Matt, would you be open to having someone who takes the other side of this debate, or is at least more concerned about the risks of rushing a vaccine that doesn’t work or hurts people, give their side of the story? I’m not able to evaluate this argument, and I bet most of your readers are in the same boat. I can identify concerns - I mean, this is a developer saying he knows about this *amazing* technology, and why isn’t everyone getting out of the way to let it come to market. That take has earned at least as much suspicion as government regulators, especially in the health area (Theranos, anyone?)
What would be especially helpful is to get beyond directional arguments (faster! slower!) to get a better sense of the specific requirements that take a lot of time and what’s the risk of not running them. E.g., he complains about having to run animal trials. How much time is involved, and what’s the risk in running them concurrently instead of in series? Can we focus on some big-ticket items to evaluate the debate? If it really is about death by 1000 cuts, ok, let’s still engage on that: where is the obstacle?
One perspective taking measure: for all the complaints about FDA, didn’t the US have the vaccine broadly available earlier than any other country? It’s not as though there is another paragon country out there we are lagging, is there? Or if there is it would be good to know.
Obvi it’s Matt’s prerogative to set the editorial agenda of his own blog, but I do want to second Theodore’s initial comment about how useful it would be to know what these types of posts are arguing against.
“…lack of basic institutional seriousness” and “self-imposed straightjackets” seem like harsh characterizations of differing values and principle stacks. The author makes it clear that his #1 goal is speed, so any barrier to that is a straight jacket, but speed is not the #1 goal of the FDA, so those straight jackets are in service of something else. I wish speed were more of a priority, maybe it should be the top priority, but I fail to see how it not being so makes them institutionally unserious.
I don't have the background to comment on the nature, scope, and appropriateness of the regulatory burden faced by biopharma with regards to vaccine development, but some of his statements about the fundamental science raise eyebrows.
(e.g., he proposes using antibody titers as a surrogate endpoint in clinical trials, when we know that short-term elevations in antibodies do not have robust correlation with long-term protection against infection or long-term protection against severe illness).
The argument against accelerating approval for pandemic vaccines needs to be distinguished from the more general libertarian position that drug approval barriers should be lowered. I've spent my career in pharma and medical devices, in commercial roles but continually working on clinical studies and interpreting their data.
Those advocating against faster pandemic vaccine rely on straw man arguments. It's difficult to find somebody to competently take this position. A protest "review," (https://tinyurl.com/3bzn6ph4) which included two authors who quit the FDA in protest of booster approvals, is representative. This and others tend to rely on three arguments: 1) vaccines are not riskless, 2) the positive data for [insert new vaccine or application] is not yet definitive, and 3) moving quickly will undermine confidence in vaccines. Instead of reckoning with the cost-benefit tradeoffs of the decision, the logic is one-sided : highlighting risks of approval but not of delay. And, not only does the "if you disagree with me, people will be more skeptical of vaccines" have no data to support it--note people who critique a lack of data while simultaneously making unsupported claims--but I've anecdotally found that most such claims in the pandemic ended up being completely wrong.
The more general libertarian view has two flavors. One, call it the "strong acceleration" view, wants to shift decision making from regulators to consumers and their physicians. One mechanism is making the FDA approve any drug approved by selected other countries (a lowest-common denominator approach to regulation). Another is expanding compassionate use programs where people can access unapproved drugs for extreme cases, but making the bar for this much lower. In my personal view, this is batshit crazy. Anybody who has seen how drug approvals flow to drug promotion and then to consumer choice would recognize the risks here. This is an area where having somebody write a post would be very valuable.
There is a much better argument, call it the "weak acceleration" view, that calls for streamlining the approval process in many ways. I'm very sympathetic to this. Arguments against it often have the straw man quality of those against pandemic vaccine acceleration.
I basically agree with you, but would like to steel man the argument for robust efficacy endpoints prior to vaccine booster rollout (I agree that safety concerns and the armchair social psychology are weak objections at this point, at least with the mRNA platform):
The original vaccine series had a huge and relatively durable effect size on major outcomes like hospitalization and death. The first booster had a much smaller absolute effect size, but did improve hospitalization and death in higher risk populations and people > 50. The second booster has shown minimal effect on any clinically significant outcomes, even in high risk populations.
It's not clear to me, in a landscape where we have relative equipoise on the efficacy of future vaccine boosting, and in a patient population (vaccinated and boosted) where the risk of serious adverse outcomes approximates that of seasonal influenza, that we owe Pfizer and Moderna the right to extract billions of dollars in perpetuity on the basis of expedited, poorly correlated surrogate markers like antibody titers.
That's an excellent point. Thank you for making it. it underscores the need to balance the risks of approval with the risks of delay. The combination of the significant benefit of the vaccines, extremely safe side effect profile, and impact of the pandemic all argued for speed and lowering bars to appeal.
But, a future vaccine with a worse safety profile or lower efficacy could change that calculus.
And I agree that the push to only use antibody titers is not clear cut. I view that as more of a "break glass in case of emergency" option. Healthcare is replete with times when surrogate markers did not correlate with health outcomes.
I'd argue that the commercial rewards for vaccine development are still too low. Covid has been the exception to the rule. But you're right that initial approval should not be able to be gamed into a perpetuity.
Yes the original covid vaccines were ready extremely quickly. But the problem is that now it's business as usual. And business as usual is a horrible regulatory state
But net of the entire performance of the govt, did any other country make the vaccine available more quickly? I’m asking for a big-picture benchmark. Takes descend very quickly into charged language (“wasn’t in any hurry”) and I’m skeptical that these are well informed. I’m open to believing things can be done faster - and I want them to, if the risk-benefit makes sense - but it would help to see the homework. It’s relevant to know how fast other regulators acted. And maybe, even if other regulators took the same amount of time, roughly, it only means that all the regulators are too slow. But if you want to make that argument, ’d like to know how you arrived at it: what risks were they too cautious about, and what is the improvement?
When pharma wants to justify high drug prices, they go on and on about the difficulty of drug development: they have to try tens of thousands of molecules to find a successful drug, and candidates fail at every step along the way. Succeed at phase 1? You might fail phase 2. Make it through phase 2? The results at phase 3 might uncover a very serious adverse even that affects 1 patient in 300. And so on. But now come the people who want things done very differently and it’s all impatience: why are we doing all these stupid tests? What a waste! We know these drugs work and are safe! I am not an expert, and I can’t really adjudicate these disputes. I would like to see experts, who are well informed, pick out what they see as the key risk-benefit pivots - where we would save a lot of time - and ask: do we need to do this? Why or why not?
To answer your first question, the UK administered the first Pfizer jab outside of clinical trials on Dec 8, and for the US it was Dec 14 (Sputnik doesn't count, sorry Russia). Not sure this is praise for the NHS, but the US and UK were more or less tied for first. Roll out, distribution, voluntary uptake, etc., are different questions of course.
Israel absolutely smoked everyone else at getting vaccines into arms. In the UK we didn't make the vaccine generally available until long after the US did, but that's because we did a much stricter rollout by age cohort, aimed at maximizing lives saved - in terms of jabs administered per capita, we've been ahead of you for most of the time.
The US paid a lot early on to get ensured early access to several hundred million doses each from several manufacturers (including some whose products were never approved but still got the money). The EU made the big different decision of negotiating a bit on price, and saving several billion dollars on drug costs, but getting them a few months after the US.
