Book Review: Vaccines, Part 2

Critical reflections on equilibrium effects, collective action problems, bias, and satisficing

Duality: Our bodies give us (and sometimes others) life. Yet, they also carry the seeds of their own destruction. Cancer cells vie to seed latticeworks of lymphatic tissue. Chinks in the immune armor fall down, parapet by parapet. Calcium filters out of bones and accumulates in arteries.

It’s sometimes said that civilization, too, contains the seeds of its own destruction. Karl Popper warned in The Open Society and Its Enemies that liberal democratic societies’ very tolerance — including to those who are intolerant — could destroy them in the paradox of tolerance. J.S. Mill warned about the tyranny of the majority, which might be seen as both subset and inverse of that problem. But numerous others, including Hannah Arendt in The Origins of Totalitarianism and James Madison (as Publius) in Federalist 41, have warned that overly zealous security measures could similarly destroy the very fabric of the society they’re ostensibly meant to protect. So it seems society is caught between the rock of liberty and the hard place of security, ever struggling to strike the right existential balance.

Vaccines sit at the intersection of these shared physical and sociopolitical vulnerabilities. And so they make us rightly afraid. Afraid because fully preventable death and disease are scary: It’s scary when people go unvaccinated against highly contagious, deadly diseases for which we have safe, effective vaccines. And it’s scary when corruption and state overreach threaten our health, too — as when we don’t have good vaccines, but are encouraged (or even required) to take them, anyway.

This fear, in turn, makes us stupid. Not that human beings need much help with that. But that stupidity doesn’t determine which norms or considerations govern behavior. Fear can produce cooperation, as in deterrence (compliant avoidance) as well as defection, as in evasion or sabotage (adversarial avoidance). These patterns belong to the realm of strategic behavior.

In both contexts — epidemiology, where bodies are on the line, and politics, where hearts and minds are — it’s best for everyone when society rests in a stable and cooperative equilibrium, the system equivalent of individual-level homeostasis. But how do we get there? Once knocked out of it, how do we stay there? Where do equilibrium effects come from, and why do we need to think about them in relation to vaccines?

Equilibrate Me, Baby

Feedback loops implicate equilibrium effects, requiring in methods terms that we leave aside causal diagramming in acyclic graphs (DAGs) in favor of directed cyclic graphs (DCGs). They are perhaps best-known from the famous dating-as-strategic-behavior scene in “A Beautiful Mind,” illustrating John Nash’s insight about Nash Equilibria, meme-ified here:

The prisoner’s dilemma is the classic textbook case, where upholding cooperative norms accrues mutual benefits, while defecting incurs mutual costs — provided both players choose the same strategy. It’s modeled as a static concept; the point is that the system tends to stability at one of two equilibria (cooperation or defection). But it starts to get us into feedback loop territory, because something (temporal and dynamic) gets us into and out of those equilibria in real life. For that something, we often see differential equations or DCGs modeling how trust, risk, and compliance fluctuate over time. Why do we care about this now?

Measles vaccination presents a textbook case of equilibrium effects. Thus, my recent adventures reading Deer’s and Gøtzsche’s relevant books. Aside from being curious about the context of the recent, fully preventable child measles deaths in Texas, and wanting to get things right for my family, I need to know more about equilibrium effects in order to wrap up groundwork for the apocryphal proposal. I need to know how to think about modeling strategic behavior feedbacks in mass security screenings, and how they stack up with other causal pathways, even though I’m afraid part of the answer may be indeterminacy: Looking at the whole picture to assess net effects may crash the causal modeling system, because you can’t graph totalitarianism. But hopefully there is a way to model net effects of proposed mass security programs anyway, to assess if they actually advance or undermine security.

These dynamics aren't just theoretical or relevant to niche interests like polygraphs and mass surveillance. They have serious practical consequences for public health and beyond — as well as for Gøtzsche’s orientation to vaccine evidence and policy. Structurally, measles vaccination is a public goods game with a cooperative equilibrium. If enough people cooperate, herd immunity protects the entire society. But if just a small fraction (5%) defect (by not vaccinating), it undermines the whole system, destroying that equilibrium.

