A Solution if We Act

Many simulations have been run in recent weeks using standard epidemiological models and the emerging consensus, as I read it, is that test, trace and isolate can be very effective. Paul Romer’s simulations are here and he notes that a COVID-19 test does not have to be especially accurate for the test, trace and isolate strategy to work. Indeed, you don’t even need to trace, if you test enough people. Linnarsson and Taipale agree writing:

We propose an additional intervention that would contribute to the control of the COVID-19 pandemic and facilitate reopening of society, based on: (1) testing every individual (2) repeatedly, and (3) self-quarantine of infected individuals. By identification and isolation of the majority of infectious individuals, including the estimated 86% who are asymptomatic or undocumented, the reproduction number R0 of SARS-CoV-2 would be reduced well below 1.0, and the epidemic would collapse….Unlike sampling-based tests, population-scale testing does not need to be very accurate: false negative rates up to 15% could be tolerated if 80% comply with testing, and false positives can be almost arbitrarily high when a high fraction of the population is already effectively quarantined.

Similarly, Berger, Herkenhoff and Mongey conclude:

Testing at a higher rate in conjunction with targeted quarantine policies can (i) dampen the economic impact of the coronavirus and (ii) reduce peak symptomatic infections—relevant for hospital capacity constraints.

This is exactly the strategy I discussed in, Mass Testing to Fix the Labor Market, where I wrote “Testing, isolating and tracing will [get the economy back on track] much faster and cheaper than dealing with a prolonged recession.”

I want to expand on the costs because it’s clear that a mass testing regime will require millions of tests. Is that cost-effective? Yes. The two types of tests we have are a RT-PCR test for COVID-19 (there are several versions) which costs something like $100 but could probably be much less as we ramp up. (We can cut costs and greatly increase throughput, for example, by pooled testing.) The second test, a blood test for antibodies, is, as best as I can tell, in the realm of $10. Both types are useful. I am going to be very conservative and say that we use a combination of tests at $75 per test. To test the entire US population, therefore, it would cost on the order of $25 billion dollars. Coincidentally, $25 billion is about what we spent on the Manhattan Project in current dollars. Thus, I am proposing a Manhattan Project for testing.

Twenty five billion dollars to test the entire US population. Now suppose the pandemic knocks 5% off US GDP over the next year or two, that’s roughly a trillion dollars lost. Or to put it differently, $3 billion a day. Thus, if mass testing reduces the number of days we are away from work by 9, it pays for itself. Let’s again be conservative and say that testing will also require a $25 billion fixed cost to build the enzyme factories and so forth, for a total cost of $50 billion. 18 days and it’s worth it.

We would also save medical costs by suppressing the virus. (The focus on ventilators has perhaps been overdone given that ventilators in no way guarantee survival–better to stop people needing ventilators.) We would also save lives. Thus, a program of mass testing seems like a no-brainer. Yet, there is no direct funding for anything like this in the $2.2 trillion CARES bill which is stunning. Here’s Austan Goolsbee:

We literally put in a tax break for retailers and restaurants to expand their capacity but not money for production of more COVID tests.

Here’s Paul Romer:

We have an economic crisis because it is not safe for people to work or consume. Our Congress just passed a bill that will spend $2.2 trillion to deal with the crisis. Can anyone identify any spending in this bill devoted to making it safe for people to work and consume?

As I wrote:

We need to attack the virus with test, isolate, and trace. More money for counter-attack!

Objections will no doubt be raised. Isn’t there a shortage of reagents? Do we have the personnel to test everyone? To which I answer, $50 billion solves a lot of problems. We won’t know how many till we try. We don’t need all of final testing capacity at once and even poor tests like simple temperature checks will help but we need to move rapidly in the right direction. The main constraint is time. Social distancing and lock downs are starting to have an effect. I expect the emergency will peak in mid-April and then things will slowly start to get improve. Even when the worst of the emergency passes, however, we will still need lots of testing. This virus will be with us and the world for some time. Let’s get on it.


Does not actually need the test part, but it requires unbending discipline regarding isolation. Particularly on the part of those traced, who need to feel that being isolated for 14 days without symptoms is the only correct response in stopping a pandemic.

The idea that adequate testing replaces tracing is kind of silly when even in Italy or Spain, not more than 2% of the population has been infected with covid19.

Hi, my wife and I have had basically zero contact with other humans for sixteen days now. How is that not as good as a negative test?

Actually, it is better than a negative test, assuming your basically zero contact means zero contact. Viruses don't care about what people think about zero contact, only that it is zero contact.

Fine. The fact that I haven’t developed symptoms suggests I didn’t have the virus BEFORE isolation, right? I was at an NHL game with 20,000 people on March 11th, 19 days ago, right before the shutdown. Can I conclude now that I didn’t pick up the virus then?

Since then, my wife and I have come within six feet of other humans zero times, we are wary of door handles and elevator buttons, I don’t touch my face on the four occasions when I left our apartment and I wash my hands vigorously upon return.

