[Epistemic status: Very weak – I’m still trying to figure all of this out. Some things in here will almost certainly be wrong. Please don’t let this overrule what government agencies or your common sense are telling you. For a more careful guide to the coronavirus and what to do about it, see here.]
For a description of why you might want to prep, see Putanumonit: Seeing The Smoke. For a description of how to prep, see this article by Kelsey. For a really intense guide by a professional prepper, see here.
But there’s such a thing as being too intense. You probably won’t need to store water – the water kept running in Wuhan. You probably won’t need a generator – Wuhan has electricity. The most important thing seems to be food (and toiletries, and other necessities). If the epidemic gets bad, you’ll want food so you can avoid going out to coronavirus-filled supermarkets. And if you get the coronavirus and are feeling sick, you’ll want food at home so you don’t have to get too far out of bed.
What about Amazon? Getting fresh groceries delivered whenever you want seems like a pretty good alternative to stocking up on canned beans. You’d need faith that it won’t get so bad that Amazon’s logistics break down. I think that faith is mostly justified – The Chinese version of Amazon seems to still be making deliveries in Wuhan. And it’s hard to imagine anything – pandemic, nuclear war, demon apocalypse – getting between Jeff Bezos and his next billion dollars. Three cheers for capitalism!
Coronavirus can remain on objects for a few days, and Amazon warehouse workers may come into work with the condition, so if you’re using Amazon as part of a self-quarantine strategy, you should probably open packages away from your main living space, using gloves, and disinfect anything inside that doesn’t have its own packaging from the manufacturer.
How bad will it be?
From the Chinese numbers, people have estimated these death rates:
The good news is that it’s pretty unlikely to kill young people. The bad news is that even young people seem to have severe cases that can land them in the hospital. This comes from former FDA commissioner Scott Gottleib – the rate of hospitalizations for flu vs. COVID-19 by age group:
This page says that the Chinese population is 70% under 49, 20% 50-65, and 10% 65+. So the graph above implies that every demographic has approximately equal hospitalization rates, which other sources suggest are 15% to 20%.
This is a weird pattern – why are so many young people getting hospitalized if almost none of them die? Either the medical system is serving these people really well (ie they would die if they didn’t go the hospital, but everyone does make it to the hospital, and the hospital saves everyone who goes there), they are being hospitalized unnecessarily (ie they would live even if they didn’t go the hospital, but they do anyway), or it’s statistical shenanigans (eg most statistics are collected at the hospital, so it looks like everybody goes to the hospital).
Are these an overestimate? Maybe most cases never come to the government’s attention? There’s some evidence for this. In South Korea, coronavirus started spreading through a tightly knit cult, so the government had to test the whole cult, including the people with no symptoms. 70% of asymptomatic members tested positive for the virus, suggesting that many cases may not come to medical attention (although some of those people might have just been in the incubation period). So hopefully the 20% hospitalization rate will prove to be a worst case scenario, and the real number will be less.
[EDIT: several people are saying that in other countries the hospitalization rate is closer to 5%; let me know if any of you find a source for this]
Self-quarantine starting now vs. later?
Rough Fermi estimate: there are 10 reported cases in the Bay Area. Optimistic scenario, we’re only missing another 20. Median scenario, most haven’t been detected and there are really 100. Pessimistic scenario, 1000.
(in a draft, I used the term “worst-case scenario”, but it turns out some people can think of some really bad scenarios. Let’s stick with “pessimistic”.)
The number of infected people doubles every 2-7 days. Let’s say optimistically 7, median 3.5, pessimistically 2.
So over the next week, we can expect there to be 20 (optimistic), 300 (median), or 10,000 (pessimistic) new cases in the Bay Area.
The Bay Area has a population of 8 million. So we can expect that 1/400,000 (optimistic), 1/30,000 (median), or 1/800 (pessimistic) of the population will be infected in the next week.
Suppose you’re trying to decide whether to quarantine your 5-person household. Worst-case scenario you multiply the number by 5. But realistically your infection risk will be correlated – you’ll be in the same areas, hanging out with the same social circle. If the epidemic spreads solely among Chinese-American schoolchildren in San Mateo, and none of you are in that demographic, then you’re all safe. So for the median and optimistic scenarios, we’ll say risk varies with square root of the number of people in your house, and for the pessimistic we’ll keep it linear. So per house risk is 1/200,000 (optimistic), 1/15,000 (middle), or 1/160 (pessimistic).
