1. Commenter John Alcorn watched the entire Swedish health minister video. John points out that the Swedish experiment definitely differs from ours in that they kept schools open. Some more of John’s take-aways.
School closures would de facto pull 20% of medical personnel away from hospitals because parents (including medical personnel) would have to stay home with children.
Sweden is halfway thru major wave of pandemic. Now seeing slowdown of contagion in Stockholm. One third of populace “has been involved” (exposed?). Summer probably will diminish contagion. But this virus, unlike SARS and MERS, won’t go away. Key will be to achieve original goal of shield or isolating the vulnerable (esp. those in elder care) much more effectively.
Skeptical of face masks (except in hospitals and nursing homes) because they tempt people who are symptomatic to go out with a mask instead of properly staying at home.
conditional on infection, death risk isn’t much greater than the flu if the individual receives timely care.
So they are approaching it as just a rapid-spreading flu that you want to keep out of nursing homes. Maybe Swedes are healthier then we are. Physically–less obese? Or mentally–less easily frightened?
And what does the health minister make of the disparity in incidence between their immigrant population and natives?
2. While trying to understand how New York’s delayed reporting of 4000 deaths is affecting things, I came across the NY health department page. I had to go to my browser settings and shrink the type to be able to see it all, but it has some interesting information. The co-morbidities that matter the most seem to me to be generally associated with obesity. I wonder if this puts me in a low-risk category, in spite of my age. It depends a lot on the Unknown Denominator, which is how many people are infected. The higher that number, and the lower the number of deaths of people my age with my BMI, the better off I am.
Poking around the site further, I found data on deaths among nursing home residents in NY state. Does anybody know how to prevent outbreaks in nursing homes? What do the Asian countries do about it?
3. German virologist Hendrick Streeck claims to have debunked the doorknob effect.
“There is no significant risk of catching the disease when you go shopping. Severe outbreaks of the infection were always a result of people being closer together over a longer period of time, for example the après- ski parties in Ischgl, Austria.” He could also not find any evidence of ‘living’ viruses on surfaces. “When we took samples from door handles, phones or toilets it has not been possible to cultivate the virus in the laboratory on the basis of these swabs….”
“To actually ‘get’ the virus it would be necessary that someone coughs into their hand, immediately touches a door knob and then straight after that another person grasps the handle and goes on to touches their face.” Streeck therefore believes that there is little chance of transmission through contact with so-called contaminated surfaces.
He bases his view on the result of a “case cluster study.” I gather that the idea is to try to determine how the people with the virus in a particular region contracted the virus. If there are no doorknob cases in a sample of one thousand people, then you are inclined to downplay doorknob effects.
4. What will be the long-term economic effects of the virus crisis? I am going to try to put my thoughts together this weekend. Meanwhile, Joel Kotkin writes,
Growing corporate concentration in the technology sector, both in the United States and Europe, will enhance the power of these companies to dominate commerce and information flows. As we stare at our screens, we are evermore subject to manipulation by a handful of “platforms” that increasingly control the means of communication. Zoom, whose daily traffic has boomed 535% over the past month, has been caught sharing data from its users with its clients widely, and without approval. Not surprisingly these platforms are most widely deployed in tech centers like the Bay Area, Seattle, and Salt Lake City as opposed to areas like Las Vegas , Tucson, or Miami where more jobs require close physical proximity.
The modern-day clerisy consisting of academics, media, scientists, nonprofit activists, and other members of the country’s credentialed bureaucracy also stand to benefit from the pandemic.
Off hand, I don’t agree with the second paragraph. I think that there is now an “essential/non-essential” divide, and a lot of the clerisy fall on the wrong side of it. But I am still pondering.
5. The story about the findings of lots of asymptomatic carriers at a Boston homeless shelter is being framed as scary news about the way the virus gets transmitted. Personally, I would have headlined the story “Homeless people show the way in developing herd immunity.”
6. Eran Bendavid and many co-authors write
These prevalence estimates represent a range between 48,000 and 81,000 people infected in Santa Clara County by early April, 50-85-fold more than the number of confirmed cases.
