Covid 8/19: Cracking the Booster

post by Zvi · 2021-08-19T14:20:01.080Z · LW · GW · 23 comments

Contents

  The Numbers
  Predictions
  Deaths
  Cases
  Vaccinations 
  Vaccine Mandates
  Mask and Testing Mandates and Other NPIs
  Delta Variant 
  Booster Shot
  Think of the Children
  In Other News
  Not Covid
None
23 comments

The news about Covid-19 is now essentially on a few distinct tracks: Vaccine effectiveness and booster shots, vaccination mandates, mandates for NPIs and other Covid-19 related crippling of the living of life, and the actual path of the pandemic. One could argue for splitting those further, or for combining the mandates.

Vaccine effectiveness against the spread of Covid has come into question, with many claiming that immunity fades dramatically with time or that the vaccines were never that effective against Delta. I’ve looked into the claims and go into detail. There’s little question that the vaccinated can spread Covid, which is in contrast to previous attempts to sell the line that they couldn’t at all, and they spread Delta somewhat more than they spread Alpha, the numbers here are disappointing relative to my expectations, but the big claims of hugely waning immunity are almost certainly greatly exaggerated. 

Vaccines remain highly effective at stopping the spread of Covid-19, and of stopping symptomatic disease, and especially in preventing death. Still, boosters are more effective than not having boosters, and I think the cost-benefit for most people favors getting a third shot if one is made available to you. Boosters are coming soon, eight months after your second dose. However, until it is highly encouraged and likely eventually mandatory, it is still for the moment mostly forbidden. Get your shot exactly when we tell you to, they insist, not a minute before.  

There are continuing debates over mask mandates at various meta levels, which led to some sentences that were fun (and tricky) to write. There’s increasing mainstream and in-group pressure to force people to wear masks regardless of whether it does anything useful in context, and to shame those who disagree and blame them along with the unvaccinated for the entire pandemic.

Oh, and also the pandemic is still growing but at a decreasing rate, and case counts are a favorite to mostly stabilize within a few weeks, which is great news. The death rate lags, so it continues to climb rapidly for now.

Let’s run the numbers.

The Numbers

Predictions

Nate Silver, the media’s official master of predictions, officially wouldn’t want to predict this thing. So even though he kind of does, officially he doesn’t. He observes the CDC models are split. Some models think this week is the peak, other models disagree and think things will continue getting worse for some time, but it’s always either one or the other. Odd, but perhaps makes some sense given the sharp peaks elsewhere.

Early this week the NIH director said there were ‘no signs of having peaked out’ which was technically true, but misleading, as I’d noted in last week’s update that there were definitely signs we might be about to peak. Standard stuff, standard line.

I sympathize with Nate, but when it gets hard is exactly when we need people like Nate to step up and take a stand the most, even if they’ll often be wrong, so I’ll give the standard Teddy Rosevelt quote and ask how I did this week.

Prediction from last week: 900k cases (+21%) and 5,028 deaths (+35%).

What’s weird about that prediction is that the percentage here is wrong – 5,028 deaths is a (+48%) increase. I have edited the last post to point this out but note the error. Looking at the context, it’s clear my intent was to predict a +35% rise and I had a spreadsheet error that gave the wrong number, so I shouldn’t get credit for the better ‘actual’ prediction here. Intent should win. The counterargument is that if you do the California correction I’ve now done, this methodology gives a prediction of 5,081 deaths, which is similar. 

Result: 872k cases (+17%) and 5,545 deaths (+48% after correcting for California last week, +63% without that correction).

For cases, we got a good result, where the slowdown was real. For deaths, we got a quite bad result, where the slowdown was a mirage and we are back on the previous trajectory, and likely overshot a bit. Both are common, so predictions have to split the difference, but on reflection the slowdown in deaths couldn’t continue, and I should have predicted at least +50% there. I also should have directly caught the California correction, which I didn’t, and that meant the count a week ago was low by at least 350 deaths there, which accounts for about half the surprise here. I’ve corrected it for the charts and numbers going forward. 

For next week, deaths should continue to rise with lagged cases, so I’m going with +45%. For cases, the slowdown in growth is presumably the new normal, and things should on average continue to slowly improve as more people get infected and vaccinated, and behaviors continue to adjust. The only danger would be if immunity is fading enough to counteract that, which I find unlikely. 

Prediction for next week: 1,000,000 cases (+14%) and 8,040 deaths (+45%).

Deaths

DateWESTMIDWESTSOUTHNORTHEASTTOTAL
Jun 24-Jun 305504597061861901
Jul 1-Jul 74593296121281528
Jul 8-Jul 145323986891451764
Jul 15-Jul 214343417321701677
Jul 22-Jul 2849138510091572042
Jul 29-Aug 469347714153042889
Aug 5-Aug 1170562921812343749
Aug 12-Aug 1891285133943885545

Steady rise across the board. Deaths lag cases by several weeks, so in a few weeks the growth rate should slow a lot, but stabilizing fully will take longer than that. 

