My Model of the New COVID Strain and US Response
Cross-posted from Putanumonit, I will likely keep the post more up-to-date there than on here as my estimates change. This was also intended for a non-LW audience, since smarter people than me have provided their estimates here and you shouldn’t overweight me because I have my own blog. I am including some new info and analysis that wasn’t in Zvi’s post, so update accordingly.
There is evidence of a new highly-infectious strain of COVID-19 emerging in the UK and other countries. This post is an elaboration of my current model of it, as promised in my review of LessWrong’s collection on Epistemology. Most of the information I have on this new COVID variant comes from this post by Zvi, the links from it, and comments. Read those if you haven’t yet.
Though I usually strive to have timeless content on my blog, I’m making another COVID exception. I’ve heard from dozens of people who were able to prepare for COVID’s first outbreak as a result of reading Seeing The Smoke, and this post may end up even timelier than the last. I anticipate both my model and the evidence to change rapidly, everything below is a snapshot of what I believe as of Christmas Day 2020. Reminder: I am not an expert on this even compared to other non-experts in the LessWrong comments.
The median prediction on Zvi’s LW post is around 60% that the new variant is at least 50% more transmissible and same for there being a third COVID wave in the US:
I would add a few things without straying too far from the crowd’s wisdom.
First, I would revise the crowd’s estimate for the first question down for no other reason than that the poll came at the end of a post titled We’re Fucked, It’s Over. If ever there was going to be a framing effect pushing estimates upwards, this would be it. There are also more alternative explanations for the information coming out of the UK than Zvi accounted for.
Instead of just saying that that there’s a 60% chance it has a 50% higher transmissibility, I would break it down something like this:
35% that the new strain is a nothingburger.
30% that it’s slightly more transmissible, say with an effective reproduction rate (rt) 20% higher.
30% that it’s significantly more transmissible, e.g. 70% higher.
5% in the Captain Malcolm Reynolds probability bucket.
I also believe it very likely that new variant is already present in the US, given that a dozen flights from London have landed in NYC just in the last week. Travel from the UK has been restricted on Christmas eve but the dust from the bolting horse has already settled by the barn door.
A variant that’s just 20% more infectious will still take over exponentially from the old variant resulting in a chart like the one we see.
Several people have tried to explain away the growth of the new variant as being due to things like catching a superspreader, but it doesn’t have to be one or the other. It could have rt=1.2 and catch a superspreader event to make it look like rt=1.5-1.7. This is why I am giving extra weight to the hypothesis of slightly higher transmissibility.
I also think there’s a huge difference between rtnew=1.2 and rtnew=1.7 (here I mean rt relative to the old strain given the same measures of containment) in terms of outcomes.
In the first case, to keep the virus suppressed (i.e. r<1) we need to take measures that would have yielded rt=0.8 for the old strain. New York sustained that number (albeit never dipping below 0.7) for two months running in the spring. Given that ~30% of the state has already been infected, that makes this all the more doable. Gyms and restaurants will close and people will grumble, but we’ll probably survive till the vaccine.
rtnew=1.7 is an entirely different case. Suppressing it would require the sort of lockdown that would yield rt=0.6 for the old strain, a number that has never been reached by any US state for any amount of time. I see no way in hell that Americans would agree to a lockdown much stricter than any we’ve had so far, especially after they’ve been promised that the worst is behind them.
Rural red tribers will not agree because Biden is in the White House. Urban blue tribers because containment in cities is much harder anyway. Young people will not agree because the virus harms them less than the lockdown. Dumb people will not agree because they will not understand the science and math and smart people, seeing that, will try to get infected early before the hospitals are overwhelmed.
No one will believe any of the “experts” at the CDC or WHO or Dr. Fauci because they have all repeatedly lied and tried to manipulate people and reversed their positions without admitting that they have. They all have negative credibility at this point with many Americans.
If the new virus is 1.7 times as infectious I believe that Americans will most likely simply give up on containment, the populace if not the government. We’ll have time to vaccinate the most vulnerable 10-20% of the population by April-May, at which point the majority of everyone else will get the new strain.
There is a small chance that the US will get its shit together vaccine-wise and beat the new strain in a race. Israel, for example, has a simple system in place to vaccinate the vast majority of its 9-million strong population by the end of March. On this timetable, even a virus with rt=1.7 will not have time to explode before herd immunity kills it.