Wow. Pointer from Tyler Cowen.
If you believe that this is true, and that it holds for the country as a whole, the implications are staggering. As of today in the United States, we are at about 700,000 confirmed cases. Multiply that by a number between 50 and 85. That would make the infection fatality rate 1 in a thousand, as opposed to the expert estimate of between 10 and 20 in a thousand. It also would say that voluntary social distancing and government-imposed lockdowns came too late to stop the spread of the virus, for better or worse (probably for worse–if the death rate is so low, we should have just let it keep spreading). It would make it seem probable that the virus was in the U.S. much sooner than we now believe.
That is too much revisionism for me to adopt, based on just the one study. But I am encouraged by an apparent trend toward more studies and more pushback against relying on computer simulations.
[UPDATE: Balaji S. Srinivasan pours some cold water on the study.]
1) So is it now safe(r) to mention Sweden in polite company? And, If you have changed your mind somewhat, thanks for having the courage to say so on the record.
As I stated earlier today:
My conclusion: Sweden’s version of “lockdown” is probably better than ours and may put them in a much better position for the next steps. [but who knows for sure]
6. It befuddles me how California cases have been contained in general although it does appear specific members of the population do have some kind of immunity in general. The big question is did immunity already exist in their bodies (so 30% simply already something similar for strong antibodies) or were they recently infected?
“Personally, I would have headlined the story “Homeless people show the way in developing herd immunity.””
Consider the USS Roosevelt. All 4,800 crew have been tested. So far, over 650 have tested positive. 400 (60%) of those were asymptomatic, and one Sailor (Chief Petty Officer, 41, which is “old” in the military) has died.
The disease spreads among slim young people like a prairie fire, but most of them don’t even know it, and those who do, just barely notice. For them, this isn’t like the flu, because the flu is much worse.
Older and chubbier people who live and work among them in enclosed areas, however, get burned by that fire. It’s not the flu, it’s knocking on death’s door, and sometimes he lets you in.
As far as variolation proposals go, dose doesn’t seem to matter much for young people. Even being cramped up with spreaders in tight spaces seems safe enough, as it doesn’t seem to matter in terms of developing bad symptoms.
“Expose all the young” (and hide the old away until the young are all immune) is looking better as a strategy.
“Expose the young; hide the old” is a thought provoking strategy. It sounds to me at least, that herd immunity equals eradication.
Suppose the demarcation is age 40 and almost nobody dies but the economy’s production is low. You set the line at 55, more people die, but production is higher. If you set the line at 60, many more people die, but production is almost normal. None of these sound like tempting solutions even though they all lead to eradication. The process for deciding the demarcation age will be a source of strife.
Suppose the demarcation is geographic. Suppose the planet has 2 countries, Sweden and New Zealand. Suppose Sweden achieves herd immunity by working as normal and hiding the at-risk people. Suppose New Zealand simply caught every case at the airport and that person was isolated and recovered or isolated and died. Some Swedes die but eradication happens. Then there can be free exchange between the two countries. This also works. Each country chose its own course according to its own circumstances and there is less strife resulting from geographic demarcation because we already accept that borders exist. I am not sure what the most relevant circumstance is. Having somewhat leaky borders might be one. An increasing number of infections within the country being another. When people say “the virus is already here” what they are making is an economic argument that if a large percentage of new infections is locally caused and a small percentage of new infections is caused by recent airport arrivals then we need to consider the value of air travel lost to restrictions. They never seem to state that explicitly but since the percentages are hard to know people are not explicit.
Individual U.S. states cannot control their borders the way a country can. The geographic strategy applies but not perfectly. Since it is imperfect there might be some anger. Infections have already increased within individual states so that is another circumstance.