On March 11, in a sea of humanity, I did none of this.

"The fact that I haven’t developed symptoms suggests I didn’t have the virus BEFORE isolation, right?" Wrong. It could be that you're someone who didn't develop noticeable symptoms, even though you might be infected.

OK. It's been 19 days. Am I still infected? Am I still contagious? There is a tiny tiny tiny chance I picked up the infection in the past 19 days, but I'm pretty sure a 0% chance I have passed it on.

nuke the house from orbit, it's the only way to be sure

That's what the Andromeda Strain wants you to think.

14 days isn't a hard limit.

There is a 1% chance that an individual will incubate for 20 days.

That said, it is unlikely that both you and your wife are 20 day incubators. If one of you is incubating, then most likely by this point you have infected your spouse and they should be having symptoms.

Well for starters suppose your wife had an asymptomatic Covid illness and was infected two days prior. She has an incubation period of 6 days. So on day 4 she becomes infectious. But you get lucky and manage not to get infected until day 8. You also have a six day incubation period (day 14) and will also be infectious for several days after.

The more people in your isolation bubble, the longer you can sustain an infection.

And lest we forget, I am not even highballing the numbers. We have case reports of symptom onset over 10 days out. Likewise, surface contamination can be infectious for several days so we could actually start the clock several days after you went into isolation.

Then, of course, there is the question of your immune function. In most epidemics there are people who are better and worse and developing immunity to the pathogen. This largely comes down to MHC genes (these are the most polymorphic of all human genes) and some people are lucky and theirs very efficiently displays viral epitopes that are highly immunogenic while other folks are unlucky and their MHCs have trouble displaying immunogenic epitopes. Long story short, if you are unlucky you may never clear the virus. Certainly this can be the case for people with immunosuppression.

Yeah, your odds of being infected are low, particularly if you were unlikely to have exposure prior to isolating, but you could actually be contagious at this moment without anything too crazy going on.

Ok thanks. I live in Chicago. To me, it seems like, before March 12, we were all swimming in a million person venue and now there are circuit breakers all over the place. I don’t think the virus had a high penetration rate as March 11 here, so I don’t see how what we are doing can’t work and smoke out infections.

You trolling BD, somebody hijack your account? Had you for a smart fellow. Can't figure out your point here.

Bonus trivia: Greece's second week of total lockdown, soon to become four weeks, seems to be working to flatten the curve. US can afford to do the same.

Not sure if you heard Ray, but worldwide deaths fell 400 yesterday. Italy has peaked, Spain may be peaking earlier than expected.

What I'm saying is: maybe the first part is the easy part. Right now, all we can do is minimize the first wave.

The short-term trajectory of the West is looking known and encouraging, so keep up with the isolation a while longer, but it's time to start thinking longer term.

China just re-closed movie theatres.

yes and treatments are coming... finally reason for some limited optimism

theater restrictions seem to be local decisions globally


In response to your question I think we could say your 16 day isolation is better than a negative test. But a negative test or your isolation is not a 100% guarantee you don't have the virus.

Of course you are assuming you've done a pure 16 day isolation. No contact? Didn't get the mail? Maybe said hi to a neighbor 15 days ago and forgot? It's possible to have a perfect isolation just like it's possible to stick to a strict diet for 16 days, many people will have slipups.

However, if we do a good, if not perfect, job, we take an infection with R>1 and make it R<1 which drives it to extinction (or at least into remission).

Exactly. It seems to me that distancing is equivalent to reducing the size of the venue, the importance of which MR pointed out.

On March 11, I engaged in dozens of interactions with door knobs, elevator buttons, and other people. I was probably touching my face non-stop and not washing my hands much. I even went to a professional hockey game, on the eve of the shutdown. I was doing everything possible to catch the virus. 19 days later, it looks like I didn't.

So yeah, I suppose it's possible I picked it up at Jewel on one of my two trips there in the past 19 days while applying the full measure of protections, in the sense that anything is possible.

Wow. Stick to finance and leave science to others BD.

Come on Ray. I'm a failed scientist, in the sense that you are a failed lawyer.

You usually have interesting things to say beyond "wow.just wow." Let's hear it.

"Hi, my wife and I have had basically zero contact with other humans for sixteen days now. How is that not as good as a negative test?"

because its entirely possible to be isolated from other humans but not be isolated from what other humans have coughed,rubbed or dripped on

I understand this, and I have developed a new-found appreciation for the menace presented by door knobs, revolving doors, elevator buttons and proximate other human beings, and I express this appreciation in my behavior.

I used to laugh at germaphobes. Not anymore.

you and we both!


I am been doing the same. It is about as good as us being tested. What we need is random sampling testing of the population so we can can identify the hotspots before people start dying. It’s one of, if not the most important of the many millions of ways. we can get back to normal as quickly as possible.

Alex, you have been an excellent commentator during those last few weeks or months... thank you.

