Covid 1/5/23: Various XBB Takes

The majority of this week’s post is various people doing their Obligatory Thread on Latest Variant, in this case XBB. All reach broadly compatible conclusions:

1. XBB is the next variant and rapidly taking over, and this is normal.

2. XBB is not more deadly.

3. Our existing protections against severe disease and death carry over.

4. Our existing protections against infection are reduced a lot.

5. XBB is not obviously going to result in more infections over time.

6. All the standard Public Health recommendations remain in effect.

Executive Summary

1. XBB is taking over.

2. XBB is no more or less dangerous than previous variants.

3. It will however extend the winter wave longer than it would have lasted.

Let's run the numbers.

The Numbers

Predictions

Predictions from Last Week: 385k cases (+16%) and 2,600 deaths (+17%)

Results: 423k cases (+27%) and 2,703 deaths (+23%).

Predictions for Next Week: 450k cases (+7%) and 2,950 deaths (+7%).

The numbers came in higher than expected, more due to the standard ‘data is funky so hedging my bets’ factor than anything on the ground. This next week we will smooth out the double reporting in the Carolinas and other back-fills, while making up for New Year’s, so it’s another funky week.

Florida failed to report for the second straight week. I’ve given them the same numbers I gave them last week.

For all the talk of a wave, so far we have not seen a substantial spike in deaths. This week’s numbers are not much higher than early December, instead they are reverting after Christmas.

North Carolina and South Carolina clearly reported two weeks worth of data, adding about 30,000 cases and 100 deaths that should instead be part of the previous week. I did not adjust for this as it is part of the standard holiday cycle. Thus, the case numbers in the South should not be alarming.

Deaths

Cases

Bob Wachter notes that Covid asymptomatic positive test rate in SF declined from 5% to 3.2% on December 30. He notes that this is likely a temporary lull, partly due to XBB.1.5. It is still an excellent sign that we were at least close to stabilizing before that.

Caitlin Rivers joins the camp that says case numbers are no longer that useful, says that our case numbers have become almost completely decoupled from other indicators.

I agree that there is danger when comparing case numbers now to case numbers a year ago. I still think that comparing to last week is highly useful, as the reporting regime will have changed little. Thus, I still see them as useful, as long as they are combined with a metric that does not have this problem, such as deaths. As I’ve noted, wastewater and hospitalizations are useful while having their own problems. Test positivity depends on who gets tested and why, and which tests are reported.

BNO reports hospitalizations are their highest since February. This was well after the peak of cases last year.

From the nowcast of about a week ago, the spread of XBB.

Here’s wastewater monitoring.

That shows the west in decline up until about a week ago, and even the Northeast only going up a factor of about 2.5.

XBB.1.5

New year and variant, who dis? (preprint)

In summary, XBB.1.5 is highly immune evasive and is spreading due to this, it is not more dangerous than previous Omicron variants, and the bivalent boosters and recent other infections (while not ideal) still prevent severe disease - if that’s all you need to know you can skip the section.

The good news is that it seems the advantage of XBB.1.5 is mostly that it is better at immune escape, and that it is if anything even less inherently dangerous than previous Omicron variants. That means that we will have a wave once again, but that our situation did not become permanently worse. Once we cycle over again, we'll be back where we started, and with one less plausible escape candidate mutation waiting for its chance.

Eric Topol asks point blank the important question: What can be done about this?

Alas, the answer he comes back with is, essentially, nothing. The boosters are still likely better than nothing despite the immune escape. That’s about it.

Once again, in an adequate civilization this is exactly when we would rush out a new updated booster for XBB.1.5, ideally not bivalent at all, so those who were vulnerable or otherwise wanted to be protected could take it.

Needless to say we are not going to do that. We are not talking about it. I have yet to hear anyone even mention the theoretical possibility that we could respond to a new variant that had immune escape by trying to vaccinate people before they could get infected. At which point, it is reasonable to ask what we are even doing. In practice, our vaccine update plan has utterly failed and is capturing almost none of the available value.

Similar:

I have no idea what policy action is being suggested.

1. Something must be done.

2. ???

