I’ve gone through a lot of introductions to this post but maybe this is the most honest one:
I am scared. Quite scared, actually. My chances of catching COVID-19 are actually quite low, and my chances of surviving it if I do are quite high, and I’m still scared. What if I get into a car accident and have to go to the ER? Will they have a bed for me? Will I leave with coronavirus? What are my pregnant friends going to do? What is anyone over 70 going to do?
My goal, and the goal of everyone on the LW staff, and I assume most everyone who’s participated in all the coronavirus threads, has been to figure out what is happening and what we can do about it. We’ve already done a lot. Posts like Seeing the Smoke got coronavirus on people’s radar faster than it otherwise would have been, aided by the numerous modeling threads backing it up. The Quarantine Preparations thread gave people a starting place to act from. The Justified Practical Advice (summary) thread let us share our expertise, in ways that led to concrete behavioral changes. More recently we examined asymptomatic transmission. I’ve had a legit, reasonably high ranking government official say they look at us to see where everyone else will be in weeks.
This is currently the LessWrong team’s top priority, and they’ve done a number of things over the recent weeks to facilitate research and action on coronavirus, including hiring me to be a point person on it. To facilitate as much progress as possible over the coming weeks, habryka and I have compiled a list of what we consider the most important questions in fighting COVID, and are asking anyone with the skill to help us answer them.
That list is at the end of this post. But first, what is the overall plan here?
Who are we trying to help?
We have three broad categories of potential beneficiaries in mind:
Individuals making choices for themselves and their loved ones, who need accurate information about the current threat level and how to lower it with existing tech.
Individuals creating the tools for the people above, meaning anything from noticing that copper tape is anti-viral to creating plans for DIY non-invasive ventilators, who need accurate information about how COVID-19 operates and where the current gaps and bottlenecks are. We’d like to help people in this group get volunteers and money when appropriate.
Organizations and institutions making decisions that affect many people, who need all the information the previous two groups do, plus more to know what the effect of their decisions will be.
How Are We Doing That?
I am managing a Coronavirus Agenda, composed of what myself and habryka think are the most important coronavirus-related questions to answer (think we missed some? Please comment). But the full agenda is kind of overwhelming, and there are benefits to coordinating multiple people around the same question, so every so often I’ll pull out Spotlight Questions to generate a critical mass of attention around the most critical questions. I want to say “every so often” will be once a week, but I feel like those kinds of commitments are for situations where I know within an order of magnitude how many people are going to die in that week. I will spotlight as often as seems merited by the situation at the time.
If your eye is caught by a question on the agenda that’s not currently spotlighted, of course pursue your interest. That’s the point of sharing the whole agenda. And if you think the agenda is missing something important, of course pursue that, and add a comment explaining it if you have time so I can add it.
Without further adieu, the spotlight questions...
Spotlight Questions
What is the impact of varying initial viral load of COVID-19?
The hypothesis that lower initial viral load leads to better outcomes, and might be worth pursuing deliberately, is a central assumption is Zvi’s post Taking Initial Viral Load Seriously. Is it true?
Economics Questions
The Full Agenda
These are the questions about coronavirus I and habryka (and in the future, commenters on this post) most want answered. We’ll be nudging LessWrong to pursue them over the coming weeks, but for clarity wanted to share the whole thing as a package.
Some of these someone has already answered, or attempted to answer, in which case I’ve linked to the (attempted) answers. I’ll continue to update as more answers come in:
How many people are infected?
Worldwide
In a location of your choosing
No one suggested a dashboard that met all of my or habryka’s goals. PlaguePlus.com is the placeholder winner for at least attempting to do estimates of the true count instead of just reporting test results, and for showing any history instead of just cumulative cases, but I’d sure love for it to be replaced by something that can show history broken down by region.
What projects need volunteers or donations?
We collected a number of suggestions and aggregations in the LessWrong Coronavirus Links DB (see Work & Donate tab), but ultimately didn’t find any that were both widely applicable and exciting to us.
What should I do if I get sick or am caring for someone sick?
My answer. This is 80/20ed, not completed.
What is my prognosis if I get COVID-19?
What is the basic science of coronavirus?
My favorite was this talk by a virology professor, it answered basically all of my questions, but requires too much background biology knowledge to be a perfect intro for everyone.