This makes civilization look very fragile: A small minority can critically endanger the whole. Can it be true? And, if it’s true, what can we do about it?

Let’s look at the feedbacks…

Feedbacks in measles vaccination

My child’s vaccine’s efficacy depends partly on our neighbors’ vaccine uptake. This suggests vaccines involve feedbacks.

We can envision four feedback loops:

• In the herd immunity (negative feedback) loop, the feedback is self-limiting or stabilizing. As vaccination increases, disease decreases, which can eventually reduce perceived threat, lowering uptake because, hey, why risk any harm from a vaccine that your child doesn’t benefit from? No vaccine is perfectly safe, and my child’s health is paramount! (Don’t be a free-rider.)
  
  The answer to “why risk it?” is: Because, if you don’t (if you defect), then outbreaks happen. Risk of infection, severe disease, hospitalization, disability, and even death increases. Vaccine hesitancy decreases, uptake increases, and the loop dies out again. Success breeds complacency, which can undermine success — but the loop is self-regulating, because the social reality of fully preventable death and disability is harsh and often highly visible.
  
  In other words: Vaccination Rate → Herd Immunity → ↘ Infection Risk → ↘ Outbreaks → ↘ Perceived Infection Risk → ↑ Vaccine Hesitancy → ↘ Vaccination Rate
  

• Sound bad? Don’t worry; it gets worse. The misinformation / disinformation feedback loop is more dangerous to public health, because it’s self-reinforcing instead of self-limiting. In this loop, myths matter. These myths include misinformation like Wakefield’s Lancet MMR-autism fraud, official misinformation like what Gøtzsche shows the CDC issues about the flu vaccine and I’ve shown the UK’s Independent SAGE issued about Covid vaccination — and hostile state disinformation including Russian vaccine propaganda in Ukraine. In addition to showing people the consequences of not vaccinating, outbreaks can generate noise and polarization, which can fuel more mistrust and misinformation. So maybe the negative herd immunity loop is not self-limiting in the current information environment, after all. Not when it’s combined with the misinformation/disinformation feedback loop we see here:
  
  Misinformation → ↑ Vaccine Hesitancy → ↓ Vaccination Rate → ↓ Herd Immunity → ↑ Infection Risk → ↑ Outbreaks → ↑ Misinformation
  

• The reactive outbreak loop shows how outbreaks can cause a short-term correction to a system set at a disadvantageous equilibrium by scaring people into vaccinating. But this feedback only lasts as long as people remember the pain, or until new myths enter the system.
  
  In other words: Outbreaks → ↑ Media Coverage → ↑ Perceived Risk → ↓ Hesitancy → ↑ Vaccination Rate → ↑ Herd Immunity → ↘ Outbreaks
  

• Finally, we should note the virtuous feedback loop: a cooperative norm can self-sustain, because mass vaccination generates mass (herd) immunity, and trust in the system endures.

  
  That means: Vaccination Rate ↑ → Herd Immunity ↑ → Infection Risk ↓ → Perceived Infection Risk ↓ (but trust in the system and norms endure) → Vaccine Hesitancy ↓ → Vaccination Rate stays ↑
  

  This equilibrium can persist for long periods. But it depends on sustained cooperative behavior.
  

These cyclic causal logics have a few implications:

• The gold-standard Harding Center Fact Boxes and other linear effect estimates for vaccine risks and benefits are importantly wrong. They fail to account for these feedbacks. It matters in the net assessment that I don’t just benefit from my vaccination; I benefit from your vaccination, and vice-versa.

• For public health, and eventually also for economic and national security reasons, we need to more fully develop a societal framework for combatting myth with fact when it comes to vaccines. Simplistic but true: dead kids lose a lifetime of earnings, and hostile state disinformation shouldn’t contribute to those losses. (This seems relevant to political propaganda, too, in light of the recent reported death of the CIA Deputy Director’s son in Ukraine fighting as a Russian enlistee.)

