There’s a whole ecosystem of ethics which forbid or strictly limit cost benefit thinking. It’s called deontology. Most folk ethical systems are deontological. The ten commandments don’t say “avoid adultery unless it would be really enjoyable and you are highly certain your wife would never find out.” Nor do they say “thou shalt not lie unless it would protect innocent people from harm .”They are categorical.

Children are taught that morality is categorical and that principled people follow moral rules even at great personal cost. This is a deep part of both Christian and many secular moral traditions.

Costs and benefits are squishy and once people are allowed to think in those terms, they usually smuggle their self interest into the equation. One might think this tryst will be really fun and, if my wife never finds out, what’s the harm? Does the federal government really need me to pay all the taxes I owe when it can print money? I can make better use of this money than the military industrial complex!

There’s a related strain of folk thought that “some things are too important to be about money.”. For instance, a family might make very frugal decisions about vacations, housing, even groceries, but bust out the credit card for an unplanned medical expense because “health is more important than money.”

That is a form of the binary thinking MY is criticizing. Once something lands into the health bucket, cost-benefit flies out the window. A family that can’t afford fresh food might shell out $5k for an mri that isn’t really necessary. This is a shitty decision even in pure health terms because fresh food and less stress would probably have better health effects than an mri which was taken mainly to protect a doctor from malpractice liability.

Refusal to optimize also happens when other big abstractions like “safety” and “democracy” get tossed around. The actuarial risk of death by terrorist attack was always minuscule, but this issue has gotten a lot more attention than highway safety or even opiates. “Democracy” is the foreign policy equivalent of “public confidence.”. Anything that can’t be justified in cost benefit terms can be framed as part of a moral crusade to advance democracy. Heroes and martyrs die for principles.

The best way to push back against naive deontological thinking is to impose its costs upon those who insist on thinking that way. A lot of people call me and want to litigate different things because of this or that principle. They never pay me. People who are afraid of jail or prison pay whenever they can. Consequences clarify the mind.

Expand full comment

FDA-er here. The critique here is absolutely spot-on - I'm gob-smacked by how Matt has lined up the relative issues and weighed the subtleties (with one exception, mentioned below). What to do about it is much less well-developed. FDA is an enormous ocean liner; it can't change on a dime. Changes like these take years and are likely to produce all sorts of unintended consequences.

The point that is missing from Matt's analysis is that once a product is approved, it can in most cases be legally used in any way a medical practitioner sees fit. If Medicare or other insurers want a cost-benefit analysis to pay for a given use (particularly if that use is not FDA-approved), they can do it themselves. If it is useful from a public good standpoint for some disinterested party to do that analysis, the best model is to create an agency designed for that purpose like NICE in the


More importantly, the prospect of a NICE-type agency is anathema to the health care industry. Congress has regularly kneecapped AHRQ, the agency we have that is closest to NICE, whenever it began to take this subject seriously. There is less chance of FDA, Medicare, AHRQ or some other agency creating a sound cost-benefit framework than there is of CDC studying gun violence as a public health problem.

Expand full comment

You write: "there is a lot of anxiety among both teachers and parents about Covid risk". Just as we rightly call out the vaccine-resistant (more often Republicans, though not all) for ignoring science and not taking the vaccine, we need to call out the irrationally anxious (borderline neurotic) among the other side.

Experience has shown that the COVID-19 virus isn't very dangerous to children, and they aren't a vector for spread. And evidence shows that remote learning is ineffective and especially bad for poor kids. There are risks and costs on both sides of the school debate, and we shouldn't cater to those who exaggerate the COVID side of the equation.

Expand full comment

The UK's solution has been to separate the "is this safe and effective" determination from the cost-benefit determination.

So the EMA (before Brexit) or the MHRA (now) determines whether it is legal to prescribe a drug and then NICE (for drugs) or JCVI (vaccines) determines whether it's cost-effective and considers availability vs demand and then prioritises who should get it first if it's supply-constrained.