Cases

DateWESTMIDWESTSOUTHNORTHEASTTOTAL
Jun 24-Jun 3023,24614,52131,7736,38875,928
Jul 1-Jul 727,41317,46040,0317,06591,969
Jul 8-Jul 1445,33827,54468,12911,368152,379
Jul 15-Jul 2165,91339,634116,93319,076241,556
Jul 22-Jul 2894,42960,502205,99231,073391,996
Jul 29-Aug 4131,19786,394323,06348,773589,427
Aug 5-Aug 11157,553110,978409,18466,686744,401
Aug 12-Aug 18183,667130,394479,21478,907872,182

The slowdown in growth is across the board, with no sign that the South is going to peak first and then we’ll get an explosion in the Northeast. If that happens, which it might, it will probably happen in the winter. 

The test positivity rate has mostly leveled off as well, with a clear phase shift:

Things aren’t quite stable yet let alone going in the right direction, but it seems more likely that the peak (at least for now) is relatively close and not too far above current levels. 

Vaccinations 

I am assuming, but am not fully confident, that third shots are not included in the dose counts above. The percentages should hold either way, and continue to show slow but steady progress. 

Here’s a regional map.

New study finds that vaccines have no effect on pregnancy (study). Mostly third trimester vaccinations since that’s mostly that’s available to study. Most pregnant women are declining the vaccine and there’s lots of FUD going around about how it could be unsafe or bad for fertility. It contains no live virus, and is doubtless being watched very carefully. The short term side effects include fever, which is a reasonable thing to have a nonzero amount of worry about, and a reason to be thankful to have a study. Somehow I doubt many minds will be changed regardless, but it’s good to at least be able to say Studies Show. 

Vaccine Mandates

You know how you convince a lot less than no people to get vaccinated or support your mandates

I’ve spent much of the last month arguing in favor of mandates and vaccinations, and I notice that this statement pissed me off quite a bit. It’s almost engineered to piss people off, to the extent it doesn’t feel like an accident, as if Fauci’s goal was to signal bad intent and further inflame tensions. 

There’s also this.

As in, you want to tell people like James that they’re being crazy, or at least have the mechanism wrong.

Rather than, you know, reinforce their narrative. 

Also likely would help if we didn’t use a giant picture of a needle on every news story.

FiveThirtyEight has a breakdown of unvaccinated America, which tells the usual story that there’s a constant pool of ‘I will never ever ever get the vaccine’ attitudes, while a steady portion of everyone else gets vaccinated. If that’s the case, then keeping our vaccination rate essentially constant is quite the accomplishment at this point, as it’s drawing from a quickly narrowing group of people. 

The most important chart there, to me, is this one:

This is not a story about lack of access or time off of work. It’s a story of people who are scared of the vaccine, mostly for factually wrong reasons, and aren’t scared of Covid.  

FiveThirtyEight attempts here to answer some of the justifications for vaccine hesitancy. It’s not perfect or complete, nor would I agree with every argument, but at least it seems likely to help rather than do further damage. A major theme of the objections addressed is the mindset that doesn’t differentiate 1% from 99% – the idea that if you could get infected after the vaccine, and you also might not get it without the vaccine, then the vaccine does nothing. Such people presumably understand this in other contexts (for example, you might crash if you drive sober rather than drinking first) but it really is a true objection here, because they want to buy a mindset rather than buy actual safety, and refusing to sell that mindset means no sale.  

Amy Coney Barrett refuses to block Indiana University’s vaccine mandate. It wasn’t a full mandate, as testing was an option, and had religious and medical exemptions. I continue to have more faith than most otherwise similarly thinking people that the Supreme Court will do mostly reasonable things and mostly enforce the rules.

San Francisco joins New York City in requiring vaccination for indoor dining.

Here’s a complete list of what requires vaccination in New York City:

You can see the room now, where someone was tasked with coming up with as many different names as possible for places requiring vaccination. What do people even do inside, anyway? We need ten different variations of indoor dining by the 11am press conference. The two sides are both trying to make the mandates look as obnoxious as possible, for different reasons.

What if something goes wrong? I see a lot of people asking questions like this:

For New York in particular, you can use the Excelsior pass as a substitute for your vaccine card. If you’re medically exempt from the vaccine and can’t take it, I think you’re simply out of luck.

What about the general question of whether you can get your vaccine card replaced if you need to? In Wisconsin the answer is clearly yes, and they say to contact the health department in the state you were vaccinated in if it didn’t happen locally. A patchwork solution isn’t ideal but I’d be very surprised if there wasn’t a way to get one in other places. Also CNN recommends scanning your card with your phone no matter what else is going on, along with some links to additional options, and that seems very wise.

New York and Washington DC require health care workers to be vaccinated (WaPo), joining California and Washington state. 

A Virginia hospital mandating vaccinations faces a nurses strike (WaPo).

Brooklyn Nets start requiring proof of vaccination at their games, as per NY rules.