"when even in Italy or Spain, not more than 2% of the population has been infected with covid19"

There are lots of people arguing the actual number is well above 2%, even without going so far as the "everybody has it" idea. See some of the comments at the linked post by Jason Y, M, and MDL, for example - https://marginalrevolution.com/marginalrevolution/2020/03/where-does-all-the-heterogeneity-come-from.html#comments

They are all pretty much unable to show their math, compared to commenters who suggest following either the current (officialish) Italian ratio of 1 confirmed to 10 unknown cases, which leads to less than 2%, or using a 1 to 100 ratio (without any supporting proof in the slightest), in which case Italy has less than 20% infected.

Math is not really hard, especially when there is actual data available, such as population and number of confirmed cases.

We will 'know' how many have died from Wuhan Flu. We 'know' the US population.

Latest numbers I saw 2,489 deaths; I use 330,000,000. That works out (2489 / 330,000,000 * 1,000,000) to 7.54 PER 1,000,000; OR 1 / 100,000.

Dr. Fauci is now giving worst case scenario of 200,000 deaths. That's horrible: 999,394 per 1,000,000, or 99,939 per 100,000 American Wuhan Flu survivors.

Its not a flu virus.

Forget it, he's rolling.

Two months ago, you thought using wrong pronoun was a national emergency.

Someone quoting from the 1978 movie Animal House is the sort of person who cares even less about pronouns than you seem to.


The problem with a testing regime is it would have to done only after a strict fourteen regime at least metropolitan wide with zero interaction across borders outside moonsuits.

The thing people aren't talking about is the window between contagious and detection, they aren't equal even even if you were tested 24x7. There is a real issue in most (all?) viruses where you are fully contagious but no test can pick it up yet and you will test clean. It's why for example post-HIV exposure you have to test 30/60/90 days to be clean.

I have no idea what that window is for coronaovirus but it's not zero and it will break any testing regime if that is your measuring stick.

Hi Peter, can you elaborate on why that is? (I am just curious. Is it because the tests test for antibodies and not the actual virus? If so, then the test will also be positive even well after the disease is gone because I am still producing antibodies, right?)

Serological individual test can be currently bought in China at 5$ per piece, buying tend of thousands. I am fairly sure that a contract for tens or even hundreds of millions can bring the price down significantly, possibly even around 1-2 $.

False positives are around 12%, and false negatives even less.

Assuming that people cannot contract the virus again, positive but healthy people would be an increasing pool of ready-to-work individuals. A database of those individuals can be maintained by a reputable firm doing it for profit (let’s say that a certification with a fake-proof ID card for 100$). As a matter of fact, you guys at Mercatus should could up with CVS or WalGreen and manage the program.

The only problem is, of course, that for some reason the FDA has not approved these tests, although they are already widely available in pharmacy in many countries.

"False positives are around 12%, and false negatives even less."

I don't know that I believe that, given the evidence from rigorous testing of other Chinese tests showing up in other countries, but the point is that effective quarantining can have large benefits even with relatively poor tests.

Indeed, I think that ties together the many reports about other countries seeing poor quality Chinese tests together with Chinese policy to lower the rate of spread. Bad tests, if done quickly and at a high rate, are more effective against a fast spreading pandemic than perfect or very good tests, produced too slowly and not done enough.

The FDA and CDC made the perfect the enemy of the good.

"Bad tests, if done quickly and at a high rate, are more effective against a fast spreading pandemic than perfect or very good tests, produced too slowly and not done enough."
that sorta depends on how bad the tests are and whether they are bad in a false positive or false negative way. and it also depends on how well
the other proven public health measures are implemented alongside
the testing

"positive but healthy people would be an increasing pool of ready-to-work individuals"
If you test positive, then you cannot possibly be ready-to-work, right? Or did I misunderstand you?

Scaling PCR testing is not as easy as scaling Economics Papers.

Not if we pool. With Abbot and Cepheid coming on board, we can do over 300K tests per day . Using 50 samples in per pool that's 15MM tests per day.

a PCR takes about 5 hours
now include collection and transportation time and staff
and the administration you need to create to organize and keep track of 300k tests

all those models are solving for a spherical horse

We should be yelling this from the rooftops. I've been wondering why we wouldn't just test the entire population 2 or 3 times say on May 1st, May 20th and June 30th, and have a strict quarantine for positives. The cost would pale in comparison to the cost of protracted shutdown + lives lost, as Alex points out, and the logistics are hard but doable—we do a census of the entire population, right?

In an ideal world, the US should also work with international organizations and other governments to expand this strategy to as many countries as possible, but I have 0 hope that that could happen.

I like testing and quarantining. It's worked well for some East Asian countries as a backbone of their strategy.

However, the famously independent American would balk at forced testing or forced quarantine. Implementing the strategy may not be feasible.

+1 Americans are so ignorant. They were against traffic light cameras for dubious grounds, the only logical reason was people brake too hard when they see them, which is easily fixable with more cameras.

So additional cameras to fix the problem the first set of cameras have caused?