I think the regulatory process was similarly slow in all countries (though as others mentioned, the UK did grant approval a week or so earlier).
It's interesting that there's basically no constituency for this. People on the right largely never thought covid was a big deal, so them taking action now all of a sudden won't happen, and people on the left seem very hesitant to criticize the FDA or the public health establishment, so they're not going to support reform if they won't admit reform is even needed.
So we're just left with bad policy. It's unfortunate.
We need smart people who believe in science to stand up and say the quiet part out loud: the FDA, NIH and CDC have all been proven incompetent. Reform and urgency are both very much needed. But the opposition from the right seems to make everyone reluctant to do that for fear of looking like a Trump supporter or anti-vaxxer or some such nonsense. This post is a great step in the right direction. I hope to see this type of column in the NYT and be mainstreamed.
Yes, assuming it's accurate, the most persuasive point in the article that this is a $1 trillion bill just laying on the sidewalk, whatever you think about it's overall significance in the scheme of things. But there're probably a lot of $1 trillion bills, or at least large ones, laying around that we don't pick up, because they're not important enough to anybody in particular.
It's important to remember that this has been true in every human society throughout history and will be true in all future ones.
There's a tension between crying endlessly over all the milk that never even came into existence because it was "pre-spilled" by various failings of this sort, which will make anyone miserable, and pushing to make concrete changes now and again, which is what's necessary to become and remain a developed country.
That we are a developed country, and a particularly rich one to boot, suggests that historically we've done fairly well at making those changes.
Exactly, this is pretty how a collective organism like a human society makes decisions about where the cutoff is between a collective problem that everyone needs to stop and focus on, and one that that can be muddled through.
Call it hive/swarm decisionmaking vs centralized decision-making. The difference is illustrated by the contrast between how the US and China have responded to covid.
I think that's mostly a presentist fantasy. We've always been like this; pluralism requires it. From the earliest days of the Republic (the colonies, even), folks have always set out to protect their slice of the pie, then the dam breaks or things really become untenable and those barriers get washed away, then new ones arise...
The wide-reaching Federal role in economic and regulatory life is newer, but I have few doubts that its current descent into sclerosis will be subject to the same interplay of popular pressure, economic sustainability, and insider pushback/protectionism. It did happen in the 1970's and 80's, then a bit of a swing back in the 90's and 00's, which gave birth to the modern regulatory state that sits atop a mostly deregulated economy.
Again, there's a balance to be struck between gibbering panic about all the "bills on the ground" that serves no one versus steady popular pressure on the organs of power to permit and even force concrete, if incremental, change.
Things mostly work, well enough, for most folks.
That's probably the best we'll ever manage until someone figures out how to make us something other than human.
The right would respond by saying that if we stopped caring about covid and sucked it up and got on with our lives, we can pick up a bill greater than $1 trillion. The left would respond by saying that why are you caring about money when it's clearly less important than people's lives.
(But yeah I agree that that's the most persuasive point.)
I would rather put it that there is no apparent political feasibility. It seems like one of many many no-brainer centrist position that in fact would have a very healthy constituency in terms of percentage of the electorate supporting some of the suggested moves (and certainly the goal of government doing more to facilitate the speedy development of safe and effective next generation of covid vaccines and therapeutics) but it's rather the polarized political landscape that makes it difficult to get the popular will to actually matter.
"Virtually everyone knows plenty of people who contracted COVID after being vaccinated and were out of commission for a week or more." Having just had COVID after being vaccinated and boosted, while getting sick wasn't super fun, easily the worst part was being essentially unable to work and being trapped in the house for *weeks* as COVID moved through our family and we desperately waited daycare's required quarantine to end, even though the kids are fully vaccinated and were symptom free shortly after testing positive. I think encouraging liberalizing daycare COVID restrictions will be very popular in a couple of months after all the little kids have had an opportunity to be vaccinated.
Did you test your kid(s) for covid every time they got sick this fall? Mine got sick from daycare so many times it just made no sense to me to try to schedule an appointment, wake up really early, drive down to the testing place, wait in line for an hour or more, etc..
She was probably sick more than a dozen separate times from Oct-March, and we had the same symptoms around half those times. I assume that one of those illnesses was covid but I guess I'll never know.
All this is to say: I'm curious if your experience with testing is similar or not. I suppose now it would be easier to do the tests than it was in the winter, so maybe I will get one for her next time she has a cold. But sometimes I wonder why so many people are still testing for covid if they're already vaccinated and boosted. For young, healthy, vaccinated people, what difference does it make if the test comes back positive or negative?
I wish we had multi-virus tests. It seems that it should be not too much harder for Abbott to make a little card with four potential stripes - one control stripe, one that turns red if you have COVID antigens, one that turns red if you have flu antigens, and one that turns red if you have RSV antigens. I don’t exactly know how much good would be achieved by being able to know which of these you have, but I do think this is one of those things where having the ability to have widespread knowledge would enable better responses.
As someone who will shortly have two children under two, I would love to be able to test visitors for RSV on their arrival to my home. That would be quite valuable to me.
I hear where you're coming from, but I think my conclusions are most like Belisarius below. They just get sick so often, for my kid she'd get a cold or fever every other week from about November to March. Maybe if I knew first exposure was less risky at age 4 than age 3 then delaying the inevitable makes some sense. But my sense is its worse for younger kids mostly because it's their first exposure, not because of their age.
Agree overall that we should stop testing because it's not like we're trying to stop transmission anymore in the US
If you're a rule follower and you test positive for the virus, close contacts are supposed to test. Kids would be close contacts. We have a very similar story to OP and the reason we tested the kids is we were supposed to as they are close contacts; and of course, camp / childcare does not want your positive kid attending even if they are asymptomatic.
We tested a lot, mostly because both kids are in the Moderna trial. We also mostly did rapids, even though those aren't technically approved for kids under two.
The test that started our quarantine was we were headed to a party with grandparents who have risk factors and my wife was symptomatic.
Ah but you were a victim of FDA not approving dirt cheap tests in early 2020. Asymptomatic testing could have avoided a lot of the cost of restrictions on (and fear of) social interactions
My experience is similar. I know plenty of people who have had Covid who were vaccinated and it was just annoying. Even in people who weren’t boosted. I would prefer Covid to a regular old rhinovirus because it was milder and I recovered fully after 3-4 days.
Yeah, the author chose to, ah... operate at a credibility deficit... with language like that (see also post-vax long-covid doomscrolling). I have no idea why. It seems fairly common, like you need to establish bona fides by being Still Rather Concerned. I think it plays the opposite.
*The worst* post-vax covid case I personally know apologized when he called into the next week's meeting, because he hadn't gotten as much done on the project as planned on account of "covid kicked my butt". He wasn't "out of commission"--he was working throughout the dang thing!
The majority of my career was in the biopharma industry doing drug regulatory affairs and safety. Mr. Collinson's column while interesting needs to be put into perspective. One major hurdle for vaccines is doing the necessary safety studies to demonstrate that the benefits outweigh risks. Lots of the side effects seen with vaccines are small, if present at all, and on the order of 1 in 10,000 or less. You cannot pick up these in a normal clinical trial of 3000 patients. This necessitates that vaccine trials be done on a far larger cohort and is one reason the FDA is more conservative than looking at a new cancer drug.
