Hacker News Reader

Comment by krona | Hacker News Reader

force everyone to shut down and get it over with quickly

On a purely factual point, the way to get it over with quickly is not to stop people coming in to close contact with one another. Everyone seems to have forgotten what the original and only purpose for the lockdown was: to avoid overwhelming hospitals.

View on HN

I don’t think that’s as purely factual as you make it out to be. Maybe that was the purpose for the US, but other countries did it for the purpose of eradication (eg New Zealand) or to get it to a point that the contact tracing workload was manageable. The hospital capacity threshold is just one of many arbitrary targets.

It's not entirely arbitrary - the fact that the US is not a small island about 1000 miles from its nearest neighbour seems to comprehensively eliminate the New Zealand-style elimination option, given the trouble they're having with it. Like, even if you assume an equally competent lockdown in the US would have the exact same per-capita chances of infections slipping through from outside, with no disadvantage from the porous southern land border with a country that has a major outbreak, it still has 100 times the population. Which means that rather than causing a new major imported cluster once in 100 days, the leaks which the New Zealand government can't identify, fix, and reckons are likely an inherent limitation that they'll just have do deal with would cause them roughly daily. That means no elimination, not even if the numbers worked out slightly better than my rough estimate. And remember, New Zealand was the last big, well-known elimination success story left standing.

100x the size is neither beneficial nor harmful when controlling a pandemic. As demonstrated by the US’s successful reduction in spread before opening things up again.

0.9 new cases per existing case and it goes away. 1.1 new cases per existing case and it grows uncontrollably. There isn’t a true middle ground.

It doesn't even make sense to talk about cases per existing case for imported cases, because by definition they're not caused by existing cases within the country. Like, maybe that could help tame down the clusters from those imported cases, but then you have to impose lockdown restrictions indefinitely and lose one of the big benefits of elimination over the "try and keep R just below 1.0" strategy - being able to lift those restrictions and avoid the ongoing economic and social damage they cause.

Also, I think that in reality this isn't actually as scale-free as the nice simplified models suggest, and having 100x the population would make it harder to stamp out spread within the country. Remember, the number of cases per existing case isn't some fundamental physical constant - it's an average over a bunch of different opportunities for transmission with wildly varying expected numbers of new cases per existing case. 100x the scale means 100x the opportunities for low-probability superspreading events that ruin your day.

Reproduction rates is below 1 mean a small number of new cases from minor outbreaks or border crossings are effectively meaningless. The US is currently reducing the number of active infections with poorly implemented social distancing and minimal mask usage. DisneyWorld for example is open and without becoming a hotspot of infections.

Reaching that point 4 months ago was a completely reasonable goal. Sure, our current situation is improved by some herd immunity, but that’s far from the only effect in play.

Canada manages pretty well with a "porous southern land border with a country that has a major outbreak". I wonder what the difference is. https://www.theatlantic.com/ideas/archive/2020/09/i-crossed-...

Reply on news.ycombinator.com