3. Therefore we must do it.

4. What would even be something?

The second best time to plant a tree is right now. Won’t help in an emergency.

Chise is here, same as every week, to remind us that immunity remains a thing.

(Sources: 1, 2, 3, 4, 5).

This says that XBB.1.5 is spreading faster than XBB.1 due to much higher hACE2 binding ability.

Here is another relatively worried report, although the most important fact - no increase in severe disease - is again confirmed.

I like this visual explanation of the history of variants, via Bloom Lab. Full thread doesn’t have a strong opinion on whether new variant will increase total cases. Does include a bunch of detailed genetic information on XBB. Makes it sound like this is mostly an evolutionary trade-off that does not obviously make things generally worse.

FDA commissioner has a thread about XBB and vaccines (study mentioned).

‘No clear evidence’ is an interesting twist on No Evidence. I notice that I very much did not update the way Califf would have liked me to when I saw it, as my brain interpreted it as a Double Weasel and based on what I know I would have expected him to be comfortable saying something far stronger. Not sure why he didn’t do so.

(Also he seems clearly behind the times on the ‘will move to other regions quickly’ bit, XBB is already everywhere.)

Yes. I agree that it is highly likely the existing vaccine and booster are helpful in preventing serious illness and death.

It is also all but certain that you and the agency you head, sir, are the reason we don’t have a more effective updated booster on the way sooner. It is on you that by the time we get a booster that works it will be too late to do much good.

I notice this kind of talk is engineered to silence the opposition and protect against possible objections to vaccination rather than seeking understanding. Do vaccines raise or lower the probably of a new variant emerging? It can go either way, depending on vaccine impact on several things including severity and spread. Using the Authoritative Dismissal like this likely only backfires.

And now, Saying the Quiet Part Out Loud.

‘We are concerned that raising doubts about the value of vaccines will continue to lead well-intended people to delay getting an updated vaccine.’

This is not the suppression of misinformation. This is flat out saying:

1. People who notice certain true things don’t act how we want them to.

2. We recommend you don’t talk about those things, or ask questions.

3. We recommend you instead act how we want you to act.

I for one appreciate the honesty.

Gounder tries to thread the needle on ‘fully vaccinated.’

She wants it to be one way. That would be good. Instead, goalposts are being moved, and most sources have moved to ‘fully vaccinated’ meaning what she calls ‘up-to-date.’ It is interesting to see a defense of the old classification system.

Also, once again, do you need more boosters? No, you need - if you want maximum protections - an updated booster that targets XBB. Which Dr. Califf is keeping illegal.

My model continues to think that vaccination reduces Long Covid risk per infection because it is proportional to severity. I continue to be unable to take seriously anyone who takes ‘42% Long Covid risk’ seriously.

This is not how I would interpret the case numbers or wastewater graph, nor would I engage in the typical ‘you could have Done More’ routine. Yes, cases in lots of places are ‘up’ because it’s early January, but clearly the primary driver at this point is XBB.

Ashish Jha thread on XBB, essentially saying we know enough that this is unlikely to be a huge setback but still don’t know much, and suggesting the usual interventions.

Marc Veldhoen thread on XBB. There’s always a new variant, hospitalizations look proportionate, our defenses against severe disease are holding.

Physical World Modeling

Matthew Healy offers a Google Document aptly called Key Covid Metrics and How to Compare Them. It also explains a bunch of his reasoning and how he thinks about the pandemic. Kudos to him for writing this up.

The major sources of information he follows, formatted the way I do such things:

• US Wastewater SARS-CoV-2 virus concentration numbers from Biobot.

• US Reported Deaths and Hospitalizations, and many epidemiological forecasts, from Covid19ForecastHub.

• US Excess Deaths and World Excess Deaths from OurWorldInData.

• Reported hospitalizations and deaths in various parts of the US from Covid Act Now.

• The truly superb blog Your Local Epidemiologist by Dr. Katelyn Jetelina.

• Many deeply informative talks by one of the world’s top experts on the evolution of viruses at Bedford Talks. Note: it’s not obvious how to go from slide to slide in Dr. Bedford’s presentations!  On a mobile device, drag left and right.  On a laptop computer, use the left and right arrow keys.