So even in a pretty pessimistic scenario, Bay Arean households have a less than 1% chance of getting the virus this week (people from elsewhere will have to redo the math with their local numbers). Next week you can check how many cases there are (and how many hidden cases you suspect there to be) and reassess.
Ways this model might fail: there are more than 1000 cases in the Bay Area, you live with more than 5 people, some of the people you live with have an unusually high risk profile (are health workers? travel a lot?), unknown unknowns.
EDIT: Daniel Filan on Less Wrong has made a more complete risk assessment model.
What’s the endgame?
The theory is: you quarantine yourself for a few weeks or months while the epidemic is going on, eventually there aren’t enough virus-naive hosts to sustain transmission, and we’re all safe until the next time some moron in China eats a bat.
This theory might not be true. CDC director Robert Redfield says:
This virus is probably with us beyond this season, beyond this year, and I think eventually the virus will find a foothold and we will get community-based transmission…The containment phase is really to give us more time.
I don’t know exactly what this means, but it sounds like he’s saying it will become something like the cold, flu, or chicken pox – a virus that’s just ambient all the time, and you’re always at risk of getting it. You can’t outlast it because it doesn’t go away.
If that’s true, the goal of quarantine would change to outlasting the crisis. There will probably be a peak of the epidemic when hospitals are overcrowded and there are no spare ventilators and you really don’t want to need medical care. If you can get past that, and catch it a year or two from now, there will be spare doctors and nurses to care for you and spare ventilators to support you if you need it.
But if that’s true, maybe it would also be acceptable to catch it early and beat the rush. If you catch it literally today, you’re near-certain to get a hospital bed; if you catch it a month from now, who knows? This is another argument against premature quarantine, although not a strong one if you expect your quarantine to work.
(This part is especially speculative, even more than the rest of this post. Please don’t go out and try to catch the coronavirus on purpose.)
No, really, what’s the endgame?
Some sources argue that recovered patients can get reinfected with coronavirus, ie people don’t build immunity after surviving it once. This would really suck. I don’t know exactly how to model this – surely it wouldn’t mean a perma-epidemic all the time – but at the very least it would mean it was around permanently and even good quarantines wouldn’t be very valuable.
Microbiologist Florian Krammer (link) thinks these are probably false positives. Some other people have said they might be very elderly or sick people with abnormally poor immune responses. Let’s hope.
Home health care
Some people in the rationalist community have put together a document with some useful suggestions. Among them:
– Most coronavirus infections will be generally mild and probably treated at home. Make sure your home has oral rehydration solutions like Gatorade or Pedialyte available to deal with dehydration from vomiting, diarrhea, or other causes, and Tylenol available to help with fever.
– If you think you’ll fret a lot about whether you need to go the hospital, get a pulse oximeter ($20 on Amazon). Readings consistently below 90-94% are pretty bad and mean you should seek immediate treatment.
– Everything that makes you feel better and healthier in general helps your immune system too, so get good sleep, eat healthy, and take Vitamin D if you’re deficient.
– The best way to avoid infection is still to wash your hands frequently and avoid touching your face. Less good but cooler ways to avoid infection include putting copper tape over everything in your house (copper kills viruses on touch). For best results, use a combination of copper, iron, and silver tape to be protected from coronavirus, fae, and werewolves simultaneously.
Don’t visit Grandma
Seen on Tumblr: the coronavirus death rate is 0.2% for people in their 30s and 15% for people in their 80s. The vast majority of our concern should be for our elderly
and for what happens to the 2020 election after we lose every presidential candidate except Pete Buttigieg.
You might want to call your elderly parents and grandparents and ask them to store enough food and water to stay in the house for a long time. You might want to talk to them about how bad things would have to get before they should stop visiting large gatherings, and how bad things would have to get before they stop leaving the house. You might want to explain to them how to use Amazon if they don’t already know how to do it, or coordinate with them so you can Amazon them things if they need it later on. You might want to make sure they have enough of their prescription meds stored up in case going to the doctor suddenly becomes harder. You might want to talk about whether they should come stay with you assuming you aren’t more exposed to the outside world than they are; if you have kids who go to school every day, your elderly relatives might be better off at their retirement home.
But have these conversations by phone! Don’t visit them! The last thing we need is everyone suddenly going and visiting the elderly much more often than usual!
