This pandemic is really making people lose their minds. I am increasingly frustrated by the tone taken in arguments about the virus, and it's basically impossible to have a rational discussion about anything. The mode of analysis being used is 1) identify the speaker's conclusion, then 2) use 1) (perhaps along with their political orientation) to categorize the person, then 3) dismiss the rest of the argument based on 2). We're all postmodernists now. The speaker rather than the speech is the focus of every analysis.
The problem is that true statements are being preemptively dismissed because they can be used to argue against the lockdown ("the lockdown" here could mean either public or private efforts to control the spread of the virus). When some good news is discovered, people are very quick to dismiss it, often with wild speculation. Sometimes the truth of a claim is preemptively disputed, because the listener knows how it will be employed. Other times, the truth of a claim is conceded, but the listener adamantly denies that the information weighs at all on various policy conclusions.
Past Infection Confers Immunity
It's bizarre to see all the wild speculation that Sars-Cov-2 doesn't behave like a typical respiratory virus. An example of this is the denial that past infection confers immunity, and the related denial that a large population of recovered victims of the virus would contribute to herd immunity.
Scott Sumner wrote an excellent post defending Rand Paul here. It was really a breath of fresh air, because my impression of Sumner is that he's been slightly more alarmist about the virus than what's warranted. Rand Paul essentially said that people who survived a bout with Covid-19 with their health in tact could breath a sigh of relief and live their lives freely. A scolding Huff Po piece essentially said, Nuh uh! Because some people have been reinfected, we can't assume you have immunity!
I find this truly moronic. Of course prior infection confers immunity, at least in the vast majority of cases. It is stunning that a major "news" source would dispute what everyone knows about viral infections. Your immune system learns to combat the virus and eventually fights it off. If this process didn't grant you some kind of immunity, you'd never get well in the first place. You'd just keep getting reinfected by the viruses that inhabit your body. (Yes, there are viruses that you don't ever recover from without medicine, like HIV. As far as I know, there are no such lingering coronaviruses.) Sumner points out that the number of reinfections has been vanishingly small. Millions of people have been infected and recovered. If reinfection were common, we'd have heard a lot more about it by now. Presumably a substantial fraction of these reinfections are real (some may be cases of people getting different viruses and thinking they've had Covid-19 twice, but I don't want to assume they're mostly false reports), but the magnitudes are tiny. At some point, we really do have to stop worrying about infinitesimal risks and focus our attention on more important things.
(By the way, all of the vaccines grant imperfect immunity, too. There appears to be a double-standard at play, and it's leading to some very low-quality arguments.)
The Huff Po piece strikes me as motivated reasoning. They don't like the implications of Paul's suggestion, so they're trying to deny a widely understood truth about how our immune systems work. If our public health scolds could manage a more nuanced version of their recommendation, maybe we could trust them more. "It's true that prior infection confers immunity in the vast majority of cases, but a policy of allowing some people to dodge the mask mandate would likely be abused." Or "Take risks as you please, but be aware that some small fraction of you will still be susceptible, either because immunity is not permanent or because you had something other than Sars-Cov-2." They need to stop assuming the mental bandwidth of public opinion is so narrow you have to lie to the public to get compliance.
I also understand that immunity may be time-limited. You might only have it for a few months or years, but not permanently. If immunity is time-limited in this sense, then one policy implication is that we should not have tried to flatten the curve. If people keep dropping out of the "immune to covid" population because their immunity degrades over time, then the virus will keep finding new hosts and keep going, possibly forever. This is another example where the policy implications of a factoid are poorly explored. It's as if people are just gathering bulleted lists of "things that favor my policy," not noticing at all what these factoids contribute to the other side of the ledger.
None of this is to say we shouldn't shelter the vulnerable. It's just that whoever is likely to eventually get the virus, we should hope they all get it over a shorter period of time rather than stretched out over a longer time. The original lockdowns in March were justified under this "flatten the curve" rationale, where the area under both curves was presumably equal, but it quickly morphed into a "quash the virus" policy. I wrote about that here. Whatever the merits are to the "immunity is time-limited" argument, it has clear policy implications for a "flatten the curve" approach. My frustration, which I am trying to express in this post, is with people who fail to acknowledge that some point of fact changes the policy calculus in a direction they don't like.