There are many reasons Americans dislike those traffic cameras. State intrusion/overreach; it might be they can be gamed; they interfere with our god-given right to bend the traffic rules.

As it happens, as a conservative I’m more upset at cheating (speeding) and indifference to the law. So I’m pro cameras ... but I can grasp the reluctance of others to share my view.

Yeah, more tech will do it. In Greece they use a radar gun to record speed over a series of intervals as well as a camera. If somebody is speeding, slams on the brakes, and gets rear ended, such data is useful in states that don't apply the mechanical rule that the driver who rear ends somebody is always at fault (i.e., it helps portion blame in contributory negligence states). As for privacy and pubic roads, they taught us in law school, which I flunked, that there's no such privacy (nor at borders, that why border guards can be such jerks).

Instances of cities pairing red light cameras with shorter yellow light times - often yellow times that fall below recommended minimums from traffice engineers and/or state DoT's - cause people to question (rightfully) whether the motivation for these cameras is traffic safety or nothing but revenue generation. See https://time.com/3505994/red-light-camera-problems-tickets/ as just one example.

'To which I answer, $50 billion solves a lot of problems.' Make it a trillion dollars, and it still won't buy you any extra time to start preparing in mid-February,

First, I agree.

Test for those with the virus or those who had it.
Those folks with be those who can go about.

Then, there are those who were healthy, stayed inside,
Didn't get or have the virus
They will be discriminated against
Will claim that the Americans With Disabilities Act
Applies to them
Because they will not be admitted or served at
A Restaurant.

There are far too many cases to test and trace. At least 10 million already, probably 100 million plus. In most people infection is asymptomatic or very mild so they never get tested or seek treatment.

Mandatory testing catches the asymptomatic.
I would imagine that colleges will require testing before students get back on campus.

Note that it may take 2 to 3 weeks before antibodies are detectable: "The first detectible serology marker is total antibody and followed by IgM and IgG, with a median seroconversion time of 15, 18 and 20 day post exposure (d.p.e) or 9, 10 and 12 days post onset, separately."


Yes, which means that everyone needs a couple of tests, spaced out several weeks apart, before they could be declared uninfected. Using antibody tests, it's entirely possible that a large number of people (those who have no or very mild symptoms) could test negative and go traipsing around without realizing they are spreading the infection.

The ideal would be both test types, with rapid results (has have just been announced by Abbott (virus test) and Schein (antibody test) would be given one after the other. If you test positive for the virus, you go into strict quarantine. If you're negative, you do the antibody test, and then you'll know if you're in the clear. If you're negative for both, then you don't need quarantine, but you have to behave very carefully and continue to practice social distancing, etc.

the study you referenced is for sars cov2
not cov 19

Is mandatory testing (and targeted quarantine) for purposes of work or school currently legal under HIPAA?


I am required to demonstrated Hep B seroconversion or document three failed vaccine series as terms of my employment. In the event that I do neither, it is grounds for summary dismissal. We actually had to fire an MA who refused to get the vaccine during onboarding.

Quarantine is even easier 42 U.S. Code § 264 allows HHS to quarantine people to prevent the spread of diseases internationally and across state lines. Most states have similar laws for within their jurisdiction.

So yeah, we can easily make proof of seroconversion a requirement for work and we can already mandate quarantine for those testing positive.

Maybe there is some legal eagle vodoo I am missing, but both of this are already extant under law to some degree.

This also should raise your belief that the poor quality Chinese tests other countries are seeing may not be any sort of malicious act, but rather the PRC was able to limit the spread by using tests with relatively high rates of false results by using a lot plus centrally quaranting

At the same time, this result should increase your belief that the FDA and CDC insistence on getting a perfect test at the sacrifice of time delay (for design, approval, ramping up) was a very bad decision, making the perfect the enemy of the good.

The worse your testing is, the more indiscriminate the quarantine must be. At a certain point, you have to admit that the bulk of the benefit comes from quarantine rather than testing.

testing is very important but it is still over-rated by sociologists who don't
understand its limitations

Combine a tester with a census-taker and let’s meet two goals!

Good idea, but do they still send census people around to people's homes? I just got a census letter in the mail directing me to go to a website and fill out a form.

Yes. If you read the letter, they say that if you don't / can't fill out the form on the web, they will eventually send a person to your address.

That's $25 billion per test. Wouldn't you want to test everyone every day, or every couple of days? 20 tests per person is $500 billion.

Still, I'd pay $75 every two days to be able to go on vacation, say. Reduce the cost to $20, and that's like parking downtown for work. I'd pay $20 a day easily to get life back to normal.

Put a stamp on your passport as proof of serological conversion and immunity to coronavirus. Make it something like the yellow fever vaccination required to travel.

Only question is how long the antibodies last for. No one has answered that yet.

They last for ~ 2 years for SARS and the other 4 Coronaviruses that give only colds

I can only really cite authorities, since as you say no one knows but...

https://www.independent.co.uk/news/world/americas/coronavirus-vaccine-mutation-covid-19-latest-a9423301.html -

Scientists now are studying more than 1,000 different samples of the virus, Peter Thielen, a molecular geneticist at the Johns Hopkins University Applied Physics Laboratory who has been studying the virus, said.