Developing that framework is exactly the project Peter C. Gøtzsche undertakes in Vaccines: Truth, Lies, and Controversy (Skyhorse Publishing, 2021). In its admirable execution of that important goal, it covers a large amount of valuable ground, but also misses some crucial pieces of the bigger picture. Chiefly, it could have been deepened in its insights by incorporating critical reflections on equilibrium effects, collective action problems, bias, and satisficing.

Do equilibrium effects justify force?

In the view of many modern states, equilibrium effects justify vaccine mandates. Measles is the best case: The highly transmissible virus can cause fully preventable, devastating harms to defenseless infants and children. Luckily, available vaccines are so safe — possibly even generally beneficial — that this looks like a no-brainer at the state level. Setting public health policy doesn’t get any easier.

As a mother, I agree. Measles deaths are rare but stupid. And someone else’s stupidity doesn’t have the right to endanger my baby. Particularly when that someone else may have been influenced by fraud (cf Wakefield) or hostile state propaganda…

Asymmetrical information warfare

Gøtzsche mostly ignores global politics, including both less-developed countries’ (often well-founded) mistrust of Western states and Western-dominated NGOs, and the alleged role of hostile states in disseminating vaccine disinformation in opponent countries (e.g., Russia in Ukraine and Serbia). This seems an odd omission given Gøtzsche’s emphasis on informed consent and awareness of misinformation. Hostile states have an interest in undermining informed consent through misinformation in order to harm adversaries by weakening them with distraction and preventable disability/death.

The omission is especially unfortunate, because Gøtzsche’s larger argument — that we should all review the evidence well enough to make our own decisions about every vaccine — is potentially consistent with what Katrina Keegan writing for Harvard’s Davis Center for Russian and Eurasian Studies called:

the Kremlin’s uniform strategy… to sow information chaos, characterized by a lack of narrative consistency, to make truth seem subjective (Fitzgerald & Brantly, 2017; Global Engagement Center, 2020; Jankowitz, 2018; Singer & Brooking, 2018; Will, 2021).

Misinformation misinformation?

But, maybe that itself is misinformation? Because “data alone say nothing at all” (Sander Greenland). Scientific evidence is often complex, ambiguous, and incomplete. As such, it requires interpretation at every step of the way, from observation and analysis to saying what the results mean (nominal interpretation).

So, maybe the U.S. establishment position that critical thinking about complex realities reflects successful Russian info ops, is itself an Orwellian misframing of critical thinking as Soviet suspect.

This reflects the deeper problem with the misinformation discourse. It is itself full of misinformation, because the designation of what is fact and what is myth in our complex world is political. There is no exit to the problem of non-neutrality.

Bodies are real

At the same time, public health is a game with body counts, hostile states have perverse incentives to knock adversaries out of cooperative equilibria, and letting them do that is not in states’ (or their citizens’) interests.

Also, we are not playing a symmetrical game in this respect. The U.S. is not (as far as I know) running information operations trying to convince other countries’ citizens to mistrust vaccines. (But I am living under a baby and could be behind the times.)

Great to encourage people to think critically. Not so great when the Kremlin is paying for it. This looks like it implicates a latter-day version of the paradox of tolerance: open societies risk being knocked out of cooperative equilibria by asymmetrical information warfare, with life-and-death consequences for their citizens.

Letting people sort this out for themselves might not work well enough to protect my baby. And then what? I’m not offering a solution, just highlighting a problem to which I think eventually thinkers like Gøtzsche must attend in this line of work.

The measles vaccine risk/benefit perception irony

It’s ironic (actually ironic) that the regular live measles vaccine and its combination MMR form appear to have beneficial all-cause mortality effects — while a societal panic continues due to concerns that they may have negative non-specific effects that do not appear to be attributed to them with evidence (e.g., causing autism and/or deaths).