What is the impact of varying initial viral load of COVID-19?Q
What are the most predictable second order disasters?
What problems are people running into when trying to work on all of this? Are there more things like the link database that we need?
What skills should I be rapidly acquiring to be most useful to this whole situation?
What mental health problems can we expect to spike hard in the next 1-6 months given people feeling shut in and helpless?
What are the basic epidemiological parameters of C19, such as incubation rate, doubling times, probability of symptomatic infections, delay from disease onset to death, probability of death among symptomatics, etc?
How much food do I need to have stored?
I’ve seen anywhere from 2 weeks to 9 months and given that neither the money nor the space is trivial to everyone, I’d really like to see model-backed estimates.
What is actual hospital elasticity? Is there an existing gathering of data on this from previous disasters?
How long should I be in isolation given the median assumptions about the world and the specifics of my area?
Which physical objects have longer supply chains and thus can be expected to be less robust to disruption?
What can we do to raise the standard of home care?
Potential answers are anything from electrolytes to DIY ventilators.
Are most NSAIDs dangerous?
Is there an asymptomatic infectious phase?
Probably. Mean incubation period is 4-9 days, but the mean serial interval (period from when person A is infected to when they infect person B) is 4-6 (and estimates are closer to 4, although averaging different studies is not really appropriate)
What are the risks of…
Accepting delivery food
Accepting packages
Using public transit
Going to work
For a variety of workplace types
Hosting a large gathering
Hosting a small gathering
Shaking hands with with an infected individual
Walking past an infected individual in a hallway
Standing or sitting 4 feet from an infected individual and having 5 minutes of conversation
Opening a piece of mail handed to you by an infected person
Opening a piece of mail left in your mailbox by an infected person 1 hour ago
Holding a grocery bag handed to you by an infected person
Picking up an item in the grocery store that was placed on the shelf by that person 1 hour ago
How do I convince others to act?
What is the value of handwashing, when you are currently healthy? How much better is WHO-approved handwashing than what we do by default?
What is the value of copper taping high-touch surfaces?
What is the value of masks, when you are currently healthy?
What is the value of goggles, when you are currently healthy?
What is the value of contact tracing? How do you do it?
What are the chances of vaccine development?
What are the chances of treatment development?
Do we actually have any chance of an approach that is not herd-immunity based? Is there still any chance at containment?
Build new vaccine production facilities.
It seems clear that focusing on the vaccine pipeline will become critical in the coming months, and we need to get ahead of it ASAP. Currently, the plan is to wait for safety approvals, then start manufacturing. That will obviously change—when we have moderate confidence that a vaccine is effective, we will want to start manufacturing, but there are several candidates, and too little productive capacity to make large quantities of several different vaccines. In fact, there is too little productive capacity to make any one vaccine in global quantities without stopping manufacture of other vaccines.
Vaccine manufacturing is very complex, and needs specialized facilities with clean rooms, sterilization facilities, very specific types of HVAC, etc. Building these is capital intensive, and there has been too little capacity for quite a while, leading to occasional vaccine shortages. I think we should be pushing large companies and governments to figure out how to create greater production capacity for vaccines. This is a global public good anyways. There are a few economic concerns for companies doing this, but right now is the perfect time to get government subsidies for such capital intensive projects.
How specific are vaccine production facilities to individual vaccines? To what extend can we build them before knowing which of our vaccines will succeed in the clinical trials?
From what I understand, we use eggs to incubate and clean-rooms to produce the final product for all of them, and I understood that vaccine producers can switch between which ones they make, with a couple month delay for incubation and switching over.
FWIW, eggs are actually specific to influenza vaccine manufacturing. Page 3 of this book chapter ( https://reader.elsevier.com/reader/sd/pii/B9780128021743000059?token=F492A74B3C4545B108379536769CF93D7F1DB89321DADE859256496F5D85CB6259372D34376809219BBBE2FFFDEF25FB ) has a really nice table showing the production process of a number of different vaccines—they are all very different from one another. This is why we need new vaccine platform technologies—i.e., tech that can be used to produce multiple different vaccines. mRNA vaccines would fall into this category and is a reason why Moderna’s mRNA vaccine candidate for COVID-19 would be so exciting if it works.