Vaccine manufacturing is not trivial. Vaccines are sterile products made under exacting conditions. However, one the process is validated, it is possible to make changes in the vaccine to adopt to new variants. This is why we are able to get new seasonal influenza vaccines each year. Public health authorities and the vaccine industry agree upon the circulating flu viruses and change the production strain. Major regulatory filings and review are not required. We are seeing this approach being applied by both manufacturers of mRNA vaccines right now. A similar approach can be taken by companies who are close to approval of adjuvant/protein vaccines.
During the first year of the pandemic, I put out a daily newsletter for former industry colleagues and friends surveying the research that was being done. I stopped once the first vaccines became available. I looked closely at vaccine development and saw a lot of interesting work going on. Some of the projects are still going on. Canada has licensed a tobacco produced protein vaccine and the ferritin nano particle vaccine developed by Walter Reed is in clinical trials right now:. Volunteers are still being sought to finish up the Phase 1 trial: https://www.wrair.army.mil/node/645
Nasal vaccines have been used in the past with varying degrees of success. A commercial influenza vaccine has been around since 2003 but is not recommended for those > 50 years of age. There is also a bacterial delivery system that will get Covid antigen to the gastric mucosal system.
A number of vaccine approaches were under development several years ago when Zika virus was thought to pose a major public health problem When the virus disappeared these projects were shelved and a major opportunity to test vaccine platforms for the development of neutralizing antibodies was lost. Had such work continued it might have provided valuable information that could have helped out Covid vaccine developers.
There are a number of other issues that can be discussed but I'll stop with these. Intrepid uses of google can find the archive of my Covid-19 newsletter fairly easily.
"Lots of the side effects seen with vaccines are small, if present at all, and on the order of 1 in 10,000 or less"
Ok, so? Tell people there might be a 1 in 10k chance of something going wrong, pay out compensation if it does happen, and then release it on the market.
Right?! Of course the only way to measure a side effect like that would be to more quickly run larger trials. So highlighting this risk is not really a counterargument to speeding up the process. And this is to say nothing of the probability of death from COVID for the elderly without a vaccine. It's basic cost-benefit analysis that the FDA fails terribly at: death > rare possible side effects. The haunting reality is that Moderna had already developed and manufactured the vaccine for testing before the first American died from COVID.
Logistical roadblocks should be fixed. There's no reason that studies should be delayed by monkeys being stuck at customs or supply chain issues with adjuvants.
But when it comes to lowering the barrier for starting human trials, I'd want to see some concrete scientific and bioethical basis for this change of procedure. I understand we all want faster role out of therapeutics! But each decision is a cost/benefit problem, and the costs of a harmful or ineffective therapeutic, especially one widely administered to healthy people, being released is potentially very high.
I think this is an area where the public health community’s believe that everything has to be either forbidden or mandatory does real harm.
You could relax human trial requirements and then say to the public, “we are releasing this but we also totally understand if you don’t want to take it; what we’re going to do is closely monitor the people who do want to take it and gather a bunch of real world data that we hope will demonstrate safety and efficacy or else raise red flags that cause us to pull back.”
It seems like there’s a common (sometimes in bad faith, but I think often not) blending of “approving this puts people at risk of harm from the treatment” and “approving this puts people at risk of a false sense of security, leading them to take risks that they wouldn’t take if the treatment was ineffective”. The former is a straightforward safety question - basically no one serious that I’m aware of argues for skipping the safety trials prior to approval. The latter is generally armchair sociology - this may be true, but it’s not like the FDA is staffed by sociologists! A very specific class of experts has gotten carte blanche over decisions that involve fields in which their training provides no inherent expertise.
I entirely agree that we need better cost/benefit analysis, but that means including the costs of *not* moving faster. Speed is not inherently good, but the reason it feels like so many of us are pushing for it is that there is a cost to slowness as well that is not actually being incorporated into decision-making.
What real harm? So far, there's been zero evidence that anything beyond two shots and a boost (+/- a bout of infection) does anything to move the needle on your durable risk of bad outcomes from future SARS Cov2 infection, including Long Covid. And truthfully, now that we're over a year out, there's still not great evidence that booster did much in people <50.
The problem we are facing is that the primary series of the mRNA vaccine worked really, really, really well - an effect size that staggers most of what we see in biomedicine. Durable marginal improvements on that effect are going to be hard to see even in carefully controlled prospective data, much less the incredible noise you're going to see with elective uptake in additional mRNA shots
This is pretty similar to the arguments routinely advanced by Scott Alexander & commenters at ACX for broad scale pharma deregulation, and the issues are the same: by abandoning prospective randomization and a priori demonstration of efficacy in favor of 'safety only' , you open the floodgates to snake oil. Even with our current FDA regime, healthcare and biomedicine has basically eaten America's productivity and GDP gains over the past 50 years. I shudder to imagine what they (ok, we) would do if they didn't actually have to prove anything worked.
This scenario where you lower the testing threshold for allowing human trials would likely require an increase in tolerance for deaths and adverse events during drug testing. Maybe you'd say it's worth it to save more people from the diseases these drugs are for? But you'd have to be honest about your increase in tolerance for bad things happening in drug trials.
I’m on the fence on challenge trials because I think there’s so much other low-hanging fruit that skipping that debate might be beneficial in terms of how quickly we can get improvements elsewhere. That said, those that I’ve seen advocating for challenge trials are not advocating them for the safety trial phase, rather for the efficacy phase. Efficacy trials are already in some sense challenge trials (if you ran an efficacy trial with no chance of viral exposure, you’d learn nothing). It’s true that you may encounter higher levels of negative outcomes on a percentage basis in a challenge trial, but you can also run them far more quickly and with a smaller pool of participants (because your variance in exposure will be lower, your required sample size will be correspondingly smaller), so you could end up with a higher rate of adverse outcomes but a smaller absolute number.
And most bioethics frameworks consider it worse for a healthy person to die from a drug they took during a trial than it is for someone with a disease to die because there isn't treatment...
I get that you're explaining rather than advocating here briross, but this "framework" is what needs to be changed. While it's important that no one should be tricked into volunteering for something where they don't have full information about, it ISN'T worse for society if people die from an experimental treatment rather than a disease, so the current framework is the problem.
If people are willing to pull the lever to help save the lives of those who may be run over by the trolley, even though this increases the chance that that the trolley may run over them, we should be thanking them rather than having rules that prevent them from making this choice.
Can you point me to an example of bioethicists doing a cost-benefit analysis (i.e. "calculating" something in a "model") and coming out against human challenge trials?
As far as I can tell, they never bother to do that; it would be very hard for any cost-benefit analysis justify the level of conservatism they desire without some obviously arbitrary parameter to make deaths in trials *much* more "expensive" than deaths from the actual disease.* As far as I can tell, bioethicists that oppose human challenge trials mostly take absolutist stands and pontificate. I've said it before: if bioethicists did cost-benifit analysis, they'd find out that the world would be a better place without bioethicists.
* To be fair: lots of people like these arbitrary parameters that are based solely on personal preferences, such as Blackstone's ratio. But different people express wildly different values for their ratio (e.g. Ben Franklin and Blackstone disagree by an order of magnitude --- and others differ by yet another order of magnitude), so I don't think that such a parameter can be meaningfully used in a cost-benefit analysis; if a personal-preference parameter can vary by several orders of magnitude, it is just a fudge factor to ensure that you get your desired result.
