Nate Silver

Nate Silver @NateSilver538

Generation 0: 2.5% newly infected Generation 1: 3.3% Generation 2: 3.7% Generation 3: 4.1% Generation 4: 4.3% Generation 5: 4.2% Generation 6: 3.9% Or something like that. Which would get you to 26% infected cumulatively. Actually higher, since some people had it before Gen 0. — PolitiTweet.org

Nate Silver @NateSilver538

A month is enough time for that 2.5% of cases to spread 5-6 times over. If cases were turning over with an R of 1.3 (maybe what you'd get with relaxed social distancing) that gradually declines because of increasing herd immunity, you'd wind up with something roughly like this: — PolitiTweet.org

Nate Silver @NateSilver538

So... I think al their data is pretty compatible. First of all, one needs to be careful about comparing stocks (total number of people ever infected) and flows (number of people infected at any one time). So let's say the *flow* in Stockholm was 2.5% of people infected as of 4/1. — PolitiTweet.org

Nate Silver @NateSilver538

2) However, it also increasingly seems that their numbers are leveling off, especially in Stockholm, which is what you would expect if 30% of the population has had it there (as their ambassador claims) providing some herd immunity. Here is their ICU data: https://t.co/I6cNalVUYL — PolitiTweet.org

Nate Silver @NateSilver538

Those comparisons matter more because lots of factors matter in death rates, not all of which reflect government policy. The Nordic/Scandinavian countries should be pretty comparable and will control for a number of those variables. — PolitiTweet.org

Nate Silver @NateSilver538

Two things about Sweden's data: 1) It has had a lot of fatalities. I don't buy the "other places [e.g. Italy] have been worse" take. Sweden's death rate per capita is 6-7x that of Norway and Finland, and 3x that of Denmark. https://t.co/9fFIJJjqhz — PolitiTweet.org

Nate Silver @NateSilver538

@JeremyKonyndyk @chrislhayes That possibly overestimates R in Washington by not accounting for the amount of testing. WA's reporting has not been very reliable on testing so it's hard to know what's happening for sure, however. https://t.co/7CxfmiXPWZ — PolitiTweet.org

Nate Silver @NateSilver538

People really need to be careful with these headlines. If restrictions are lifted on Monday and cases spike on a Tuesday, it is *not* because restrictions were lifted. In fact, you won't see the effects until ~2 weeks later. https://t.co/XB3O1ek7GT — PolitiTweet.org

Nate Silver @NateSilver538

@PollsterPatrick @PatrickRuffini But I'd say hospitalizations, the positive test *rate* and deaths are the 3 reasonably robust indicators. We could debate about which order to put them in. Hopefully they all tell similar stories (they basically do in NY). — PolitiTweet.org

Nate Silver @NateSilver538

@PollsterPatrick @PatrickRuffini Mortality isn't perfect, either. It's lagging, and there are a lot of issues with respect to how we're counting deaths. I think measures related to hospitalizations should theoretically be the most robust measures, but they aren't always robustly reported in practice. — PolitiTweet.org

Nate Silver @NateSilver538

@PollsterPatrick @PatrickRuffini The positive testing rate can have issues, but it is leaps and bounds better than the raw number of positive tests. It is a useful metric to a first approximation, while the raw number (without context on test volume) is a useless one to a first approximation. — PolitiTweet.org

Nate Silver @NateSilver538

Korea now seems to think the so-called "re-infections" it found are actually just false positives. A lot of smart epidemiologists thought this was likely the case all along, so not a huge surprise, but passing this along for people who freaked out earlier. https://t.co/C3iIjxDr7S — PolitiTweet.org

Nate Silver @NateSilver538

This is a steeper decline than we're seeing in other places. It may reflect the fact that enough NYCers had COVID-19 (the state estimates 25%) to provide some benefit from herd immunity. Not likely to be the case in most other places. — PolitiTweet.org

Nate Silver @NateSilver538

The positive test rate continues to decline in New York City in a way that's pretty reassuring. https://t.co/8fGrpKxCUq — PolitiTweet.org