Younger People Are At Negligible Risk
I wrote this post a couple of months ago, complaining that people aren't noticing that young people are at negligible risk. Or perhaps they'd grant that (the data are inescapable), but they fail to concede the obvious policy conclusion: that the relatively young and healthy should be able to go about their lives in a business-as-usual fashion. The focus should be on protecting the relatively old and the vulnerable. Certainly this implies putting some restrictions on what healthy people can do. For example, they should be isolating from the vulnerable people in their lives, and they should mask up if they're going into a mixed space.
I don't know what's going on here. Does anyone think the risks to young people are unacceptable? This is unlikely, considering that their (negligible) risk of dying from Covid-19 is smaller than most risks we routinely subject them to. There might be some people who think it's much worse if their child dies of covid than if they die of a much more likely hazard, say an automobile accident on a frivolous outing. If people want to prioritize their own risks this way, that's their call, but such irrational risk preferences shouldn't influence public policy.
Or is the issue that the elderly and vulnerable are more at risk if there are a bunch of diseased young people intermixing and spreading the virus? I also consider this position unserious. The vulnerable people are already isolating and taking protective measures. Now is the time for young people to live their lives, to spread the virus, to acquire herd immunity. At some point people lose their patience with the lockdown. Vulnerable people have been locked up for months without any human contact, and many of these folks are in the last few years of their lives. They could emerge into a world where most young people have already had a bout with Covid-19 and are probably immune, or they could emerge into a world of negligible immunity, where most healthy people are still potential spreaders. Misguided public health policy has ensured the latter scenario.
(By the way, I was completely against school closures from the beginning. I think we had enough information about the age-mortality curve even at the time to recognize that school children were not at risk. Emily Oster has since collected a ton of data on school districts and viral outbreaks, and schools just don't appear to be major spreaders. Families should have been free to remove their kids from school if they were concerned about the virus, say if they had a vulnerable person in the home, and staff and teachers who weren't comfortable with the risk should have been furloughed or given some other arrangements. But totally closing the schools was unjustified and incredibly disruptive to people's lives.)
The only serious objection to the "herd immunity for the healthy" strategy is the possibility of long-term unknown health effects, which I'll treat below.
Long-term Health Impacts of Sars-Cov-2
I haven't done my homework on how to quantify the known long-term harms of Sars-Cov-2. Presumably there is enough information for someone to determine the likelihood and seriousness of various sequelae that have shown up so far. As we should. That's perfectly legitimate, and such lingering harms should affect our cost-benefit calculus for whatever mitigation measures, private or public, we wish to undertake. If any readers have done their homework and believe that known harms (or reasonable extrapolations from what's known) justify lockdowns or extreme caution from the young and healthy, please feel free to share your work.
What I keep seeing is a broad appeal to unknown harms of the virus. There are right and wrong ways to reason about decision-making under uncertain risks, and I keep seeing the wrong way. Here's what I mean. Say you're at work and I say, "Are you sure your toilet stopped flushing this morning?" You might start to worry. "No, I'm not entirely certain. I don't usually check that. I just walk away after flushing and assume it will stop, as it always does." I didn't actually do anything that made a catastrophic flood of your home any more likely, but I put the uncomfortable thought in your head. The problem is that our stupid monkey brains treat risks as more likely just because someone mentions them. There's an evolutionary just-so story here. In an ancestral village, if someone is talking about a lion or an armed stranger from another tribe, it's probably because they saw one. It probably pays to be extra vigilant about the risks people in your village are discussing, because our ancestors probably spent their time discussing real, imminent risks. In contrast, if I just make up some hazard on the spot and speculate that it might happen, that doesn't actually make it any more likely to happen. You probably shouldn't worry about it. You can't be driving home every time you think about some potential catastrophe and feel the urge to check up on it. On the other hand, if my wife texts me and asks the same question, I really should feel worried. She probably has some information I don't about the functioning of the household plumbing. There is a big difference between informed risk management and wild speculation.*
When someone says, "We don't know the long-term health effects of Covid-19", it is technically true. But we can't base massive policy responses to mere possibilities that come to our attention because someone blurts them out. We should infer whatever we can from our experience with other respiratory viruses, perhaps other coronaviruses or seasonal flus with similar symptoms. But if the vast majority of people who recover look like they're basically in the clear, we shouldn't speculate too much about some phantom hazard that's going to emerge in the coming years or decades.