You could do that, but for an OTC product like a vaccine, where the public's access does not run through a learned intermediary prescribing physician, who can explain the nuances of the scientific uncertainty and risk-benefit calculation to lay patients in terms they can understand to make sure there's informed consent and make an independent decision, it would necessary to provide some kind product liability protection for manufacturers. The normal prescription drug or vaccine liability protections wouldn't apply in the situation you're describing, and manufacturers would be at risk of major mass tort liability that could easily swamp the entire profit of the product even if it the end of the day it turns out the lawsuits are bogus.
The upside of a pan variant vaccine seems pretty worth a challenge trial, so long as the participants know that the risks, even if low, aren’t zero. Hell, going to the moon had far more risk for arguably far less upside.
It seems like we're in a position where we let ethicists dictate what is acceptable to society, rather than society laying the foundation for ethicists.
It also comes down to the bioethics framework. It's like the trolley problem--do you pull the lever and kill one person to save five? The current framework is very conservative about pulling such a lever.
I don't deny the complexity of the problem along multiple dimensions, but I've struggled to find math behind the various covid approvals or requests for additional data. If you're going to ask for a trial to be extended several months to gather additional data (or just telegraph it will not be approved without it) you should be transparent and rigorous about your assumptions. What risks that data would reduce and the cost in lives/dollars of the delay laid out in a comprehensive manner.
This I don't know. I'm just going based on what I learned in my research ethics courses. I don't know how the math is actually calculated, and how precise of a science weighting these costs/benefits is.
What I’m interested in reading in this area is the intersection of science and policy. How does regulatory policy shape what’s possible with regard to progress with new Covid vaccines, and what are the levers we could push to change policy, speed innovation, AND still produce a safe and efficacious product? This collaborative post is a nice push in that direction, thank you.
Following this suggestion, I subscribed to Topol and read the last two columns. He's obviously a very smart and technically well-informed guy. However . . .
In his May 15 post "The Covid Capitulation" (https://erictopol.substack.com/p/the-covid-capitulation) he describes how we're in the midst of a dangerous new COVID wave and we're capitulating to it. Almost two months later, one can say that we're not in the midst of a new wave, even though of course we'd like to see deaths go down much more (https://www.worldometers.info/coronavirus/country/us/). So no explosion from the BA.2 variants.
His most recent post on June 27 "The BA.5 Story" (https://erictopol.substack.com/p/the-ba5-story) paints a worrisome picture of how bad a BA.5 wave will be. Maybe! I hope not, but unlike him I'm no expert. But even still I'd tell the experts to be careful about predicting the next dangerous wave because if it doesn't happen, your future predictions of the same will be heavily discounted.
Because there's no way by which we stop that. Maybe, maybe, if we do absolutely every goddamned thing right from a public policy perspective, starting with Trump, we reduce the death toll to 600,000. Near-universal, bi-partisan masking and limited public interactions through early-to-mid 2021, a blind sprint to vaccinate the elderly and vulnerable followed by everyone else, vaccination rates hitting Icelandic levels (mid-80's), prompt FDA streamlining to roll out variant boosters...
But there is just no world in which our technological toolkit is sufficient to *stop* COVID, regardless of policy. And viewed in that context, all of the agonizing coming from the public health people has been worse than counterproductive. They've been mostly wrong on everything except the initial 2020 call to panic, and it's impossible to know how correct that was because we followed it. Basically every time they've made a call since, it's been incorrect.
China's plan to pretend we do have the toolkit to stop COVID, a.k.a. to ride Dynamic Zero COVID down in flames, doesn't seem great either. Even Australia's admittedly brilliant transition from zero COVID to "vaccinate everything that moves and get on with it" wasn't within reach for the United States. We have actual borders and there was no legal framework to enforce Australian-style lockdowns every time something slipped across them.
Seriously ? Mortality rates from COVID in us are far worse than in most peer countries. They could have been much worse still without initial lockdowns and without the (admittedly belated) mask mandates. Hundreds of thousands of lives in either direction are nothing to scoff at, in my book.
This specific CEO is also known as one of the best people at execution in tech. He’s not making a scientific case as much as he is making a case about removing self imposed obstacles that impede Getting Shit Done.
A good system generally needs a productive tension between the “getting shit done” people and the worry warts. Throw one side out of a system and expect the “get it done” people to look past risks that turn out to matter, maybe a lot. The worry warts will probably overlook the costs of inaction and crossing every t and dotting every i. That a GSD wants to remove obstacles is interesting and I’d like to hear from people who value the obstacles what he is leaving out.
He’s also the founder of the Arc Institute at Stanford. Not that this necessarily increases his knowledge, but his wife is a former CZI Biohub investigator (the virus hunters from The Premonition) and Stanford biochemist. I’m admittedly biased as I’ve followed Patrick Collison closely for a while and am involved in various efforts that he’s supported, but he’s quite clearly got more expertise here than reducing him to “payments CEO” implies.
Are nanoparticle-based vaccines safe though? I've seen a lot of hype around nanoparticle delivery for a number of drug classes (with good reason!) but there are serious questions about how the body clears them. Which again, is why we run large clinical trials in the first place, to tease out these effects before they're shown to be harmful.
Meta-comment here: I'm familiar with Patrick Collison and the good work he's done, but I'll restate a comment I've made previously: it would be beneficial to have some guest posts written by some actual scientists. It's fine for Mr. Collison to tout this work, but I'd rather read someone poking holes in the argument and then showing that the benefits outweigh the risks, rather than ignoring the risks entirely.
I agree that more guest posts (and best of all, a back and forth between those on the opposite sides) would be valuable.
I have a meta question though - what does it mean to be an “actual scientist”? I think there is a defensible case that Patrick Collison does not meet a reasonable bar, but let’s actually look at what he’s done (as opposed to his lack of a degree) - he has spent the past two years working on Fast Grants, reviewing literature, funding and following clinical trials (disclosure: I worked with an org that helped recruit participants for one of them), founded a scientific institute at Stanford, reads science textbooks, sometimes tracks down the raw data of studies and replicates the analysis, etc. He doesn’t have a degree but I would give him >50% odds at being able to pass at least undergraduate science exams with fairly minimal additional study. The bar for “actual scientist” should probably be based on “has some verifiable scientific knowledge” and “has actually practiced science”. He seems to meet both bars in a binary sense, the question is really about how much of each is required, unless you insist on credentialism.
Maybe? I suppose that's fair and I realize I'm going to open the door to a lot of nitpicking on who constitutes a scientist and who doesn't. I would argue that Patrick Collison is extremely scientifically literate (if you took a slice of only the most educated set of the populace, he'd probably still be in the 99th percentile). I'm wondering if he himself would consider himself a scientist; my guess is that he probably wouldn't (I'm not sure he "has actually practiced science"). But I don't know and it really wasn't the point I was trying to make. I'd like to see a guest post from Eric Topol or Angela Rasmussen or Derek Lowe or any number of top scientific bloggers who are very good at mass communication and to whom most of Yglesias's readership hasn't been exposed.
My overall frustration though, with this post and the several Yglesias COVID trial-related posts over the past few years is a lack of acknowledgment as to *why* clinical trials are the way they are. It's really not because of excess bureaucracy! It's because we've set up the burden of proof to be on the drugmaker to demonstrate both safety and efficacy before they are allowed to market their drug.