That’s not quite right. I can’t get to that book right now, but measles and mumps for MMR are also done in Chicken eggs, IIRC, as are Herpes and Poxviruses, while cell lines and other media can be used to grow other viruses—but the remainder of the facilities are still similar, and can be repurposed.
But I agree that we do need new platform technologies.
Hmm, well that book chapter claims measles and mumps vaccines are produced in chick embryo cell culture, which is different from propagation on chicken eggs. My quick Googling revealed that we don’t have a licensed herpes vaccine, and that while there might be one or two smallpox vaccines that are produced in chicken eggs, many are done in cell culture.
You might be right about the broader (and more important) point about ease of facilities repurposing, however—I don’t know enough to say, although the table in the book chapter makes me doubtful, given that pretty much all steps in the manufacturing process (production, isolation, purification, formulation) seem unique to each vaccine.
yellow fever vaccine is one that springs to mind that also uses eggs in production
Good to know, thanks!
I’ve heard that the eggs used are special, more sterile than usual (you don’t want the chicken to have other diseases now, do you?), and usually require ordering at least a year in advance.
(Came up when I was researching flu-vaccine development.)
Some other vaccine production methods involve cell cultures, but the output of different cell cultures is pretty wildly variable and the preferred cell culture is different depending on the specific virus. This is probably a more expensive means of production. You may be able to scale it up faster and with less early prep-work, however.
Fair warning: While there have been coronavirus vaccines that have just worked, there have also been a lot of them that seemed to make the course of infection worse, probably due to antibody-dependent enhancement or a similar phenomenon. The set that were somewhat challenging to develop vaccines for seemed to include SARS-1. The lengthy process of animal testing would probably spot this, but it may make getting a reliable vaccine slower and harder than it would be with viruses that don’t have this problem.
Why do you need the eggs in the first place? Couldn’t you just feed animo acids that you get when you electrolyse proteins instead of having the proteins from the eggs?
...I’m confused about what method you’re even trying to gesture at.
They’re viruses*, they need a full set of environmentally-provided cell machinery to replicate or produce proteins: ribosomes, transcription machinery (ex: t-RNAs), ATP, the works. They need cells, so you’d need need at least a cell culture. All of biology has heavily optimized protein assembly lines, you’re not going to beat it acellularly.
The cells near the outside of an egg are probably used because they’re an elegant and self-contained little solution to sterilization (against everything but your virus) and the quality-control problems you’d have to contend with otherwise. It’s not really about the protein content, mostly.
(Cell culture is probably more expensive than eggs because 1) bioreactors are kinda expensive, 2) bioreactors are a bit of a pain to maintain, and sterilization is hard, two problems that using an egg pretty neatly solves, and 3) which cell culture will work best is surprisingly hard to predict, you basically have to test it experimentally.)
* Well, technically it’s weakened viruses, or single-gene plasmids, or something similar. The need for cells still holds either way.
Eggs do have a lot of ovalbumin where it’s not really desireable for that to end up in your final vaccine but I don’t think this is a discussion to have at a point where our key issue is scaling up vaccine production.
If you have to order the steralized eggs a year in advance, and we want our COVID-19 vaccine before a year is over, that suggests to me that we also have other problems.
If I understand the work Moderna is doing for their COVID-19 vaccine and read the paper where they describe their framework, it seems to me that they use human cell lines:
Just like Moderna, CureVac which is another of the companies that want to produce a COVID-19 vaccine also focuses on delievering mRNA and not viruses. I didn’t immediately find information about how CureVac gets their mRNA but it wouldn’t surprise me if they also don’t use eggs.
Whoah, lipid-coated mRNA vaccines, not as an intermediate step but as the actual delivery method? That’s actually new to me! Sounds like it’s mRNAs coding for some subset of the viral proteins, which probably get assembled into proteins in your cells and then get used as something for antibodies to respond against. mRNAs should then just degrade themselves with time.
I have no idea what the most efficient method for producing those is; I am very used to vaccines being protein-based. This probably is in the realm where it’s simple enough that modifying PCR-protocols to produce RNA instead might actually work reasonably well, although RNA is generally more fragile and error-prone and that could be a problem.