• The Twitter feeds of several dozen experts in various aspects of Virology, Epidemiology, Infectious Diseases, and related fields whose Tweets I have found to be especially valuable, with a warning that one needs to be able to discern who is worth listening to and who is not, and this is hard.

He then discusses how he thinks about these metrics.

His first metric is wastewater data. He attributes the rising ratio of wastewater data to reported cases to the rise in home tests that are not reported. I would also include that people are getting less sick and are less concerned about it, many more cases are outright asymptomatic, and many more cases are not even being tested.

I also wonder about whether different variants create different relative amounts of virus in wastewater. It seems plausible, for example, based on the details we know, that a given Omicron case could generate a larger wastewater signature than a pre-Omicron case. I’ve seen no attempts to measure this and find out. I haven’t talked about that concern. I’ve had it for a while. It’s been one reason I haven’t relied as much on wastewater measures.

Another reason is that I have a system I have learned to understand and navigate, which I then supplement with wastewater data, and found it easier to allow path dependence here. I was confident (and still am) that if there is a big divergence between measures I will notice and think about it.

He also looks at hospitalizations. I used to do that a lot, then I got frustrated by the degree to which capacity constraints and policy decisions were warping that data, and concluded it had stopped being useful.

Whereas he does his best to avoid case count data, I presume due to the lack of reporting and data integrity issues. If I was starting from scratch without my experience navigating those numbers, I might do that as well at this point.

He estimates 50% of America will be infected by Omicron in its first year or so.

He mostly thinks excess deaths are a better measure than reported Covid deaths, since almost all deaths are reported while cause of death is not as reliable, and takes it as relatively safe to say that excess deaths equal Covid deaths.

One note he makes is that most excess deaths post-vaccine were in red states, and he estimates that Trump ‘embracing scientific reality and strongly urging people to get vaccinated’ could have saved 400k lives. I don’t think it was within Trump’s power to close the majority of this gap, because I don’t think people would have listened even to him, unless other big changes are also made, such as him remaining in office, or everyone working together and releasing the vaccine pre-election, or complete sea changes in messaging and distribution from public health. Otherwise, a majority of what happened was effectively baked in.

He then highlights some papers.

1. He cites this paper as important, emphasizing the roll of county-level political affiliation and its correlation to excess deaths, and that state governments that enforced their restrictions narrowed these gaps.

2. He cites this paper that says that developing countries had double the death rate from Covid when adjusted for population age, the pre-vaccination infection fatality rates (IFR, not CFR) between developed and developing countries were similar at 0.5%. I would also consider quality of data and case reporting here.

3. Finally, he cites this study of individual level data in Ohio and Florida, showing excess death rates for Republicans were 76% higher than those for Democrats, widening from 22% pre-vaccine to 153% post-vaccine, concentrated in counties with low vaccination rates, so most of this effect was presumably due to vaccinations. This is a big argument for the value of vaccination, while also being somewhat of an argument against the importance of NPIs.

An excellent contrast. I wouldn’t be mandating masks anywhere at this point, but if I did terminals would be near the top of the list. I’d certainly be handing out KN95s all over in such places to whoever wants them, it would doubtless net save money. If you want to make air travel safer, focus on the terminals.

I should also note that I don’t think Issa is right about the magnitude of these effects. It’s easy to focus on conspicuous failures and hot spots that involve relatively few people, rather than focusing on the everyday interactions and other dynamics that matter far more. The next few weeks are likely to be somewhat ugly, but that would mostly be due to timing and XBB, and only a tiny bit about airports. There simply aren’t enough travelers involved to matter that much here.

Analysis of latest study on surgical versus N95 masks, finds non-inferiority for surgical masks, which in practice performed only slightly worse well within margin of error. There are still lots of reasons to presume N95s are the superior technology if all you care about is prevention, but this closes my estimate of the effectiveness gap, and the degree of discomfort matters. A doctor who feels, as one is quoted as saying, that their ‘head is in a vice’ by the end of the day is not going to be on top of their game.

Booster Boosting

The good news is that even the University of California is now willing to let you opt out of Covid vaccination. The bad news is that to do so you have to attest to being ‘aware of the following facts.’