Does lockdown result in eradication? It seems if everybody is isolated they either recover or die. This means perfect lockdowns mean eventual eradication. I would picture this as almost zero economic production for 4 weeks and ideally total eradication. If you have a leaky border you have a lockdown of lower effectiveness so that makes the strategy of seeking herd immunity relatively more attractive. If you achieve herd immunity within a particular U.S. state, you still have to make sure visitors to the state do not visit a long-term care home.
Do leaders need to call for a lockdown? If people are scared they will not visit restaurants anyway so the lockdown does not make any difference in some cases. When leadership call for a lockdown they’re just putting their thumb on the scale and it’s like saying let’s have a little more lockdown and a little less production.
I’m inclined to agree with another internet commenter that the U.S. is going for herd immunity. Just keep all but the most vetted persons away from long-term care homes and other at-risk people. It may be that the New Zealand strategy works because they decided to draw their circle around the collection of their islands instead of drawing their circle around the elderly. Viewed this way it is not really a different strategy; just a bigger circle. Viewed this way the relevant decision is what is the optimum circle for my circumstances.
Will May 15 become a Schelling Point?
Seven U.S. states extend coronavirus shutdown to May 15
It has the benefit of being the beginning of summer. No need to restart schools. People itching to get outside. Outside generally being less dangerous than inside.
By coincidence, Memorial Day weekend comes really early this year, May 23-25. That’s the traditional “start of summer”, for instance, it’s a common time to open public pools and amusement parks. People go on a lot of trips. There are also lots of public events: parades, fireworks, barbecues. Tourism kicks off, wedding frequency takes off.
May 15th buys some time to prep, or maybe a little slack to delay one more week.
In Massachusetts, schools normally go until June 20 or so (depending on when Labor Day fell and how many days the schools were closed for snow). State officials are floating the idea, “We’d really like students to come back for a while before summer.” One complication is that teacher contracts almost always run from September 1-June 30. And working outside the contract period simply isn’t done.
There are now dozens of preprints pointing to potentially massive undercounting of actual number of cases, in line with the Santa Clara study. Someone is starting to put them all in one place here: https://docs.google.com/spreadsheets/d/1Kz9cSk_49-lZsL1AGlP8yV5yIZlBnTbMZ307lxepuXk/edit#gid=0
Here’s a twitter thread summarizing the evidence as of a week ago (there have been a bunch of additional papers since then): https://threadreaderapp.com/thread/1248655370257608704.html
I think it is easy to believe that the true number of infections is 5x the number of reported cases. But 50x puts us in a different realm altogether.
Danish CDC says 30-80X undercounting
Scotland’s number was up to 130X
I expect to see more big numbers as more of the antibody studies are published.
It boggles my mind that NYC doesn’t appear to be doing a study like the Santa Clara County one. They have ~120k positive tests so far and a very high positive rate for new tests still. If that underestimated the true had-been-infected rate by 50x, NYC would already be at herd immunity (6M out of 8M infected) and could safely reopen. Now I doubt very much that that’s the case, but it is just insane to be making decisions about extending or easing restrictions without even knowing how far off they are from that, when it’s this straightforward to get the data.
#6, The Santa Clara County study, done April 4. Very clever and expeditious. They recruited a sample stratified by age, sex, zip code, etc, via facebook ads, and did 3,300 drive-thru serology tests in one day in 3 locations. The population-weighted estimate of the number infected or previously infected as of April 4 was 45,000 to 75,000, out of a county population of 1.8 million (big counties in California). Total deaths in Santa Clara County so far as of April 16 are 69. Pretty much all of the people who died would have been infected by April 4. So if we take the study’s high estimate of infected, 75,000, the case fatality rate is .0085 percent, or about 8.5 per 10,000 cases, about the same as the flu, or the low estimate, 45,000, it is .144 percent, or about 14.4 deaths per 10,000, about 50% more lethal than the flu.
Make the estimates of infected or previously infected 48,000 to 81,000. Other figures are right.
“German virologist Hendrick Streeck claims to have debunked the doorknob effect.”
Does this imply rigorous handwashing is not as important as we’ve been led to believe?
