In comments Frank is explaining his theory of how the pandemic is worsening.
I’ve been looking at the data from a different perspective than you have. You are looking at new cases (in a few locations), whereas I have been following the slope of the log2 of cumulative cases vs time. I am averaging the noise over the last 10, 20 or 30 days and comparing these slopes to see how much the doubling time has changed. So I’m focusing on the bigger picture – total cases have roughly doubled NATIONALLY in the last month – not on what has happened in the last week of fairly noisy data in a subset of locations. The doubling period was slightly shorter for the last 30-day period (24 days to double) than the last 10 days (36 days), so things are getting better. (I’m also almost a week behind now.)
I haven’t been posting everything I looked at. But I’ve been looking at lots of data– including state, county, national and international. Given the argument in comments, I thought I’d post the US new daily death, and new daily mild, severe, critical cases. The small dots are daily values, the large are 7 day averages. (Note: I have no idea how mild, severe and critical are defined.)
I leave it to the reader to decide if I’m looking at “noisy” data, or “a few locations”. My thought is that as we were discussing what US policies might be, US data would be appropriate. I could, of course, post cummulative deaths, which are the integral of this graph. But I think daily deaths, and a semi-log scale are more useful for understanding the evolution of a contagion.
For the past week, the appearance of the slope of daily deaths on a semi-log graphs is negative. If this were a “well mixed” contagion (which it is not) that would suggest an R<1, which is what we’d want. Would I like a steeper negative slope? Hell yeah! But a negative slope means the contagion is not getting worse.
Might it go back to getting worse? Sure: mistakes could be made. We don’t entirely know what to do to control the contagion. But I think it’s is important to recognize that right now it’s not getting worse in the US over all and in many hard hit places, it is slowly getting better.
Consider 1% infection rate, with a test that is 90% accurate.
There will be an 11-1 ratio of noninfected to infected among people who test positive, if everyone is tested. South Korea with its extensive testing regime may be giving a fatality rate that is too low by an order of magnitude.
If the test is 99% accurate, you still have an even split among positive tests.
A 15% infection rate with 90% accuracy would be 53% are positive, while 95% accuracy would be 77% positive among positive tests.
MikeN,
One of the reasons I like death rates is we very rarely think someone is dead when they are not dead.
.
(Yes. I’ve heard people complain about attribution. But at least we do know those people are dead.)
Lucia,
Consolidating different regions at different stages of spread is fraught with problems. In some places the rate will be rising, in some places almost constant, and in some places (especially where LOTS of people have already been infected) the rate is still falling pretty quickly.
.
Here is the weird thing: NYC, with a far higher death toll (per million) than just about anywhere on Earth, and with a large fraction of people already either exposed and recovered or naturally resistant, could probably ease restrictions with little consequence, but will not, while places with relatively fewer cases, and so at much higher risk of rising cases and deaths, are rapidly easing restrictions already.
SteveF
Absolutely. To some extent, the reason things are falling is those hit earliest were places with large population concentrations. NY, LA, ChiTown, Detroit, etc. are approaching or past their peaks. We are now seeing the contagion move to somewhat lower population concentrations.
We are going to see outbreaks in other cities. But because of the way this transmits and the greater state of readiness, they probably won’t be hit as hard. Also, the populations are simply smaller and there won’t be one big correlated block of people being hit at a particular time.
Obviously, there still is a danger in getting 2nd peaks in any and every city. We don’t know how big a danger because, there may be pre-existing immunity from people who caught “corona-colds” rather than mere “rhino-colds”. We don’t know how big that might be nor if it’s evenly distributed in the country or different demographic groups. (We can probably guess it’s better distributed among young people who cluster is schools! 🙂
Lucia,
“Obviously, there still is a danger in getting 2nd peaks in any and every city.”
.
I agree that is for certain true in most places. But maybe not NYC or Milan. It’s a bit like the old Bond film: ‘You only die once’.
.
The T-cell study that showed the widespread presence of existing reactivity to coronavirus spike protein (was it ~35%?) did not break the positives down by age, so no telling for sure the rates across populations… but kids do get lots of colds.
I don’t think there has been a single region that hit anywhere near herd immunity. There is still a good possibility that there is some built in cross immunity in the population. It’s unfortunate there has been no progress on this and other basic questions. The science community is incapable of acting very quickly, and ironically they seem rather proud of it.
.
I’m pretty disappointed so far, but we shall see if they save the day with a vaccine or effective treatment. I think a 1000+ deaths per day is enough to go into high(er) risk science. Things like intentionally exposing volunteers to the virus to get better information sooner is currently unethical.
Tom Scharf,
I assume intentionally exposing volunteers to the “cold” version of the corona virus is also unethical. I’d volunteer if they were doing it! 🙂
Sh** is hitting the fan in Brazil. 20K cases per day and climbing. The US peak was around 35K. They seem likely to get to 1M before it is all over. Who knows how accurate Brazil counts are, likely larger under counts than the US is my guess. Their leader is a bit of a wacko.
An update on college admissions scandal:
“Actress Lori Loughlin and her husband, fashion designer Mossimo Giannulli, have agreed to plead guilty to fraud conspiracy charges in connection with their involvement in the nationwide college admissions cheating scandal.”
“Prosecutors will recommend two months in prison for Ms. Loughlin, and five months for Mr. Giannulli, as well as fines of $150,000 and $250,000, respectively. They also agreed to two years of supervised release and community service.”
.
The current count is 23 people serving jail time for this. About half the cases are still pending I think. This may be overkill in the grand scheme, but this was a “must send a message” type of prosecution.
Tom Scharf,
Brazil is on track to take number one in new cases/day and new deaths per day real soon now. Social distancing is simply not an option for a lot of Brazilians. Peru, Chile and Mexico aren’t looking very good either. I think the numbers from Mexico are highly questionable too.
Tom Scharf (Comment #185107): “I don’t think there has been a single region that hit anywhere near herd immunity.”
.
Not so. New York was at 20% a few weeks ago. It might well be that the drop off there is due to expanding herd immunity. And if 20% is enough for herd immunity there, it might take even less elsewhere.
