80,000 Deaths...

[lilpixieofterror]

So you've finally realised that your claim that China knew for months was claptrap. A pity you don't have the integrity to admit you were wrong.

Reading is hard:

...spent well over a month...

What does that phrase mean?

[Mountain Man]

So you've finally realised that your claim that China knew for months was claptrap. A pity you don't have the integrity to admit you were wrong.

Because apparently "well over a month" does not mean months.

[oxmixmudd]

maybe, but they also are not taking into account any effective treatments that might be found either. There are a lot of missing data in these models.

I guess my point was that we can only achieve that low a mortality if we do not try to open up too early and if we maintain a non-normal level of vigilance about the virus for quite some time to come, I wasn't commenting on how those models can be improved.

The tendency I see right now is for people to think that these changes in the models either mean the current measures were not necessary, or that they can be relaxed soon.

[Juvenal]

The models from the beginning have always been way too high, but I guess the "experts" were operating with a "plan for the worst, hope for the best" mentality.

In IC-L 9, the estimates used "conservative" figures, and explicitly clarified that meant "pessimistic." At the same time, many took the 2.2 million figure as an estimate, even though it was explicitly described as unlikely because it depended upon no changes in behavior, which everyone has always known would not happen. That was unfortunate.

But the larger takeaway is that if we relax the NPIs, as we are sure to do, because we can't live in lockdown forever, we can expect multiple waves during the 12 to 18 months or perhaps longer before we have a vaccine. And if those waves aren't met with re-introduction of targeted NPIs we'll be heading back to an estimated million deaths in the US.

[Mountain Man]

In IC-L 9, the estimates used "conservative" figures, and explicitly clarified that meant "pessimistic." At the same time, many took the 2.2 million figure as an estimate, even though it was explicitly described as unlikely because it depended upon no changes in behavior, which everyone has always known would not happen. That was unfortunate.

But the larger takeaway is that if we relax the NPIs, as we are sure to do, because we can't live in lockdown forever, we can expect multiple waves during the 12 to 18 months or perhaps longer before we have a vaccine. And if those waves aren't met with re-introduction of targeted NPIs we'll be heading back to an estimated million deaths in the US.

Thing is, even without any behavioral changes, there was no guarantee of hitting that 2.2 million mark because it's all speculative. I suppose simple logic would suggest that the less people interact, the less the virus will spread, but there is really no way of determining to what extent that is true. For all we know, the social distancing has only had a minimal impact on the spread of the virus, and that things were never going to get as bad as the models (and the boatload of assumptions behind them) suggested.

[Juvenal]

Thing is, even without any behavioral changes, there was no guarantee of hitting that 2.2 million mark because it's all speculative. I suppose simple logic would suggest that the less people interact, the less the virus will spread, but there is really no way of determining to what extent that is true. For all we know, the social distancing has only had a minimal impact on the spread of the virus, and that things were never going to get as bad as the models (and the boatload of assumptions behind them) suggested.

There were two variables involved in the 2.2 million figure, the initial reproduction rate, R0, and the case fatality rate, both of which are estimates. The estimated R0 of 2.4 is likely within 10 percent, but the CFR can't be pinned down much closer than between 0.25 and 1 percent.

There is no doubt that social distancing has reduced R, however. That's a basic conclusion, because it directly impacts the number of contacts. Cutting the number of contacts in half will cut R in half.

[Sparko]

There were two variables involved in the 2.2 million figure, the initial reproduction rate, R0, and the case fatality rate, both of which are estimates. The estimated R0 of 2.4 is likely within 10 percent, but the CFR can't be pinned down much closer than between 0.25 and 1 percent.

There is no doubt that social distancing has reduced R, however. That's a basic conclusion, because it directly impacts the number of contacts. Cutting the number of contacts in half will cut R in half.

The new study, published in the Emerging Infectious Diseases journal, shifts the R0 for COVID-19 from about 2.2 to about 5.7.
https://www.forbes.com/sites/tarahae.../#45ed043e29a6

[Mountain Man]

The new study, published in the Emerging Infectious Diseases journal, shifts the R0 for COVID-19 from about 2.2 to about 5.7.
https://www.forbes.com/sites/tarahae.../#45ed043e29a6

People were starting to become a little optimistic, so it's time for the "experts" to roll out the latest "SCARY!" numbers.

[demi-conservative]

People were starting to become a little optimistic, so it's time for the "experts" to roll out the latest "SCARY!" numbers.

If R0 is closer to 6, that implies that there are so many undetected cases and the actual fatality rate is much lower than the 1% value they love to frighten with.

[Mountain Man]

If R0 is closer to 6, that implies that there are so many undetected cases and the actual fatality rate is much lower than the 1% value they love to frighten with.

They count on people assuming that the rate of deaths will keep pace with the rate of infections and that we're sailing towards the original "millions dead" projections.

[mikewhitney]

If R0 is closer to 6, that implies that there are so many undetected cases and the actual fatality rate is much lower than the 1% value they love to frighten with.

Right. The numbers don't work.

[demi-conservative]

Right. The numbers don't work.

The latest narrative by NYT etc is that the virus arrived in New York City at least one month before the 'first case' in 1 March. This would of course imply that the number of cases are understated from at least 10 to 100 times, and the fatality rate overstated by that amount.

It would also mean all the models for the USA are completely useless since the curve started at least one month earlier. Everything is off.

[Juvenal]

The new study, published in the Emerging Infectious Diseases journal, shifts the R0 for COVID-19 from about 2.2 to about 5.7.
https://www.forbes.com/sites/tarahae.../#45ed043e29a6

That shouldn't make much difference under the IC-L 9 suppression scenarios. By lowering the triggers for resuming social distancing — which has minimal effect on the time under relaxed social distancing — we should still be able to limit fatalities to a single digit percentage of those we'd otherwise expect. It would make it more difficult to restrict fatalities to the 80,000 figure, but that may still be within reach, even under the pessimistic compliance assumptions of the report.

[Juvenal]

The new study, published in the Emerging Infectious Diseases journal, shifts the R0 for COVID-19 from about 2.2 to about 5.7.
https://www.forbes.com/sites/tarahae.../#45ed043e29a6

Like most of us getting our heads around epidemiological math without the benefit of study under an epidemiologist, these concepts clarify over time.

When I first started reading, I didn't understand that the "contacts" we're trying to reduce using social distancing are the same "contacts" that shoe-leather epidemiologists try to map out via "contact tracing." I didn't understand that a contact is more than just a person who gets inside our six foot bubble, it's the time they spend there, too.

When I first started reading, like most folks, I didn't know what the 0 in R0 meant, and misused the term repeatedly. It means the initial, t=0, value for R. I didn't even know what the "R" stood for. It means "reproduction" rate.

And more than that, I didn't understand that both R and R0 vary by location, because the base level of contacts varies by location, and reproduction, R, is directly proportional to contacts, at any time, including time zero.

The R0 in the above study is for Wuhan, a densely populated city that should expect a higher R0 than Montana, say, where population gets measured in square miles per person rather than persons per square mile. R0 depends on the base level number of contacts. That number is bigger in big cities. It's bigger in New York. It's bigger in Wuhan.

The study measured reproduction in Wuhan. It found an R0 more than twice what we should expect for its weighted average over geography. Thanks for bringing this up.

This is going to change my thinking about how SD should be rolled out or relaxed across the country.

[firstfloor]

80,000 dead in less than 10 days from now. The results of action to avoid infection will take several weeks to show up in the data. The objective is to increase the doubling time and get the basic reproduction number below 1, when it will gradually die out.