There is good reason for imagining novel hazards and coming up with hedges against them. I think it makes sense to spend, say, a few billion dollars a year monitoring space for extinction-threatening asteroids. It's worth doing extensive testing on GMOs** to reassure people who believe they're harmful, even if we think we understand their nature well enough to rule out any proposed harms. Nature is full of surprises, for sure. So we should be setting aside resources to limit our "tail risk." We're materially endowed enough to indulge a lot of this kind of hedging, and I think we should take it seriously. What we can't do is simply posit that some existential harm might exist and then, "given the very high stakes", marshal all of society's resources to combat it. I hear this kind of argument in climate debates all the time. (The debate, by the way, is almost never over whether the change is real, but how extreme it is, how harmful it's likely to be, and to what degree any proposed mitigation is likely to pay off.) In this Econtalk, Daniel Botkin calls this a "child's game", where I can just make up a scenario that's more extreme than the one you made up. We shouldn't take wildly speculative risks seriously; we should try to quantify them and hedge against them with a prudent use of our limited resources. I should also say that this silly game of "making up hazards" is symmetrical. Should we spend 10% of our material resources mitigating climate impacts? What if that slows economic growth such that we're not equipped to repel an asteroid strike 100 years from now, whereas under the counterfactual (in this case, spending significantly less than 10% on climate hedging) we're fully equipped? There is no such thing as "the precautionary principle" when mistakes in all directions are costly. We can't just play this silly game of "a hazard exists because I say it might exist."
We can play this "child's game" with the vaccine, too. One can simply state that the vaccine has "unknown potential long-term harms." It's unfalsifiable. We can reason about such a proposition. We can explore what we know about how the immune system works, and why that makes any mechanism for long-term harm likely or unlikely. Aren't some of the more dangerous effects of Covid-19 caused by the body's immune response to it, rather than being caused by the virus itself? (The "cytokine storm" was a major news story for a hot minute a while ago.) Doesn't a vaccine stimulate the immune system in a way that simulates what the live virus would do? Now, it's also true that your body's bout with the live virus would be longer lasting and invoke a more extreme immune response. At the same time, let's not pretend we can use first principles to rule out unknown harms. Nature is indeed full of surprises. But if we follow this rabbit hole too far, we will become paralyzed with fear and lose any ability to make informed decisions. In this recent Dark Horse podcast, Bret and Heather discuss the "unknown unknowns" question regarding the virus itself and the vaccine. The discussion is at times insightful. They discuss the prospect of monitoring side-effects of the virus or vaccine as they emerge, and then monitoring and quantifying the specific side-effects that become known. All of this is good, and we should certainly be monitoring and quantifying known or likely hazards. But one can carry the "we can't rule this hazard out" argument too far. (There is some very interesting stuff in the Dark Horse episode about the delivery mechanism of the vaccine and the mechanics of how it "trains" your immune system.)
I have heard that some "asymptomatic" cases resulted in lung damage, which showed up on an x-ray. I have several reactions to this. First of all, "asymptomatic"? Really? Does that word mean what you think it means? I'd call lung damage a pretty serious symptom. On the other hand, if it's going unnoticed by the patient, I'm skeptical that there's material harm, even if something shows up on a scan. I'm also not sure how they connected the lung damage to Covid-19, versus some other cause. I know that doctors will sometimes use x-rays or other scanning technology on someone suffering chronic back pain and find a bulging disk or some other identifiable condition. The problem is that they will find similar insults to the spine doing the same scans on samples of healthy people. So it's hard to causally connect these conditions to the back pain. I don't know if something similar is happening here, but I do know that causal inference is hard to tease out.