(Grandma, if you’re reading this, the reason I haven’t called you about this yet is that Dad is investigating options and he will be contacting you soon).
Also, although the original post was titled “Don’t visit Grandma”, technically it’s Grandpa you should be most concerned about – elderly men seem to have twice the death rate of elderly women. [EDIT: Michael Keenan points out this might just be because elderly men in China smoke a lot and have bad lungs]
The view from China
A Chinese-speaker on the subreddit sums up what they’ve gleaned from Chinese social media. For example:
People are talking less about Wuhan and more about the rest of the world in recent days. The worst of the overload in Wuhan is over, and the quarantine in the rest of the country is keeping spread more or less manageable. People are talking more about the usual scandals and fandom stuff. Meanwhile, people are horrified at, say, Japan’s mismanagement of its earlier cases. I’ve seen the attitude “we sacrificed Wuhan to slow the spread and you guys are just wasting your window of opportunity?” And comments that maybe the CCP isn’t that incompetent in comparison after all.
The framing of sacrificing Wuhan to save the rest of the world is pretty poignant, and partly true; they did buy a lot of time. And cruachanmor describes the ways the Chinese quarantine can basically be called a success:
The fact is that the Chinese efforts at containment do seem to have worked – cases have been dropping consistently now for nearly a couple of weeks. Back in the first week of February the Chinese were diagnosing between 3,000 and 4,000 cases a day (and that was lab diagnosis only) on an alarmingly upward trend, yesterday there were just 415 (lab+clinical), all but 17 in Hubei, with 600 or so suspected (which follows a similar pattern). That’s really quite remarkable – and despite the common internet belief that China is lying through it’s teeth that’s unlikely because more open societies like Singapore and Thailand have followed similar patterns. If there were a more widespread infection then this wouldn’t be the pattern. Indeed at this point the stats would suggest there’s now significantly less people in China with an active infection than there was two weeks ago – this is not “dubious efficacy and inherent harms of China’s historically unprecedented crackdown.” – well that is unless the only thing you care about is the amount of product coming out of Chinese factories.
Whilst the virus may now be in more countries, epidemiologicaly there’s little difference between spreading around the world and spreading in China – which does after all contain 20% of the world population.
Of course that’s not to say we won’t see a pandemic, but The Atlantic is way off on this. In the UK for instance a 40% infection rate with a 2% death rate would be half a million dead. If any democratic government threw up it’s hands and took the ‘fuck it, we’ll let it burn through’ approach suggested by the Atlantic whilst at the same time China does manage to contain it then said elected government is not going to be elected long.
No country that has to respond to the concerns of its citizens (and that includes China where the CCP needs to retain legitimacy) is going to let this go. Even Iran, as there’s a lot of pent up discontent by the younger generation there who are not going to look kindly on their government after a major epidemic.
The story I’m hearing from most smart people is that China has done amazing work and mostly halted their epidemic. This is very impressive, but it’s unclear what it will achieve long-term; they’ve basically turned off their whole country, and they’re going to have to turn it on again sometime. Once that happens, the coronavirus can just pick up where it left off, either from a few people in Wuhan who they missed, or from foreign travelers. I guess they’re gambling they can put out every fire as soon as it starts using the logistical capacity they built for Wuhan, and maybe they’re even right. Nobody’s too optimistic that democratic countries can follow their lead, though.
No problem so bad overregulation can’t make it worse
So far the government has bungled its coronavirus response pretty egregiously.
Most hospitals have the equipment in house to detect coronavirus. But the FDA banned them from using it. They said all coronavirus tests needed FDA approval, and refused to approve anything except the official test made by the CDC.
Unfortunately, the official CDC test was defective. The test itself worked, but one component in the test kit was broken. Most hospitals had their own supply of this component and could have substituted it in, but the way the FDA approved the CDC test banned them from doing this.
The CDC tried as hard as it could to fix their broken tests quickly, but they weren’t able to do it fast enough to satisfy demand. In order to ration the scarce tests, they mandated that hospitals only test people who had recently been to China, or been in close contact with someone who had.
This was a disaster. For example, here’s a story from a person who traveled to Japan, where the coronavirus is active. He came back to the US, started developing symptoms, and went to a hospital. The hospital said since he hadn’t been to China, they couldn’t test him, and sent him home (he voluntarily quarantined after discharge, so thanks).