You’d be using nucleotides, not amino acids, but mRNA from DNA is a short-enough assembly line that you might not need cells to do it.
(Protein production has a lot of dependencies. mRNA transcription should basically just require your DNA of interest, nucleotides (x4), and a transcriptase protein. Maybe add a transcription factor or two.)
HeLa definitely is a human cell line (although that was for Ebola, they may end up using a different cell line). That’s good, that probably scales up easily.
From last year: From CEPI awards US$ 34M contract to CureVac to advance The RNA Printer™
It seems that the third mRNA vaccine company is BioNTech.
It seems that Johnson & Johnson is still developing a vaccine the traditional way:
There’s a forth company with Inovio Pharm that also develops a COVID-19 vaccine. It’s technology is based on delievering DNA based.
I have the impression that the mRNA/DNA ways of vaccine delievery allow for faster development of a vaccine then the old fashioned protein based way.
At scale? Not easily—eggs are cheaper, more effective, and easier to deal with.
The peer-reviewed literature has several papers talking about GI symptoms of COVID19, and there are several GI cells that are ACE2+ that are plausible targets. What I am wondering is the following a potential vaccine strategy?
innoculate with live strain in GI tract to avoid respiratory infection
Interesting—I’d ask Robin Hanson if that fits with his variolation suggestion.
Related to that observation I have wondered, but never posted/asked, if how one gets infected might influence severity as well. If I touch a contaminated surface and then rub my eye or then eat a sandwich without washing my hand is that more likely to end up somewhere other than my upper and lower respiratory systems?
If the same type of facility works for almost every kind of vaccine, do we think there would be interest in constructing the facilities as a speculative venture? Consider:
1. The economy is in chaos and may remain so, which I expect to produce unusually affordable access to design firms, construction crews, raw materials, and land.
2. There will be a strong incentive for regulators/inspectors to move with best speed, and the current administration at least in the US has a track record of being friendly to shortcuts.
3. If the facilities are already built, this allows a limit to the risk the companies producing the vaccines need to absorb in order to increase supply.
4. We could squeeze out unscrupulous opportunists.
Contact Tracing at Scale!
One thing we need, that the Less Wrong community could likely help with, is contact tracing capability at scale. I know of one such project in the US—https://www.covid-watch.org/ The Covid Watch project, based out of Stanford.
I think the major tech companies need to set up and throw a ton of engineering and design resources at contact tracing efforts. They currently control the software supply chain to most mobile devices on earth, and thus are ideally placed to help track the spread of infections.
The more testing we have, the more effective contact tracing will be, so this needs to be paired with an increase in testing world-wide, as previously mentioned in the thread.
https://www.lesswrong.com/posts/fxfsc4SWKfpnDHY97/landfish-lab?commentId=4ftZGNxtNRiwgXTbf
My collection of links to the projects I know about in this space and some news coverage of them.
There might be benefits to having an privacy sensitive open-source solution like the one proposed in Covid-Watch over a Google/Facebook solution.
I would strongly encourage people to try brainstorming some questions. Even if you don’t come up with anything directly useful you might jog someone else’s creativity. Remember to go for quantity over quality on your first pass.
Epidemiology questions that, while we probably can’t do much about, would be useful to try to ad hoc model given how bad official info has been so far:
Are estimates of doubling time off from bad modeling of rapid test ramping making it seem faster than it is?
What is actual hospital elasticity? Is there an existing gathering of data on this from previous disasters?
How long do human trials need to be before they are rolled out to the majority of the population? Just to the extremely vulnerable? What is the gears level model here?
What granularity of travel restriction makes the most sense? In general, how can cities and counties act knowing that federal response may (will continue to be) be too slow?
Which physical objects have longer supply chains and thus can be expected to be less robust to disruption?
What mental health problems can we expect to spike hard in the next 1-6 months given people feeling shut in and helpless?
What are the most predictable second order disasters?
Does moral hazard show up anywhere here?
What’s most likely to be ignored during this? Civil liberties? Already seen discussion of that. What’s even more ignored?
I’ve seen people from a Stanford lab asking on facebook about being put in touch with someone from an MIT lab. How can lab cross talk increase?
If UV 210nm turns out to be effective, how can you build your own flashlight/lightsaber (from the virus’ perspective) out of off the shelf parts?