The Pac-12 football conference reverted to the old rule that a school unable to field a team must forfeit the game. 

The electric company gets in on the act.

Texas nursing homes, on the other hand, do not get in on the act.

Nor do any localities in Texas, since the Texas Supreme Court has (for now) ruled that the executive mandate against mandates is mandatory, and therefore mask mandates are forbidden, although Biden is trying to use Federal coercion to change that.

He’s also going to lay down the law on nursing homes, requiring vaccine mandates if they want to keep getting Medicare and Medicaid funding. I hate the tactic, but if there’s one place you really, really want a vaccine mandate, it would have to be nursing homes. 

Incentives matter, small amounts of money can be highly motivating (such as this study from the flu vaccine), so be careful how you set them.

Mask and Testing Mandates and Other NPIs

Andrew reminds us of the key question.

There is a golden middle where population-level NPIs (non-pharmaceutical interventions) are great, which is where you can stop Covid if and only if you use the NPIs for a limited amount of time. 

If you would have stopped Covid anyway then obviously you didn’t need the NPIs.

If you can’t stop Covid that way, or you can only do so until you relax the controls, then all you’ve done is buy yourself some time that didn’t do you any good. The time needs to change things for the better, such as by getting people vaccinated. Yet Australia seems to be in no hurry to vaccinate, and places like America have already vaccinated most of the people they are going to vaccinate.

The other case is the ‘flatten the curve’ argument, where the time you purchase stabilizes the medical system. That only makes sense if you stabilize at a high level that churns through cases, otherwise what’s the point? Australia is halting things at zero.

NPIs can look good during the crucial two-week blame period, but then you’re in the same situation two weeks later, over and over again.

I disagree with Misha, I think this happened, because that’s mostly what I would have expected. 

My model is that most children who would care about such a plea were already wearing masks, and the last thing kids want to be seen doing is caring about the teacher’s health in front of their peers.  

New Zealand locks down for three days after one positive Covid test. If you’re going to play this game, by all means play to win it, but it’s not a long term solution. I’m curious how often this turns out to be a false positive.  

On a related note, reports that Amazon pulled their Lord of the Rings production out of New Zealand to the United Kingdom due to not wanting to deal with all the Covid quarantine and other restrictions.

Via MR, thread by Andy Slavitt, Former Biden White House Sr Advisor for COVID Response, laying out the need to live with Covid long term and proposing eminently reasonable actions for doing the best we can with that. Here’s his takeaways at the end:

I definitely agree that we should adopt a permanent ‘stay home if you’re sick’ norm, which we should have had anyway. The question is what we are buying, in this scenario, with all these extra precautions, the same way we’d ask what our flu precautions are buying to see if they pass a cost-benefit test, if we did cost-benefit tests. 

MR also shares this study of mask mandates. Given how non-linear the dynamics are in such situations, and the role of control systems, saying that a mandate saved a certain number of lives seems like not a reasonable way to measure whether mask mandates work, but I’m not sure what other options are available. I haven’t looked at the study myself.  

Los Angeles to require masks at outdoor concerts and festivals, regardless of vaccine status, but not require vaccination. To be fair, it looks like the threshold for this is 10,000 people, usually crammed together tightly, so even outdoors I can see taking extra precautions. 

A call for the end to mass testing in the UK, since it can’t stop the virus in any case. As always, when gathering information, one must figure out the value of information and compare it to the costs. Enough mass testing to generate good statistics seems clearly still worthwhile to me, but beyond that, what changes based on what you find? Does it do you any good?

Survey of ‘experts’ says most wouldn’t go to the gym. Most wouldn’t go to the gym anyway. Gyms are a weird case, because the main point of them is health benefits, and there are (mostly) other ways to get the same effect. Not sure what I’m going to do on this once I’m back in the city.

Delta Variant 

The attributes of Delta are the biggest and most important unknowns. 

How much more infectious is the Delta variant? How much less effective are the vaccines against it than against Alpha or the original strain? How much deadlier is it, especially to the vaccinated or to children? How much faster does it cycle through? Do its higher viral loads mess with test results?

These can all be considered as part of a series of equations.

  1. We know almost exactly how fast Delta displaced Alpha.
  2. We largely know how fast cases and deaths have been growing.
  3. We know what percentage of various countries/states are vaccinated.
  4. Alpha’s serial interval is ~5 days, Delta is likely faster and more like ~3 days.

There’s a lot of room to argue details on #2-#4. 

#2: We only know about reported cases and deaths, and our positive test percentage is down a lot so it’s possible we are missing a higher percentage of cases. However, if Delta is deadlier than Alpha, then we can’t suddenly be missing a lot more cases unless we’re also suddenly ignoring deaths, and the testing slowdown doesn’t apply to other places. One could however argue, as we will below, that if vaccinated people are a larger percentage of infections, and we’re missing a lot more of their cases because they’re less serious or don’t get tested, then that could throw this off.

