Nicely done with a very important caveat - social distancing (in its various options) is the only tool in our toolkit but NOT the only something that stops it or slows it down.
There is the joke of a married couple who are both scientists who send out a family newsletter 3 months after the birth of their first child, which they titled “Experiment Number One” and included in it a graph and analysis of the growth of their child to date: “To date Experiment Number One has doubled in rate over 3 months and has a current weight of 6.2 kg. Doubling its weight every three months we therefore project that by 10 years of age Experiment Number One will weigh roughly 6,816,972,092,211 kg.”
Of course the math of modeling deaths at any particular day into an epidemic is not quite so straightforward as continuing its initial curve up forever.
The reality is that we have no real knowledge of what the curve would look like in absence of applying any of the social distancing tools because we have no real idea of the some basic critical inputs that go into the models. Will this baby keep growing at this rate forever, or begin to slow down, when? With babies we know the doubling rate progressively slows down after 3 months; the experiment has been run with others of the species before. We’ve never had a baby of this exact species before and we don’t know too much about it yet. Lacking that information we still have to act and the least poor choice is to at least slow it down as best we can until we have better information and as we increase our surge capacity from its current fairly pathetic levels.
Cuomo has been advocating widely testing to see who is already immune by now and allowing them to be part of the first cohort to return to work. Many of those found will not ever have been tested by swab or even presumptively diagnosed. Some will not have even suspected they had COVID-19; some may never have had any symptoms.
His plan not only would not only be a great way to restart the economy with the lowest risk pool; it would provide the modelers with one of the critical missing inputs of how the undocumented and not very ill portion is, and thus how fast the pool of “Vulnerable” drops, and how quickly the pool of “Resolved” grows for the models. IF it turns out that with this testing there are a large number of undocumented asymptomatic to mild infections for every one currently identified, maybe ten to one who knows, THEN a plan in which transitions to using more focused social distancing as the specific tool is one that can be applied with more confidence and one that prevents the morbidity and mortality of the current “isolate them all” approach.
Do medical personnel have access to enough tests now so that they can carry out widespread testing. I can’t seem to keep up with all the bonfires currently burning/smoldering/ready to burst into flame.
The widespread testing for evidence of having HAD the infection is not the swab test (using reverse PCR); it is testing for antibodies (also referred to as “serologies” or “IgG and IgM”). Weeks ago the CDC briefing described these tests as ready to roll out within the week. No idea what happened after that. It’s a completely different platform and getting enough done to be able to state that X% of 20 to 30s and Y% of 30 to 40s (etc.) in NYC show evidence of having been infected, compared to A% and B% of identified cases, does not take tens of thousands of tests run, any more than a poll needs tens of thousands to estimate voting results, at least within some broad confidence intervals.
Googling I can find from a week ago this Science mention that they have already been used in studies on small scales.
Also needed is more conclusive proof that children are minimally contagious when they have COVID-19. If we start with 24% of the American population functionally in the Resolved bucket (not spreading it) the model would result a completely different course and some actions are completely needless or even harmful.
Although… if you want to find out if a person is safe to return to the workforce, you’d need an antibody test to prove they’ve had it, and a PCR test to prove they are no longer infectious, right?
Current standard already is to consider safe to return once 72 hours of fever-free and symptoms consistently improving. I don’t know exactly what data that recommendation has been based off of. Nevertheless for other viral infections it is established that some low level viral shedding long outlasts the period of meaningful contagiousness. The standard for influenza in comparison is safe to return after 24 hours of fever-free even though it is documented that some viral shedding can last for up to ten days.
I’m curious as to how that works when there’s more than one strain of the virus and there have been folks who have recovered and then gotten re-infected.
The national bloodbank here in the Netherlands is supposed to be testing all donations on antibodies this week. If they actually pull it off, it should give an idea of the spread of the disease by the end of the week (based on 10000 data points).
We are pretty much testing only people that are sent to hospital. That’s why the “mortality rate” is close to 5%. These rates are meaningless, as is the 10 % rate in Italy… but still people keep throwing it around.
What, R[sub]0[/sub] is going to suddenly start working differently?
Look, the rate of spread of this thing is like the rate of spread of any other disease. It’s all about how many new people each infected person infects. That’s R[sub]0[/sub]. Absent social distancing, that’s going to be a constant until enough people have developed a resistance to the disease already to reduce the number of uninfected people each infected person runs into. To noticeably reduce R[sub]0[/sub] that way, a good portion of the population has to have developed resistance.
We’re nowhere near that point yet.
And until we are, the exponential rate of growth will continue as before, modulo the extent that social distancing reduces the number of contacts each infected person has.
So yeah, I think I can extrapolate along an exponential curve for the next freakin’ two and a half weeks. Based, as before, on the assumption that social distancing is only minimally effective - and using the difference in actual deaths from what the exponential curve predicts as a measure of the effectiveness of social distancing.
One country I’ve been wondering about is Russia, from which little has been heard (at least in the mainstream US press) since the crisis began. It appears their response, or lack of one, has been strikingly similar to that of the US in the early days of the epidemic: mostly claim the outbreak is under control and deny there is an issue. Well, Russia (meaning Putin) apparently is coming round to the notion that they may have a problem:
The article mentions that Putin has declared next week to be a paid stay-at-home holiday, but that seems a laughably tepid response and otherwise there seems to be no long-term plan to deal with the outbreak as yet.
Deaths per million per state in the US as of this morning (death numbers from Worldometer, state populations from Wikipedia). Just the top ones, with country numbers for perspective.
Italy : 124
Spain : 89
Andorra : 39
Iran : 27
Netherlands : 25
Switzerland : 22
France : 20
New York : 19.8
Belgium : 19
Washington : 17.3
Cayman Islands : 15
Louisiana : 14
Luxembourg : 13
Vermont : 12.8
Wales : 8.7
Sweden : 7
Denmark : 7
UK : 7
New Jersey : 7
Guam : 6
Curaçao : 6
Iceland : 6
Channel Islands : 6
Portugal : 6
Connecticut : 5.3
Nevada : 5.2
Austria : 5
Georgia : 4.5
Michigan : 4.3
District of Columbia : 4.3
Colorado : 3.3