The same is not true of the US, which is currently projected to reach majority immunization no earlier than the late summer by refusing to do basic things to rush the process.
And what’s the Captain Mal bucket?
The new variant being more virulent in addition to being more infectious. Another newer strain that’s even worse. Or, in a nightmare scenario, a highly-transmissible mutation that also renders the Pfizer and Moderna vaccines ineffective.
If COVID has taught us one thing, it’s that it’s all usually worse than we thought.
So the bottom line is that I currently estimate a 30-35% chance of us being truly fucked, in the sense of dozens of millions more Americans getting COVID in 2021 and hundreds of thousands more dead, along with local attempts at extreme lockdowns. I’m not going to speculate on the further implications of this for things like financial markets or the fabric of our civilization. I also anticipate that I’ll be revising these numbers in the next few days.
Merry Christmas.
Something is odd about how people are analyzing vaccine effects.
My guess is that the US is doing a tolerable enough job of vaccinating the vulnerable first that the IFR will start dropping by around 50% per month, for new infections, starting in a week or so (until we get to limits due to people refusing to be vaccinated?). So even if we get another wave with infection rates 2 or 3 times the recent levels, it seems unlikely to keep the death and hospitalization rates from declining.
It will also become gradually easier to keep r low due to the increasing number of people who are immune.
I’m surprised you put this so high. Would you accept my $30-150 to your $10-50 that this doesn’t turn out to be a nothingburger?
With most of this I agree. Two remarks:
“In the first case, to keep the virus suppressed (i.e. r<1) we need to take measures that would have yielded rt=0.8 for the old strain. New York sustained that number (albeit never dipping below 0.7) for two months running in the spring.”
The important word here is “spring”. New York went below 0.8 at the end of March. Whether something like this could be achieved in winter is a different question (whether that distinction matters depends on how fast you assume the new variant to spread). Furthermore it depends on the prevalence of the virus in the adjacent regions which could be different than in spring 2020.
″...at which point the majority of everyone else will get the new strain.”
Í agree that in that case the majority of everyone not vaccinated will get the strain. But nevertheless a minority will not get it, and I think this minority could be huge in absolute numbers. I think almost everyone would get the new strain who
has to use public tranport, or
works in a place with constant face-to-face customer or coworker contact, or
has children in school or daycare.
I think the majority of those will not get the new strain who
work from home, and
have reduced indoor contacts outside the household to almost zero .
Because for those groups the risk is by magnitudes lower than the risk for the groups mentioned before. This is important to emphasize because otherwise—depending on the news about the new virus strain the next couple of weeks—it could lead to resignation and thereby to a self-fullfilling prophecy even for those who could still protect themselves against the new strain (assuming, of course, that the vaccine works against the new virus strain, too).
Thanks for explaining your assessment of the situation.
Are there polls supporting this view? (Negative credibility would mean that people assume the opposite of what is said by these people is true, right?)
Update. Doing a quick search led to this:
“The U.S. public’s overall trust in Fauci, the National Institutes of Health’s top infectious disease doctor, has declined 10% since April. Republicans have particularly soured on him: His favorables dropped nearly 30% among Republicans since April. Democrats’ confidence in Fauci, meanwhile, has increased from 80% to 86% since April.” (Statnews, Sep 10)
“79% of Democrats said Fauci has done a good or excellent job handling the pandemic, compared with 56% of independents and 54% of Republicans.” “Voters have consistently rated the WHO, the CDC and their state governors above lawmakers and the president.” (Oct 14, Morning Consult)
I would be very interested in knowing whether you use more up-to-date polls for your statement.
This post (and the discussion in its comments) were interesting reads, thanks.
I noticed a small typo you might want to correct. The “will not agree” is missing from
Is this valid reasoning? Intuitively I’d expect current measures to be more effective for a more infectious strain, so that it would require a lockdown that would yield something closer to 0.8 for the old strain.
I suspect that some measures will remain close to 100% effective, like not seeing family members and friends. Similarly some things physically can’t become more infective, like being infected by your partner.
In that case I would expect that what currently are very high risk interactions (e.g. partner) and very low risk interactions (e.g. masked, 6ft+ grocery shopping) will still be almost as risky for a more infectious strain.
