More on NY's outbreak and subways and testing
NY's outbreak was out of control already before they had enough testing to figure that out, as testing was almost non-existent until about 3/7 (when less than a few hundred tests had been run - NY testing wasn't even approved by the CDC until 3/2), well after the outbreak truly started; see the graphic below for the very few positives before about 3/10 really (when NY hit 100 new cases in a day), when testing really got going.
So, even though NY has now done more tests per capita than any country, massive testing was started too late to stop the exponential spread, especially since social distancing wasn't strongly in place before about mid-March, when it all started (NY gatherings were limited on 3/12 and schools were closed on 3/16, but the full stay at home order wasn't until 3/22) - especially on the very crowded subways and trains. Other states had later outbreaks than NY and had enough testing to keep better on top of that (like CA).
Speaking of NY, a cool paper from MIT (thanks
@Knight Shift) just came out analyzing the NY outbreak data and the likely correlation to subway ridership and transmissions. Population density and the subways/buses have been discussed a lot in this thread, so was wondering when we'd see a paper like this. No doubt that the subways were the biggest transmission vector imaginable and should have been shut down first, before the schools or anything else, but that would have been a very hard sell before there were deaths. Nowhere else in the US has the population density or the commuter density that NYC does (and that also seeded the major outbreaks in NE NJ and SW CT).
And the first NY death wasn't until 3/14, so closing subways and the city before then would've been a hard sell, but it would've been the thing to do, in hindsight. There were 60 a week later and nearly 1000 by 3/28, which is not surprising since deaths lag infections by 2-4 weeks, which is why NY really needed to be shut down in early March. And as an aside, disinfecting the subways is like rearranging the deck chairs on the Titanic - almost useless compared to keeping people from being on top of each other in the subways (and buses/trains, too), where the vast majority of infections occurs.
http://web.mit.edu/jeffrey/harris/HarrisJE_WP2_COVID19_NYC_13-Apr-2020.pdf
10,000 infections in NYC on 3/1?, more on Pueyo article, and why, again, testing is so critical...
More on the Pueyo "Learning to Dance" article in Medium and today's Times article showing a Northeastern University analysis that NYC likely had 10,000 infections as early as 3/1 (and the rest of the US was far more infected than thought, also); these dovetail with the NYC subway analysis above, too.
https://www.nytimes.com/2020/04/23/...action=click&module=Spotlight&pgtype=Homepage
Feel like I've been screaming for weeks about how our lack of early, aggressive testing capability was the most important element in letting this outbreak get way out of control, especially in the the NYC metro area (<200 tests in NY through 3/6!). If NYC had 10,000 infections by 3/1, as per this new analysis of the data, that's roughly 1000 per 1MM in population and as I've noted a few times, some experts believe that aggressive interventions need to be in place before 200 cases per 1MM in order to control an outbreak.
And the only way to know if you have an exponentially growing outbreak is to have aggressive early testing (like South Korea and Taiwan had) and the US was so far from that it's sad, given the testing debacle. I have been saying we needed to be in lockdown starting early March, but this analysis probably says it should have been late February, when we weren't even testing and had no data and no way would anyone have supported shutting anything down without data. Yes, NY (and NJ) now lead the world in tests per 1MM, but that's only because we started so late in the outbreak and our testing was always playing catch up to the outbreak.
The graphic below from Pueyo's "Learning to Dance" Medium article shows how important early testing is to get ahead of an outbreak, as reflected in the low % of positive tests in countries like SK, Taiwan, and Hong Kong. The US number is bad, but the NY/NJ area numbers are far worse as our area was in the 40-50% of tests being positive for weeks at the height of the outbreak, as we were so far behind. The second graphic shows where various countries are on the "hammer" (heavy social distancing to control major outbreaks) vs. the "dance" (relaxing harsh social distancing in a careful way to avoid a second outbreak) - and we're clearly nowhere near ready to "relax" social distancing yet, based on that curve, even in states with less of an outbreak, so far, vs. the worst states.
Coronavirus: Learning How to Dance
And one more time, but if we had done what South Korea had done (and Seoul is almost as densely populated as NYC, so don't say it's not possible), we could probably be near their 5 deaths/1MM and not the 144 deaths/1MM in the US - or sadly the ~1000 deaths/1MM we have in NY or ~500 deaths/1MM in NJ. Even at 10 deaths per 1MM, that would be 3300 in the US vs. the ~47,000 we have now (and 200 in NY vs. the ~20,000 we have now).
I know none of this solves the problem of what to do now, but it should at least shout at people/politicians to not repeat this mistake across the country by relaxing social distancing without having massive testing and efficient contact tracing/quarantining and a mask culture in place. Repeating the folly we just went through would be Einstein's classic definition of insanity: doing the same thing over and over again, but expecting different results.
Minor comment: I do think the Northeastern study probably overstates SF's outbreak (9300 on 3/1) relative to NYC's (10,700 on 3/1), since the outcomes diverged so heavily after that, with SF having much less of an epidemic and far less deaths. I don't think their 3/15 shelter-in-place order (vs. NY's on 3/22, even though NYC had shut schools, restaurants, bars and such by 3/16) should have made that much of a difference in outcomes. But that's arguing minutiae vs. the big picture of there likely being major outbreaks underway by mid/late February. Also, as per the subway/mass transit study, the far greater population density and "mixing/contacts" of commuters in NYC would drive a much faster outbreak vs. SF, as has been seen.