20 March 2020
I’ve been overwhelmed (in the positive sense) by the number of thoughtful e-mails I’ve received with ideas, resources, good articles/reports, suggestions and questions. I can’t do justice to all of this in a single update.
Yesterday I posed the question: “What are the arguments against the current ‘expected case;'” and in particular, what are the best arguments that the human/health and economic impact may be less severe than markets and the media are currently suggesting?
I’m still digesting the input I’ve received in the last two days and need another day or two to organise my thoughts on this.
In the meantime, for today, I’ll share a hodgepodge of interesting materials from a variety of sources. In the spirit of engaging with counterarguments, they’ll be lopsided towards those who argue that things aren’t as bad as they look.
Interesting updates from a variety of sources
Taleb et al. disagree with Ferguson
A strongly worded (would you expect anything else from Taleb?), but largely polemical (rather than introducing new evidence), piece arguing that lockdown followed by track & trace followed by tight border controls can be effective in controlling COVID-19; basically, why Ferguson is wrong.
Israeli virologist: we’re overreacting
“A leading Israeli virologist on Sunday urged world leaders to calm their citizens about the coronavirus pandemic, saying people were being whipped into unnecessary panic.”
What do the superforecasters say?
I’ve referred in earlier posts to Philip Tetlock and his seminal book, Superforecasting. (Indeed, we have a superforecaster friend who contributes to these updates!). Two offshoots of his work, the Good Judgement Project and Metaculus, are running forecasting tournaments on COVID-19:
Good Judgement’s questions all have an evaluation date of 31 March 2021, one year from now. The algorithmically aggregated probabilities [corrected from my early misstatement that these %s reprented the share of forecasters] are as follows:
Reported cases – 50%: Between 53 and 530 million cases reported by then
– 14%: worse than that — implying less than 14% probability of global attack rate will be above 6%!
– 36%: fewer than that
Deaths: 50%: 800K-8M WW
– Only 13% predict 8M+.
So in summary, forecasting tournaments don’t think it’s as bad as some are predicting from the perspective of confirmed cases and total deaths. (Still, 800K-8M deaths worldwide is a lot, compared to around 500K for flu.)
Ackerman’s CNBC monologue
If you haven’t already watched it, this is well worth a few minutes:
In fact, if you listen closely, Ackman is arguing something close to or even less severe than the current “expected case” / received wisdom: that intense intervention, early, can reduce both the health/human and economic impact.
Young people aren’t immune after all
This is a great example of how looking a mix/proportions among cohorts, rather than ratios within cohorts, is misleading. By the chart above, 39% of total severe (hospitalized) cases were patients 54 and under. So that means it doesn’t disproportionately cause severe cases in the elderly, right? Wrong.
If you look at the table below instead, it’s clear that on an in-cohort percentage basis, the disease disproportionately causes serious cases in older cohorts. No cases under 20 required ICU admission, and only 2-4.2% of cases 20-44.
A serological test may be coming soon!
From the Johns Hopkins daily COVID-19 newsletter
(Why does such a test matter? Current PCR testing based on saliva swabs can only detect the current presence of the virus; it can’t tell us who has had COVID-19 and recovered, nor whether an individual has antibodies that could confer resistance.)
SEROLOGICAL TESTING As diagnostic testing for SARS-CoV-2 infection ramps up in the United States, many questions remain regarding the number of cases and asymptomatic infections that are going undetected, both in the United States and around the world. The PCR tests currently used to diagnose COVID-19 patients are effective at identifying active infections by detecting virus currently present in the specimens, but they are not able to determine whether an individual was previously infected after the patient has recovered. For this, serological tests are needed. Serological tests identify the presence of antibodies, which were generated as a result of prior infection. A study published on March 18 (pre-print) describes the development and initial testing of an ELISA serological test by researchers at the Icahn School of Medicine at Mount Sinai, in collaboration with colleagues from multiple international institutions. Based on tests using human samples from both uninfected individuals and recovered COVID-19 patients, their preliminary findings indicate that the new serological test can effectively detect the target antibodies. Additionally, the researchers note that the test “do[es] not require handling of infectious virus” and that production is “amenable to scaling,” which could allow for rapid production in order to conduct larger population surveys.
Our World in Data switches from WHO data to EDC dataI’ve mentioned the outstanding Our World in Data website many times before. I was fascinated to see that they have stopped using WHO data and have moved to the EDC data set instead. I’ve also been very frustrated with the WHO data for reasons ranging from high latency (lag in reporting), to errors in reporting, to inconsistency with presumably authoritative sources. https://ourworldindata.org/coronavirus
Sobering interview with a doctor who helped defeat smallpox.