#3: The base percentage is good but what we need is the effective vaccination percentage for various purposes, all of which are different. For deaths, age matters a lot, and the vaccination rate is effectively higher. For cases, how often are we detecting cases in people at various ages and with varying vaccination status? For infections and thus growth rate, how much do children matter, and how much do the vaccinated matter (which is more of a thing we’re trying to solve for, but it impacts our answer here as well)? 

#4: The faster Delta replicates the less additionally infectious it needs to be in order to displace Alpha and grow cases the way it did. The drop from 5 days to 3 is a big game.

The rate at which Delta took over from Alpha and then grew, in various places, is the central constraint. You only have ‘so much infectiousness’ to go around. The data is consistent with Delta being 50% or so more infectious than Alpha across the board. If you then put additional growth in infectiousness in one place, to the extent that it’s big enough to matter, you have to take that growth from another place. The average is an average. It’s not a floor. 

As a working example, here is a long thread doing math on ‘breakthrough’ infections and deaths, along with a Mayo Clinic study. Starts with the executive summary:

Before we go further, suppose this is right. What would that correspond to in terms of vaccine effectiveness if true?

Let’s say 15% of new infections are in vaccinated individuals. Depending on what percentage of the population counts as vaccinated, you get a different answer. 

At the time this statement was made about 70% of the adult population had at least one dose. If you use that number and they have 15% of infections, the vaccines would be 93% effective against infection, or 89% effective if they’re 20% of infections. If we cut their share down to a conservative 60% to account for partial vaccinations and children, and say 15% of infections are among the vaccinated, we get 88% effectiveness against infection. So it’s not like the 10%-20% range is either an update or scary versus our previous expectations.

For death, if we assumed the vaccinated and unvaccinated populations were identically distributed, we’d improve from 88% to 95% effective against death. But that’s not right, because vaccinations increase dramatically with age and we can essentially ignore children entirely. 

At a minimum we’re effectively looking at 75% here where it counts, which would make the vaccines 98% effective against death. That is disappointing compared to what we had before, but that’s quite the statement about how effective they were before. It’s still amazingly great.

(Standard disclaimers that there’s a bunch of confounders I’m not dealing with here, none of this is exact, but it’s for intuition pumping and Fermi estimation rather than an exact answer.)

Then we can compare that with the body of the thread, along with other data coming in.

What we find here is a comparison of the early numbers and reports, which had almost no breakthrough infections let alone deaths, and the new numbers that aren’t as insanely great.

Then of course we get a reference to Topol once again, because the world is small. Sigh.  

Even the Mayo Clinic gets in on the act (study, note that Delta was not their intended target for the study, it started too early).

The ‘as far as’ is based on the numbers for Pfizer, which did much worse than Moderna here:

I’m going to go ahead and say this is probably all hopelessly confounded. It’s all observational, they didn’t periodically test populations. The size of the difference between Pfizer and Moderna here is absurdly high. And, yeah, no. 

One big motivation for that: That obviously doesn’t make any sense in the context of the headline claims. If vaccines were only 60% effective, even if we only count 50% of the effective population as vaccinated we would still get 30% of cases as breakthrough cases. Also, if Pfizer vaccinations are only 60% effective, then as per the calculation last week Delta would represent a more than doubling of cases every four days versus Alpha until behavior adjusted, and we would have seen the Delta variant take over from 1% to over 50% of cases within a span of under three weeks, and from 1% to 99% within six weeks. 

So if we’re not seeing anything like those results in general, then either we need to explain what’s going on somehow or claims of vaccine effectiveness this low are Obvious Nonsense. These are the first checks any reasonable person would run, and I’m confused why I’ve literally not seen anyone else do the calculation. 

The prevalence of Delta in their study in May was 0.7% and in July it was ‘over 70%.’ If we assume a constant growth rate, what does that imply? I built a toy spreadsheet to see, with two free variables of daily growth rate of Delta relative to Alpha, and the initial share on May 1. We get 0.10% Delta share on May 1 and a daily growth rate of 11% of Delta relative to Alpha, which passes sanity checks. 

Which once again places us back at Delta being 50% or so more infectious than Alpha among the entire United States population. But we also think it’s 50% more infectious among the unvaccinated population alone. Can’t have it both ways. There are only so many ways out of this, as it’s a math problem. For completeness:

  1. [HYPOTHESIS] Serial interval for Covid is longer than we think.
  2. [HYPOTHESIS] Delta isn’t that much more infectious than Alpha for the unvaccinated.
  3. [HYPOTHESIS] Behavior adjusted radically during this period and somehow explains it.
  4. [HYPOTHESIS] Regional dynamics explain this or something, maybe?
  5. [HYPOTHESIS] Vaccinated people aren’t getting tested and their cases are missed.
  6. [HYPOTHESIS] Vaccinated people aren’t infectious, or at least much less than others.
  7. [HYPOTHESIS] Which cases are vaccinated aren’t tracked properly so there are a lot more ‘breakthroughs’ than the data say.

