Monday, January 18, 2021

Lessons from the Pandemic

It's generally recognized that our (American) response to the Covid-19 pandemic was disastrous. But I think far fewer appreciate the full scale of the disaster, or the most significant causal levers by which the worst effects could have been avoided.  (Yes, Trump was bad.  But his public health disinformation and politicization of masking -- while obviously bad -- may prove relatively trivial compared to the mammoth failings of our public health institutions and medical establishment.)  Much of the pandemic's harm could have been mitigated had our institutions been properly guided by the most basic norms of cost-benefit analysis.  Consider:

(1) The dangers of blocking innovation by default.  In ordinary circumstances, the status quo is relatively safe and so untested medical innovations present asymmetric risks.  That is, until they are proven safe and effective, it may be reasonable to assume that the potential risks of an untested product outweigh its potential benefits, and so block public access to such products until they pass stringent testing requirements. (There are arguments to be made that FDA regulations are excessively onerous even in ordinary circumstances, but I remain neutral on that question here.  I take it that there is at least a reasonable case to be made in the FDA's defense ordinarily.  No such case for the FDA's stringency seems possible in a pandemic.)

A pandemic reverses the asymmetry of risk.  Now it is the status quo that is immensely dangerous, and a typical sort of medical intervention (an experimental drug or vaccine, say) is comparatively less so.  The potential benefits of innovation likely outweigh the potential risks for many individuals, and vastly so on a societal scale, where the value of information is immense.  So the FDA's usual regulations should have been streamlined or suspended for potential pandemic solutions (in the same way that any ethics barriers beyond the minimum baseline of informed consent should have been suspended for pandemic research).  This should be the first thing the government does in the face of a new pandemic. By blocking access to experimental vaccines at the start of the pandemic, the FDA should be regarded as causally responsible for every Covid death that is occurring now (and many that occurred previously).

Just think: if any willing member of the public could have purchased themselves a shot of the experimental Moderna vaccine back in the first half of 2020, its effectiveness would have been proven much sooner, and production and distribution ramped up accordingly, bringing about an end to the pandemic many months sooner than we will actually achieve.  The sheer scale of the avoidable harms suffered here is almost impossible to over-state.  (If the FDA managed to prevent a Thalidomide-scale disaster every year for several decades, it still would not be sufficient to outweigh the harm of extending this pandemic by many months.  But of course the real choice facing us is not so "all or nothing".  There's no reason we can't reap the benefits of FDA protection -- if it is a benefit -- in ordinary circumstances, while sensibly suspending policies that are very obviously inapt in the face of a pandemic.)

Of course, we couldn't know in advance which (if any) experimental vaccines would work.  Even so, the expected value of my recommended policy (encouraging experimental vaccination followed by low-dose viral inoculation to confirm immunity) strikes me as clearly positive, even just given what we knew back in March. (If you disagree, please comment there -- and show your working.)  If nothing else, consider how many lives would have been saved simply by requiring immunity certification for anyone working in elder-care.  Providing targeted immunity to high-risk transmission vectors in a pandemic should be an obvious policy priority. I'm appalled that it proved to be beyond our medical policy establishment.

I've focused here on the error of blocking opportunities for early immunity (whether through experimental vaccines or viral inoculation -- and of course the failure to run vaccine challenge trials also belongs on this list), but the underlying lesson applies to many other errors in pandemic policy, including banning early Covid tests (in Feb 2020), banning quick tests throughout the summer, etc.  In future pandemics, the FDA should only be allowed to ban a pandemic-alleviating product after first producing a cost-benefit analysis to justify their intervention.  In a pandemic, innovation must be permitted by default.


(2) Misguided perfectionism: Closely related to the above mistake is the implicit assumption that it's somehow better to do (or allow) nothing than to do (or allow) something imperfect.  Letting the perfect be the enemy of the good in a pandemic is disastrous.  Blocking quick Covid tests for having lower accuracy than slow ones is an obvious example of this form of stupidity.  Deciding in advance that a vaccine must prove at least 50% effective in trials to receive FDA approval is another. (Obviously a 40% effective vaccine would be better than nothing!  Fortunately it didn't come to that in the end, but this policy introduced extra risk of disastrous outcomes for no gain whatsoever.)