There are only about four to 10 genetic differences between the strains that have infected people in the U.S. and the original virus that spread in Wuhan, he said.

“That’s a relatively small number of mutations for having passed through a large number of people,” Thielen said. “At this point the mutation rate of the virus would suggest that the vaccine developed for Sars-CoV-2 would be a single vaccine, rather than a new vaccine every year like the flu vaccine.

It would be more like the measles or chickenpox vaccines, he said – something that would likely confer immunity for a long time.

“I would expect a vaccine for coronavirus would have a similar profile to those vaccines. It’s great news,” Thielen said.”

Does not seem like a brutal mutator without much conferred immunity.

That doesn't necessarily answer the question of how long anyone would produce antibodies of course, only how long antibodies would be likely effective. But I don't know how you could answer that.

The antibody question has been answered for SARS-CoV and the other 4 more benign coronaviruses
Sars-Cov2 is pretty simlar in structure and MO to SARS-Cov. It uses its spike protein to attach the ACE2 receptor but SARS-Cov2 is a bit more effective at entry


Trace and isolate requires a high-trust compliant society. Societies that encourage price-gouging by black market hoarders are not high-trust compliant societies.

So mass testing is simply a waste of time.


In the US, New York City remains the undisputed epicenter of the national outbreak as the number of new cases out of the Seattle area has noticeably declined. An area that produced 37 of the first 50 fatalities in the US has seen deaths drop off markedly, while hospitals have been mercifully underwhelmed. While each infected person was spreading the virus to an average of 2.7 other people earlier in March, that number appears to have dropped, with one projection suggesting that it was now down to 1.4, according to the New York Times.

Spreading ratio, in abstract tree this determines the band limit on the tree trunk. Remember, in the balanced model, we are producing anti-bodies at this rate, two weeks later. The issue is how much of this k finite band limit is shared in the trading pit between the virus and anti-body. When a virus enters a community what i the chance of it surviving in a dense set of immune humans? The band limit leads to the 'cluster', on cluster hold enough humans to establish a sustainable balance, I call it the neighborhood. One neighborhood can collapse the virus in two weeks, after the flattening of the curve. One neighborhood can crash the anti-bodies in one immune time, which we are hoping is seasonal.

The dynamics.
Right now the balance is taking long shot guesses as to the neighborhood size, the majority of us, four weeks ago, we at zero, hence the explosion. The virus-antibody war is occurring in two or three phases, each about two to four weeks apart. So the neighborhood size is about the number of people seeing battle over a two week period, note it is dominated by incubation (or remission) time. In my home town, phase two looks to equilibriate in the central part, which has high mixing, a neighborhood of 150 thousand, and our wait time was three weeks in this agricultural town. When phase three hits, it will spread North of us to suburbia, but central Fresno will have a significant number f immune, the additional infections will be milder In equilibrium, the battle is waltzing over neighborhoods, about one in forty suffering as their immunity wore out..

I love droning on this.
At equilibrium, take a map of the USA and color the neighborhoods, with uncertainty, two colors. Black for virus white for anti-body safe. Note, they are slightly asymmetyric in their boundary conditions.

Back up from your map, zoom out and the map will turn a uniform gray with some fuzziness. The virus/antibody balance has 'two colored' the graph, to a good approximation. And that is where the math gets fun, the idea of graphs in the limit as uncertainty tends to the minimum amount, a calculus of graphs, a way to spectrally decompose logistic networks, look inside, when you see the complete sequence of transactions in a chain. This is what anstract tree is about.

at this stage of the viral pandemic
"antibody safe" should be considered a hypothesis not a probability

BTW the Imperial team now says that 6 million Italians have been infected as of March 28. Bringing the fatality rate there down to 0.16% from 11% (!!!!!!!)


Which led to the Italian Army being required to haul away the bodies, day after day.

When the evidence for a ratio of 1 known case to 80-100 unknown cases arises, don't forget to link to it too.

Evidence is given in the paper linked. Have you read it?
It is certainly not definitive yet. And the 6 million Italians infected are there medium figures, it can be many times more or many times less.
By the way, if you look at the figures for Spain, that gives you an even lower death rate, similar to a mild version of flu.

It is amusing that tis comes from the same team whose preceding reports were used by the worst fear-mongerers. Still, their arguments and estimates and evidence need to be discussed and taken with caution, like anyone's.

side question, do you or anyone else understand why prior_approval believes the fact that the Italians have engaged the army to help deal with the huge excess of deaths (which everyone acknowledges already) is some kind of rebuttal on whether the current deaths are due to a low IFR in a high infected population, rather than a high IFR in a low infected population? I can't work it out.

Another thing on this topic, assuming for the sake of discussion that the 6 million is correct, then about 50% of the probably would be Lombardy. 3 million in a population of 10-11 million, close to 30 local % infection rate...