I think we can safely dismiss 1 through 4:

  1. This would contradict a lot of other data points and there’s no one proposing it. Delta is considered to move faster than Alpha, so if anything this variable makes our problem bigger rather than solving it.
  2. Delta’s mechanism is much larger viral loads, we have lots of data points showing it’s more infectious in the unvaccinated, and it spread like wildfire to take over India where no one was vaccinated. Seems highly implausible.
  3. If we separate out the Alpha and Delta case counts we don’t see any radical adjustments and I don’t see any way this solves the puzzle even if we found them – I listed this for completeness.
  4. Delta took over at roughly the same time across the country, I don’t see how this could solve the problem.
  5. Could be true, given that we think ‘breakthrough’ cases are mostly asymptomatic or at least not serious. Logically this combines well with (6), since (5) solves the ‘where are the infections’ issue but not the growth rate, and (6) solves the growth rate but doesn’t explain where the infections are.
  6. Could also be true to varying degrees. To the extent that someone never gets infectious and isn’t detected, the case doesn’t really ‘count.’ 
  7. The earlier numbers for breakthrough infections were so low that it does make one wonder about this possibility. This can substitute for (5).

So the only way to make sense of this is some combination of (5) and (7), plus a large dose of (7). In this scenario, we’ve been missing most of the breakthrough infections the whole time, either because they don’t get tested and/or we don’t write down that they’re breakthroughs, and mostly they’re not serious infections. The missing cases would be invisible, neither spreading disease nor causing noticeable problems. Not exactly a nightmare scenario, and those infections would in turn strengthen immunity going forward.

Thing is, that scenario reconciles the population data with the vaccines losing effectiveness, but it’s not consistent with the study data, because this predicts that you’d see in the study what you see in the population. This was all observational, so the missing infections shouldn’t have been detected. On top of that, hospitalization rates conditional on infection didn’t change in the study, so a bunch of missing harmless infections doesn’t work at all here. 

Mason here cites Israeli data that on its face suggests vaccine effectiveness declined down to 16% over time, with dramatic and rapid drops. I don’t consider this credible, and presume it comes from the same mistake that was previously made in Israel of assuming all the cohorts are otherwise the same. Scientifically I don’t find this plausible, and also all the arguments about growth rates apply here as well. Still, it would be wrong to silently not include such information. 

A less courteous view of the Israeli data (his full post):

Here’s the chart with and without adjusting for age, yay Simpson’s Paradox:

Then you need to adjust for other things, including that vaccination rates are higher in the cities, which was the core problem last time. I still don’t love the numbers we’re seeing here, even after adjustments, but they’re not super scary or huge outliers. 

Most of all, let’s not overthink this and lose sight of the obvious.

This was on July 29, so not fully current, but cases were already mostly Delta, and the vaccinated were the majority of adults and a large majority of the elderly and most vulnerable  Eyes on the prize.

 

There remain large unknowns in all this, but my best guesses on the state of Delta have not changed much. Here’s my current model.

  1. Can vaccinated people spread the virus if they catch it? Yes, of course they can, but less often than a similar unvaccinated person, mostly due to differences in severity and duration. This is too much nuance for the press.
  2. How effective is the Pfizer vaccine against symptomatic Delta? Over my range of possible answers, mean 86%, median 89%.
  3. How effective is the Pfizer vaccine against death from Delta? 99%+.
  4. Does the Pfizer vaccine lose effectiveness from those numbers, with time? Probably some, but nothing like the extreme numbers being suggested.
  5. Should you get a booster shot? If there’s sufficient supply and they’ll give it to you if you ask for it, I’d do it. 
  6. How much more infectious is Delta than Alpha among the unvaccinated? About 50%.
  7. What is the serial interval of Delta? Probably 3 days, maybe 4. For Alpha it’s 5.

Booster Shot

In case we need them, we’ve already purchased everyone’s booster shots. Looks like we’re going to need them.

Pfizer’s earnings report seems to have contained its latest data on booster shots, which then got covered on CNN. I don’t think anything meaningfully fishy happened, but it’s a very strange place to put your scientific data.

Booster shots are now available to the immunocompromised, and there are reports that they’re considering calling for boosters for everyone soon. Of course, until the time when the booster is the Only Responsible Thing To Do, it’s forbidden, even unthinkable to many, because that’s how things work and they think talk of boosters will discourage vaccinations.

The real nonsensical thing, however, is that if you accidentally got J&J as your first shot, well, whoops (MR). No second dose for you, let alone a third, no matter how immunocompromised you are. 

Whereas it looks like they’re not requiring proof of underlying medical conditions to get the third shot, so, as MR puts it, solve for the equilibrium.

I notice that this is sufficiently absurd that it crosses my ‘you’re not supposed to lie’ threshold. If you’re in this situation, and you want a second shot, I think it’s fine for you to do what you need to do.