Nate Silver @NateSilver538

@ForecasterEnten Actually a long way from 100% effective, but enough to provide some real help. I'm talking treatments, not vaccines though. Hopefully at some point we get both treatments and a vaccine. Treatments likely come first. — PolitiTweet.org

Nate Silver @NateSilver538

Non-rhetorical question. Let's say that by fall, a combination of drugs can reduce COVID-19 fatality rates by (30%?) (50%?) (75%?) And that administration is reasonably widespread (most people who need them get them), but with gaps. How much does this change society's behavior? — PolitiTweet.org

Nate Silver @NateSilver538

@chrislhayes Yeah, totally agree with that. — PolitiTweet.org

Nate Silver @NateSilver538

@thecity2 We can all dream! — PolitiTweet.org

Nate Silver @NateSilver538

I haven't really seen any stories that seem to have entirely the right handle on Sweden. It's a complicated case and likely requires some fairly nuanced conclusions. Its policies could turn out to be smart under one set of assumptions but tragic if those assumptions are violated. — PolitiTweet.org

Nate Silver @NateSilver538

@chrislhayes Yeah, part of what places need to figure out is how many excess fatalities have been caused by people avoiding/unable to be hospitalized for conditions that aren't COVID. Those deaths are still part of the cost of COVID. But maybe they can manage trade-offs better going forward. — PolitiTweet.org

Nate Silver @NateSilver538

@JADubin5 That might be part of it, tho the hospitalization rate also seems to be lower than expected even in places where hospitals are not overwhelmed. It may just be that the disease takes a fairly binary set of outcomes, i.e. either "treatable at home" or "life-threatening". — PolitiTweet.org

Nate Silver @NateSilver538

@CitizenCohn Yep, I've seen that too. And I'm not even saying they made wrong decisions given what we knew at the time. But they might balance these decisions a little differently going forward. — PolitiTweet.org

Nate Silver @NateSilver538

And again, the *fatality* rates have been roughly as bad as models feared/assumed. (Emphasis on roughly since there's some range of debate.) But hospitalization rates have been lower than most projections. In general, there has been some reluctance to talk about this. — PolitiTweet.org

Nate Silver @NateSilver538

Cuomo's press conferences have also featured slides where he's talked about how actual hospitalizations were considerably lower than all of the models the state had been looking at. https://t.co/aIzf6qRodU — PolitiTweet.org

Nate Silver @NateSilver538

Yeah, IMHE lowered various hospitalization numbers to around 1/3 to 1/5 of where they had been originally to make them match the actual data they were seeing. https://t.co/2uTXs2rJyV — PolitiTweet.org

Nate Silver @NateSilver538

You might say "better safe than sorry", and I agree with that to some extent. It's clear that hospitalizations need a larger reservoir of PPE, beds, etc. However, there are trade-offs if you're also keeping other people out of hospitals who could really benefit from treatment. — PolitiTweet.org

Nate Silver @NateSilver538

We know that COVID-19 cases *can* overwhelm hospitals if there are a great number of cases in one place at one time. See NYC and Lombardy. Even in NY, though, per Cuomo, the peak wasn't as high as the state's projections and some temporary hospitals were never/very lightly used. — PolitiTweet.org

Nate Silver @NateSilver538

By comparison, the Imperial College study projected a hospitalization rate of 4.4%. And that was on the low end of projections. The NYT/U. of Toronto model let you choose between 5% and 20%, which was a common range of assumptions at the time. https://t.co/dKv9t4i27C — PolitiTweet.org

Nate Silver @NateSilver538

In New York, about 2% of *infections* have resulted in hospitalizations, if you buy the state's (IMO relatively credible) serological data saying 15% of NYS / 25% of NYC got COVID-19 at some point. The rate is about the same in NYC and non-NYC parts of the state. — PolitiTweet.org

Nate Silver @NateSilver538

There's room to ask whether hospitals have been too aggressive about delaying elective procedures, which can have serious health consequences. COVID-19 has been roughly as deadly as people feared. But, it has NOT resulted in as high a hospitalization rate as most projections. — PolitiTweet.org