Regarding the Santa Clara study:
“Participants were recruited using Facebook ads targeting a representative sample of the county by demographic and geographic characteristics.”
OK, so the targeting was probably well sampled, but there’s still self-selection bias for who actually got tested, right? Who is more likely to want a test:
a) Someone who got a flu shot in November but experienced flu/coronavirus symptoms and couldn’t get a test while sick
b) Someone who never had symptoms
The answer is a by a mile, right? I’m sticking with the true infection rate is probably ~10X the confirmed cases.
With 30% daily growth, it would take 66 days to go from 1 case in the US to 35m (50x 700k).
If 30% growth ended on say March 10 (and there were literally no new cases since then), then the first US case would only have to come in around January 4th (assuming a single initial seed, which is highly unlikely for a country like the US). The first official case was January 20th but if we’re undercounting by a factor of 50 today (when testing has dramatically improved), then it’s exceedingly unlikely we got the first case. (And of course there must’ve been at least a few doublings between March 10 and today).
Overall it seems like accepting 35 million cases is not that hard to do.
This is a good site dedicated to synthesizing research data for clinicians.
https://www.uptodate.com/contents/coronavirus-disease-2019-covid-19-epidemiology-virology-clinical-features-diagnosis-and-prevention#H3174740477
Direct prolonged contact with symptomatic individuals is the transmission vector. There are a few cases of clusters from family, work and social group gatherings where a symptomatic individual attended, such as two Illinois clusters at family gatherings where there were communal food, embraces and prolonged face-to-face conversations with individuals who later tested positive.
“The risk of transmission with more indirect contact (eg, passing someone with infection on the street, handling items that were previously handled by someone with infection) is not well established and is likely low.
“Environmental contamination — Virus present on contaminated surfaces may be another source of infection if susceptible individuals touch these surfaces and then transfer infectious virus to mucous membranes in the mouth, eyes, or nose. The frequency and relative importance of this type of transmission remain unclear. It may be more likely to be a potential source of infection in settings where there is heavy viral contamination (eg, in an infected individual’s household or in health care settings).”
Washing hands is good practice in these times, but it along with social distancing has been used more as ritual, i.e., busy work for individuals than science. Risk is low for surface transmission especially when anyone snotting that much would be removed from a public space such as a store now.
As for social distancing, it is based on the supposition that when people sneeze, cough or talk (and they should add sing) they are spitting into the other person’s mouth and nose without the recipient noticing and avoiding. Well, the “scientific” phrase is “droplets” but that is just spittle plain and simple. Oddly, the “experts” deny up and down that virus may desiccate to nuclei in low humidity and remain in enclosed spaces for (2 hours in lab conditions) to be breathed in over prolonged exposure. This is the area where masks were originally recommended against (along with the delay needed to move the commercial PPE masks to the medical system)
The real question is with all these experts why have none applied their expertise to refine the transmission avoidance protocols beyond the “childish rules they rushed out that offer no nuance”.
We could wear cloth face coverings to stop our spittle and catch that of others, and frequently cleanse our hands and likely achieve as good or better transmission prevention than the current “rules”.
Our experts are probably good in the lab and even in policy, but it quite apparent they are not good “operationally” in how to remain an operating society (fighting force) while taking precautions.
Robin Hanson notes that the biomedical experts he’s met are detail-oriented but relatively poor abstract reasoners. I would add “cripplingly overspecialized”.
Ok, sports fans, I just had a brief correspondence with Eran Bendavid, who led the Santa Clara County study. He believes it can be replicated pretty much anywhere, on a similar scale (3,330 subjects) for $250,000.
What are the barriers to doing a bunch more?
#6. Wow, these results seem incredible! I too was struck by the implications of this study. By the way, I think you misplaced a decimal point in your analysis. The study suggests that the true infection rate for Santa Clara County is about 2%. That is 2 out of 100, not 1 out of every 1,000. And it looks like implied true infection rate is actually higher than that. That is because the study also calculates under-reporting of about 50x – 85x, and there are about 700,000 diagnosed U.S. cases. Multiplying the lower bound of 50x by 700,000 equals 35 million actual cases, or about 10% of the population. But incidence of U.S. testing has improved quite a lot since early April, so call the real number about 5%. Either way you slice it, that is much higher than I would have guessed before reading this article. Perhaps we really are headed towards some sort of heard immunity in future months if actual incidence gets to around 20%.