Soon, you’ll be able to get a third shot eight months after your second shot. Biden plans to do this ‘when it is available (WaPo)’ as if he wasn’t the President of the United States and couldn’t get his hands on a third dose or perhaps face a slightly unusual cost-benefit calculation that might justify doing it a bit early. It’s bizarre the times that Biden thinks the geometries bind him, versus when he thinks they do not, I can mostly predict it intuitively but it’s still very strange. 

Think of the Children

Kids are at higher risk from Delta than Alpha, as one would expect.

The thing is, tripling something only matters three times as much as the original number. The actual increase here is still very small. 

How bad is it really? Scott Gottlieb points out we’re not gathering the data (REACT data).

On the one hand, yes, absolutely we should be gathering the data. On the other hand, data does not stop being true because it’s from across the pond, and child data can be compared to the adult data, so why not look at what Britain’s REACT data found? I didn’t see any takeaways on child infections, and it doesn’t seem like the best use of time for me to look right now, but someone should definitely look. 

Periodic reminder: Whether mask mandates prevent Covid transmission in schools is one of those things that’s too important to know, because it would not be ‘ethical’ to study it properly even though some places already have mandates and some do not. Still, if you want to issue a study that says masks prevent Covid transmission in schools, it would be useful to actually compare schools where they wore masks to schools where they did not wear masks. Or else, one might say you have provided exactly zero evidence

As a periodic reminder, very few people are saying anything about preventing individuals from doing NPIs to protect themselves, or telling them that they have to dine indoors or go to concerts. No one is proposing ‘banning masks in schools.’

I do realize that saying this is false is a highly uncharitable interpretation of what Biden said, I mean everyone know what he meant, take him seriously not literally and all that, but it’s also not what anyone in question is doing, and the distinction is important. Mandating a lack of mandates is not to forbid the underlying action. 

The arguments are mainly over whether to (A) mandate mandating masks or (B) mandate not mandating masks, without that much support for (C) neither mandating nor not mandating mask mandates and letting people decide under what rules they will associate, either in local public venues like schools and/or local private venues like gyms or restaurants. Thus, Florida and Texas mandate not mandating masks in schools and threaten to withhold funding, and others try to sue or coerce them into mandating mandating masks instead, or at least take Biden’s tactic of mandating not mandating a lack of mandates (WaPo) including using the civil rights act. 

You gotta love attempts like this (from the WaPo article on Biden’s booster shot):

I do think masks are an important escalation in the dystopian quotient of the American school, but it is nice to remind oneself that the existing bar was not so low.

And it’s a reminder that I can’t find an example to contradict the claim that ‘no one is banning masks in schools’ but I have no doubt that someone somewhere is doing exactly that. Schools tell you how to dress, it’s rather standard. A large percentage of schools banned masks in 2019 and assumed anyone wearing one was up to no good, and any claims regarding health benefits were doubtless mostly ignored. Everything being either mandatory or forbidden is the mindset here, and also central to what the schools are trying to instill in the children.

Periodic reminder: Air filters in schools were worthwhile before Covid anyway, estimated 0.15 SD impact on test scores, and also we shouldn’t need to cite test scores here, free our kids or at least let them breathe in the meantime.

Periodic reminder: Schools are what they are, and some kids hate them enough they prefer the dystopian nightmare that we refer to as ‘remote learning’ in order to be free of an even worse nightmare. 

Scott Alexander has a post this week called “Kids Can Recover From Missing Even Quite a Lot of School” pointing out that losing out on a bunch of school seems to do surprisingly little, even measured in school test scores, as long as you don’t miss school during the month when they’re spending your life cramming for the test before you forget everything again. Excused absences don’t do anything, even though being sick is usually rather bad for almost everything, so one could wonder whether they’re actively good for test scores. Scott doesn’t explicitly ask the obvious question, which is “does school actually do anything useful while it’s traumatizing kids to obey authority rather than examine physical reality, and ending their lives one minute at a time?” or alternatively “But Can Kids Recover From Not Missing Even Quite a Lot of School?” but he does point out that school didn’t teach him anything useful and now he’s Scott Alexander, and it didn’t teach me much of anything either, so search your experiences and draw your own conclusions, and then draw your own secondary conclusions about the pandemic. 

(Great Minds Think Alike note for next story: Between when I wrote this section and when I hit publish on Thursday, Scott Alexander came out with his links post, in which he has almost exactly the same take I do. 

Young children learn about the world by interacting with it. What would happen if suddenly all the adults around them were covering the lower half of their faces? Someone really ought to do a study, except no, no one will ever do that.

Whoops:

This is the only proposed method of possibly getting a study done, and I don’t think even this would work in context:

There is a taxonomy whereby there are two kinds of things in the world. There are things that are Risky until proven Safe by a Proper Scientific Study, and things that are Safe until proven Risky. Authorities and “experts” choose, based on framing, context and their incentives, which way to present a given thing, and a lot of the talking past each other comes from this disagreement over priors. To the extent that it’s an honest disagreement and continuous rather than a boolean, this is reasonable, and to the extent it’s not, it’s not. 