The good news still vastly outweighs the bad news. If you are going to the University of California, you are doubtless used to and accepting of being forced to consistently attest to things that you do not believe.

Permanent Midnight

What’s better?

The comments include quite a lot of very explicit support for Permanent Midnight. Remember. Also remember to give thanks that Public Health types did not win.

They continue their quest to make all other voices illegitimate.

In Other Covid News

China

Not Covid, seems worth including anyway.

Morocco outright bans travelers from China due to Covid risk. This won’t meaningfully help Morocco avoid Covid. Still, China did kind of ban travelers to China for several years, so… fair.

I continue to not understand what people mean when they say, as in this video clip from MSNBC, that call for ‘collaboration’ or claim that ‘Xi needs to slow things down’ or otherwise that there is anything China could still do. It’s over. It’s done. Whatever is going to happen essentially can no longer be changed.

Remember

Bryan Caplan looks back on the libertarian reaction to Covid. I agree with his framing of there being essentially two camps, those who advocated for freedom to protect ourselves from Covid (he calls this ‘tech freedom’) in ways that make sense, and those that advocated for freedom against Covid protections (he calls this ‘personal freedom’ although that doesn’t seem like a clarifying term to me) that didn’t make sense. As he notes, these are fully compatible positions - freedom to do what you want and freedom to not do what you don’t want.

Yet most people only advocated mostly one or the other, and the two camps got into arguments. Caplan says that both camps ‘lodged some fair complaints against the other.’ The complaints by the tech freedom side were largely about vaccine skepticism. He doesn’t mention these, but I’d throw in complaints about their frequently advocacy for things like Ivermectin, unjustified skepticism of the effectiveness of NPIs, and often portraying the ‘focused protection’ or non-intervention scenarios as far less deadly than they would likely have been.

I certainly broadly agree that our use of NPIs stayed far too intense for far too long, and that the expectation this will happen again is the strongest argument against allowing NPIs (or any other such ‘temporary’ intervention) to take hold in the future. Yet I think there was a serious intellectual integrity problem here.

Whereas the complaint against the tech faction seems to be that they… were complaining about vaccine skepticism or a failure to support the crusade against NPIs sufficiently strongly? I’d agree that this was often an issue, and for my own part I did not come around quickly enough when the time came, but it’s a very different style of complaint.

Ultimately Bryan cites the freedom-from-interventions cause as more important going forward, because we were exceedingly lucky to get vaccines so fast this time, and without them we risk sliding into years of tyranny. In this model, keeping life worth living is the higher priority.

There are several counterfactuals and potential future scenarios here where we likely disagree on how things work.

I expect to usually have a vaccine available very quickly in these situations if we actually care about creating it. I don’t think we got as lucky in this particular way as Bryan thinks especially now that we have mRNA vaccines, and the alternative scenario where we get vaccinated starting in May or June of 2020 is worth keeping in mind, and I think the FDA is likely to also shut out many other vital tools that would clearly return us to a more normal state faster.

I also don’t expect a Permanent Midnight scenario to take hold so easily. There is a reason China eventually gave up. The cost of maintaining these kinds of restrictions rises with time, and anyone calling for them to be sustained will get voted out, and I expect people patience to if anything be lower next time. To get such a scenario, we’d need a virus that was much deadlier, especially to younger people, that also was in the strange ground where interventions were neither successful nor hopeless.

In that scenario, I’d still be much more interested in pushing for faster vaccine and treatment development, and better tests and NPIs, than I’d be interested in fighting harder against the NPIs so we could more or less ‘accept the deaths’ and move on. The space where the opposite holds might exist, but it seems quite small to me.

Tyler Cowen looks back and finds rapid Covid tests not so important. Certainly they were not in the same league as vaccinations. I think Tyler is selling the tests short here, largely from failing to think on the right margins.

In many cases, tests enabled gatherings and events that would otherwise have not happened, because people would have not been willing to take part. This likely includes a large number of schools, as well as things like family gatherings. It allowed many high value offices to reopen. This change seems super valuable.