Compare Dr. Ladapo's argument in the WSJ that "Doctors should follow the evidence for promising therapies. Instead they demand certainty." (Steve Kirsch expands on the complaint.) Again, this is a very basic form of irrationality that we're seeing from the medical establishment. 

Misguided perfectionism has also damaged the vaccine rollout due to prioritizing complex allocation schemes over ensuring that as many people are vaccinated as quickly as possible. (Some are letting doses spoil rather than "risk" vaccinating anyone "out of turn"!)

More examples are discussed here.


(3) Agency bias. Sometimes the pressure to do nothing seems to stem from inflating fears of potential downside, while disregarding missed potential gains.  Relatedly, we tend to blame people for harms that stem from action (whether performed or allowed), and ignore or downplay harms that stem from inaction (& so are seen as built into the status quo).  While this bias leads to avoiding policies perceived as "risky", I don't call it "risk aversion" because while some risks are inflated, others are irrationally neglected.  It seems closely related to the omission-commission error. Whatever its roots, it strikes me as a very deep-rooted psychological bias, and one that is plausibly behind much bad thinking about the pandemic.

One recent example of this mistake: holding second doses of vaccine in reserve instead of giving out first doses first (and trusting that stockpiles would be replenished in time to provide booster shots before initial immunity waned).  It's easy to imagine how "first doses first" could go wrong.  But it's harder to see how that slight risk could mathematically outweigh the likely benefits of quickly vaccinating twice as many people, in "expected value" terms.  As I previously noted (in relation to inoculation), it's not enough to just flag a potential down-side of a policy proposal.  Every option in a pandemic has downsides.  We need to assess which option is the least bad, or the most promising, in expectation.

Many of the world's problems (e.g. California's wildfires) may ultimately be traced back to a kind of asymmetrically-biased blame-aversion incentivizing a bad status quo over even mildly "risky" solutions that would obviously be worth trying according to a neutral cost-benefit analysis.  Foolish inaction is frustrating enough in normal circumstances.  It is outright disastrous in a pandemic.

(See also my companion post on the epistemic analogue of this asymmetric bias.)


(4) Status-quo bias. This is really just a summary of the previous points.  But I cannot possibly emphasize enough what a mistake it is to privilege the status-quo in a pandemic.  It's just nuts.  Quietly maintaining the status quo in a pandemic kills thousands upon thousands of people, and indirectly harms millions more.  Yet everyone behaves as though it's somehow intolerably "reckless" to even consider unconventional policies that have any potential downside (no matter how disproportionately greater their potential upside).  Meanwhile, the only people I see outraged about the FDA's obstructionism are libertarians who are always outraged by the FDA.  How is it not obvious to all that obstructing medical progress is the single greatest threat in a pandemic?  (If only this could inspire a fraction of the outrage that was directed at ordinary people for going to the beach...)


(5) Other failures of cost-benefit analysis.  This all seems to come down to a failure to even attempt a proper cost-benefit analysis.  This failure also took other forms.  One of the most striking involved the blind prioritization of physical health over social, economic, and mental welfare.  One saw this in the commonly-voiced idea that it was somehow "indecent" to question whether lockdowns might do more harm than good all things considered, for example.  (Not to mention the Covid "security theatre" of closing parks!)  N.B. I'm not here claiming that lockdowns were all bad.  I'm claiming that cost-benefit analysis was needed to answer the question, and it's bad of people to deny this.

Aside: I'm aware of studies showing that people are biased against experimentation, and more risk-averse for others than for themselves (both via Marginal Revolution). I'm sure there must be similar studies on what we might call "health bias" -- prioritizing quantity over quality of life, and direct physical health threats over indirect effects on welfare -- any pointers would be most welcome.