It seems like it might be important to educate people about this scientific evidence. But the authorities’ response is instead to generally demand compliance instead of understanding. Here, Gøtzsche’s insistence that people should educate themselves stands him in good stead on the side of science, transparency, and integrity.

A logical next step would be to then formulate some sort of public policy routinizing this education, or process of seeking out good information sources to educate oneself. If you want people to take a vaccine that they’re scared of because of fraud and misinformation, and that vaccine turns out to actually probably have unrelated benefits, then maybe public health organizations should take up the task of letting people know, or encouraging them to inform themselves about it?

Why do we have to jump, instead, from needing more cooperation (herd immunity) — to use of force (mandates), without thinking about organizing some better science communication in-between? Of course this gap is not Gøtzsche’s fault. But it is a political problem that warrants a political solution. This book is generally short on recognizing the former and proposing the latter.

Collective action problems abound

In the same vein, Gøtzsche doesn’t propose political solutions to political problems. He barely even seems to recognize collective action problems as such. This is a potentially fatal flaw in the project.

If everyone risks unnecessary harm from now-standard vaccine co-administration, then the rational solution isn’t for all parents to read the literature for themselves and take on the substantial extra transaction costs of taking their kids in several extra times for several extra shots, if they can even source singleton instead of combination shots to replace the standard coverage. Rather, the rational solution is for society to find a better standard protocol.

This could include something like delegating doctors’ exclusive vaccine authority to other healthcare workers who then spend a lot of time making home visits on the public dime. Something like this is often proposed and sometimes successful, e.g., at least getting pharmacists vaccine administration authority in some EU countries. But more hierarchical societies like Germany have resisted this change. So I’m still waiting for my vaccine deliverypeople to show up with the latest medical information, shots, and chocolate.

As it stands, avoiding vaccine co-administration like Gøtzsche recommends is currently not feasible due to availability concerns. It would also put a huge burden on parents and patients, a burden that’s bound to fall disproportionately on moms.

Similarly, trying to sequence a live vaccine last in order to benefit from non-specific effects, and trying to do so in a way that is (maximally) safe for disproportionately affected groups (infants, young children, reproductive-age women, people over age 65, and the immunocompromised), seems like a complex problem that individuals shouldn’t have to try to solve on our own.

The same goes for vaccine safety monitoring more broadly. For someone who’s quite clear on how corrupt the medical, public health, and media establishments can be, Gøtzsche seems weirdly naive about trusting existing safety monitoring systems instead of proposing that we establish ones with more transparency and checks. Under current systems, patients themselves don’t seem to be able to check if their harms were correctly recorded, be they reporting alleged adverse events relating to vaccines, drugs, therapies, or other medical interventions. That seems like a recipe for bad safety monitoring that could be structurally improved, so reformers should at least say that.

These are collective action problems, and politics exists to solve those. It’s fine if you don’t believe human societies are very good at that (they’re not). But then at least acknowledge who these failures hurt. In this case, they’re hurting especially vulnerable groups to whom society in general and the medical profession in particular owe a heightened duty of care and effort to do no harm.

Satisficing

I didn’t have time to read this book before starting my kids on the suboptimal, standard German child vaccination schedule.

But it’s not just me. No one has enough time to study everything we need to know in the required depth in order to optimize potentially life-and-death decisions for ourselves and our kids. It’s horrible, but it’s true.

The volume of information implied by the complexity of modern society simply precludes it, dooming us to death by a thousand cuts in the mother of all collective action problems.

That’s also my main criticism of this excellent book. Gøtzsche concludes:

Vaccines differ markedly in their benefits and harms, and for some, it can be questioned whether they are worthwhile using. We therefore need to study carefully each vaccine separately if we want to arrive at a rational decision whether we should get vaccinated or should have our children vaccinated (p. 211).