My understanding was that Team Science says at least 60% to reach herd immunity. It may be possible that herd immunity is reached at lower numbers but that is not something that is mainstream from what I have read.
Tom Scharf (Comment #185115): “My understanding was that Team Science says at least 60% to reach herd immunity. It may be possible that herd immunity is reached at lower numbers but that is not something that is mainstream from what I have read.”
.
Yes, but mainstream seems to treat R0 as if it were a property of the virus. But if that were true, there would be no point to social distancing, let alone lockdowns.
.
But R0 is obviously higher in New York City than in Albuquerque. And it is obviously different in different subgroups within any city. And the early stages of the epidemic is obviously disproportionately among the subgroups with the highest R0. So an R0 estimated from the early stages of an epidemic in a virgin population will be much higher than the R0 averaged over society as a whole.
.
The upper limit of R0 is perhaps as large as 3; that would imply that herd immunity would require 67% to have been infected. But the real value is smaller. If the average is 2, then 50% would have to be infected for herd immunity. If the average is 1.5, then only 33% would have to be infected. If half the population is already immune, those numbers are reduced to 25% and 16%.
.
Of course, social distancing has reduced R0. If the epidemic peaked with 20% infected under a social distancing regime, then relaxing restrictions would let it start again. So at least in NYC, more than 20% infected will be needed to maintain herd immunity as things return to normal. I am guessing that even there, much less than 60% will do the job.
I’ve been reliably informed that a study for a marketplace in Iquitos, Peru, of over 1000 people showed ~70% with antibodies.
In a homogeneous population with an Ro of 3, the new case rate peaks at 67%, otherwise known as the HIT, but the total infected will be higher, if I remember correctly. For an Ro of 2.4, 81% of the population would end up having been infected, according to Ferguson. But we are reasonably certain at this point that the population is not homogeneous.
That was the point of Nic Lewis’ post at Climate Etc. Looking at the data from Stockholm County, it looks like the Herd Immunity Threshold there is about 17%. It’s unlikely, IMO, that the HIT is anywhere near as high as 60% anywhere.
But even with the ending of hard lockdowns, we’re not going back to ‘normal’, as it were. There still won’t be masses of people going to sports events, for example, for a long time, if ever. So it’s unlikely that we’ll see R going much above 1 except for local clusters from super spreaders like the one in South Korea who seems to have infected over 50 people in one night.
Dewitt
Yes. Quite a bit higher using an SEIR model. I can’t remember the theoreitcal limit– but I ran the model and I think it’s about 80-90%. There is quite a big of overshoot.
.
Of course, that all assumes people don’t all self quarantine and “Ro” stays the same during the whole contagion. But that’s the same assumption that gives the herd immunity value for the “top” of the epidemic.
True herd immunity for a homogeneous population– which would be the value where Reffective <1 when we go back to our pre-Covid level of interactions would be the (1-1/Ro) level, where the fraction of immune for any reasons is (1-1/Ro).
But really, in the long run, to protect those in nursing homes, we are going to need the vacinne. Whatever Ro is in nursing homes or assisted living facilities, just having the virus introduced results in it whipping through. With no vacinnes, no one can visit their loved ones, you can’t have any group activities and so on.
Lucia,
“With no vacinnes, no one can visit their loved ones, you can’t have any group activities and so on.”
.
An accurate, fast swab test with very low false negative rate is needed to screen people for virus shedding. It is unclear if any have yet been developed, with conflicting reports of accuracy.
.
If such a test became available, visiting elderly care facilities would become possible.
SteveF,
Maybe… but who wants to have a swab shoved up the nose every time they visit their mom, mother in law and so on? I don’t. I’d much rather get a vaccine, and have them get one!
But yeah, you are right. We could do it that way.
Lucia,
A vaccine given to the general public is probably a year or more away, even if one becomes available for select people a bit sooner. That’s a long time for an aging relative to wait for a visit.
lucia (Comment #185123): “but who wants to have a swab shoved up the nose every time they visit their mom, mother in law and so on?”
.
Not to mention the risk of exposure to the person taking the swab and the PPE used. But there is a saliva test: just spit in a container, put the cap on, and hand it over. Then wait a day or so for the result. The wait is a big problem.
Nursing homes don’t have the instruments to do the fast tests. Even if mandated, it will likely take a long time to make enough of them.
We might or might not eventually have a vaccine.
——-
Do we actually know that nursing home residents are at signifcant risk from visitors? After all, the outbreaks almost entirely occurred with visitors locked out. Transmission of the Wuhan virus requires close contact. Compared to visitors, the staff have *much* closer contact with residents. Maybe quality masks are enough.
Mike M.
An infected staff member represents a larger threat than an individual visitor. But there are potentially a lot of visitors. These include relatives, but also volunteers who want to come sing, bring in therapy dogs or do any number of things that potentially relieve the boredom of living in a nursing home or assisted living.
.
I’m hopeful about a vaccine. Which means diddly squat about likelihood we’ll get one.
What is the evidence against the proposition that the lockdowns increased cases and deaths?
MikeN,
South Korea? Japan? Singapore? Hong Kong? It depends on what you mean by evidence.
Lucia,
“I’m hopeful about a vaccine.”
.
I am also. The coronavirus is going to mutate, but it is not likely to mutate as fast as influenza… which manages to stay ahead of vaccine production. Influenza also has a known wide range of animal hosts, including domesticated foul and pigs, which act as reservoirs/developers of new infectious strains. (https://www.scientificamerican.com/article/influenzas-wild-origins-in-the-animals-around-us/)
A band new new coronavirus appears to be unusual. It is likely to have originated in bats, not domesticated animals. So there seems a good chance repeated introduction of new coronavirus strains can be more easily avoided than new influenza strains. Which is not to say the virus won’t evolve fast enough in humans to make vaccines less effective, but I think the weight of evidence (including T-cell cross-reactivity form another coronavirus stain that causes common colds) suggests a vaccine may be a lot more effective than flu vaccines.
DeWitt Payne (Comment #185130): “South Korea? Japan? Singapore? Hong Kong? It depends on what you mean by evidence.”
.