So yes, by all means the government should be investing billions of dollars to speed up manufacturing at risk. That's money well spent even taking into account the drugs/vaccines that will fail trials. And yes, it's inexcusable that the FDA hasn't given the green light to the second generation Omicron-specific vaccines. Where I object is statements like, "For many vaccine candidates, we could run human trials concurrent with primate trials (once basic safety data has been obtained)." The parenthetical there is doing a LOT of work, and especially so for new platforms. Safety data necessarily needs at least a somewhat large trial to have the statistical power to be meaningful, and you can't simulate time. That's enormously frustrating but all the models in the world can't replace real world data.
I agree with your first paragraph (which is why I tried to preface with the note about it being a meta-question), and apologies to you for getting this comment (I could have replied to a number of commenters asking the same question). It’s really just a thought that bubbled up while reading all the requests for “actual scientists” and the few trying to reduce Patrick Collison to a generic tech founder.
> the several Yglesias COVID trial-related posts over the past few years is a lack of acknowledgment as to *why* clinical trials are the way they are. It's really not because of excess bureaucracy! It's because we've set up the burden of proof to be on the drugmaker to demonstrate both safety and efficacy before they are allowed to market their drug.
Your second sentence is both true and not at odds with your first statement being false. The system we set up is *highly bureaucratic*, intentionally. The argument is that this system is not the optimal one! It’s true that the reason we have such a bureaucratic system is that we have set up a system that requires proof of safety and efficacy, *as determined by a bureaucracy*. You even admit that it’s inexcusable to have not approved the new Omicron vaccines! The entire reason for that is the bureaucratic process. Sure, real-world data is important, but people like Patrick Collison are arguing for *actually allowing the collection of real world data*. Ultimately, someone is going to decide to just route around this and go somewhere that allows them to actually move forward (although I guess the US being the primary consumer mitigates this to some extent, and is another example of the US being able to just kinda muddle through on its inherent advantages).
It’s impossible (to me, I guess) to look at what is happening, say, with monkeypox vaccines, where we literally own millions, but only thousands were available because the storage facility certification had lapsed (IIRC that it lapsed, not had never been certified), the FDA never got around to recertifying it, but then the EU certified it and the FDA announced they’d honor said EU certification, but never got around to *actually doing whatever bureaucratic paperwork to officially honor it*, and then claim that excess bureaucracy isn’t a problem. We used to have a Lyme disease vaccine! Even after the COVID vaccine trial results were announced, it took *weeks* for the FDA to get around to reviewing it. Why is an American university going to South Korea to trial a new platform? Why has Novavax been available outside the US for a year (even if one concedes the initial delay in the US was justified, those issues have long since been resolved, not to mention the real-world evidence we’ve seen from people who _actually received it_)? Why do I have to buy the good Japanese sunscreen from 3rd party sellers on Amazon that list it as a “watery essence”? The evidence is endless about this.
I may not have been entirely fair in addressing your specific example of a statement that frustrates you, so I’ll try to be better here. You’re frustrated with statements like "For many vaccine candidates, we could run human trials concurrent with primate trials (once basic safety data has been obtained)." I am reading your specific frustration as the parenthetical obscuring that obtaining basic safety data is time-consuming. Ok, so let’s say we grant that, and even go so far as granting that it takes the majority of the time devoted to trials (this is actually the opposite of how I understand it, but I have low confidence that I am correct about that). Doesn’t that still mean that running the subsequent trials concurrently would save time?
I will admit to being a bit confused as to why we’d still need the primate trials at all if we’re running the human trials at the same time - do the primate trials allow us to study something adjacent or inform future vaccines?
I have very little to object to in either of these replies. I'll only say that I didn't mean to imply that excess bureaucracy was *not a problem*, simply that the safeguards built into clinical trials are there for good reasons (the old Chesterton's Fence analogy).
I place new versions of mRNA and other vaccines already demonstrated to work and be safe in the same category as seasonal flu vaccines. The FDA already has a framework to deal with this and I share the frustration that this process hasn't been expedited.
Where I believe we need to be more cautious is with new technologies and/or modes of action. Is our current clinical trial framework optimal for this? Almost certainly not. But I don't think it's as far off as the libertarian voices would have us believe.
I think that is almost certainly true. A generally safe bet is that the most extreme complaints are likely overstated in reality (or, at the very least, need to make the concessions that reality is not necessarily optimal). I personally believe that the FDA and CDC are far gone enough that we should basically just be collecting complaints, stack ranking them by how easy they are to fix (even if that means conceding to political considerations), and moving down it. I think that is a POV shared by many vocal people, and is a POV that leads to underrating the value of explaining how we got here (and why it’s no longer appropriate). I fall victim to that too.
Well, we know there are small but definable risks for myocarditis and Guillain-Barré Syndrome for the mRNA vaccines, for example. There are always risks and potential safety issues with any new therapy.
Huh? You're arguing that side effects from drugs or vaccines are unexpected? On the contrary, Guillain-Barré Syndrome is a well-known side effect from many vaccines and is something that clinical trials specifically look for. And again, the reason to run larger clinical trials is specifically to identify (and quantify) these safety risks. 1% is different from 0.1% or 0.0001%, etc.
More importantly, the human body is really, really, really complicated. A drug or vaccine that targets a protein or pathway of interest will almost certainly hit other pathways as well. Those of us in drug discovery know to look for these complications and do whatever we can to mitigate them.
The current mRNA vaccines are delivered via nanoparticles encasing the mRNA. I couldn’t tell from the level of description given here whether this is exactly the same as the new ones (just the new ones have chosen fractional protein targets instead of full proteins) or if there’s a different delivery method.
Of course, "nanoparticles" encompasses many different *types* of nanoparticles, all of which are going to behave a little differently. But still, having the mRNA vaccines delivered that way safely is a good precedent to set.
Nanoparticle-based vaccines cover a huge realm of different modalities whether of biological (e.g., self-assembling protein nanoparticles, virus-like particles, exosomes) or synthetic (e.g., ferritin) origin. Some of these groups have other assets in later stage clinical trials and so the modality risk has been addressed somewhat in those studies if we (and regulators) want to make that logical "bridge".
Again, it will vary substantially between groups and for different approaches, but generally the issue of regulatory hurdles are big one that prevent the rapid deployment of new assets from the same platform. Hopefully the FDA is loosening up to this, so that we can address safety on a modality-by-modality basis and move to more quickly deploy differentiated vaccines.
Matt, would you be open to having someone who takes the other side of this debate, or is at least more concerned about the risks of rushing a vaccine that doesn’t work or hurts people, give their side of the story? I’m not able to evaluate this argument, and I bet most of your readers are in the same boat. I can identify concerns - I mean, this is a developer saying he knows about this *amazing* technology, and why isn’t everyone getting out of the way to let it come to market. That take has earned at least as much suspicion as government regulators, especially in the health area (Theranos, anyone?)
What would be especially helpful is to get beyond directional arguments (faster! slower!) to get a better sense of the specific requirements that take a lot of time and what’s the risk of not running them. E.g., he complains about having to run animal trials. How much time is involved, and what’s the risk in running them concurrently instead of in series? Can we focus on some big-ticket items to evaluate the debate? If it really is about death by 1000 cuts, ok, let’s still engage on that: where is the obstacle?
One perspective taking measure: for all the complaints about FDA, didn’t the US have the vaccine broadly available earlier than any other country? It’s not as though there is another paragon country out there we are lagging, is there? Or if there is it would be good to know.
Obvi it’s Matt’s prerogative to set the editorial agenda of his own blog, but I do want to second Theodore’s initial comment about how useful it would be to know what these types of posts are arguing against.