(6) Epistemic obtuseness. The medical establishment's demand for certainty (mentioned under #2 above) is one kind of epistemic obtuseness. There are others worth mentioning.  A big one is the assumption that we can have "no evidence" for P until a trial specifically testing for P has been conducted. (As Robert Wiblin jokes, "We have no data on whether the Pfizer vaccine turns people into elves 12 months after they take it.")  The WHO trumpeted that there was "no evidence" that Covid antibodies conferred any immunity, when in fact we had perfectly good (albeit uncertain) evidence of this based on (i) what we know of similar viruses, and (ii) the absence of large numbers of confirmed reinfections that we would expect to see after X months of a raging global pandemic if recovery did not confer any immunity whatsoever. (The latter evidence obviously got stronger the larger the value of X became.)


(7) Mundane practical failures.  This post has focused on what I think are philosophically interesting lessons from the pandemic -- mistakes that stem from systematic biases in our thinking, for example.  There are more mundane errors too: failures to plan for the logistics of vaccine distribution, and other "maddeningly obvious" stuff.  As a philosopher, I don't have any special expertise to add there, but readers are very welcome to contribute in the comments with whatever they take the biggest mistakes (and associated lessons) of the pandemic to be...


UPDATE: See also my follow-up post on why, even if you disagree with me (on various empirical details) you should nonetheless agree with me (on the more general need to guard against excessive conservatism in a pandemic).

12 comments:

  1. This is an excellent summary and I endorse every point. And the fact that you're so clearly right on all counts and we're not freaking out about it is causing me even more distress. I would also add that this problem goes far beyond the USA. The vaccine we're all hoping to receive was created in one weekend in February 2020, and yet no country decided to move straight to scale manufacturing in parallel with challenge trials. (If they had, they would now can have been done with COVID for months.) From inside the US we see all the derping our country has done, but then we look out and notice that competent countries like Switzerland and Japan are not ahead of us in putting an end to COVID. If anything, their derping around is even harder for me to explain than ours is. We basically all shit the bed. And I'm starting to worry that this is a sign that global leaders and elites have transcended all national boundaries, and this allows a global groupthink to lead ever country to make the same mistakes at the same time. There can't be the historically usual local experimentation. On the margins Sweden was a bit more lax and Russia was a bit quicker with vaccine approval, but even regarding each of those small response variations there was a freakout of "mainstream" global outrage. Isn't this a sign that we created a global epistemic and ethical orthodoxy? They are clearly capable of some terrible decisions, but if they get even worse, where can we go to escape them? North Korea? Mars?

    ReplyDelete
    Replies
    1. Yes, I take your point that it's both striking and unfortunate that there wasn't more "local experimentation" / variation on the global stage. (Quite apart from the Millian epistemic benefits of "experiments in living", there's also the important point that circumstances can differ -- e.g. economic costs are vastly graver in developing countries -- which can make a big difference to which policies would actually prove optimal in different localities.)

      Delete
  2. On "health bias", Robin Hanson's paper "Why health is not special" may be of interest (though you may be aware of it already).

    ReplyDelete
  3. In thinking about releasing experimental vaccines earlier, one factor that does need to be considered is that COVID isn't a one-shot game. Suppose that there had been large scale use of an experimental vaccine that turned out to have some serious complications. How much would that have bolstered the anti-vax movement in the country, and how many future deaths would result from the greater public reluctance to make use of safe and effective vaccines like MMR? The calculus may still come out in the direction you say that it does, but the math may be more complicated.

    ReplyDelete
    Replies
    1. Yes, quite right. I would hope that we could find ways to mitigate that risk -- an explicitly labelled "UNTESTED, EXPERIMENTAL" vaccine shouldn't, in principle, be taken to reflect on thoroughly tested and proven vaccines. Alas, anti-vaxxers aren't known for their comportment with reasonable epistemic principles. Perhaps the necessary separation in the public mind could be enhanced by, e.g., officially denoting all experimental treatments as part of a "clinical trial" still, not a "public" release -- it's just that any member of the public is allowed to join this ever-expanding "trial". It's a problem worth considering, anyway, I agree. (I just hope that people don't treat such risks as a reason to dismiss accelerationist proposals out of hand!)