However if the bulk of infections happen because of moderate risk interactions, then moderate risk interactions becoming high risk interactions would significantly increase overal cases.
I had a long discussion on this very topic, and wanted to share my thoughts somewhere. So why not here.
Disclaimer: I am not an expert on any of this.
The scaling assumption (if the new strain has an R of 1.7 when the old one has an R of 1, then we need countermeasures pulling the old one down to 0.6 to get the new one to 0.6 * 1.7 = 1) is almost certainly too pessimistic an estimate, but I have no clue by how much. A lot of high risk events (going to a concert, partying with 10+ people in a closed room for an entire night, having a multiple hour Christmas dinner with the entire family) will become less than linearly more risky. I interpreted the “70%” (after some initial confusion) to represent an increase in risk per event or unit time of exposure. But if you are sharing the same air with possibly contagious people for a long period of time your risk is all the way on the saturated end of the geometric distribution, and it simply can’t go above 100%. So high risk events will likely stay high risk events.
At the same time, I expect a lot of medium and low risk events to become almost proportionally more risky. This includes events like having one or two people over for dinner while keeping the room properly ventilated, going to supermarkets, going to the office and using public transport. Something that has been bugging me is that the increase in R-value has been deduced from the actual increased rate at which it spreads, so it is simply not possible that every activity has less than 70% (or whatever number you believe in) increased risk, since that is apparently the population average under the UK lockdown level 2 conditions. So some of this nonlinearity has already been factored in, making it very difficult to say what stronger lockdowns would mean.
In conclusion, I think it is possible that even if the new variant is 70% more transmissible that lockdown conditions that would have pushed the old strain down to 0.7 or only 0.8 might be sufficient to contain this new strain, and of course if the new strain is less transmissible than this we have even more leeway. At the same time I have absolutely no clue how to get a reliable estimate of the “old R needed”.
I don’t think it was that easy to get to the saturated end with the old strain. As I remember, the chance of catching COVID from a sick person in your household was only around 20-30%, and at superspreader events it was still just a small minority of total attendees that were infected.
I also thought this, but was told this was not the case (without sources though). If you are right then the scaling assumption is probably close to accurate. I tried briefly looking for more information on this but found it too complicated to judge (for example, papers summarizing contact tracing results in order to determine the relative importance of superspreader events are too complicated for me to undo their selection effects—in particular the ones I saw limited to confirmed cases, or sometimes even confirmed cases with known source).
EDIT: if I check microCOVID for example, they state that the chance of catching it during a 1 hour dinner with another person who has been confirmed to have COVID is probably between 0.2% and 20%, The relevant event risks for group spread (as opposed to personal risk evaluations) are conditional on at least one person present having COVID. So is this interval a small chance or a large chance? I wouldn’t be surprised if ~10% is significantly high that the linearity assumption becomes questionable, and a 1 hour dinner is far from the most risky event people are participating in.
Suppose, as you say, some of this nonlinearity is already factored into the 70% estimate, that would imply that the ‘real’ number is even higher. For some interaction, like having a face to face conversation without any protection, the probability of an infection may have increased by 100% or even more.
I’m also not an expert. Intuitively this seems like a big step with just a handful of mutations.
I agree that this means particular interactions would have a larger risk increase than the 70% cited (again, or whatever average you believe in).
In the 24-minute video in Zvi’s weekly summary Vincent Racaniello makes the same point (along with many other good points), with the important additional fact that he is an expert (as far as I can tell?). The problem is that this leaves us in the market for an alternative explanation of the UK data, both their absolute increase in cases as well as the relative growth of this particular variant as a fraction of all sequenced COVID samples. There are multiple possible but unlikely explanations, such as superspreaders, ‘mild’ superspreaders along with a ‘mild’ increase in infectiousness, or even downright inflated numbers due to mistakes or political motives. To me all of these sound implausible, but if the biological prior on a mutation causing such extreme differences is sufficiently low they might still be likely a postiori explanations.
I commented something similar on Zvi’s summary, but I don’t know how to link to comments on posts. It has a few more links motivating the above.
As mentioned on Twitter, I don’t buy this. I think we’d get more infections and deaths, but once hospitals are overwhelmed, society’s negative feedback loop will kick in and we’ll get R back close to 1.
I believe that lots of individuals could be a lot more cautious than they already are, and I don’t think people will stand for hospitals being overwhelmed.