I don’t know how concerned I should be about this particular problem, but I’m confident the correct answer is not to be unconcerned, and especially not to be unconcerned due to there not being a study on it. 

Also, at least some data is in on how babies are doing these days, and Kerry’s perspective seems directionally wise, although I agree with the study authors that the social isolation is the dominant factor here rather than masks (study).

In Other News

FDA Delenda Est, but it could always be worse. We could have the TGA.

Fluvoxamine reduces hospitalization from Covid by 31% in preliminary results, Ivermectin found to have no effect. Additional coverage here. Sample sizes between one and two thousand. This is a cheap generic and known to be relatively safe, so this seems like enough to justify using it. 

Zeynep thread on Covid origins, it seems the WHO let the Chinese tell them what hypotheses could and couldn’t be in their report. Sounds about right.

The Governor of Texas tests positive for Covid, after testing negative every day or quite a while. Daily testing is an interesting idea.

Latest Vitamin D study, not directly on Covid or directly measuring deficiency, but showing that moving from a high-D area to a low-D area (due to less sunlight) is associated with worse outcomes. I don’t mention Vitamin D as much as I should, as it’s one of the practical things an individual can do that has high expected value in terms of preventing or helping with Covid, that would be a good idea even without Covid. And yet I struggle to remember to take it. 

A reminder that while you do not want to catch Covid, permanently crippling our way of life is not a reasonable price to pay for that, and we need to stand up to ‘health experts’ who think such things are reasonable and run them out of town on a rail. Such talk does not encourage reasonable short term behavior, and potentially lays the groundwork for the destruction of our way of life. Life beckons. Live it. Which, as Nate Silver notes, most people are correctly doing.

The fall plans meme is over and we have a winner, congratulations Cedric:

Magic banned Oko. That was wise.

Not Covid

If you don’t think environmental review is completely out of hand, New York City tried to pass congestion pricing, and it’s going to be held up for sixteen months to start for “environmental review” despite its impact there being purely beneficial.

If you think that’s bad, the city of San Francisco is going to build a tunnel for a train line and its baseline budget for environmental review is (all caps mandatory, there are rules) ONE BILLION DOLLARS. With a B.

Failure to do cost-benefit analysis and scope insensitivity, in one Twitter poll:

23 comments

Comments sorted by top scores.

comment by Gunnar_Zarncke · 2021-08-19T22:03:13.265Z · LW(p) · GW(p)

I really like your weekly Covid updates and have read all of them basically since the beginning. But at least for me, they have become too long and/or too difficult to scan for the relevant/new information. I found it (the part after "Here’s my current model.") in the middle of a lot of text and screenshots but even that didn't add anything really new for me (having read all your other posts). 

For me, it would be more valuable if you could somehow break it down into "key facts" and "nice to know". 

Replies from: Zvi, Pattern
comment by Zvi · 2021-08-20T10:25:22.430Z · LW(p) · GW(p)

Key facts yeah I should so that. I try to make the opening kinda have that but it's not explicit.

Alas I do not have the time to make it shorter, other than having a higher bar for inclusion (and I do make cuts for space). Only so many hours in the week, not my day job with high opportunity costs, etc.

Replies from: Dagon
comment by Dagon · 2021-08-20T21:28:05.377Z · LW(p) · GW(p)

Honestly, I read most but not all of these - I want to reinforce Gunnar's "I really like your updates" opener more than anything else.  

comment by Pattern · 2021-08-21T05:39:22.273Z · LW(p) · GW(p)

There are these people called "commenters" that can do that.

comment by 25Hour (aaron-kaufman) · 2021-08-21T04:00:32.237Z · LW(p) · GW(p)

I look forward to Thursdays specifically for these updates.

comment by aphyer · 2021-08-19T15:58:28.009Z · LW(p) · GW(p)

Your seven hypotheses alternate between being numbered 1-7 and being referred to as A-G, which is rather confusing.

Replies from: Zvi
comment by Zvi · 2021-08-19T17:58:33.334Z · LW(p) · GW(p)

Yeah, fixed in original along with 2 other issues, mods please reimport to be sure I got all of it.

Replies from: habryka4
comment by PatrickDFarley · 2021-08-20T16:32:17.178Z · LW(p) · GW(p)

The two sides are both trying to make the mandates look as obnoxious as possible, for different reasons.

This is such a thing, I see it all the time, and it is both completely obvious and apparently not noticed by anybody else.

When your goal is to signal to your side rather than convince the other side, and their goal is to signal rather than be convinced, you get this perverse symbiosis of everybody saying ridiculous things on purpose.

Replies from: gabriel-holmes
comment by tkpwaeub (gabriel-holmes) · 2021-08-21T10:13:51.092Z · LW(p) · GW(p)

Alternatively, Team Anti-Vaxx keeps prying open the Overton Window for "my" side (oops).

comment by Craig Fratrik (craig-fratrik) · 2021-08-20T15:07:53.837Z · LW(p) · GW(p)

Typo:

"after testing negative every day or quite a while. Daily testing is an interesting idea."

should be

after testing negative every day for

comment by philip_b (crabman) · 2021-08-20T13:41:31.034Z · LW(p) · GW(p)

I play MTG but I don't understand the Oko, Thief of Crowns meme. Can you explain?

