your death rate from COVID-19 would be higher than allowing everyone to intermix to their heart's content? Huh. That's kind of a strange theory, but OK. Yeah, I included those in there because I thought you wanted an explanation of when deaths should be decreasing, not to discredit your "quarantining causes order-of-magnitude spikes in COVID-19 deaths" theory. My mistake.