Satisficing is the more rational strategy than optimizing here, for social as well as limited time and attentional reasons. I satisficed by getting my baby a vaccine she didn’t need (hepatitis B, included in the standard combination shot). I love nerding out in the methods weeds almost as much as I love my babies; I just don’t have the time or energy to fight the might of industry incentives, state-level priorities, and social norms when it comes to making a custom child vaccination schedule for my family in order to minimize extremely small risks of possible iatrogenesis from an unnecessary vaccine. The rational move here is the irrational one of going along with good enough.

It’s also what we want at least 95% of the population to do with the measles vaccine. If we have to work harder as a society to get back into the cooperative equilibrium there, it’s not because everyone needs to read more science. It’s because criminals like Wakefield actively disrupted the system — and walked free to keep doing it.

Bias pervades

Everybody’s got bias, including scientists doing bias research. Intelligent, educated people like Gøtzsche should recognize that, and at least nod to their own perspectives and the limitations they likely generate. But it’s always easier from the outside. Here are a few of the unreflective biases that undercut Gøtzsche’s arguments.

Doctor privilege

As a doctor, Gøtzsche decides what vaccines are right or wrong for him, and acts accordingly. Good for him.

However, not all of us medical doctors, and so not all of us have that privilege. For most of us, negotiating with healthcare practitioners who think they know better than us what we need is part of healthcare.

This wrinkle troubles Gøtzsche’s entire project. So what if I can review the evidence and decide I should have the pneumococcal vaccine for recurrent sinusitis, and the MMR booster for the non-specific effects, since following the current guidance to get the TDaP repeatedly in pregnancy may well have screwed up my already wonky immune system? I still have to convince someone to give them to me. In a world of vaccine hesitancy, strict Prussian separation of duties (“That’s Not My Depahtment”), and hierarchical dominance, that’s not easy. It would be really cool if doctors whose job it is to know this literature actually knew it, or were open to discussing it. But that’s generally not the case. I don’t think I’m the only patient in the world staring at this access chasm. It would be nice if an entire book on vaccines gave the tiniest glint of recognition that most of us have to find a way across it.

Cultural bias

Gøtzsche assumes we need less price signaling instead of more in medicine in order to solve corruption problems, calling for the U.S. to adopt European-style publicly financed healthcare (p. 142). Others argue the exact opposite: more price signaling = less corruption.

This has the appearance of a cultural bias, where the author likes his country (Denmark) and its system (socialized medicine), and assumes it would solve the rest of the world’s problems if they would just be more like him/them.

Anti-lockdown bias

Gøtzsche assumes Covid lockdowns only increased maternal and child mortality in low- and middle-income countries (p. 133). He fails to note that lockdowns also protected some patients from harm from unnecessary medical interventions and their transaction costs. For instance, prenatal care norms often require pregnant women to travel long distances in temperatures that are unsafe for them and their child, disproportionately impacting poor women in places like the U.S. where the weather can be extreme and travel means access is highly socio-economically stratified.

A substantial proportion of pregnant women also experience fatigue, and lockdowns may have led them to stay home more, overexert less — and experience less preterm birth as a result. There were a large number of studies associating lockdowns with preterm birth reductions (see, Calvert et al 2023).

Gøtzsche concludes it’s likely the net loss of life far exceeds the gain from lockdowns (p. 134). But this appears to potentially be a statement about the logical consequences of his priors (lockdowns: bad), rather than reflecting a systematically logical or quantitative assessment about how the different posited causal effects actually stack up. We would expect deaths associated with preterm birth to be particularly high in poorer countries, and good data on these deaths to also be harder to come by.

Anti-masks bias

Gøtzsche sides with fellow Cochrane veteran epidemiologist Tom Jefferson (masks: bad), but mischaracterizes the evidence from Jefferson’s controversial masks review, writing that it “did not find that the use of face masks reduced the occurrence of flu-like illness” (p. 97). Gøtzsche then misrepresents this as “the science indicated they didn’t work,” as did the report’s first author. This is a logical error, as I’ve previously argued. Absence of evidence isn’t evidence of absence, in Altman’s classic formulation.