But even worse, the policy ruled out by fiat ever being able to detect when the epidemic spread to the US. So in mid-February, when a patient with no history of travel to China came to a hospital in California with coronavirus symptoms, the doctors had to ask the CDC for special permission to test. The CDC dithered for four days before granting the permission, during which nobody put any work into containing the disease. Finally the test came back positive – after some health workers had already been infected.
There were many points where this could have been avoided. A better CDC could have made tests that worked from the beginning, or ramped up production of working tests faster, or come up with smarter criteria for rationing tests. But it would have been even better to have a system where things don’t have to go perfectly, and where a few mistakes don’t choke up the entire response to an epidemic for weeks. If we hadn’t let our culture reach the point where governments ban things by default and review at leisure, and where individual iniative is frowned upon in favor of waiting for official permission to do the right thing, we could have recovered from all of these mistakes. Hospitals would have used their existing tests which they already have more than enough of, doctors would have had permission to test suspicious cases at their discretion, and we would have had a chance to catch infections early before they could spread. If the government didn’t already regulate adrenaline, buspirone, insulin, and genetic testing to the point of near-unavailability, maybe people would have thought it was weirder, or raised more of a fuss, when they started doing it for coronavirus tests.
If you don’t trust me, trust former FDA director Scott Gottleib, who explains the situation here in an unusually candid communication from an ex-government official talking about his former agency. His Twitter feed is a great source of information in general.
And here’s a more careful analysis of some of the laws around diagnostic testing and how they contributed to the current crisis. And by more careful, I mean it ends with “Bottom line: the FDA is going to kill us all”.
The efficient market hypothesis is the real victim
I know a bunch of people who sold or shorted stocks when the virus started hitting China hard, based on the assumption that a pandemic sounded bad for business and business hadn’t priced this in yet. Jacob of Putanumonit and Wei Dai on Less Wrong both mention doing this and making healthy amounts of money. Good work on their part – though given the recent market crash, trying to replicate their success now seems suspiciously like a buy high, sell low strategy.
Masks and respirators
People are coming down really hard against face masks for coronavirus. Wirecutter: You Don’t Need A Face Mask For Coronavirus> Marketwatch: US Health Officials Say Americans Shouldn’t Wear Face Masks To Prevent Coronavirus – Here Are Three Other Reasons Not To Wear One. Fastcompany: Five Reasons Not To Wear A Surgical Mask To Stop Coronavirus (number five is “Just don’t.”) The CDC and Department of Health and Human Services have both officially recommended against mask-wearing.
Some of the reasons given are idiotic: the virus is not common in the US yet, so you are wrong to worry. This reminds me of all the people saying that AIs are not currently superintelligent, so any discussion that AIs might become superintelligent is just fearmongering. Who are these people? How are they still alive? How do they avoid driving off cliffs? They’re heading towards the cliff face, and their passengers scream at them: “YOU’RE DRIVING TOWARDS A CLIFF!”. And they calmly respond with “We are not falling off the cliff yet, we’re on perfectly level ground, there’s no reason to panic.”
Other reasons are superficially better, but collapse under scrutiny. Masks, the articles say, only help people who are sick avoid transmitting the disease, so healthy people don’t need to wear them. But the coronavirus can be asymptomatic for weeks, so if you want contagious people to wear them, part of your target demographic will think they’re healthy. Also, if you become sick, you don’t want to have to go out to the store to buy a mask or wait for it to arrive from Amazon, you want to have it immediately.
Other reasons make more sense. Standard-issue surgical masks may not help much, especially if you aren’t trained in their use. The coronavirus isn’t airborne; wearing a mask while walking outside is unnecessary – it’s when you’re kissing your spouse or chatting around the water cooler or touching a doorknob that you really want to worry. Any mask weak enough to stay comfortable while wearing for long periods, or even weak enough to talk through, is probably too weak to work. There’s a good NYT article about some of these issues here. If people with masks risk compensate even a few percent, that’s enough to make them a net negative.
And other reasons are really truly excellent: if random people buy up all the masks, there won’t be enough left for health care workers or the very sick. I agree this one is a good point.
But the thing is, I already own a P100 respirator. I bought it during fire season last year, aka the-air-is-unbreathable season. Living in California is full of excitement, and after a couple of years you end up prepared for lots of stuff. And the other day, I wore it on the BART – a densely-packed subway full of people who are constantly breathing in your face. And my friends told me – haven’t you heard that the government says masks don’t work?