“…lack of basic institutional seriousness” and “self-imposed straightjackets” seem like harsh characterizations of differing values and principle stacks. The author makes it clear that his #1 goal is speed, so any barrier to that is a straight jacket, but speed is not the #1 goal of the FDA, so those straight jackets are in service of something else. I wish speed were more of a priority, maybe it should be the top priority, but I fail to see how it not being so makes them institutionally unserious.
I'd say that the problem is that FDA does not seem to consider delayed benefits (and costs) in the way it designs its cost benefit analyses.
I don't have the background to comment on the nature, scope, and appropriateness of the regulatory burden faced by biopharma with regards to vaccine development, but some of his statements about the fundamental science raise eyebrows.
(e.g., he proposes using antibody titers as a surrogate endpoint in clinical trials, when we know that short-term elevations in antibodies do not have robust correlation with long-term protection against infection or long-term protection against severe illness).
The argument against accelerating approval for pandemic vaccines needs to be distinguished from the more general libertarian position that drug approval barriers should be lowered. I've spent my career in pharma and medical devices, in commercial roles but continually working on clinical studies and interpreting their data.
Those advocating against faster pandemic vaccine rely on straw man arguments. It's difficult to find somebody to competently take this position. A protest "review," (https://tinyurl.com/3bzn6ph4) which included two authors who quit the FDA in protest of booster approvals, is representative. This and others tend to rely on three arguments: 1) vaccines are not riskless, 2) the positive data for [insert new vaccine or application] is not yet definitive, and 3) moving quickly will undermine confidence in vaccines. Instead of reckoning with the cost-benefit tradeoffs of the decision, the logic is one-sided : highlighting risks of approval but not of delay. And, not only does the "if you disagree with me, people will be more skeptical of vaccines" have no data to support it--note people who critique a lack of data while simultaneously making unsupported claims--but I've anecdotally found that most such claims in the pandemic ended up being completely wrong.
The more general libertarian view has two flavors. One, call it the "strong acceleration" view, wants to shift decision making from regulators to consumers and their physicians. One mechanism is making the FDA approve any drug approved by selected other countries (a lowest-common denominator approach to regulation). Another is expanding compassionate use programs where people can access unapproved drugs for extreme cases, but making the bar for this much lower. In my personal view, this is batshit crazy. Anybody who has seen how drug approvals flow to drug promotion and then to consumer choice would recognize the risks here. This is an area where having somebody write a post would be very valuable.
There is a much better argument, call it the "weak acceleration" view, that calls for streamlining the approval process in many ways. I'm very sympathetic to this. Arguments against it often have the straw man quality of those against pandemic vaccine acceleration.
I basically agree with you, but would like to steel man the argument for robust efficacy endpoints prior to vaccine booster rollout (I agree that safety concerns and the armchair social psychology are weak objections at this point, at least with the mRNA platform):
The original vaccine series had a huge and relatively durable effect size on major outcomes like hospitalization and death. The first booster had a much smaller absolute effect size, but did improve hospitalization and death in higher risk populations and people > 50. The second booster has shown minimal effect on any clinically significant outcomes, even in high risk populations.
It's not clear to me, in a landscape where we have relative equipoise on the efficacy of future vaccine boosting, and in a patient population (vaccinated and boosted) where the risk of serious adverse outcomes approximates that of seasonal influenza, that we owe Pfizer and Moderna the right to extract billions of dollars in perpetuity on the basis of expedited, poorly correlated surrogate markers like antibody titers.
That's an excellent point. Thank you for making it. it underscores the need to balance the risks of approval with the risks of delay. The combination of the significant benefit of the vaccines, extremely safe side effect profile, and impact of the pandemic all argued for speed and lowering bars to appeal.
But, a future vaccine with a worse safety profile or lower efficacy could change that calculus.
And I agree that the push to only use antibody titers is not clear cut. I view that as more of a "break glass in case of emergency" option. Healthcare is replete with times when surrogate markers did not correlate with health outcomes.
I'd argue that the commercial rewards for vaccine development are still too low. Covid has been the exception to the rule. But you're right that initial approval should not be able to be gamed into a perpetuity.
Yes the original covid vaccines were ready extremely quickly. But the problem is that now it's business as usual. And business as usual is a horrible regulatory state
But net of the entire performance of the govt, did any other country make the vaccine available more quickly? I’m asking for a big-picture benchmark. Takes descend very quickly into charged language (“wasn’t in any hurry”) and I’m skeptical that these are well informed. I’m open to believing things can be done faster - and I want them to, if the risk-benefit makes sense - but it would help to see the homework. It’s relevant to know how fast other regulators acted. And maybe, even if other regulators took the same amount of time, roughly, it only means that all the regulators are too slow. But if you want to make that argument, ’d like to know how you arrived at it: what risks were they too cautious about, and what is the improvement?
When pharma wants to justify high drug prices, they go on and on about the difficulty of drug development: they have to try tens of thousands of molecules to find a successful drug, and candidates fail at every step along the way. Succeed at phase 1? You might fail phase 2. Make it through phase 2? The results at phase 3 might uncover a very serious adverse even that affects 1 patient in 300. And so on. But now come the people who want things done very differently and it’s all impatience: why are we doing all these stupid tests? What a waste! We know these drugs work and are safe! I am not an expert, and I can’t really adjudicate these disputes. I would like to see experts, who are well informed, pick out what they see as the key risk-benefit pivots - where we would save a lot of time - and ask: do we need to do this? Why or why not?
To answer your first question, the UK administered the first Pfizer jab outside of clinical trials on Dec 8, and for the US it was Dec 14 (Sputnik doesn't count, sorry Russia). Not sure this is praise for the NHS, but the US and UK were more or less tied for first. Roll out, distribution, voluntary uptake, etc., are different questions of course.
Israel absolutely smoked everyone else at getting vaccines into arms. In the UK we didn't make the vaccine generally available until long after the US did, but that's because we did a much stricter rollout by age cohort, aimed at maximizing lives saved - in terms of jabs administered per capita, we've been ahead of you for most of the time.
The US paid a lot early on to get ensured early access to several hundred million doses each from several manufacturers (including some whose products were never approved but still got the money). The EU made the big different decision of negotiating a bit on price, and saving several billion dollars on drug costs, but getting them a few months after the US.
I think the regulatory process was similarly slow in all countries (though as others mentioned, the UK did grant approval a week or so earlier).
And all things that the UK regulators did.
It's interesting that there's basically no constituency for this. People on the right largely never thought covid was a big deal, so them taking action now all of a sudden won't happen, and people on the left seem very hesitant to criticize the FDA or the public health establishment, so they're not going to support reform if they won't admit reform is even needed.
So we're just left with bad policy. It's unfortunate.
We need smart people who believe in science to stand up and say the quiet part out loud: the FDA, NIH and CDC have all been proven incompetent. Reform and urgency are both very much needed. But the opposition from the right seems to make everyone reluctant to do that for fear of looking like a Trump supporter or anti-vaxxer or some such nonsense. This post is a great step in the right direction. I hope to see this type of column in the NYT and be mainstreamed.
Well you can be nicer and just say they fail to use the best methodology in making their decisions, something like rule utilitarianism.