      Delete
  4. [Adam Omelianchuk writes:]

    I suppose I will register my level of disagreement about each of the points.

    First this: "So the FDA's usual regulations should have been streamlined or suspended for potential pandemic solutions..."
    This can be a double-edge sword. I know challenge trials are in view here, but more is (and was!) encompassed by the"usual regulations" that *were* loosened. For example, the EUAs that bypassed the requirement of a randomized-control trial to prove efficacy were given for hydroxychloroquine and convalescent plasma. The evidence standards were lowered so patients could get access and try those things for "expected utility" reasons. It took longer than it should have to find out that neither offer a statistically significant benefit. A lot resources were wasted on these things because the regulations were streamlined.

    What is said above is enough to undermine the second claim that "Letting the perfect be the enemy of the good in a pandemic is disastrous." We should at least it raise the question as to what we are talking about. We can't assume the resources it would take to make a vaccine that is 40% efficacious would be worth it given that we now have one that is 95% efficacious and it is *very* difficult to get medical supplies, personnel, and distribution networks in place to get shots into arms, much less in the arms of those who need them the most. Nor can we assume that such a vaccine would be in demand especially in a culture that is increasingly distrustful of vaccination (and the reasons for distrust are complex, sometimes stupid, sometimes not so stupid [especially if efficacy is low and side-effects are nasty], racialized, and not easily overcome). Nor do we have any good reason to trust that current "stockpiles [meant for a 2-dose regimen] would be replenished in time to provide booster shots before initial immunity waned" (yes, this is in the third point, but it is relevant). Indeed, as one virologist remarked, doing what the author recommends here sets us up (perfectly?) for an evolutionary process that would serve to *strengthen* the virus. Nor is 95% efficacy "perfect" (100%?) so I take this point to be a strawman at best.

    I pass over the third point, because I am not sure I understand the argument (and this may be my fault -- it has been a long day).

    The fourth point is better illustrated in people not wanting to mask up and socially distance which are far cheaper, easier, and proven methods for curbing infectious disease (see the current influenza numbers for example) than doing scientific studies for creating a vaccine, even if with altruistic volunteers.

    Fifth, I agree but also think this is more complicated. Public health officials have been undisciplined in how they reason, shifting between a precautionary approach and a cost-benefit approach. In my view, the precautionary principle weld sway, especially in the first months of the pandemic. I don't think they were even doing cost-benefit analysis up until the protests over racism began (see the open-letter in support of the protests).

    Sixth, see what I said above in the first two points.

    Seventh, I get the passion here, and largely agree.

    I can't presume to say what our "biggest mistake" was... right now I am thinking it was how we handled testing in the beginning of the pandemic. Small mistakes like that at the beginning turn into huge nightmares later on. Also this vaccine distribution thing is killing me (well not really, I was lucky to get 2 does because I work for a hospital, but still).

    - Adam

    ReplyDelete
    Replies
    1. Hi Adam, sure, there are downsides to streamlining. But you haven't given any reason to think that these come anywhere close to outweighing the benefits of speeding the vaccine rollout by many months.

      I'm curious about the suggestion that this had specifically epistemic costs -- taking "longer than it should have to find out that neither offer a statistically significant benefit". Why did this take longer? Did they stop doing RCTs once emergency use was granted? If so, that seems a mistake. Care should be taken to streamline the rules in a way that supports (rather than undermines) data acquisition. Initial access might only be offered by means of (widespread!) clinical trial participation, for example, with randomization and associated data collection, etc., as needed. Do you see some insuperable difficulty here?