I am pretty sure that there were no lockdowns in South Korea, Japan, and Hong Kong. I think that Singapore eventually did go to lockdown. The last I looked, they were getting more new cases per capita than the U.S. But I think that was the cause of the lockdown, not a result of it.
The history of trying to make coronavirus vaccines is not encouraging. The problem is not mutation; those virus have pretty stable genomes. It is partly that it is hard to produce an effective immune response in the upper respiratory system. Also, it seems that immunity against coronaviruses is not long lasting. The biggest problem might be that the deadly thing with the Wuhan virus is the immune system response; the whole point of a vaccine is to induce an immune response. So even if a vaccine looks good in testing, its use will have to be ramped up very gradually.
.
Some info here: https://www.abc.net.au/news/health/2020-04-17/coronavirus-vaccine-ian-frazer/12146616
Mike M,
So, if having an antibody titer against Covid19 and T-cells primed against the virus, which is the expected result of a vaccine, is not going to provide protection, are you suggesting the people who have had and cleared the virus over the last few months are just going to be re-infected in a short time? Not a rhetorical question. If that is true, (and I honestly think it is utter rubbish) then every policy and every effort to control the spread is pointless, as is the effort to develop a vaccine. Do you really believe that? (not rhetorical)
“no lockdowns in South Korea, Japan, and Hong Kong.”
Well not “full” lockdown. These are places burnt by SARS epidemic and went into masks and social distancing without government orders pretty quickly. Japan’s State of Emergency aimed to reduce social contact by 70-80% and had a population that by and large would do what they told. Ditto for compliance in others. SARS also left Korea and HK well prepared and with powerful contact-tracing capabililities already in place. HK 158 tests per active case; S Korea 71; (USA is 8). NZ couldnt get close to their levels of test and trace until near end of lockdown. Singapore was similarly prepared and initially did well, but seem to be paying price of ignoring their migrant workers living in crowded quarters.
Japan to me is odd one. Their testing would suggest severe under-reporting but no sign of mass dying despite an aging population. Maybe their social distancing is enough.
R0 is just transmissivity before any measure in place and depends obviously on virus capabilities but also heavily on cultural norms with respect to opportunities to transmit. eg kiss versus handshake versus namaste versus nod gruffly for greeting. Clearly varies enormously from place to place. Australia is was 3, here it was 4. I struggle to understand why difference between NZ and Aus. Aus cities are bigger and denser; cultural norms among Europeans pretty similar.
SteveF,
I have seen the claim that immunity against other types of coronavirus does not last. I have no independent opinion on the subject. It sounds strange to me.
I have also seen that cited in the context of why we can not count on herd immunity resulting from the virus moving through the population. They then conclude that we really need a vaccine. That makes absolutely no sense to me.
What is happening with the lockdowns is that people are in crowds less, but are more crowded at home or at grocery stores operating less hours, combined with less outdoor exposure and less fresh air.
Anecdotally, I know of people who got sick after lockdowns, including in the early days of school closure where I suspect remaining staff were closer together than before. Combined with what appears to be kids largely immune and not spreading the disease to adults(not sure how they can spread to kids but not adults), then it’s possible the lockdowns are not effective and perhaps counterproductive.
Even without shutdown orders, people were already avoiding crowds, in some cases wearing masks- which at the time was being called unnecessary,
MikeN
They weren’t wearing masks around here before shutdown. They did wear the before they were ordered to wear masks. But that order came quite a bit later.
.
The were still going to work, not staying at home.. The St Patricks day parades around here were cancelled. I’m not sure I’d call people not going there “avoiding” them. You can’t “avoid” something that was not happening.
The google mobility data I’ve looked at largely contradicts the theory that people stopped going places in any major way before municipalities, governors and so on put various lockdowns, stay at homes and other orders in place. Mere suggestions didn’t do it.
.
As far as I can tell, the idea that people were staying home without the lockdown in any significant degree is just made up. But if you have data, I’d be glad to look.
.
Cases and death’s have dropped in the wake of lockdowns/stay at homes and generally in the time frame one would expect based on the incubation and contagion period for this virus.
.
As far as I can tell there is absolutely no evidence lockdowns are ineffective or counter-productive. At least, none in either (a) death data (b) case data.
Phil Scadden (Comment #185141)
May 21st, 2020 at 8:06 pm
Phil,
Here in Australia, ‘transmission’ rate is actually quite low. About 2/3 of our cases had either acquired the virus on cruise ships, or were people returning from other countries and bringing the virus back with them. Most of our transmission cases were able to be contact traced back directly to those people who brought the virus in.
Mike M
So have I. But the articles never define a time period for not lasting. I’m pretty sure immunity lasts more than 2 days. Is it 3 months? That’s not trivial.
If a vaccine worked for 3 months, I’d be willing to get a booster every 3 months. I already get flu shots every 12 months. I get them at my doctors office, but you can get them at the pharmacy. This is not a big deal to me. We wouldn’t get a wide of coverage, but if at any time 1/4 of the population was immune due to a vaccine every 12 months, that would still slow pandemics. If most people didn’t get vaccinated, but medical workers, people in meat factories, public facing workers and people in assisted living and nursing homes got it, that would likely make a huge difference.
This is especially true if the theory is that some communities are particular vulnerable. We know who they are.
If such a vaccine existed at an accessible price, we’d certainly never have to shut down the economy.
MikeN, that has to depend on what the lockdown rules are. Unless you are doing strenuous self-isolation at home, your housemates are going to get it no matter. It is too infectous. The point is stop any transmission out of the bubble.
During lockdown here, grocery stores had spaced queues with markers for 2m, limits on people in shop while vunerable were supported for home delivery instead of going out. Masks encouraged and sanitiser everywhere. I dont think there was a single case of transmissions from shopping. Exercise “in your bubble” outside was encouraged but no sitting on park benches or using playground. I guess feasibility of that depends on population density as bubbles had to avoid each other. Rules here were heavily influenced by information and practice from Asia.
I find it amazing to suggest lockdowns were counterproductive when they have brought disease under control when done hard, early and were obeyed. We went from 80+ a day shortly after a lockdown to basically contained to known clusters in 4 weeks. I think is 2 cases in last 10 days, both in managed cluster. Our region was worst hit originally, but now is 32 days since last known case detected and no active cases. With 1500 cases in a population of 5 million, dont tell me that is herd immunity. Similar picture in Australia. Does need societies prepared to cooperatively for common good.