Replies from: Measure, Pattern
comment by Measure · 2021-08-20T13:53:10.056Z · LW(p) · GW(p)

Oko, Thief of Crowns was poorly balanced for the standard meta he was released into and ended up being very strong/oppressive due to being hard to interact with. The meme is that you wanted to just [play normal Magic]/[live your life], but you ended up having to focus on how to deal with [Oko]/[Delta].

Replies from: ChristianKl
comment by ChristianKl · 2021-08-20T14:13:01.381Z · LW(p) · GW(p)

It wasn't oppressive in the standard meta into which we was released but only after Field of the Dead got banned. 

Replies from: Measure
comment by Measure · 2021-08-20T16:46:34.278Z · LW(p) · GW(p)

Fair enough. I wasn't playing standard at the time, so my information is second-hand.

comment by Pattern · 2021-08-21T05:44:43.582Z · LW(p) · GW(p)

Base Meme:

My Summer* Plans:

[random picture of literally anything, I haven't seen a lot of high quality versions of this meme]

Delta:

[another random picture, seriously would be way better if it these were that general versus Darth Vader force choking him]


The Oko Thief of crowns is an actually good version of the meme/image macro:

My Summer plans:

[picture of the back of a magic card]

Delta:

[a magical thief, that has stolen...your summer]


*Summer plans, Fall plans,...same meme.

comment by AnthonyC · 2021-08-19T19:06:58.577Z · LW(p) · GW(p)

he does point out that school didn’t teach him anything useful and now he’s Scott Alexander, and it didn’t teach me much of anything either, so search your experiences and draw your own conclusions, and then draw your own secondary conclusions about the pandemic. 

Well, it took 15 years, but school did eventually teach me to stop trusting authority figures...

I don’t mention Vitamin D as much as I should, as it’s one of the practical things an individual can do that has high expected value in terms of preventing or helping with Covid, that would be a good idea even without Covid. And yet I struggle to remember to take it. 

Do you take anything else daily, successfully? I find weekly pill boxes labeled with the weekdays helpful, I fill them once a week. Although, vitamin D doesn't necessarily have to be taken daily, it builds up slowly in the body over time so you could easily double up after days you miss. One time I had an actual diagnosed deficiency, and the prescription vitamin D was a huge dose (50k IU) once a week for 8 weeks.

Replies from: Zvi
comment by Zvi · 2021-08-19T21:15:42.693Z · LW(p) · GW(p)

I don't take anything else. I do know I should likely be taking at least calcium and probably a few others, but in practice I just do D, and take bigger doses when I do remember in order to compensate, since as far as I know it's fine to do that.

comment by Sammy Martin (SDM) · 2021-08-24T13:43:45.864Z · LW(p) · GW(p)

You might be interested to know that a new report from Public Health England on a giant observational study of 300k plus vaccinated people, broken down by age and type of vaccine, basically confirmed your guesstimate numbers to within a few percent (VE for Pfizer 2 dose Vs Delta is about 85% and wanes such that it about halves in efficacy after 100 days, and is twice as good in younger age groups). They even have nice graphs showing efficacy over time,

Thread summary: https://twitter.com/john_actuary/status/1428269230798123008?s=20

Actual paper: https://t.co/9NtGXNHyKW?amp=1

comment by CraigMichael · 2021-08-20T07:41:39.415Z · LW(p) · GW(p)

I felt a nervous mentioning fluvoxamine a few months ago to someone that was potentially going to buy drugs for his country. https://www.lesswrong.com/posts/kPeiMbyBsvMMcs62e/should-i-advocate-for-people-to-not-buy-ivermectin-as-a#ZoHjQeNWwABRsAbMG [LW(p) · GW(p)]

Just took a big sigh of relief that my advice didn’t suck.

comment by Norman Borlaug · 2021-08-23T21:26:01.896Z · LW(p) · GW(p)

Is there any reason you don't provide positive tests per capita by region instead of positive test by region? Seems like the former would be more useful for comparison purposes.

comment by RedMan · 2021-08-22T22:09:54.054Z · LW(p) · GW(p)

What's the math on the 'default unvaccinated' treatment plan of

  1. get a covid test if sick
  2. if test is positive, treat with fluvoxamine and regeneron mabs?

Is there a point where the 'reduced vaccine effectiveness' curve crosses that default treatment?

Obviously a vaccinated person could do that protocol in addition to the vaccine, so this isn't a true comparison.

comment by Pattern · 2021-08-21T05:38:25.042Z · LW(p) · GW(p)

16.7% of people polled(biased sample) appear comfortable with murder by means of magic. All it takes is a couple ceilings that are only 15.5 inches high, or lower, while the space is occupied.