It is perfectly legal to pay for a private dose of a vaccine in the UK. You can't actually do so because Pfizer and Moderna have sold every dose they manufacture to governments and private suppliers don't have any.

Expand full comment

It's hard to think of a more misguided idea than the notion FDA should refuse to approve "me too" drugs in the same therapeutic classes without proof they're more effective than the first one - unless maybe it's the idea that the way to get more affordable housing is to prohibit the construction of expensive new housing.

When there's only one drug in a class, whether it's statins for cholesterol or SGLT2 inhibitors for diabetes, or whatever, the manufacturer has more leverage to charge higher prices. When there are therapeutically equivalent alternatives, PBM and other payers that make formularies can bargain the price down by preferring the drug from the manufacturer that agrees to the biggest price concession.

Expand full comment

This is a pretty minor nit, but it still bugged me. Regarding Canada's one-shot strategy, MY writes: "But then you can see how first shots first worked by looking at how they caught up to and then surpassed us."

That's not why Canada is passing us by. It's because we ran out of people who wanted to get vaccinated. We hit a peak of 3.7 million shots/day in mid-April; we're under 600,000/day now. If we were populated by Canadians, we'd have a lot more vaccinated people now.

Expand full comment

Why has historically gungho America been infected by proceduralism in the past few years?

Expand full comment

My husband called around lots of places and got the first available vaccine - J&J (I was slightly less of a go-getter and got Moderna about 2 weeks later). Unfortunately, Walgreens (and seemingly other places) aren't approved to give out Pfizer/Moderna to previously vaccinated individuals(he tried last night - the pharmacist said they would have done it themselves but the system didn't allow it). Neither of us wanted him to lie about his vaccination status - for ethical/legal/or medical contraindication reasons (not that we think there are any of the last category).

So it may be good for J&J patients to get a dose of something else, but seems to currently not be allowed at the easy-to-access vaccine places (the local Walgreen's also made sense as a place that seems to have sufficient doses - he would not be taking them from someone who needs them more (Barring export issues to other countries - but that doesn't seem to be the limiting factor right now)

Expand full comment

I think you have this wrong. I would think you have enough experience as a writer to understand how you are haunted by real or perceived bad takes. Well that works for everyone and in every business. And that is the way it works for the FDA too. And the CDC and the WHO as well. Accrued bad takes and mistakes are NEVER forgotten. Which is why the decision making processes for drug approvals are inherently conservative. And the FDA by the way is a much admired and copied institution around the world. They do, of course, have a process for dealing with exceptions and that is EUAs. And so they should.

It is probably a good thing to remember that many drugs with good efficacy (say statins) wind up getting withdrawn from the market and result in monster class action suits even though clinical trials did not reveal and could not statistically reveal lethal side effects. These often do not become evident until millions of prescriptions have been written. I seriously doubt that the long term confidence in the FDA is going to be enhanced by loosening standards and creating more fuckups. The lawyers will love it though.

Expand full comment

How are the FDA and CDC not facing Dodd Frank style reforms?

Expand full comment

A few weeks after saying that the CDC shouldn't be doing cost benefit analysis and just putting out the science, you're arguing that the FDA should. This is wrong.

The FDA should be approving drugs. For an amazing vaccine, perhaps accelerated full approval makes sense, but unfortunately we have a frequentist rather than a bayesian org. (Your/Rosenthal's critiques of the FDA approving marginally useful expensive drugs is on point, but not useful for the EUA argument).

The CDC is literally there for *Control*. It should be responsible for doing what is required for controlling epidemics.

Critically, I think, this is the question. What if we had only Sinopharm or some other much more crappy much more side effect ridden vaccine? In a pandemic would *still* want to require everyone to take it. The value to the country as a whole is worth much more than the individual side effects. Once the pandemic is under control, you wouldn't want the vaccine to be used though.