WebMD says 5-20% of people get the flu... but that's "the flu" so who knows how many strains it covers in a given year. And 2-4 colds? Each?!! Yeesh, okay maybe 30% is believable then.


More epidemiology departments will likely veer towards the iceberg model of true infection (majority of infections invisible; "underwater") as it's just the most likely theoretical model (for reasons which anyone'd have to pretty pigheaded not to grasp, at this stage of discussion on MR). May somehow not be true, when tested, but most likely from everything we know.

I don't know if IFR will really get that low, once all is in. Deaths are almost certainly underestimated for now, though I'll feel uncomfortable if they start treating every 79 year old geezer who dies 4 months after infection as a c-19 death. But certainly if these bear up will be relatively low.

For Italy, as 28th March, even taking the low end estimate of pop infected (3.2%, 1,935,875 cases), and the high end estimate of real deaths (13,000, vs observed 9,136), current IFR would be 0.67%.

(Which is of course a lag on real IFR, though the further you get from infection, for most of the elderly population that die that harder it is to actually say it was Covid19 that is the real, single cause of death.)

You need eventual deaths to outstrip current measured deaths by quite high multiples, and infection rates to be lower than their lowest bound, to get to even 2%.

(Taking low bound % pop infected and high bound deaths for all countries, "real" current median IFR for all countries is 0.35%).

So this is the future now? From concerns about complacency and decadence to a medical police state in under two months? We wanted flying cars, we got 5 minute COVID tests. What happened to cost-benefit analysis, what happened to Baye’s Theorem? The “Give me liberty or give me death” strain of American thought really does seem to be dead, doesn’t it?

The modern coddled generation of Americans don't know the meaning of sacrifice.

Considering how few Americans were alive in 1945, this should be a banal observation, not a sneer.

I’m just wondering why we’d trust that the same people who didn’t run the numbers before (it’s obvious that aggressive testing and tracing with social distancing in January or early February probably would have contained the disease) have run the numbers now. It seems ridiculous to me to assume containment is possible now that the disease is global, you are only spacing out the outbreaks at great economic cost. Supposedly this is to save lives by stopping overflow at the hospitals, but what percentage of folks actually recover due to medical intervention given that we have no effective treatments beyond standard measures (which could be done at home)? I’m sure it’s tempting to think their must be a method or payoff to this madness, but that’s not necessarily true by any means. The fact is that the livelihoods of the many are being exchanged to prolong the lives of the few, and it seems to me decisions are being driven more by emotion than ethics.

The missed opportunity was not enacting distancing/isolation earlier. It's not obvious that this would solve the long-term issue of containing the virus, but it would have reduced the short-term impact.

Best case, we could have been like South Korea. But they haven't put the cat back in the bag 100% either.

Longer term, everything is speculative and testing will help a lot, but the astounding bureaucratic snafu around testing to date is not the key missed opportunity.

would surely have sealed the borders in Jan, had we known what we know now... but hindsight is poor insight

This is an incredible opportunity to teach and learn economics - and in particular about "money" - which is at the core the "medium of exchange" - I give you $X for A,B,C - if there is no A, B or C - no amount of money will be useful.  Congress passed TRILLIONS, the EU has passed billions in rescue packages - but yet people with the levers for power demand that people pretty much stay home - So who works to make A, B and C?  There is also a misunderstanding of "essential" versus "non essential" - imagine if a physician or nurse has a plumbing or an electrical or a car problem - how will they get to work -?  and so on and so on
We are incredibly connected and unless we look at the forest and not the trees we will lose the forest AND the trees - large scale testing as indicated here makes a lot of sense - Imagine taking say $50 billion to save the world - a small price.

"I want to expand on the costs because it’s clear that a mass testing regime will require millions of tests. "

A little perspective. The US has been testing over 110,00 per day for the past 4 days. That number has been climbing. So, we are headed in that direction already.

For mass testing to work, people who get a positive result need to voluntarily self-quarantine. There’s reason to think many people won’t comply, especially since one can have the disease and exhibit mild or no symptoms. (Will an asymptomatic 25 year old beer bro who tests positive but feels fine stay at home while his girlfriend goes to the club? Will an asymptomatic 40 year old line cook with two kids stay home and not get paid?) I’m still worried but Romer’s simulations suggest that even with low compliance, this may still work. Someone who tests positive but doesn’t comply is just like a false negative. Romer shows that even if there is an 80% false negative rate, the strategy will be very effective. In other words, this works with bad tests and good people or with good tests and bad people. Definitely worth a few billion.

Pay them, conditional on some data check - electronic tag probably and ideally through multiple lines of data - all registered phones as well, digicam check every X hours.

But voluntarily is not non-negotiable - impound them under house arrest if you can't trust them at all (say they're a heroin addict who can't control going out for their fix).

Better and cheaper than population wide house arrest. Less of an imposition on liberty too.