Oh, and the pubs opened again yesterday – the last businesses to open. Nervous time now to see how it all goes, but the lockdown gave time to ramp up testing and contact tracing. Health dept believes we can achieve asia-level speeds for tracing and isolation now.
Skeptical – I dont have reference for that – it was quoted by our health department citing contact-tracing data from Aus govt early in the process. The 4 no. was from our contact tracing data, also early in process. I do see NSW modelling suggesting close to 3. And, yes of course most could be tracked to people bringing it in, but the question is about how many people did each case infect.
Lucia, I’m mostly going off what I saw of restaurant traffic being way down in Virginia, though this was after a local school closure, while the governor followed a few days later. Not sure if this was happening before.
In Illinois, I know of some restaurants that had business running around 10-20% prior to shutdown. Now they are trying to do to-go with pot roast family dinners for $20 and eventually trying buffet with all food delivered by waiters.
I agree cases and deaths went down after lockdown. They also went up as total deaths was about 40 when these were starting. Not sure what the appropriate time lag is to consider here for proper comparison of deaths before and after. The increase in cases could be caused by more testing alone.
MikeN,
Well…. yes. Local traffic dropped after a local closure.
How far prior? Obviously, they lost business a bit “prior” to the “stay at home” in Illinois because bars and restaurants were officially closed for dine in 5 days before the “stay at home” order. People couldn’t dine in restaurants once they were closed.
I think it’s also obvious they would have a huge drop in business when huge numbers of mayors collectively banned St. Patricks day parade. That is a huge celebration weekend. People cancelled their plants to load up the trains, go watch the Chicago river get died green, drink way too much and throw up their guts on the way home. Locals in Naperville also couldn’t watch the Naperville parade, and so wouldn’t have gone to the riverwalk and joined people for beer corned beef and cabbage. But this isn’t just people deciding to not go. The event was cancelled by the mayor.
Beyond that: Our restaurant traffic dropped before the quarantine because they were closed for dine in before the quarantine.
I’ve checked the dates at Wikipedia which is actually a good place for getting dates of many relevant events. I’ve been making graphs and putting the dates when important closures happen.
Provided we bear in mind all the significant closures, the mobility data do not show any significant changes before closures ordered by government officials start. Mobility shows dramatic moves between the date when the closures are announced and the date when a final “stay at home” is in effect.
A study from Wuhan in which they randomly tested and watched progress suggests the average time to death from infection to death is roughly 18 days. So you expect to see some change roughly 18 days later. That’s when we do so changes in slope of “Ro” on graphs in locations were the stay at home was fairly “sudden” (like Illinois where it’s over ~6 days.). Where Governors spread out actions (e.g. Georgia spread out actions over ~18), you don’t really see as sudden of a break.
People were discussing this amount of time before most US lockdowns, and.. gooly gee…. that’s about how much time it took to notice a change in Ro on the graphs.
FWIW For Illinois: the dramatic mobility changes are between the cancellation of St. Patricks day parades in Chicago, Naperville, and pretty much everywhere and the day when the Quarantine took effect. Between that time, there were series of closures: Restaurant take out, public schools and so on. This is over about 6 days.
People may have changed some plans before that. But overall, it was not enough to detect on google mobility.
lucia (Comment #185144): “As far as I can tell, the idea that people were staying home without the lockdown in any significant degree is just made up. But if you have data, I’d be glad to look.”
.
Your own graph clearly shows that the change in people staying at home in Illinois was almost complete by the time the stay-at-home order was issued.
http://rankexploits.com/musings/2020/effect-of-quarantine/
The same is true in almost all states, even the ones that never issued stay-at-home orders: https://fivethirtyeight.com/features/americans-didnt-wait-for-their-governors-to-tell-them-to-stay-home-because-of-covid-19/
People changed their behavior when they got scared.
——
lucia: “Cases and death’s have dropped in the wake of lockdowns/stay at homes and generally in the time frame one would expect based on the incubation and contagion period for this virus.”
.
It looks to me like the change in infection rate agrees better with the change in behavior rather than formal orders. But it is hard to tell because we don’t know when people got infected.
While looking for the link above, I came across something interesting. On April 27-28, 538 surveyed 16 experts as to what they expected to happen in Georgia.
https://fivethirtyeight.com/features/infectious-disease-experts-expect-a-surge-in-georgias-covid-19-cases/
Specifically, they were asked to predict the new daily case average for the week ending on May 16.
The actual number was 659 new cases per day. For the most recent 7 days (through May 21), it was 649. When Georgia started to reopen, new cases were just over 700 per day.
In fairness to the experts, it looks like 6 would have been not at all surprised by what has transpired in Georgia. But it looks like 538’s method of averaging biases their result towards extreme predictions.
Phil Scadden,
It sounds to me like the common denominator in places that have had success in containing the virus is an effective system of testing, contact tracing, and isolation. It remains to be seen if that is sustainable with such a widespread virus.
.
The U.S. badly screwed up testing. It seems that our governors believe they have the authority to order everyone into effective house arrest, but that they do not have the authority to order testing or contact tracing or isolation. So far as I can tell, even where contact tracing has been done, it seems to have been undertaken as a sort of academic exercise without any authority to take action.
MikeM
I have since added other lines. As you will note above, I said the drop happened between the cancellation of St. Pat’s day parade and the actual drop dead date for the quarantine.
First of course they were completely at stay at home by the drop dead date. That’s once they were absolutely required to stay at home. They were given time to prepare.
But what that first graph I made does not show all the critical acts by government:
1) Cancellation the St. Pat’s day parade.
2) Closing of public schools.
3) Closing of restaurants and bars.
4) Actual final drop dead date for the quarantine (shown.)
People didn’t just “stay home”. They stayed home because the government (either mayors or the Governor) began shutting major things down.
In fact my graph clearly shows they only started staying home after the cancellation of the St. Pats parade which is ordinarily a major event around here, and involves puking on the train ride home.