To do this, in the "FDA cost effectiveness model" you'd want the FDA to approve it, and then disapprove it. This makes zero sense. The FDA doesn't even have people thinking about epidemics. The CDC, on the other hand, does have a lot of people thinking about epidemics, and should be capable of requiring a vaccine for the greater good in an emergency.

The problem is the politics - as soon as you're past the emergency, once there is no pandemic, especially if you start vaccinating early and the pandemic doesn't happen (as would be ideal) then the people who did have the side effects from the crappy vaccine start feeling quite grumpy about the whole thing.

And that's why we can't deal with an emergency. The problem is us.

Expand full comment

I think your main point (convergence of dichotomous political critics) is legitimate, and I agree with your focus on cost/benefit analysis. But something is certainly wrong, and I think it goes beyond that. Consider the outrageous failures to approve good tests at the onset of the pandemic. I'm still unclear whether primary blame lies with FDA or CDC, but it's a Fail far more consequential than the errors that led to 9/11.

Is the malfunction intrinsic to bureaucracy, or is there a clear and solvable problem? The current dysfunction even in the face of the triumph of the mRNA vaccines may allow the kind of analysis of failure that will lead to meaningful future improvements.

Expand full comment

Whether a drug provides enough benefit to merit approval and its pricing are related but separate questions. FDA approval gives license for companies to set prices based on the FDA approval -> CMS coverage -> payor coverage and payment equation, as MY states. I don't think that FDA consideration of pricing would improve their decision-making; better would be to allow Medicare to play a negotiating role. Also, it's not a pure free-for-all. Commercial insurers choose what formulary status to give a drug. Whether a drug is covered as first-, second-, or third-line therapy by insurers has a massive impact on sales. Drug companies will generally reduce prices if they fear being assigned second- or third-line formulary status. While the resource-allocation mentality of international regulators probably does lead to different drug evaluations, it doesn't suggest making the FDA think about economic cost-benefit analysis. The UK cost benefit work is done by NHS when it decides what to cover and what to pay. That is totally distinct from the drug approval, which has been done by the European Medicines Agency (but will change to the UK post-brexit).

There are good questions as to whether the FDA could be better at weighing the medical cost-benefit of each drug it evaluates. The agency weighs these differently across different disease states and therapeutic classes. Partly this is because it relies on experts in each field (oncology, infectious disease, etc.) who bring their own biases. But patient groups, institutional bias, play a role. Some are criticizing the FDA for being risk adverse to prior failures, but is that really unfair? Wouldn't we want them to tighten standards when they find that previous standards led to ill effects?

What I most wish for is a bit better judgement from FDA. They are strict about trial protocols because there is a long history of drug companies post-hoc digging for statistically-significant findings to support claims (p-hacking their way to coverage). The FDA is right to guard against this. But then they apply this to the AZ vaccine trial, where the issue is completely different. Ultimately we have to trust regulators to exercise good judgment.

Expand full comment

Adding to your thought that "information is a public good". I'm also frustrated that the FDA hasn't funded (do they ever actually fund studies?) a study to determine optimal dosing sizes and intervals. Since over 6 billion people on earth haven't been vaccinated yet (people that might incubate the Zeta variant), this would be incredibly valuable information to have.

Expand full comment

Cost benefit and risk analysis are two big victims of this pandemic, and probably modern society as a whole.

When 1 in a million hits the headlines (whatever the subject), people go bonkers.

Half the country thinks they are dire danger from Covid. Half the people think they are at zero risk.

It explains why people think crime is going up (well it is now.... but it wasn't for the longest time).

Or why people go nuts over AR-15s despite the real danger being glocks.

The only modern exception I can think of is professional sports. They have moneyball down.

Expand full comment

"There is a very strong presumption among public sector entities that you cannot mandate a vaccine based on an EUA."

On what is that very strong presumption based? Does it come from the FDA itself? Why is the problem not with the presumption (or whatever it is the presumption based on)?

Expand full comment