An N, P, or R rated respirator, worn properly, in specific high risk situations, can be an appropriate part of a safety strategy. I think an accurate treatment of the topic would admit this, while also stressing the reasons most masks might not appropriate for most people in most situations. The statements and articles I’ve read don’t seem up to this level of subtlety. Instead, they seem focused on getting people to do what they consider the right thing (not hoard masks and panic) at the cost of oversimplifying the situation, sometimes up to the point of mistruth.
Their goal is understandable, and maybe this kind of simple messaging is the right choice during a pandemic. But the thing is, I doubt any government preparedness czar called up the major media companies and started their pitch with “I’m going to ask you to make an unprecedented sacrifice today”. I doubt there were hours of soul-searching in newrooms and government PR departments as people considered whether to take this step. I think this happened instantly and seamlessly because it’s what everybody has been doing all the time for years.
So no, don’t use masks or respirators unless you know what you’re doing and are sure it isn’t inconveniencing anyone else. But also, try to pay attention to the forces shaping your informational environment.
(update: South Korea is going full speed in the other direction and making it a top priority to give everybody masks. See the Reddit thread about it here. I hope someone compares results once this is all over).
There are a couple of compounds that probably help a little against the common cold. Some cold viruses are coronaviruses, and lots of respiratory infections share common mechanisms. Could these be helpful against coronavirus too?
For example, most studies suggest that zinc probably shortens the duration of cold symptoms by 25 – 50%. The studies didn’t really look at severity, but plausibly it affects that too. It’s unclear exactly how it works, but some people think it inhibits viral replication. It also might be involved in immune system function, and there’s some growing evidence that adequate zinc levels prevent pneumonia.
None of this evidence is great. And even if it were, there’s no guarantee it would transfer to coronavirus. Taking zinc to help coronavirus is a really long shot.
Still, it’s weird how explicitly anti-zinc a lot of sources are. Consumer Reports gives 6 reasons (why are these things always six?) that you shouldn’t use zinc for a cold – number one is because, although it will shorten the cold, it won’t prevent it entirely. That’s some galaxy-brain-level reasoning there. Wall Street Journal just warns people that zinc is not known to prevent coronavirus and so you should wash your hands instead, and the Telegraph warns about zinc in the context of scam coronavirus cures we should watch out for.
Again, I understand the impetus. Maybe people are morons, and if you say there’s a chance zinc might be helpful, they’ll go out, down a whole bottle of zinc pills, and then fly to Wuhan and gorge on batburgers because they think they’re invincible. Then when they get it anyway they’ll blame it on you. And there are so many scammers selling so many dumb supplements that it feels irresponsible to do anything which might aid or abet them. There’s no good way of saying “maybe there’s a 10%-20% chance, with lots of Knightian uncertainty, that zinc could make your coronavirus case a little less severe, but zinc pills are cheap and safe, so you might want to try it as long as you stick to safe levels”.
But again, I can’t help feeling like my information environment is being optimized to prevent stupid people from panicking, and not to help me make good decisions. I would love if the average news site I went to had an interview with an immunologist who gave their honest probability estimate for whether zinc would be helpful or not. In the real world, all we can do is make dumb guesses on our own.
Edit: knzhou has better information on this, including a microbiologist’s opinion.
Metaculus is optimistic
Metaculus is a prediction engine run by some cool people at UC Santa Cruz. It crowdsources predictions made in different ways, then studies various algorithms for aggregating them together. Unlike PredictIt, which is almost useless for anything outside the presidential primaries and Trump’s tweeting habits, Metaculus is doing a great job covering the coronavirus epidemic and the questions that are important to me (sometimes too good – I wasn’t expecting them to tell me how likely is is that the Effective Altruism Global conference next month gets cancelled).
As of me writing this they’re estimating that coronavirus kills 140,000 people worldwide in 2020 (25th percentile estimate is 33,000, 75th percentile estimate is 1,100,000). This confuses me – the flu kills a few hundred thouasnd people yearly, so the site seems to think COVID-19 will be less bad than an average flu season. Maybe they expect that people will demand higher standards of evidence in attributing a death to the coronavirus than people have making wild guesses about how many people die of the flu? Or maybe they predict a very high chance it stays contained?
The best source for continuing information about the progress of the epidemic is the Johns Hopkins coronavirus dashboard. And feel free to use the comments here to share any other information or questions you might have.