MikeM
Well.. initially our governors didn’t have tests required to order testing. So whether they had authority is irrelevant. (Most still don’t have enough tests to do that. )
The are in many states trying to gear up contact tracing. But effective contact testing is precluded if you don’t have tests. This is especially true for a disease that with many asymptomatic carriers. Yeah… so you contact someone. But you can’t figure out if they ever had the disease, have it nor or whatever. Now you have no idea if you should try to trace their contacts. It just doesn’t work.
I don’t think we can infer they thought they didn’t have “authority” merely because they did not do something that was literally impossible. Both contact tracing and widescale testing require tests.
To have done contact tracing in the US in March would have required us to first invent a time machine, go back and fix the test the CDC rolled out and then do contact tracing.
lucia (Comment #185165),
Your graph has two lines. A significant change in behavior was already occurring before the first line and was complete by the second line. But I don’t know what those dates are or how they compare to the events you identified:
The first three of those events certainly had both a direct impact and an indirect impact by increasing public concern. Without knowing where they fall on the graph, it is not possible to say which ones mattered. But it appears that the actual stay-at-home order was pretty much superfluous.
lucia (Comment #185166),
Yes, in some places effective contact tracing might never have been an option. But many states had plenty of testing in place by the time there were a significant number of cases. That was certainly the case in New Mexico. The state claimed to be doing contact tracing right from the beginning. But for whatever reason, it seems to have been not all that effective. I think that is because they just told people “you should stay home”, which is what they were telling everyone.
Maybe some states were more effective. It would be good to know.
Looking back I think two things for the US:
.
1. Badly mangled initial testing, CDC/FDA.
2. NYC
.
Even though the original outbreak was in WA, that now looks minor in the grand scheme. NYC was the super-spreader event for the entire country, at least the eastern half. Once that outbreak was in full force, it.was.over. They would have had to pull a Wuhan style city wide quarantine and lock down to stop it, just not viable in the US unless this was smallpox.
.
I will spare everyone the comments of the irony of Cuomo being treated as a hero now. Wait, I just said it, didn’t I? The disparity between how the media treats the governors of FL/NY is glaring when the states have similar populations and an outbreak that is 10x larger in one state and is responsible for spreading it through the entire eastern US.
.
As for now, it is over in the sense we have a full on national outbreak. It’s going to have to either burn itself out or we get a vaccine. I can only imagine the cat fight that is going to break out nationally/globally over early access to a vaccine. It will get ugly.
From an alternate perspective, I only see three things for the US:
.
1. Trump
2. See #1
3. See #2
Mike M
So, it appears you admit that your previous claim that governors thought they couldn’t do that wasn’t entirely correct.
Well… perhaps lack of test kits early on when they didn’t exist.
Do you have a link providing the date they started? I can only find links disucssing it in midApril. That doesn’t mean they didn’t start earlier. But midApri is not “right from the start”. It would be after the contagion is circulating in the country.
lucia,
New Mexico was doing contact tracing but it had no teeth. They apparently could not require people to cooperate, let alone be tested or isolated. The first person to die of the Wuhan virus in the state was diagnosed as probably Wuhan, refused the test, and went home.
.
New Mexico’s first positive test was reported on March 11. From March 11-16 an average of 86 tests were reported per day, of which 3% were positive. That should be adequate for basic contact tracing, at least if that was a priority. In fact, the reports of the positive tests during that time implied that they had traced the source of each new case. As of March 14, there were 10 confirmed cases.
.
New Mexico started drive through testing on the weekend of March 14-15. It took a couple of days for those tests to be reported; they seem to have shown up on March 17. Through March 16, a total of 583 tests were reported. On March 17, 689 were reported. For the week starting March 17, an average of 769 tests were reported per day. On a per capita basis, that is about 50% more than South Korea ever did. Less than 1% of those tests were positive.
If testing resulted in a mandatory away from home quarantine, many people would not get tested until symptoms were severe. That is counter productive in the US.
Poultry Plant testing in NC:
https://www.tysonfoods.com/news/news-releases/2020/5/tyson-foods-inc-releases-results-covid-19-testing-wilkesboro-nc-plant
.
“2,244 team members and contractors who work at the facility and were tested, 570 tested positive, the majority of whom did not show any symptoms and otherwise would not have been identified.”
Mike,
When did the events in your first paragraph happen?
Which means that unless they were contact tracing, March 11 was already too late. Unknown people in the site already had it. They didn’t know who they were, so they couldn’t contact trace.
.
This is my point about starting contact tracing too late. It may be sad, but by the time it was evident with 1 positive test, it was already to spread to “contact trace”.
.
That’s different from NZ and South Korea.
lucia (Comment #185186): “This is my point about starting contact tracing too late. It may be sad, but by the time it was evident with 1 positive test, it was already to spread to “contact traceâ€.
.
“That’s different from NZ and South Korea.”
.
I don’t understand. You seem to be claiming that NZ and South Korea started contact tracing before they had any positive tests, but that is obviously impossible.
No. S. Koreas started doing things before the virus had spread. That’s not the same as waiting until it has spread.
South Korea mobilized as soon as they knew things were arriving. They started monitoring airports, arrivals and so on. The result is their contact tracing and response was in place before the virus had quietly spread under the radar.
You seem to be assuming that when only 1 person has been detected, the virus hasn’t already spread. In NM when that 1 person was detected, there were probably many who had not been detected and spreading. That wasn’t the case in S. Korea due to already ongoing monitoring, interception at airports and travel restrictions on entry. NM didn’t have any of those. So: S. Korea’s intial contact tracing was before spread. NM’s was not.
Seattle attempted contact tracing on the first US case. Everyone who was in the airport shuttle was tested, and all came back negative. Perhaps it spread in the airport or on the plane, or maybe there was another carrier on the flight from Wuhan.
Lucia, the drop to 10-20% for restaurants was before the shutdown of restaurants for dine in. What is usually at least a 1000 guests turned into 200 at best. I’m not familiar with what additional order came a week later.
lucia (Comment #185194): “No. S. Koreas started doing things before the virus had spread. That’s not the same as waiting until it has spread.”
.