An outstanding interview with Larry Brilliant. What a bio! “Brilliant, a technology patent holder, has been the CEO of public companies and venture backed start-ups. He was the inaugural Executive Director of Google.org, […] the first CEO of Skoll Global Threats Fund […] Brilliant currently serves as the Chairman of the Board of Ending Pandemics, and is also on the boards of the Skoll Foundation, Salesforce.org, The Seva Foundation, and Dharma Platform.”
I’ve excerpted a few of the most interesting sections below.
Since it’s novel, we’re still learning about it. Do you believe that if someone gets it and recovers, that person thereafter has immunity?
So I don’t see anything in this virus, even though it’s novel, [that contradicts that]. There are cases where people think that they’ve gotten it again, [but] that’s more likely to be a test failure than it is an actual reinfection. But there’s going to be tens of millions of us or hundreds of millions of us or more who will get this virus before it’s all over, and with large numbers like that, almost anything where you ask “Does this happen?” can happen.
Is this the worst outbreak you’ve ever seen?
It’s the most dangerous pandemic in our lifetime.
By slowing it down or flattening it, we’re not going to decrease the total number of cases, we’re going to postpone many cases, until we get a vaccine—which we will, because there’s nothing in the virology of this vaccine that makes me frightened that we won’t get a vaccine in 12 to 18 months. Eventually, we will get to the epidemiologist gold ring. […] That means, A, a large enough quantity of us have caught the disease and become immune. And B, we have a vaccine. The combination of A plus B is enough to create herd immunity, which is around 70 or 80 percent.
Now that we’ve missed the opportunity for early testing, is it too late for testing to make a difference?
Absolutely not. Tests would make a measurable difference. We should be doing a stochastic process random probability sample of the country to find out where the hell the virus really is. Because we don’t know. Maybe Mississippi is reporting no cases because it’s not looking. How would they know? Zimbabwe reports zero cases because they don’t have testing capability, not because they don’t have the virus. We need something that looks like a home pregnancy test, that you can do at home.
Are you scared?
I’m in the age group that has a one in seven mortality rate if I get it. If you’re not worried, you’re not paying attention. But I’m not scared. I firmly believe that the steps that we’re taking will extend the time that it takes for the virus to make the rounds. I think that, in turn, will increase the likelihood that we will have a vaccine or we will have a prophylactic antiviral in time to cut off, reduce, or truncate the spread. Everybody needs to remember: This is not a zombie apocalypse. It’s not a mass extinction event.
Should we be wearing masks?
The N95 mask itself is extremely wonderful. The pores in the mask are three microns wide. The virus is one micron wide. So you get people who say, well, it’s not going to work. But you try having three big, huge football players who are rushing for lunch through a door at lunchtime—they’re not going to get through. In the latest data I saw, the mask provided 5x protection. That’s really good. But we have to keep the hospitals going and we have to keep the health professionals able to come to work and be safe. So masks should go where they’re needed the most: in taking care of patients.
How will we know when we’re through this?
The world is not going to begin to look normal until three things have happened. One, we figure out whether the distribution of this virus looks like an iceberg, which is one-seventh above the water, or a pyramid, where we see everything. If we’re only seeing right now one-seventh of the actual disease because we’re not testing enough, and we’re just blind to it, then we’re in a world of hurt. Two, we have a treatment that works, a vaccine or antiviral. And three, maybe most important, we begin to see large numbers of people—in particular nurses, home health care providers, doctors, policemen, firemen, and teachers who have had the disease—are immune, and we have tested them to know that they are not infectious any longer. And we have a system that identifies them, either a concert wristband or a card with their photograph and some kind of a stamp on it. Then we can be comfortable sending our children back to school, because we know the teacher is not infectious.
From a major biotech investor
A professional investor shared this on the condition that the source not be divulged (the summary is mine); as you’ll see, it’s firmly in the camp of “this is manageable, not the zombie apocolypse”
One specialist investor in biotech has shared with their partners that they think drug companies are at extremely attractive valuations. They make several interesting points. First, they think the impacts to drug companies of the crisis (both economics and health) are small — mostly delays in clinical trials and some temporary disruption to sales . Secondly, large biotech companies have experienced one of the largest drops in valuation in history. Third, many biotech companies have a strong cash position and therefore the ability to not only weather, but to invest through the crisis. Fourth, they think that after a period of strong control measures, following China and South Korea’s playbook, the US and other countries will be able to gradually relax controls, implementing “track & trace” with quarantine.