      "We can't assume the resources it would take to make a vaccine that is 40% efficacious would be worth it given that we now have one that is 95% efficacious"

      Huh? I'm not "assuming" any such thing. Just pointing out that if we didn't have better alternatives, a 40% vaccine could be very welcome indeed.

      To state the obvious, you do not need to blindly pre-commit to rejecting a 40%-effective vaccine in order to make sensible prioritization decisions when the time comes, based on whether or not better options are available.

      "Nor do we have any good reason to trust that current stockpiles [meant for a 2-dose regimen] would be replenished in time to provide booster shots before initial immunity waned"

      Here's a very simple one: vaccine production is ramping up, not ramping down. Rather than just brutely asserting ignorance, it's more helpful to assess probabilities: just how likely is it that FDF (now official Biden policy, btw) will result in waning immunity? (Decompose into two questions. Firstly: What's the best-justified probability distribution regarding how long first-dose immunity lasts before waning? And second: How likely should we regard it that a problem in vaccine production leaves us unable to provide booster shots before that date?) Next, just how bad would that be, compared to having half the population be entirely unprotected? What's the likelihood that your virologist's nightmare outcome of vaccine-resistant mutation occurs? Again, I want to see numbers here, not worst-case-scenario scaremongering.

      Delete
    2. [Adam asked me to post the following on his behalf: part 1/2]

      > "sure, there are downsides to streamlining. But you haven't given any reason to think that these come anywhere close to outweighing the benefits..."

      No, I did not, and I did not intend to. I am sorry I was not clear about that. I chaired an IRB and have a good sense of what the Common Rule would allow and have kept an open mind about challenge studies. I think there are real concerns about the long-term adverse effects of COVID-19 even in young people and am disturbed by the hand-waving proponents have towards these risks. Nontheless, current day regulations do allow for a reasonable balance between benefits to subject + benefits to others against the risk to the subject. I think the altruism displayed by people like Josh Morrison and others is admirable and if a study could be designed with appropriate safety measures in place, I would seriously consider voting in favor of it. I would be worried about some who have conditions like type 2 diabetes not disclosing their medical histories because they want to be a part of it, but I am not dismissive because of this type of worry. There are ways to screen for this stuff. I also worry that the benefit to science and society might be quite limited. Since they are designed to recruit only young, healthy people, they will likely discover information only about how the vaccine affects young, healthy people, and it would seem that we would want, for reasons of justice, information for how they might affect the elderly or those with diabetes. I suspect the reply will be that including this concern for justice is an instance of "the perfect being the enemy of the good", the good being that a vaccine that is safe and (somewhat?) effective for young, healthy people is better than nothing. I get what you are saying, but political acceptability matters when justice concerns are involved and it is something we must account for, especially in vaccine development and distribution. Public trust is very important and a good we need for the long-term.

      In any event, my comment was a more general point about being careful about what regulations we are weakening and why, something I will explain in the next comment.


      > "Did they stop doing RCTs once emergency use was granted? If so, that seems a mistake..."

      Yes, EUAs do not require RCTs to be completed. I don't want to say the difficulty is absolutely insuperable but the problems are much more difficult to solve than we realize. The problem is that emergency use is just that: the right to use the agent under "emergency" conditions. It is an impotent right if its use cannot be authorized until the test agent has been studied properly which is contrary to the interests of those who assert the right to try. What has happened is that "individuals" -- in the form of doctor-patient dyads -- have pursued their self-interest by weakening regulations so as to make themselves better off, and, as a result, made us all worse off because of the degradation of scientific standards. It is a tragedy of the commons! Hence, your recommendations above are actually ways to strengthen regulations, not weaken them. I think we can find some agreement here. Perhaps this recent proposal should be seriously considered (though see commentaries).

      [continued...]

      Delete
    3. [Part 2/2 of Adam's reply...]

      > "Just pointing out that if we didn't have better alternatives, a 40% vaccine could be very welcome indeed."

      Fair enough.


      > "you do not need to blindly pre-commit to rejecting a 40%-effective vaccine in order to make sensible prioritization decisions when the time comes..."