The review had limitations. For instance, population-level mask mandates in a cultural context where large subsets of people don’t have access to properly-fitting masks or desire to comply, and there are not well-functioning compliance aids (for people who want well-fitting masks but can’t find them) or enforcement checks (for people who don’t) were never going to tell us whether masks worked, or not.

But we’re allowed (even sometimes encouraged) to think about causality in science. If masks do work by mechanically stopping large viral loads from being transmitted in social settings, then this looks like another possible instance of equilibrium effects in epidemiology: When basically everyone does it, it protects the weak (infants and small children who can’t mask; the immunocompromised or ill who may bear disproportionate infection sequelae risks), along with the rest of society. When a relatively small number of people defect, it may endanger the especially vulnerable, along with the rest of the herd. Thus, some prominent pro-mask messaging along these lines, e.g., the Czech PSA “I protect you/you protect me.”

Rather than acknowledging that masks could work by mechanically lowering viral load transmitted, but that we could need high levels of compliance and well-fitting masks (as in Taiwan) to see the effect, Gøtzsche further exhibits confirmation bias in his critique of mask evidence. He argues “the apparent effect is likely due to the fact that people who voluntarily choose to wear face masks are more careful with hand hygiene and with keeping their distance from other people than people who do not wear face masks — a classic error in observational research called confounding.”

He misses that the same critique could similarly apply to his own read of the evidence. Voluntary mask wearing could work better because people could do it better (getting masks that fit and wearing them properly), while studies of mask mandates could be confounded by poorly-fitting equipment and non-compliance, and not tell us anything about the efficacy of masks themselves.

Fringe theory favoritism

Why are we to accept Gøtzsche’s assessment that one subgroup finding is trash, but another is gold? He dismisses self-proclaimed CDC whistleblower Thompson’s finding that MMR vaccination before age 3 substantially increased black boys’ autism risk. In contrast, he promotes Benn and Aaby’s findings that females benefit from live vaccines and risk harm from non-live ones in the form of non-specific effects. Both push fringe theories about what the evidence suggests about unintended subgroup vaccine effects.

There are probably good methodological and causally logical arguments for this apparent inconsistency in Gøtzsche’s application of scientific evidentiary standards. But I would have to read a lot more primary source material than I have time to, to know what they are. That’s why I bought this book; it’s the author’s job to draw those arguments out here. Lacking this fuller explanation, the contrast in how these subgroup findings are treated appears suggestive of a double standard and needs to be specifically addressed.

It doesn’t help that non-specific vaccine effects are under-theorized. If, as Gøtzsche suggests, we should be concerned about them in the context of possible lasting benefit from live and harm from non-live vaccination last (the “sequencing matters” argument), then why should we not be concerned about such non-specific vaccine effects in the neurodevelopmental context?

Overall, the book left me feeling (uneasily) that I should take the time to learn more about the logic and evidence underpinning the vaccine-autism scare — in marked contrast with Deer’s The Doctor Who Fooled the World, which Gøtzsche recommends and which left me feeling confident that the basis for the concern was entirely fraudulent. (Time is precious, so I use heuristics here, satisficing where Gøtzsche recommends optimizing: everyone including Gøtzsche recommends measles vaccination, so my course of action as a mother is clear.)

Bias about bias

Gøtzsche says health authorities fail to distinguish science from politics (p. 97). But this is ironic, since he, too, blurs the line — as we all do. Non-neutrality is a posture at worst, an unattainable aspiration at best. What looks like a matter of “what science says” is often a question of who we believe — a “turtles all the way down” problem of delegated trust in ascertaining and interpreting facts.

This is a very serious flaw, because it changes the nature of the core problem from “What does the science say?” to “Who do we trust?” His book answers the former question. Individuals reading it, and society at large, have to answer both.