Yes, assuming it's accurate, the most persuasive point in the article that this is a $1 trillion bill just laying on the sidewalk, whatever you think about it's overall significance in the scheme of things. But there're probably a lot of $1 trillion bills, or at least large ones, laying around that we don't pick up, because they're not important enough to anybody in particular.
It's important to remember that this has been true in every human society throughout history and will be true in all future ones.
There's a tension between crying endlessly over all the milk that never even came into existence because it was "pre-spilled" by various failings of this sort, which will make anyone miserable, and pushing to make concrete changes now and again, which is what's necessary to become and remain a developed country.
That we are a developed country, and a particularly rich one to boot, suggests that historically we've done fairly well at making those changes.
Exactly, this is pretty how a collective organism like a human society makes decisions about where the cutoff is between a collective problem that everyone needs to stop and focus on, and one that that can be muddled through.
Call it hive/swarm decisionmaking vs centralized decision-making. The difference is illustrated by the contrast between how the US and China have responded to covid.
But on the margin are we pushing as hard as we used to in order to overcome the obstacles to progress?
Again, sounds an awful lot like a presentist fantasy to imagine things were once vastly different.
A major premise of this blog is that, in relation to bureaucratic efficiency, the past really *was* vastly different.
Yes, but I would argue we are no longer so good at it
I think that's mostly a presentist fantasy. We've always been like this; pluralism requires it. From the earliest days of the Republic (the colonies, even), folks have always set out to protect their slice of the pie, then the dam breaks or things really become untenable and those barriers get washed away, then new ones arise...
The wide-reaching Federal role in economic and regulatory life is newer, but I have few doubts that its current descent into sclerosis will be subject to the same interplay of popular pressure, economic sustainability, and insider pushback/protectionism. It did happen in the 1970's and 80's, then a bit of a swing back in the 90's and 00's, which gave birth to the modern regulatory state that sits atop a mostly deregulated economy.
Again, there's a balance to be struck between gibbering panic about all the "bills on the ground" that serves no one versus steady popular pressure on the organs of power to permit and even force concrete, if incremental, change.
Things mostly work, well enough, for most folks.
That's probably the best we'll ever manage until someone figures out how to make us something other than human.
The right would respond by saying that if we stopped caring about covid and sucked it up and got on with our lives, we can pick up a bill greater than $1 trillion. The left would respond by saying that why are you caring about money when it's clearly less important than people's lives.
(But yeah I agree that that's the most persuasive point.)
The point is to care optimally about COVID; neither refuse to get vaccinated nor to close outdoor playgrounds.
I would rather put it that there is no apparent political feasibility. It seems like one of many many no-brainer centrist position that in fact would have a very healthy constituency in terms of percentage of the electorate supporting some of the suggested moves (and certainly the goal of government doing more to facilitate the speedy development of safe and effective next generation of covid vaccines and therapeutics) but it's rather the polarized political landscape that makes it difficult to get the popular will to actually matter.
The point of this Substack is to reform the Left on this as on many other things.
Id volunteer
"Virtually everyone knows plenty of people who contracted COVID after being vaccinated and were out of commission for a week or more." Having just had COVID after being vaccinated and boosted, while getting sick wasn't super fun, easily the worst part was being essentially unable to work and being trapped in the house for *weeks* as COVID moved through our family and we desperately waited daycare's required quarantine to end, even though the kids are fully vaccinated and were symptom free shortly after testing positive. I think encouraging liberalizing daycare COVID restrictions will be very popular in a couple of months after all the little kids have had an opportunity to be vaccinated.
Did you test your kid(s) for covid every time they got sick this fall? Mine got sick from daycare so many times it just made no sense to me to try to schedule an appointment, wake up really early, drive down to the testing place, wait in line for an hour or more, etc..
She was probably sick more than a dozen separate times from Oct-March, and we had the same symptoms around half those times. I assume that one of those illnesses was covid but I guess I'll never know.
All this is to say: I'm curious if your experience with testing is similar or not. I suppose now it would be easier to do the tests than it was in the winter, so maybe I will get one for her next time she has a cold. But sometimes I wonder why so many people are still testing for covid if they're already vaccinated and boosted. For young, healthy, vaccinated people, what difference does it make if the test comes back positive or negative?
I wish we had multi-virus tests. It seems that it should be not too much harder for Abbott to make a little card with four potential stripes - one control stripe, one that turns red if you have COVID antigens, one that turns red if you have flu antigens, and one that turns red if you have RSV antigens. I don’t exactly know how much good would be achieved by being able to know which of these you have, but I do think this is one of those things where having the ability to have widespread knowledge would enable better responses.
As someone who will shortly have two children under two, I would love to be able to test visitors for RSV on their arrival to my home. That would be quite valuable to me.
As someone with 4 kids under 11, and one still under 2...you are never going to avoid RSV unless you live in a bubble.
Same for many of the other childhood illness.
And RSV in particular sucked for my kids. 3 of them had 104+ fevers from it at some point ~2-3.
I hear where you're coming from, but I think my conclusions are most like Belisarius below. They just get sick so often, for my kid she'd get a cold or fever every other week from about November to March. Maybe if I knew first exposure was less risky at age 4 than age 3 then delaying the inevitable makes some sense. But my sense is its worse for younger kids mostly because it's their first exposure, not because of their age.
Seems good to do what you can to avoid newborn exposure to RSV. I’m far less concerned about my robust toddler.
That's a fair point. When you said two children under two, I didnt think that one might be a newborn.
It's probably worth some effort to delay, even though success is...unlikely.
Especially if toddler is in preschool/daycare.
Going back to an earlier discussion; general surveillance testing of endemic respiratory illnesses is the single greatest need we’ve got.
We have no idea what the endgame of a pandemic actually looks like.
I *suspect* we’re vastly closer than the expert community believes, but we have no data.
Agree overall that we should stop testing because it's not like we're trying to stop transmission anymore in the US
If you're a rule follower and you test positive for the virus, close contacts are supposed to test. Kids would be close contacts. We have a very similar story to OP and the reason we tested the kids is we were supposed to as they are close contacts; and of course, camp / childcare does not want your positive kid attending even if they are asymptomatic.
We tested a lot, mostly because both kids are in the Moderna trial. We also mostly did rapids, even though those aren't technically approved for kids under two.
The test that started our quarantine was we were headed to a party with grandparents who have risk factors and my wife was symptomatic.
Makes sense
Ah but you were a victim of FDA not approving dirt cheap tests in early 2020. Asymptomatic testing could have avoided a lot of the cost of restrictions on (and fear of) social interactions
My experience is similar. I know plenty of people who have had Covid who were vaccinated and it was just annoying. Even in people who weren’t boosted. I would prefer Covid to a regular old rhinovirus because it was milder and I recovered fully after 3-4 days.
Yeah, the author chose to, ah... operate at a credibility deficit... with language like that (see also post-vax long-covid doomscrolling). I have no idea why. It seems fairly common, like you need to establish bona fides by being Still Rather Concerned. I think it plays the opposite.
*The worst* post-vax covid case I personally know apologized when he called into the next week's meeting, because he hadn't gotten as much done on the project as planned on account of "covid kicked my butt". He wasn't "out of commission"--he was working throughout the dang thing!
FWIW I was vaxxed and Omicron put me out of commission for 3-4 days
The majority of my career was in the biopharma industry doing drug regulatory affairs and safety. Mr. Collinson's column while interesting needs to be put into perspective. One major hurdle for vaccines is doing the necessary safety studies to demonstrate that the benefits outweigh risks. Lots of the side effects seen with vaccines are small, if present at all, and on the order of 1 in 10,000 or less. You cannot pick up these in a normal clinical trial of 3000 patients. This necessitates that vaccine trials be done on a far larger cohort and is one reason the FDA is more conservative than looking at a new cancer drug.