The idea that ‘reassuring Nervous Nellies’ is a low value use is something I’d disagree with. Making the Nellies less Nervous is highly valuable, although it must be weighed against the cost and annoyances of needlessly imposed tests on others.

In other cases, a rapid test substituted for a PCR test that would have cost the system hundreds of dollars and often required great inconvenience, while the rapid test in practice was more useful.

Even if you ‘already pretty much knew’ you had Covid, for many reasons you often needed to confirm this, and the cost of having doubts by yourself or others remained high. The gain here is not so minimal.

The bottom line here is that, in practice given our approach to Covid, availability of cheap and quick rapid tests was a very large welfare improvement in lived experience and productivity. I am confident our children would have been in school far less often without such tests.

What about the prospects for mass testing or periodic testing? That depends on solving for the equilibrium, and what other things then happen that balance out the decrease in risk from the tests, and how much lower overall infection levels are before they stabilize, since full suppression was never viable. I do think we missed opportunity here.

Thus, I continue to think the value of rapid tests, and the value lost through our failures on that front, was quite high.

Other Medical and Research News

A study from 2015 on how often families with small children get respiratory infections.

The correlation here between children is very high. Either there is a bug ‘going around’ or there is not. Once you have two children, the chance one of them will get whatever is going around is already very high, so adding several additional children does not change things much, except insofar as they then infect each other, but even that effect seems likely to be small here.

Reminder: Claims that ‘the number one cause of obesity is genetic’ (study) and there is nothing anyone could do to prevent ‘this disease’ do not pass the laugh or smell tests. This is Obvious Nonsense .

On top of the data, there is the simple fact that not too long ago few people were fat. If the primary cause of being overweight was genetic, we would not suddenly go from mostly not fat people to quite a lot of fat people. That is not to blame anyone, simply to say that whatever is happening, this isn’t it.

GDPR so restrictive it is driving research out of Europe (FT).

If you have to deal with Europeans, Because Internet, it is hard for outsiders to understand how huge a pain in the ass and drag on productivity it is to have to follow GDPR. Perhaps something where you have a choice, like clinical trials, will reveal the depths of this damage.

Canadian Assisted Suicide

Richard Hanania makes the case for Canada’s assisted suicide program as moral progress, and, well… I appreciate the frankness.

It is true, for example, that some people might feel “pressured” to commit suicide because they don’t want to be burdens on their families or the government. I don’t think there’s anything wrong with this — in practically every other kind of situation, it is usually considered pro-social to care about the impact your life has on others.

Oh, sure. Seems fine to have society pressuring people to kill themselves because they are a burden to others. As long as it’s voluntary. Here he says ‘feel “pressured”’ yet the argument logically extends to actual exertions of pressure including by officials. Which are very much A Thing.

Relatedly, I argue against the “culture of life” argument on the grounds that we place too much value on human life and it would be better if we placed less on it, a view that conservatives implicitly hold on topics like covid restrictions.

All right then.

Given the amount of misinformation that is out there on the topic, it’s important to start with some numbers. According to the Canadian government, there were 10,064 MAID cases in 2021, which represented 3.3% of all deaths in the country.

He says this is not a lot. To me it sounds like a lot.

Canada recently expanded its assisted suicide program to cover those without a terminal condition, a fact that has gotten a lot of attention, but in 2021 only 2% of deaths fell under this new category.

So out of every 10,000 people that die in Canada, 6.6 killed themselves through MAID without a terminal condition. Again, that sounds like rather a lot.

About 81% of MAID recipients had received palliative care, and 88% were eligible for it.

I read this as ‘12% of MAID recipients were eligible to kill themselves but not to get palliative care’ and also I presume a lot of the 7% that were eligible but didn’t get it had issues with practical access and waiting lists. Of every 10,000 deaths in Canada, this is saying that 37 killed themselves through MAID while ineligible for palliative care. So there is an 0.3% chance this will happen to YOU.

Once again, this sounds pretty terrible?

Richard’s core argument boils down to Yay Individual Liberty, and that there are a lot of people out there suffering quite a lot. These are important points. My main takeaway, however, was that yes the whole thing already, in its current state, seems rather terrifying.