      Well, I wouldn't want to do that. Here's what I was thinking. Let's suppose the vaccine manufacturers each faced a choice: Make a 40% effective vaccine for everyone (let's put aside the "safe only for young people" concern above) that becomes available in June or a 95% effective one in December. Why think it would be better to opt for the 40% one when a few months later we can get 95% one? The resources needed to get the 40-percenter into arms would take away from the resources to develop and distribute the 95-percenter, and it certainly isn't obvious that would be the best way to go. Is there a statistical model that decisively shows that opting for the 40-percetner would be the way to go? I get that the manufactures could not have known that they could get a 95-percenter by December (no one did!) and would have to make the decision under uncertainty. Should they have been guided by precautionary reasoning and opted for the 40-percenter? As a matter of ethics, it may be best for one manufacturer to make the 40-percenter and another to make the 95-percenter, but from a business perspective no manufacture is going to intentionally spend resources on creating an inferior product that will be replaced by a superior one some months later (unless one wants to go out of business).


      > "vaccine production is ramping up..."

      It is ramping up (good!) but the bottlenecks in our distribution system stubbornly remain (bad!), which can lead to vials not being used (very bad). I am not optimistic that these bottlenecks will be fixed in time so that one doesn't have to wait 9-12 weeks between shots. I suspect these waiting times will be longer in the places that score high on the Social Vulnerability Index. As for Biden's policy, the latest news says, "Though Mr. Biden has indicated his administration would release more doses as they became available and keep fewer in reserve, he said last week that he would not change the recommended timing for second doses: 21 days after the first dose for Pfizer’s vaccine, and 28 days for Moderna’s." Fauci counseled as much 2 weeks ago. This is because we simply don't know what sort of immunity one-does-only provides and for how long; it wasn't tested, so we can't even assume it's even 40%. Hence, it's hard to offer the kind of probability assessment that you demand. I don't mean to be "scaremongering" and I take your points about being overly cautious, but given what has already happened in the pandemic, the worst-case scenario does not seem remote to me. If I find out more about the sort probability calculus that you demand, I will let you know. In the meantime, however, we should think seriously about reserving two doses for those that live/work in congregate living situations (nursing homes, group residential housing, prisons) rather than using all of them up at once and hoping more will come in time.

      I get the sense that our disagreements involve attitudes of optimism and pessimism here (not the only issues!), and that is interesting. Thank you for this thoughtful and challenging exchange. It helps to re-examine my pessimistic biases.

      - Adam

      Delete
    4. My response [part 1/2]:

      > "I think there are real concerns about the long-term adverse effects of COVID-19 even in young people and am disturbed by the hand-waving proponents have towards these risks."

      I appreciate your subsequent comments indicating that you're open to considering challenge trials. But this quoted remark exhibits just the kind of moral conservatism I'm criticizing. I grant that Covid is risky for everyone. No "hand-waving" here. My point is that in a pandemic where that risk is already widespread, the seriousness of the virus cannot POSSIBLY be a reason against challenge trials (with willing volunteers). Quite the opposite. The worse the risk from the virus, the MORE IMPORTANT it is to get a vaccine ASAP. I would add, in light of worries about demographic variation (if experts judge this info to be sufficiently important -- if there's a significant chance, based on prior knowledge of other vaccines, that a vaccine might work only on the young but not for the old, for example), that challenge trials may ethically involve volunteers from any demographic, including the elderly and other high-risk groups. Some might die as a result of participation, and it's of course crucial that any participants be fully aware of this, and willing to risk their lives in this way. But the alternative is VASTLY MORE people dying EVERY DAY as we are now seeing -- victims, mind you, who did NOT volunteer. Given a choice between risking a small number of volunteers, or a vastly larger number of non-volunteers, it's a moral atrocity that conventional medical ethics defaults towards endangering the latter. However "disturbed" we may be by the former, we ought to be vastly more disturbed by the latter. Sadly few people seem to appreciate this basic moral point.