Romers model that suggests a mass screening test with 80%false negative rate has a net beneficial affect should be looked at very carefully before it is accepted as true. this is why uncle paulie woulda made a shit biologist and ended up working for the newyorktimes.com

I would like to hear about any "trace" teams that are already in place as well as the number of trace teams in training. What data and tools might they have at their disposal? What laws will reinforce their decisions?

"A Solution if We Act".

For almost all of us, a better message is "a solution if we sit fucking tight for another seek or so and monitor developments."

There is every reason to expect that if most of use continue isolating and being careful we will get over the first hump.

After that, it'll be trickier, and the ability to monitor and respond quickly to new outbreaks will become important.

Lots of testing will help with the second wave, but it seems beside the point in getting through the first phase, other than being the focus of a lot of hysteria.

Amen. I’m still pretty upset and shocked that this hasn’t even breached federal policy makers’ discussions about how to address the pandemic. Instead you have Trump (need I say more?) and then Democrats adding in carbon emissions policy. Seriously?

"I expect the emergency will peak in mid-April and then things will slowly start to get improve. "

Here's where you go awry. In a couple of weeks we might see a peak in new infections in the current small number of hotspots, but by then there will almost certainly be a dozen new hotspots, especially in those states who are taking a wait-and-see attitude about implementing measures. And in those hotspots we have now, the peak in deaths will come several weeks after the peak in new cases, and the rate of new cases will take a long time to fall to near zero.

I think this optimism that things will be as bad as they're going to get by mid-April is wildly optimistic.

I don't think the facts support this line of events. Sure there will be other hot spots, but it's doubtful they'll be as bad as NYC.

The two factors that contributed to that are:

First, very high population density with a heavy reliance on mass transit.

Second, limited available testing in the cluster formulative stages.

Testing is 5 orders of magnitude higher (110K+ per day) that it was a month ago. A few states are, for unknown reasons, not doing much testing, but most of the country is running tests as fast as they can. Furthermore, very few areas of the country are as densely populated as NYC.

FYI. looking at the current testing by state, the only state that seems far behind in testing is Delaware.

That's using a rudimentary Confirmed infected divided by tests run metric.

That Strasbourg and the Alsace does not resemble NYC very much at all. It does resemble the DC metro area fairly well in a number of ways, though with a considerably lower population density.

Everyone tells themselves that covid will be different where they are. Till now the only difference has been in how well a society is able to handle a looming and then actual pandemic. Italy, being first in Europe with community spread basically got very, very unlucky. Spain and France have less excuse. And each nation confronted with rising numbers of cases has even less excuse (Iran is a prime example), unless they spent the time in intensive preparations.

I couldn't find exact statistics for use of mass transit in Strasbourg, but in looking I found reports such as this one saying it "is considered to be one of the best examples of an integrated, multi-mode transport system in Europe". ( http://hankdittmar.com/wp-content/uploads/2015/09/StrasbourgReport_revi.pdf ). That makes me think that use of public transit there is dramatically higher than most places in the U.S.

Before testing everybody, start by testing a random sample with a fairly high quality test. This will tell you what you’re facing.
Percentage infected and mortality rate of Covid-19 .
We still don’t know that yet.

This is all well and good but we are in the middle of the hurricane and Alex is telling us if we used so and so shingles the roof would be much more secure. In the next week can you ramp up enough tests so that we can test for Covid as fast and as easily as diabetics can check their blood sugar or people can check their pulse or blood oxygen?

I don't see any way around a month and a half or so of lockdown. If we can finally get a few hundred million tests, we could move to the S. Korea model of test, contract trace, isolate to maintain R<1 to drive the virus into remission until a good vaccine is online.

The question I'd like to have an answer to is
"what is the differential mortality rate, among the people that doctors would like to hospitalize, between those that are hospitalized, and those that are not due to lack of resource?"

Same questions with "hospitalized" replaced by "put in ICU" or by "put under a respirator" (and also by "put under ECMO", but this more marginal).

Any idea among (true) doctors here?

All the real doctors in an actual position to answer that in a place like Milan or Madrid are far too busy to read MR. And have been, for weeks.

Why do we need to test at all? Please identify the mistakes in the following or identify the badly mis-calibrated assumptions.

Suppose the incubation period averages 5 days. Suppose R nought is 2.5 and represents the spread happening over the next 9 days. Suppose 14 days after infection the person is either hospitalized+isolated or recovered+immune.

The R nought calculation is comprised of asymptomatic people, say 50%, spreading it to 1 other person over 9 days and symptomatic people, the other 50%, wrongly assuming they have some other illness but actually spreading SARS-Cov-2 to 4 other persons over 9 days. R nought equals 50% times 1 plus 50% times 4 and equals 2.5.

Suppose public pressure and newly adopted conventions and new laws cause the asymptomatic to spread to 0.5 other persons over 9 days but that person contributes almost as much to the economy as they did before the outbreak. I am assuming the new behaviors have minimal economic impact like more time spent washing hands and having to wear a common textile mask (not N95) when in crowded places like trains. I am assuming asymptomatic people are not capable of spreading the infection nearly as much as symptomatic people because, for example, they are not coughing. Not coughing is one of the qualities of an asymptomatic person.