I admit that much of my comment is speculative, and relies upon some iffy assumptions.
I was not estimating the true infection rate. I was taking the 35 million (hypothetical) cases of infection and comparing it to about 35,000 deaths to get about 1 death per thousand infections
Note the critique of the study at https://medium.com/@balajis/peer-review-of-covid-19-antibody-seroprevalence-in-santa-clara-county-california-1f6382258c25
Pointer from Tyler Cowen.
Points 2A and 2B are related to what I wrote above.
You have a German, Dutch and Danish serological studies, using different methods but reaching essentially the same conclusions as the Stanford study. That makes all attempts to discredit the notion of much lower lethality by discrediting this one much harder.
Danish
https://www.dr.dk/nyheder/indland/doedelighed-skal-formentlig-taelles-i-promiller-danske-blodproever-kaster-nyt-lys
(IFR 0.16% – almost identical to Stanford study)
Dutch
https://www.reuters.com/article/us-health-coronavirus-netherlands-study/dutch-study-suggests-3-of-population-may-have-coronavirus-antibodies-idUSKCN21Y102
(IFR= 0.4%)
German
https://www.technologyreview.com/2020/04/09/999015/blood-tests-show-15-of-people-are-now-immune-to-covid-19-in-one-town-in-germany/
(IFR=0.37%)
“The modern-day clerisy consisting of academics, media, scientists, nonprofit activists, and other members of the country’s credentialed bureaucracy also stand to benefit from the pandemic.”
This is half right. CDC and FDA will benefit. I expect many of the aspects of their responses which exacerbated the crisis will perversely harden. I think there will be an impetus to plow more money into public health which will benefit non-profits and govt alike. Universities will be interesting. The idea of charging someone upwards of $60k a year to learn nothing of use (a benign interpretation) seems unsustainable. But before the coronavirus it appeared to have life left. I wonder if we haven’t reached a tipping point, perhaps a decade earlier than if Coronavirus hadn’t occurred. People are essentially undergoing a forced boot camp on getting things done remotely, this includes learning. Combine this with the fact that a lot of private colleges don’t have big endowments and those endowments are worth much less today than they were two months ago. I could be wrong but I think we are about to witness a big shift in higher ed.
Johan Giesecke, Former State Epidemiologist, Sweden (and currently a consultant on Sweden’s strategy, has given a remarkable, candid interview (30 minutes): https://www.youtube.com/watch?v=bfN2JWifLCY
Here are select excerpts:
“The Swedish Government decided early, in January, that the measures we should take, against the pandemic, should be evidence-based. When you start looking around at the measures being taken now by other countries, you find that of them very few of them have the shred of evidence-based.”
“Universities are closed. […] Nursing homes are closed to visitors.”
“[Swedish policy] is very similar to the one the UK had before there was a famous paper by the Imperial College, by the modelers who made models for infectious diseases, that came out on the day after you made a U-turn in England.”
“The [Swedish] strategy is to protect the old and the frail, to try to minimize their risk of getting infected, and to take care of them if they get infected.”
“We were very pleased we were having the same policy as the UK. That gave some credibility to what we were doing. But then Mr. Johnson made his 180-degree turn.”
“I think [the Imperial College paper] is not very good. […] Models for infectious disease spread are very popular, […] they’re good for teaching, they seldom tell you the truth. […] It rests on the assumptions. And the assumptions in that article have been heavily criticized. […] It’s not peer-reviewed […] It’s just an internal departmental report from Imperial. It’s fascinating—I don’t think any scientific endeavor has made such an impression on the world as that rather debatable paper. […] One thing that the model has missed is that it assumes that hospital capacity will remain the same—and that’s not happening anywhere. In Sweden, we’ve tripled our intensive care capacity.”