      (Peter Singer & I made this argument in greater depth in our 'Pandemic Ethics' paper.)


      > "emergency use... is an impotent right if its use cannot be authorized until the test agent has been studied properly which is contrary to the interests of those who assert the right to try."

      I was thinking it could be authorized more widely as soon as the trials have a sufficient number of participants -- you do not need to wait until the results of the trial come through. Prior to the trials having enough participants, the "right to try" should be understood as the right to try through participating in a trial. I absolutely agree with you that it's crucial not to undermine data-gathering efforts. So in that sense, yes, I'm certainly not blanket opposed to regulations! The aim should be to set up a regulatory framework that gets us vital info ASAP. To that end, it can make sense to regulate how people can access an experimental treatment -- to ensure randomization, tracking, etc. So they certainly don't have a right to receive an unproven treatment rather than placebo, as that would undermine the science. But nobody should be denied so much as a chance of experimental treatment: whether through participation in a placebo-controlled trial, or -- if the trial already has enough participants -- through direct access (which should also involve tracking outcomes for the patient -- useful info in the aggregate even if not placebo-controlled).

      Delete
    5. [My response, part 2/2]

      > "Well, I wouldn't want to [blindly pre-commit]"

      Good! Sounds like we are agreed on that point then, since that's the only thing I was criticizing in relation to the 50% efficacy threshold. (As you say, "no manufacturer is going to intentionally spend resources on creating an inferior product," so there's little risk of wasted resources as a result of approving a 50% vaccine. If better options are available, the inferior approved products simply won't be invested in. You don't need to legally ban inferior products in order to avoid having them wastefully take over the world. The default outcome for unwanted products is that they are not produced.)


      > "[Biden] would not change the recommended timing for second doses..."

      Is anyone advocating for changing the "recommended timing"? I mean, the recommended timing to receive a first dose is ASAP. If we don't have enough doses to meet both recommendations, we don't need to revise our ideas of what is recommended (as ideal); we just need to prioritize or decide which recommendation is most important to meet, first. So. As I understand it, there are two issues: (1) whether to hold doses in reserve, and (2) whether to prioritize correctly-timed second doses or giving out more first doses. I think that (1) is most clearly criticizable (at least as a general rule -- there might be special cases in which it is defensible), and that for (2) it also probably makes sense to prioritize first doses -- but of course I'm not a medical expert, so I'd want to hear more about the track records of other vaccines.


      > "This is because we simply don't know what sort of immunity one-does-only provides and for how long; it wasn't tested, so we can't even assume it's even 40%. Hence, it's hard to offer the kind of probability assessment that you demand."

      As per my blog's point 6, direct testing is not the only relevant evidence. What proportion of other vaccines stop working if not boosted within a month? Two months? etc. This is all relevant evidence, and worth taking into account rather than flying blind. Crucially, without considering such evidence, there is no basis for thinking that prioritizing second doses is important. In the absence of evidence, it's just as much a guess as prioritizing first doses.


      > "I get the sense that our disagreements involve attitudes of optimism and pessimism here (not the only issues!), and that is interesting. Thank you for this thoughtful and challenging exchange. It helps to re-examine by pessimistic biases."

      Absolutely! One of my main points [#3] is that I do think that it's common to respond to uncertainty with a kind of pessimistic risk-aversion (treating unknowns as a decisive reason not to try something different or unconventional). And in other circumstances that can make sense. But I really think that in a pandemic, this can end up increasing the objective level of risk we face, because the status quo is so inherently risky: it could be worth taking some gambles to try to avoid thousands more daily deaths. Of course it depends on the details. Not all gambles are worth taking, for sure! But given this standard psychological bias, we've some reason to think that we most likely need to "recalibrate" in a less-conservative direction.

      Delete

Visitors: check my comments policy first.
Non-Blogger users: If the comment form isn't working for you, email me your comment and I can post it on your behalf. (If your comment is too long, first try breaking it into two parts.)