I don’t know just what you mean by “spread”. In any event, the fact that the South Koreans started some actions before it spread does not mean that they kept it from spreading. On Feb. 18 they had 31 cases, one week later they had 977. It was the super spreader event at that time that kicked them into high gear. Even so, one more week and they had 5328.
.
I think your point is well made as regards New York City, where they may have had as many as 10,000 infected by the time of their first official case. The resulting seeding of cases all over the country made it difficult everywhere. But I don’t think that had to be the case in New Mexico.
.
At the other extreme, I think your point also may apply to Taiwan and Hong Kong. And Singapore, until it all went wrong.
Well NZ most certainly did not have an effective contact-tracing system when epidemic started. A local manual system used to dealing with measles or meningitis. It remains to be seen how well the one rapidly built during lockdown will be but so far it is working very effectively. South Korea did have a system and it swung into action immediately.
Contact tracing is slow, intensive work. If you have an explosion of cases, then it can be overwhelmed, especially if small and under-resourced in first place. SARS survivor countries throw massive resources at it. Tracing and survellience testing is effective when caseload is within system capacity. Lockdown is what you do till you have capacity that matches caseload.
My personal opinion, without a shred of data to support it yet bar the 1918 flu analysis, is that very hard lockdown (R <0.4 rather than R~0.8) minimizes economic damage because you get faster drop in caseload and shorter lockdown. I will read post-pandemic analyses with interest.
Oh, and my money is on Singapore wresting back control of spread within a couple of weeks.
Spread== Infected people were distributed in the state (and neighboring states.) Meaning there would be clusters and people who were not related by specific contact with each other. That’s the opposite of a situation where a single “super” spreader is found before it’s otherwise spread.
Not 100%. But it’s still a different situation from NM were people were able to travel in and out of NM to elsewhere with no barrier. Obviously, some did.
In contrast Patient 31 was likely the cause 61% of cases before march 18. That means other people weren’t. The virus didn’t spread before her which made tracing her contacts effective.
https://graphics.reuters.com/CHINA-HEALTH-SOUTHKOREA-CLUSTERS/0100B5G33SB/index.html
Contact tracing is much harder when there are many small unlinked clusters. One huge cluster is not “spread”. It’s one huge cluster.
MikeN
Where are you? Do you have a link to corroborate this drop (and to show it was broad and affected lots of restaurants? and exactly when this happened? Real questions.
For example: have no doubt fewer people went to restaurants after the St. Pat’s day parade was cancelled in Illinois. If nothing else: they didn’t go to the parade and so didn’t go out to eat after the parade.
But in any case: 10-20% is tough on the restaurants. But it’s really rather small. It’s hardly an indication of everyone staying home en-mass. It’s not the drop to 0% we had over the course of merely days when they stay at homes were put in place.
Given that estimated fraction you propose I’d suggest to avoid being misleading, you’d need to edit your statemen to read
Ok.. I read that “of 10-20%” not “to”.
My other questions stand.
Phil Scadden (Comment #185203)
May 22nd, 2020 at 4:15 pm
Italy never got R below 0.8 and they had a very hard lock down. You can’t directly compare this virus to the 1918 flu… and you can’t even directly compare one country to another with this virus, as population density changes R.
I have no links, just referring to what I saw in Virginia around Mar 12-20. They were announcing cancellation of some school events even prior to this date. In Illinois I know of actual sales data from a handful of restaurants in the Chicago suburbs, including one you could reach by going about 8 miles west on Ogden.
MikeN,
Is the Illinois data before Sunday March 15? Because after is definitely after states and municipalities began closing things. (St. Pat’s parade.)
I do btw, think sit down restaurants and bars are going to be one of the things that has a huge difficulty getting back up and running.
I looked at Google Mobility data for New Mexico and found the following.
.
Retail and Recreation clearly dropped starting on March 13. The transition was essentially complete by March 24.
.
Grocery and Pharmacy was rising strongly on March 11 and peaked on March 13. It dipped below the baseline starting on March 24, then stayed there.
.
Residential was rising strongly on March 13. It might have started a bit earlier or that might be noise. The transition was essentially complete by Wednesday, March 25. After the transition, there was a strong weekly cycle, with less of a change on weekends than on weekdays.
.
Workplace started to drop on March 12, but not really pronounced until Monday, March 16. The transition was essentially complete by Tuesday, March 24. After the transition, people going to work on the weekend was reduced much less than on weekdays. Makes sense, most people who go to work on the weekend can not work from home.
———–
Key government actions:
March 11 – Public Health Emergency, school closure announced.
March 16 – restaurants, bars, breweries, etc. to operate at no greater than 50% of maximum occupancy.
March 19 – no gatherings of more than 10 people; restaurants etc. limited to take-out and delivery; closed malls, theaters, health clubs, etc.
March 24 – stay-at-home order.
So public response clearly led government action.
MikeM
You indicate school closures were announced March 11. Governments announcing policies is government action.
But R&R drops March 13. That’s after public measures are starting. Grocery and pharma rose– that’s people’s response to getting ready to lock down.
I’m not seeing how you are diagnosing this as people responding after. It’s people responding nearly instantaneously as government actions are announced.
This isn’t government acting after people. It’s all people doing things just after government takes an action (which may be announcing a policy and a deadline.) The people finishing up as government deadlines approach. (Similar: lots of people finish up their taxes on April 15, but some before. Some filing on April 1 isn’t “leading” government action!)
Lucia, I don’t know if the drop in restaurant volume was in response to school closure. It seems likely, but I don’t remember visiting anything in the time frame just prior. There had been some other cancellations announced in the days before the school closure.
Illinois data was from before Mar 15. I remember it as being low for two weeks, but seems implausible. Weren’t restaurants shut down on Mar 16, announced over the weekend? When they announce cancellation of St Patrick’s Day parade?
MikeN,
I’m not debating “because of”. I’m debating “before”, “at the same time” and “after”. March 11 is before March 13, not after.
You mean the particular restaurant? Perhaps.
I’m not sure when they announced it. Obviously, before Sunday May 15. People knew the night before. But given my having no record of that, the cancellation could have been as late as Saturday May 14 or as early as March 8.