Suppose public pressure and newly adopted conventions and new laws cause the symptomatic to almost immediately isolate without knowing the nature of their illness. Suppose the symptomatic spread it to 1 other person in the brief period before self imposed isolation. This means what was 4 new infections propagated from one symptomatic person becoming just 1 new propagated infection. I am assuming the symptomatic are more capable of spreading the virus than the asymptomatic even when new behaviors are practiced. The most important new behavior is one that affects symptomatic persons only and that is the almost immediate isolation without perfect knowledge.

Thus R nought is now 50% times 0.5 plus 50% times 1 and equals 0.75. Thus the growth of infected cases decreases.

R* is a figure that always has a cultural context attached. The most contagious disease in the world will measure R*=0 if you have everyone in space suits all the time. This disease seems to have a 'natural' R* of 2 or 3. If we pay more attention to washing hands, wearing masks, avoiding unnecessary crowds we can decrease that. If we do testing, tracing and spot isolation, we can decrease it more. If we get R*<1 the virus will burn itself out. The closer we drive it to 0, the faster it will burn itself out.

Note that if asymptomatic persons do not propagate at all, the scenario is even better.

Note that the main benefit comes from symptomatic people being quick to isolate. If they are not quick; things do not necessarily have to be terrible because they too are washing hands and wearing a common textile (not N95) mask when in crowded places like trains. These behaviors are mitigating their slowness to isolate.

There is something missing from the economic part of the scenario. There is economic impact from people with the common cold self-isolating when they do not have SARS-Cov-2. This is not so bad because may be some people prefer this.

It seems ironic the stock market depreciated a lot just because we can't do what we were supposed to do in the past. People should wash their hands and people should not go about with flu symptoms and may be not with common cold symptoms either. At OECD levels of per capita GDP why aren't all water faucets infra-red or light beam activated 20 years ago. In more affluent countries we should think about more automatic doors. Buy Stanley (SWK, 2.9% dividend)? Who else makes these things? On the other hand , if we had more automatic faucets and doors we would have less herd immunity and might be adverse in connection to the endemic diseases.

He could add using PPE for citizens where individuals put movable boxes around themselves.


Another easy variation would be to add exclusion boxes like using area isolation and making the boxes larger and smaller while adding some teleportation and leakage between boxes to the triangles that could be positive but unknowns for between tests.

My hypothesis would be that area boxes that aren't near 100% leak (complete economic shutdown) and teleportation proof will fail and individual PPE will work under all conditions by decreasing the local R(o) value to below one by decreasing the spreading transfer probability per contact when the correct PPE usage gets into the 50 to 70% range.

The transfer is one on one local and the solution must include individual testing or protection or both.

Sorry Alex.... Nancy thought it was more important to funnel money to the Kennedy Center. Maybe next pandemic, though!

Who peforms the tests? Testing of everyone seems like a virtual guarantee that every health care worker is going to get exposed.
Or, do I get to jab a Q-tip up my own nose?

That's $25 billion to test each person ONCE. Test and trace, even in the paper cited above, requires each person to be tested REPEATEDLY. If everyone is only tested once, even a single false negative is enough to allow the virus to break containment. Test-and-trace assumes repeated testing, always, and doing so on the whole population is not $25 billion but $25 billion times the number of times we have to do this (10, 100, 1000?)

I'm going to guess that the cost of testing a person three times is far less than the cost of that person being out of work for an indefinite number of weeks.

Repeated testing at places of work will be quite cheap. Once we get control of the situation, however, we don't need to test everyone but rather quickly apply lots of resources to any outbreaks.

To this point, testing has not been particularly useful except to tell us how badly NY screwed up in encouraging people to congregate a month after Daegu had abandoned the same "see no evil" strategy. Anyone symptomatic should have been self-quarantining anyway, unknowable what proportion of those actually infected were ever tested in any region, or are now immune. Error rates? Who knows? Throwing darts in the dark.

Mass testing is a different story, particularly when it identifies the immune. Probably only need to do it once or twice, treatments and vaccines will be along, to both ease symptoms and prevent infections.

Clock is ticking against the Wuhan virus as more people become immune and and better treatments make it into broad use. Initial HCQ results are mostly encouraging but if they prove illusory others will follow... these are not normal times, tests must be accelerated until at least one working treatment is found, then we can pump the brakes a bit.

Think of what happened in whole genome sequencing. The costs went down from billions, to tens of thousands, to thousands, to hundreds in a matter of years. And at 30kb, testing for SARS-CoV-2 is a much easier problem in some ways. Let people innovate!

A side note, in Kremer's Patent Buyouts, footnote 3 says that Romer was suggesting a buyout of PCR decades ago.

Good luck getting people to do this test repeatedly. https://twitter.com/DrPeckPNP/status/1244062665535864832?s=20

Comments for this post are closed