Which continuing failures of the FDA are highly predictable? What can be done to mitigate that expectation at the hospital and lab level?
How can models take into account reference classes. e.g. Many models are averaging naively which means essentially all the data points are from the least controlled regions with the widest error bars.
Thanks Romeo.
Setting a 5 minute timer:
How will this effect markets / supply chains etc, assuming it lasts for different lengths of time.
How likely are various containment interventions by governments?
How does de-escalating quarantine / lock-down in countries that have instituted those work? Is there a chance the the virus will bounce back after de-escalation?
How long do I have to wait before interacting with an object / location in order to make sure it is safe? (eg could I rent an airbnb, or a rental car, several days in advance, and then use it without risk of catching the virus?)
How bad is this really for people in my age group? I would love to have information from people I know, who catch it.
Pushing to get to a total of 10 items:
How much health risk is there to social isolation?
Do I still need to have extreme hand-washing / disinfectant procedures if I’m not leaving the house?
What could be done to help emergency workers and other parts of the medical system not get sick?
Are there resources on maintaining a balanced diet of non-perishable foods?
What skills should I be rapidly acquiring to be most useful to this whole situation?
And one more.
Am I better off if I drive to some cabin out in the boonies?
The requirements on multiplying ventilator use through sharing is
1. Equal tube lengths.
2. Equal lung capacity.
3. Equal lung resistance.
4. Same patient weight (approx)
the question is can any of these requirements be broken though clever use of 3d printed valves or other JIT solutions?
No answer here but a subquestion might be what are the essentials for an effective “hospital bed” for a COVID-19 patient? What are the binding/constraining elements? We know ventilators for critical cases are one. Others? What about those for serious versus critical—if we can treat serious cases well but in some makeshift hospital room (say an empty hotel) does that help us limit the demand for ICU space?
Second thought here. You have investors like Ackman suggesting a slow bleed process may well kill hotel owners. Is there an opportunity to address two things as once? If government (and insurance companies) can support quarantining and treating less serious cases in hotels then the industry gets some relief and society perhaps gets both better allocation of medical resources and improved quarantines.
Scaling up testing seems to be critical. With easy, fast and ubiquitous testing, huge numbers of individuals could be tested as a matter of routine, and infected people could begin self-isolating before showing symptoms. With truly adequate testing policies, the goal of true “containment” could potentially be achieved, without the need to resort to complete economic lockdown, which causes its own devastating consequences in the long term.
Cheap, fast, free testing, possibly with an incentive to get tested regularly even if you don’t feel sick, could move us beyond flattening the curve and into actual containment.
Even a test with relatively poor accuracy helps, in terms of flattening the curve, provided it is widely distributed.
So I might phrase this as a set of questions:
Should I get tested, if testing is available?
How do we best institute wide-scale testing?
How do we most quickly enact wide-scale testing?
Relevant thread: https://www.lesswrong.com/posts/pjLgE2efAozz82JmR/sars-cov-2-pool-testing-algorithm-puzzle
I’d love to work on this if someone can put me in contact with a medical professional who understands how these tests work.
Whenever you ask people to create a contact it would make sense to be explicit about why the contact would be valuable, and what good will come out of it.
I want to develop a web app that will make group testing fast and easy. This problem happens to relate closely to my machine learning research interests, and I have an algorithm in mind that I’m excited about. However, the first step to developing software is always to talk to potential users and understand their needs in order to make sure your software will actually solve them. You can share my linkedin profile if you think that will help.
Why doesn’t Japan have a huge outbreak already? (924 reported cases today, according to the Johns Hopkins tracker): https://www.bloomberg.com/news/articles/2020-03-19/a-coronavirus-explosion-was-expected-in-japan-where-is-it
Why does India have so few cases? (160 reported cases today): https://www.weforum.org/agenda/2020/03/quarantine-india-covid-19-coronavirus/
For each country – what proportion of newly reported cases comes from ramping up testing, and what proportion comes from newly infected people?
Will the economic impact of coronavirus be inflationary or deflationary on net? (for USD)
It would be great to have a list with the current teams that are working on a COVID-19 vaccine. Is such a list out there or otherwise, does someone want to create one?