I went to a dance party Sat March 7 (it’s on facebook. ) That event was packed. Several friends were talking about organizing a group to go to the St. Pat’s parade the next weekend. Chenoa was especially eager to go. But I don’t know when in between it was cancelled. (Some of these things are hard to google and find!)
I’m pretty sure TONS of schools were online by March 16. How? Bitpaper.io posted they were having unprecedented loads and their servers couldn’t take it. Bitpaper’s strength is mostly their terrific whiteboard, so I think that load would mostly be school usages. (Not necessarily high school though). I use that a lot and it had been g_r_o_a_n_i_n_g under the load for “a while”. (No… I didn’t record when it first started to slow.)
Not sure what you mean by Mar 11 vs Mar 13. Schools were closed on Mar 12. Restaurant shutdown was at least a week later. I think it’s possible the slowdown in dining happened before Mar 12, but don’t know.
In Illinois I know of the dropoff not just for one restaurant but several locations of the same restaurant.
Sorry… Got M and N confused. Mike M wrote:
That’s where I got March 11 for the announcement — it’s for NM.
Yes. In Illinois, schools closed after the St. Pat’s day parade would have happened.
NBA shutdown happened a little before the closures, and was probably the cause of many of these shutdowns.
Before that NCAA had announced no crowds at games.
I would imagine NBA closing could have an effect on restaurants and bars, especially of the “sports” variety where people gather to watch games. I don’t know what fraction of restaurants that is.
Governor Pritzker of IL I judge is getting a pass from the public and the media on keeping the economy shut down and the results of that shut down in terms of daily deaths. That is not surprising in IL where I see from other blogs that the majority are very willing to unquestionably “follow the science” – which is a common term used in these discussions- and follow the political dictates and again without question. The retort for any questioning of policies is often greed over lives. The nursing home deaths from that side of the issue are simply ignored. Pritzker like many of his fellow Democrat governs tends to place the blame at the feet of the bad orange man wherever possible. If the ruler can do no wrong in the eyes of the majority of the media and the public, I do not see any motivation for any changes in policies whether they work or not.
Reality is that we still do not know whether we are in store for a second large surge with the lifting of the more severe restrictions and a surge that could come in the colder seasons and be enhanced by those related conditions. I suspect that policies makers do not even know how to mitigate nursing home deaths or deaths to the more vulnerable. If it takes herd immunity to eventually limit the cases and deaths it would be important to determine how that threshold is or at least could be affected by factors such as non homogeneity in individual or groups of individuals susceptibility to infection and propensity to infect others. We know there has to be difference; it is just a matter how much and eventually identifying who these individuals and groups are. A couple of studies cited by Nic Lewis in his recent Climate Etc thread indicated that 20% of the infected population accounts for 80% of the Covid cases with the existence of so-called super spreaders. If we myopically only look at government lockdowns and evidence for rationalizing it we will then be limited to a single strategy which will have long term consequences that too few are willing to the face head on.
Furthermore until we remove all the restrictions for getting the economy running again we will not know where we stand wih regards to the extent of a second or higher number of surges.
I am beginning to think that more than ever our governments, and the ruling intelligentsia, believe or want to believe that the Federal Reserve can print unlimited money for buying up government and now private debt and at the same time keep interest rates and inflation at low levels. In other words the negative effects of the shutdown on the economy will be covered by the Federal Reserve’s money printing and the Federal government spending. And for big government advocates there is the bonus of knowing that all this will make the public more dependent on the government.
I think NBA closings sent a message of seriousness that pushed other closings which were still being discussed.
People may have also been affected by Tom Hanks and Rita Wilson.
Do you mean the government actions? I don’t doubt that Governors, Mayors and so on were influenced by seeing groups like the NBA close.
Kenneth
We too can be Venezuala!
Kenneth,
Strictly speaking, the Fed doesn’t print money. That’s only done by the Treasury. The Fed creates credit which does expand the money supply, but it doesn’t actually print currency.
Kenneth, as someone who advocates a slower re-opening of the country than you, I’d like to respond to your latest comment.
First, I don’t think that you or others agreeing with you are greedy or choosing money over life on our behalf. I think you are advocating in good faith for what you think is best–we just disagree on what’s best.
Second, I’ve had enough experience following the climate change conversation to not blindly accept what the scientists are saying–and it’s clear that scientists like Fauci and Birx (while acting in good faith) were not infallible in their pronouncements. My personal stance, such as it is, is I hope informed by what has happened on the ground here and in other countries.
As far as the death demographics divide, the 20,000 dead under the age of 44 in two months is sufficient for me to call for a dramatic policy response. The 80,000 oldies like myself do too, but those who wish to write us off find their position somewhat disclosed by the massive death toll of so many who are young and could have been expected to live much longer lives absent Covid-19.
We are in no danger of becoming Venezuela. As we all should have learned during the Climate Crusades, there are some quantities that are better expressed as percentages of a total. Borrowing in our own currency with very low interest rates is not particularly threatening, as President Trump demonstrated with the de facto borrowing needed to finance his $1 trillion tax cut.
It will fall to the next Democratic president to do what Democrats have had to do so frequently in the recent past–claw our country’s financial situation back to stability after years of deterioration under Republican administrations. I fear the mood of those most progressive may be angry enough to make it tougher this time around.
Team Science isn’t doing so well, NYT:
https://www.nytimes.com/2020/05/25/us/politics/coronavirus-red-blue-states.html
.
“Counties won by President Trump in 2016 have reported just 27 percent of the virus infections and 21 percent of the deaths — even though 45 percent of Americans live in these communities, a New York Times analysis has found.â€
.
The reasons for this are very likely independent of ideology, the point being that the preaching, moralizing, and condescension coming from this crowd are NOT independent of ideology and quite tiresome when they haven’t cleaned up their own house. This breeds distrust of those who self elect themselves as defenders of science when they are in fact just partisans throwing out cheap shots. Science itself does not do enough to separate themselves from the fray when their organizations choose to take overtly public positions on cultural issues that science has little to say about. Partisans who hold jobs in science are … partisans. Science education is not an inoculation from ideology.
.