This view weakens this book as a work of translational science, because it means Gøtzsche fails to recognize the validity of other good-faith interpretations than his own — differences in opinion that science can’t resolve. For instance, addressing the question “What Should We Do in the Future?” (p. 140), he says that, in a future Covid-like pandemic, “We should act like Taiwan, with aggressive early contact tracing, testing of suspected or possible cases, and isolation outside the home” (p. 141). He hates masking (to the extent that he mentions that it gives you pimples), so he doesn’t mention it here — even though he elsewhere acknowledged that part of the Taiwanese success story may well have been making masks that specifically fit schoolchildren.

In another instance, he argues against mandatory vaccinations (Chapter 3) before admitting “that the threat to other people, not least one’s own children who cannot make decisions about vaccines for themselves, might become so large that I would favor mandatory vaccinations of some kind. Hopefully without using force, which I find repugnant” (p. 212). Questions about mandatory vaccination enforcement are political, the state being defined by a monopoly on legitimate use of force. It is nonsensical to say we need to separate science from politics in this context.

A better argument is that we want states and health organizations (e.g., WHO) to get it right, and sometimes they get it wrong. He gives a compelling example in a very interesting segment on University of California President Jane Napolitano’s July 31, 2020 executive order mandating flu shots for all UC students, faculty, and staff as a condition of continued employment/enrollment (p. 127-8). (It’s stashed in the lengthy Covid chapter, when it may have fit better in the flu chapter.)

But this is an argument for structural safeguards against state abuse of mandatory vaccination powers. Not an argument against such mandates.

Conclusion

In the end, Vaccines unintentionally reinforces the value- and perspective-laden nature of the controversies Gøtzsche insists we should see as either scientific or political. Its call for individual discernment assumes that we can and should optimize in a complex, busy world that requires satisficing. The real challenge, which it neither identifies nor grapples with, is designing trustworthy institutions and stable, cooperative norms that effectively protect societies from predators — be they individual frauds like Wakefield, pharmaceutical companies hawking Covid mRNA vaccines with repeated reports of contamination problems and unproven long-term safety, or hostile states keen to fund anyone who seems willing to sow confusion in adversary open societies.

In a world full of information (some of it good), measles vaccination sits on a pole of a continuum of public goods games where we want stable cooperation, and might not be able to get it anymore. The system has been destabilized, and that’s not likely to stop. We can’t solve that problem one individual at a time. And we shouldn’t have to try. (Although, it’s certainly possible that we can’t solve it, at all.)

That’s not to say that the book winds up convincing me that the vaccine mandates against which Gøtzsche argues are a good thing. He cites evidence suggesting they don’t work. So it seems we need to know why not, and what does. It’s possible that mandates cause blowback, increasing mistrust, resistance, or misinformation along with compliance.

It also seems reasonable to counter that vaccine mandates are not dichotomous, since there are so many possible ways to formulate and enforce them. It’s not like with a bike helmet, where cops can ticket cyclists for not wearing one. No state, to my knowledge, has a police officer going around with a doctor enforcing vaccine mandates by going door-to-door asking to see proof of vaccination, and giving people a shot on the spot when they can’t prove they’ve already had it. This is not a dismissal of Gøtzsche’s representation of the evidence on vaccine mandate efficacy (which I haven’t read myself), but a methodological caution. I don’t know how existing vaccine mandate analyses have been done, but they’ve probably been done wrong…

Mandates are intended to bring back a stable, cooperative equilibrium. This implicates feedbacks. We need to think about the cyclic causal structure of this problem before doing statistical analyses; that’s the current state of science that ultimately someone doing the sort of work Gøtzsche has picked out for himself needs to rise to.

Otherwise, we’re left with an impossible project of individually knowing the evidence well enough to make informed decisions about too many things to possibly do it all — in a world of other people who also can’t do this impossible task. And their inevitable failure can endanger our children. The alternative is to use what we know about the structure of this problem to solve it.