[Q: What has flattened the curve, if not the lockdown?]: “One thing is immunity. The other is that the people who are frail and old will die first; and when that group of people is sort of thinned out, you will get less deaths as well. The other thing is that when you start your ‘exit strategy,’ […] you’ll have some other deaths that we had already. […] [Cognizance of an] increasing number of deaths will be part of which strategy should be kept and not. […] When I first heard […] about the different draconic measures that were taken, I asked myself, How are they gonna climb down from that one? When they will open the schools again, What should be the criterion to open the schools? Did anyone of the strong and very decisive politicians in Europe think about, How do we get out of this?, when they introduced it? ”
“The nursing homes in Norway are quite small, whereas in Sweden the nursing homes are quite big, with hundreds of people. Which means that if you get the virus into one nursing home in Norway, it will affect far fewer people.”
“We should have this discussion a year from now. […] I think that the difference between countries will be quite small in the end. […] I think what we’re seeing is a tsunami of a usually quite mild disease which is sweeping over Europe, and some countries do this, and some countries do that, and some countries don’t do that, and in the end there will be very little difference. […] Most people who get it will never even notice that they were infected. […] I think it will be like a severe influenza season […] which would be an order of 0.1 percent [IFR].”
[Q: Once we get mass antibody testing in place, how many people will we find have had the infection?]: “At least half.”
“What I’m saying is that people who will die a few months later are dying now, and that’s taking months from their lives. […] Comparing that to the effects of the lockdown, what am I most afraid of? It’s dictatorial trends in Eastern Europe. […] It may pop up in more established democracies as well. I think the ramifications can be huge from this.”
[Q: What about fatality rate in nursing homes?]: “Sweden failed, we were not on our toes enough to really shield the old people. We should have banned visitors earlier. Many of the people working in nursing homes are from other countries, they’re refugees or asylum seekers in Sweden, their Swedish may not be perfect, they may not always understand the information that has been shared to the population. There are many things that we could have done better a couple of months ago.”
“I don’t think you can stop it, it’s spreading. You can stop it for some time. […] Taiwan, I don’t know about Taiwan, they were quite successful. […] We’re saying: Protect the old, try to slow the spread of the epidemic a bit, so that the health care system will manage when we have […] many severely ill people. Tsunami is not a bad [analogy], it will roll over Europe, no matter what you do.”
“But how long in a democracy do you think people keep a lockdown? […] You can do it in China.”
[Q: Why has this pandemic led countries to adopt lockdowns?]: “New disease, a lot of people dying, we don’t know really what will happen, and this fear of contagion, I think, is almost genetic in people—and showing political strength, decisiveness, force, very important for politicians.”
“Do you think you could keep the lockdown, to protect the old people until we have good drugs and good vaccines, six months, a year, 18 months?”
“I think it will be like a severe influenza. […] lf the influenza came around as a new disease, you never had it before, but suddenly this new disease called ‘influenza’ popped up, we would have exactly the same reaction as we have now.”
“A lot of influenza deaths are not recorded in that way in a normal influenza season. […] This may be double as much, but not ten times as much.”
[Q: Should the UK cancel lockdown?] “No, you can’t do that, then you would have a wave of cases. Then you really have a peak—one week, two weeks later. […] No, you really have to climb down, one rung of the ladder at a time, and probably start [by lifting] school closures, maybe.”
“In Sweden we have one million children between [ages] zero and ten. They need to be looked after. If they’re not in school, then someone has to stay at home with them. One of my friends is a nurse and head of an emergency ward here in Stockholm. She prays every morning when she wakes up that the government will not close junior school, because then she loses her staff.”
[Q: Can we suppress the disease and wait for a vaccine?] “No, it will take too long, it won’t work in a democracy.”
[Q: Will the Sweden persist in its strategy?]: “I don’t think it will be tougher.”
Thanks for your hard work in putting the transcript together and for the link to the video! Fascinating!