There are numerous contemptuous articles on churches wanting to reopen as the virus burns through blue communities. If only we could be more wholesome and virtuous like the good people in NYC.
“It will fall to the next Democratic president to do what Democrats have had to do so frequently in the recent past–claw our country’s financial situation back to stability after years of deterioration under Republican administrations.”
.
Laugh.out.loud.
.
I imagine they will do this with the Green New Deal and other magical thinking.
Tom Scharf (hi), “The data show that, since World War II, the economy has performed substantially better under Democratic presidents. On average, real (inflation-adjusted) GDP has grown about 1.6 times faster under Democrats than under Republicans.
…The findings for private-sector job growth are even more striking: businesses have added jobs at a nearly 2.5 times faster rate under Democrats than under Republicans, on average. In fact, the private-sector job growth gap between Democrats and Republicans is even greater than the gap when including government jobs.”
https://www.jec.senate.gov/public/index.cfm/democrats/2016/6/the-economy-under-democratic-vs-republican-presidents
Or, as Bill Clinton said in 2012, “It turns out that advancing equal opportunity and economic empowerment is both morally right and good economics, because discrimination, poverty, and ignorance restrict growth, while investments in education, infrastructure, and scientific and technological research increase it, creating more good jobs and new wealth for all of us.”
Republicans pass tax cuts financed by borrowing. Democrats reverse the tax cuts to bring the budget back in line.
We have more than 40 years of the same cycle reoccurring.
Thomas Fuller,
That’s mildly interesting. Does this mean that you think progressives support policies that are economically destructive or damaging?
.
Don’t worry about too much about the progressives. I worried about them too, until Sanders flamed out. Democrats will genuflect and pay lip service to the far left, but I don’t think the majority aren’t going to get behind them anytime soon.
Hi Mark
Yes. For example the Green New Deal as proposed would be quite damaging to the U.S. economy. However, the damages vary according to which version of the GND are looked at, especially with regard to healthcare.
Many of the line items in the GND… well, it’s not really a proposal, more of a wish list, are economically feasible and should in fact be implemented. But a guaranteed income without limits and a too-quick conversion to Medicare for All would be a bit spendy and we should really look twice at those items.
Thanks Thomas.
I actually spent quite a bit of time looking at the Green New Deal here: https://thegreennewwave.com/
DeWitt Payne (Comment #185311)
Yes, it is the treasury that actually prints the money and I recently read a piece where an economist was called out for saying otherwise. A more common phrase used for these actions of the Federal Reserve is creating money out of thin air.
The current economic situation created by government reaction to Covid-19 and the government’s and Federal Reserve’s reaction to their own spending and money creation in such a ho- hum manner is very much in line with acceptance of Modern Monetary Theory. I would be very surprised if most Democrats do not start using MMT to rationalize profligate spending and more than a few Republicans doing the same using perhaps a different terminology. Or doing it,at least, until the magic is no longer magic.
Thomas Fuller (Comment #185312): “the 20,000 dead under the age of 44”.
.
That number is wildly in error. In Pennsylvania, the number dead of age 44 and under is matched by the number dead of age 100 and older.
MikeM, you’re correct. I should have written ‘under the age of 54.’
https://www.cdc.gov/nchs/nvss/vsrr/covid_weekly/index.htm
Thomas,
Looking at that website, I see COVID-19 deaths for 54 and under around 5K of 70K total, as of 5/16. About 1700 were 44 and under.
Source:
https://data.cdc.gov/NCHS/Provisional-COVID-19-Death-Counts-by-Sex-Age-and-S/9bhg-hcku/data
Thomas Fuller,
Once again, sheltering in place and bending the curve is not a solution to COVID-19. It is a palliative to keep the health care system from being overwhelmed. In the absence of an effective vaccine, the total number of infections to reach herd immunity will be about the same. The area under the flattened curve will be about the same. It will just take longer. And shelter in place for the general population cannot be maintained for as long as you apparently would like. People won’t stand for it, as should be obvious even to you from the demonstrations and the newly crowded beaches.
Yes the daily case rate is likely to increase, but probably not at anywhere near the rate of increase in March. The Swedes recognize this. You apparently don’t.
I posted this a while back, but it seems it needs to be repeated.
I see partisan economic policies as a mosquito on the back of an elephant. The larger economic forces at play, both domestic and international, play a much larger role in actual outcomes than the policies of either administration. In the grand scheme the actual economic policies * executed * by either party when in office (not their rhetoric) are very similar. It’s US economic policy which is uber-capitalist when compared to Europe and many other places.
.
The timing of who is in office when the pandemic or 2008 meltdown occurred was largely irrelevant to the events occurring. Even perfectly wonderful socialist governments took it on the chin. Feel free to use hindsight and believe that your party’s pixie dust would have stopped all that from happening. I can also believe massive structural leftward change can turn the US in Venezuela, as well as isolating yourself completely from the world to be self sufficient works great if you want to be North Korea.
The US deficit doesn’t matter, at least according to whatever party is currently holding power in DC, ha ha.
The deaths by age group for the virus is curiously similar to the death rates of these age groups without the virus. Not sure what that means, but perhaps this is a common thing.
Tom Scharf,
See, you only say that because you’re an ignorant deplorable racist like me. If only we could see the light. Sigh…
/SARC
HaroldW (Comment #185330): “Looking at that website, I see COVID-19 deaths for 54 and under around 5K of 70K total, as of 5/16. About 1700 were 44 and under.”
.
Yes. The same here: https://en.wikipedia.org/wiki/COVID-19_pandemic_in_the_United_States#Demographics
Fuller’s 20% seems to be for 64 and under. 44 and under is a little over one tenth of those.
Tom Scharf,
Chance of death from all diseases pretty much always increases exponentially with age. Covid 19 is no different.
Hi, yeah–y’all are right, read the graph wrong.
From the team behind the IHME model:
“With mobility rising throughout the US over the last several weeks, our team had expected to see large increases in reported COVID-19 cases and deaths in more recent days. After all, the time lag between heightened mobility and potential rise in COVID-19 infections is approximately two weeks. Yet such a surge has yet to materialize, suggesting that increases in human mobility alone may not fully capture risk of transmission.”