See here: https://www.statnews.com/2020/03/19/an-updated-guide-to-the-coronavirus-drugs-and-vaccines-in-development/
Why haven’t we ever created a vaccine for a coronavirus before?
Is coronavirus vaccine development more limited by need for technological innovation or economic incentive?
There was a twitter thread I didn’t save that said:
1. we have vaccines for cat and dog CVs
2. Human CVs are unrewarding to vaccinate against because they only cause 30% of colds, so you can only advertise a reduction, not total prevention, of colds.
Same virus family. Different pathogenesis and shouldn’t be directly compared but more for information:
There are vaccines for coronavirus for dogs and cats. They are not commonly used for multiple reasons.
The main species that get vaccinated are cattle.
Bovine coronavirus (BCoV) is an important livestock pathogen with a high prevalence worldwide. The virus causes respiratory disease and diarrhea in calves and winter dysentery in adult cattle.
Bovine coronavirus disease info.
Vaccine methods:
pregnant cows (to create antibodies to pass immunity to calves via colostrum) info on a product available. (multi-virus vaccine)
intranasal (IN) vaccination of calves with a modified live BCoV
(It’s been years since I’ve worked with cattle but don’t think the situation has changed)
Does hydroxychloroquine + azithromycin effectively treat COVID-19?
See Gautret et al. 2020, a small trial of this (not randomized) that found a big effect.
I looked into this a bit with a friend who’s an MD, and it turns out that this paper isn’t very good.
Study not randomized, groups not balanced by disease severity, several treatment-group patients excluded from the data after trial started because they got worse (some went to ICU; one died).
From p. 10 of the paper:
That paper is indeed a piece of crap.
This being said, there is other preliminary data from Asia that chloroquine and hydroxychloroquine could hasten recovery, and there were multiple biochemical reasons to suspect it could help which are the reasons it was being used in the first place. I would call the French studies nearly useless to determine actual efficacy, but I am still fairly optimistic they will have at least some positive effect.
To Address the Problem: “How do I convince others to act?”
By now it seems clear that social distancing and shelter-in-place protocols are the most effective for reducing the spread of infection. I don’t know about other regions, but compliance in the US is unfortunately low. If increasing compliance is desirable, even when balanced against economic concerns, how do we encourage it?
Part of the problem is that people have to seek out information to become informed. Time and energy have to be invested for a person to figure out how important it is to stay home, and what sources of information are reliable.
Proposed Solution: Hospitals and medical groups should write letters to their entire mailing list pleading with people to stay home if possible. A message from your doctor’s office is far more persuasive than a general government announcement or news report. It’s local, personal, and credible. Everyone opens an email from their doctor.
Medical providers can explain the staff and resource shortages they face. They can explain that if everyone stays off the road as much as possible, this reduces accidents and frees up first-responders and scarce emergency room capacity (how significant would this be?). They can encourage a moratorium on other risky activities like extreme sports, even though those don’t violate social distancing rules (how significant would this be?).
This proposal is virtually costless, near effortless, can be implemented immediately, and would hopefully be effective.
Is it worthwhile to focus on getting medical providers to do this? If so, how do we reach out to them and maximize the number who do it ASAP?
What sources are governments using for decision-making?
The biggest impacts seem to me to be via influencing government. The UK government, for instance, is still very reticent to enforce widespread testing or mandatory quarantine. Their ‘quarantine guidance’ for households with symptoms looks like this, which seems patently foolish for a number of reasons.
Influencing governments’ decision making is high-impact and potentially tractable via getting modelling and trial data to them. The UK Government publish their ‘scientific basis for decision making’ but it appears to be weeks out of date and unreferenced.
With that in mind, how do we get better decision-making information into government? What theory of change can we find for influencing policy makers? I believe this should be primarily targeted towards larger organisations and researchers who can have more direct influence, but may be useful for individuals as well.
Sir Patrick Vallance seems to be the key figure behind the UK policy. The guy was a professor of medicine in the past and who heads the Government Office for Science. Their policy is likely much more driven by modeling then the policy of other countries where the policies are decided by politicians instead of people with that kind of credentials.
To the extend that they have data on that page that’s weeks out of date it’s likely because the page has little to do with their actual decision making processes.
Vallance might still be wrong, but I think it’s wrong to model him as being simply misinformed.