I am prepared to do an about-face on this one if something more concrete emerges regarding Sars-Cov-2. But for now I am drawing a line in the sand against Pascal's Mugging. The prospect of unknown long-term harms exists for everything: other novel viruses, new medications, vaccines, new chemical additives in our food, different dieting patterns (as in following the latest dieting fad), spending nine months in relative isolation from family and co-workers, etc. Treating unknown hazards as decisive in any cost-benefit calculus puts us all in straight-jackets. On the other hand, if anyone is doing informed speculation based on known or likely harms, I applaud them. But I want them to show their work.
Death Certificates Often Contain Errors
There have been plenty of dumb conspiracy theories about the virus, so I understand why reasonable people are on edge about this stuff. But the low-quality arguments proffered by conspiracy theorists have at times led reasonable people to overstate their case.
Early on, some people were commenting on the death certificates, the issue of "deaths with" versus "deaths from" covid, the possibility of a doctor wrongly attributing a cause of death (in which case the death certificate really would imply that Covid-19 was the cause, but would perhaps be in error). I initially (early March) thought this could be a big deal. But then the excess deaths data started coming in. Covid-19 is deadly enough to make the all-cause mortality higher than it should have been, comparing the same week from prior years. I wrote about that here. And I noted here that our worst flu season in recent memory, 2017-2018, shows a spike in overall deaths. So that was a definitive nail in the coffin of the "cause of deaths are just being misidentified" hypothesis. Also, the experience of places like New York and northern Italy were inconsistent with this virus being some kind of mass hysteria. It was clear by then that this really was a novel and dangerous threat.
On the other hand, I saw some dumb responses on social media. "Oh, so your judgment is better than a doctor's? So you, as a keyboard warrior, are overruling a doctor's judgment?" This is silly, because doctors often disagree about a cause of death. (See studies here and here; I quote them in this post.) Misidentifying the cause of death is common. It is entirely possible that a cause of deaths gets misidentified in a systematic way. It is possible for this to lead to a spurious trend in the official statistics, and for the "public health" establishment to mount a wrong-headed response to this trend. We need to be alert to this possibility. We shouldn't dismiss it just because it happens to be an arrow in the quiver of conspiracy nuts.
There are plenty of other poor quality arguments regarding the virus. I'm sure I've indulged some myself. But these are the ones that have been bothering me lately.
________________________________
*I actually did come home one day to find that a toilet had been flushing all day. The chain that connected the flapper valve to the handle came off and blocked the flapper valve from sealing. Luckily, it wasn't blocked or anything. It could have been worse. It was the downstairs toilet, which relies on a pump to actually dispose of the water. That pump burned out a few weeks ago and we had to replace it. If it had burnt out the same day that the chain came undone, it would have been a significant flood event in my basement. So in my case this is a known hazard that I worry about and have to occasionally check on. But I shouldn't break out in cold sweat if someone suggests a novel hazard might exist, nor should I drop whatever I'm doing to go check on it.
**On GMOs and tail risk, Nassim Taleb went on a ridiculous tirade against them a few years ago. The flavor is captured in this Econtalk. He posited that they might be an existential threat, "ecocide" is what he called it. He insisted that he didn't even have to come up with a plausible mechanism by which this could happen. (Seriously!) Many experts on GMOs insisted that it couldn't happen, and gave various reasons based on how these things are tested and how they work. Some pointed out that a similar mechanism happens in nature, with long segments of DNA from one species being transported to another (by a virus, as it happens). Taleb had an analogy ready involving a roulette wheel maker and a statistician. Something to the tune of: the statistician knows something about the nature of the roulette wheel that the mere craftsman doesn't understand. It's kind of ironic, because roulette wheel makers are well aware of the statistical properties of the game and are very careful to ensure the wheel is well-balanced. Any imbalance will be detected and exploited, and the results could be ruinous for a casino. If you're trying to quantify tail risk, talk to experts. Ignore random hypotheses that "X might happen", same as you would ignore someone asking if your toilet stopped flushing.
No comments:
Post a Comment