Vaccine manufacturing is not trivial. Vaccines are sterile products made under exacting conditions. However, one the process is validated, it is possible to make changes in the vaccine to adopt to new variants. This is why we are able to get new seasonal influenza vaccines each year. Public health authorities and the vaccine industry agree upon the circulating flu viruses and change the production strain. Major regulatory filings and review are not required. We are seeing this approach being applied by both manufacturers of mRNA vaccines right now. A similar approach can be taken by companies who are close to approval of adjuvant/protein vaccines.
During the first year of the pandemic, I put out a daily newsletter for former industry colleagues and friends surveying the research that was being done. I stopped once the first vaccines became available. I looked closely at vaccine development and saw a lot of interesting work going on. Some of the projects are still going on. Canada has licensed a tobacco produced protein vaccine and the ferritin nano particle vaccine developed by Walter Reed is in clinical trials right now:. Volunteers are still being sought to finish up the Phase 1 trial: https://www.wrair.army.mil/node/645
Nasal vaccines have been used in the past with varying degrees of success. A commercial influenza vaccine has been around since 2003 but is not recommended for those > 50 years of age. There is also a bacterial delivery system that will get Covid antigen to the gastric mucosal system.
A number of vaccine approaches were under development several years ago when Zika virus was thought to pose a major public health problem When the virus disappeared these projects were shelved and a major opportunity to test vaccine platforms for the development of neutralizing antibodies was lost. Had such work continued it might have provided valuable information that could have helped out Covid vaccine developers.
There are a number of other issues that can be discussed but I'll stop with these. Intrepid uses of google can find the archive of my Covid-19 newsletter fairly easily.
"Lots of the side effects seen with vaccines are small, if present at all, and on the order of 1 in 10,000 or less"
Ok, so? Tell people there might be a 1 in 10k chance of something going wrong, pay out compensation if it does happen, and then release it on the market.
Right?! Of course the only way to measure a side effect like that would be to more quickly run larger trials. So highlighting this risk is not really a counterargument to speeding up the process. And this is to say nothing of the probability of death from COVID for the elderly without a vaccine. It's basic cost-benefit analysis that the FDA fails terribly at: death > rare possible side effects. The haunting reality is that Moderna had already developed and manufactured the vaccine for testing before the first American died from COVID.
https://johnhcochrane.blogspot.com/2020/12/free-market-vaccines.html
Logistical roadblocks should be fixed. There's no reason that studies should be delayed by monkeys being stuck at customs or supply chain issues with adjuvants.
But when it comes to lowering the barrier for starting human trials, I'd want to see some concrete scientific and bioethical basis for this change of procedure. I understand we all want faster role out of therapeutics! But each decision is a cost/benefit problem, and the costs of a harmful or ineffective therapeutic, especially one widely administered to healthy people, being released is potentially very high.
We need some clear cost/benefit analyses.
I think this is an area where the public health community’s believe that everything has to be either forbidden or mandatory does real harm.
You could relax human trial requirements and then say to the public, “we are releasing this but we also totally understand if you don’t want to take it; what we’re going to do is closely monitor the people who do want to take it and gather a bunch of real world data that we hope will demonstrate safety and efficacy or else raise red flags that cause us to pull back.”
it’s amazing how well freedom can work when there are a few, common sense side rails and people acting in good faith
It seems like there’s a common (sometimes in bad faith, but I think often not) blending of “approving this puts people at risk of harm from the treatment” and “approving this puts people at risk of a false sense of security, leading them to take risks that they wouldn’t take if the treatment was ineffective”. The former is a straightforward safety question - basically no one serious that I’m aware of argues for skipping the safety trials prior to approval. The latter is generally armchair sociology - this may be true, but it’s not like the FDA is staffed by sociologists! A very specific class of experts has gotten carte blanche over decisions that involve fields in which their training provides no inherent expertise.
I entirely agree that we need better cost/benefit analysis, but that means including the costs of *not* moving faster. Speed is not inherently good, but the reason it feels like so many of us are pushing for it is that there is a cost to slowness as well that is not actually being incorporated into decision-making.
What real harm? So far, there's been zero evidence that anything beyond two shots and a boost (+/- a bout of infection) does anything to move the needle on your durable risk of bad outcomes from future SARS Cov2 infection, including Long Covid. And truthfully, now that we're over a year out, there's still not great evidence that booster did much in people <50.
The problem we are facing is that the primary series of the mRNA vaccine worked really, really, really well - an effect size that staggers most of what we see in biomedicine. Durable marginal improvements on that effect are going to be hard to see even in carefully controlled prospective data, much less the incredible noise you're going to see with elective uptake in additional mRNA shots
This is pretty similar to the arguments routinely advanced by Scott Alexander & commenters at ACX for broad scale pharma deregulation, and the issues are the same: by abandoning prospective randomization and a priori demonstration of efficacy in favor of 'safety only' , you open the floodgates to snake oil. Even with our current FDA regime, healthcare and biomedicine has basically eaten America's productivity and GDP gains over the past 50 years. I shudder to imagine what they (ok, we) would do if they didn't actually have to prove anything worked.
This scenario where you lower the testing threshold for allowing human trials would likely require an increase in tolerance for deaths and adverse events during drug testing. Maybe you'd say it's worth it to save more people from the diseases these drugs are for? But you'd have to be honest about your increase in tolerance for bad things happening in drug trials.
I’m on the fence on challenge trials because I think there’s so much other low-hanging fruit that skipping that debate might be beneficial in terms of how quickly we can get improvements elsewhere. That said, those that I’ve seen advocating for challenge trials are not advocating them for the safety trial phase, rather for the efficacy phase. Efficacy trials are already in some sense challenge trials (if you ran an efficacy trial with no chance of viral exposure, you’d learn nothing). It’s true that you may encounter higher levels of negative outcomes on a percentage basis in a challenge trial, but you can also run them far more quickly and with a smaller pool of participants (because your variance in exposure will be lower, your required sample size will be correspondingly smaller), so you could end up with a higher rate of adverse outcomes but a smaller absolute number.
And most bioethics frameworks consider it worse for a healthy person to die from a drug they took during a trial than it is for someone with a disease to die because there isn't treatment...
Why?
*How much* worse is it?
I don't have an answer. I'm sure bioethicists have ways to calculate that into a model, but I don't know how much worse it is.
I get that you're explaining rather than advocating here briross, but this "framework" is what needs to be changed. While it's important that no one should be tricked into volunteering for something where they don't have full information about, it ISN'T worse for society if people die from an experimental treatment rather than a disease, so the current framework is the problem.
If people are willing to pull the lever to help save the lives of those who may be run over by the trolley, even though this increases the chance that that the trolley may run over them, we should be thanking them rather than having rules that prevent them from making this choice.
Can you point me to an example of bioethicists doing a cost-benefit analysis (i.e. "calculating" something in a "model") and coming out against human challenge trials?
As far as I can tell, they never bother to do that; it would be very hard for any cost-benefit analysis justify the level of conservatism they desire without some obviously arbitrary parameter to make deaths in trials *much* more "expensive" than deaths from the actual disease.* As far as I can tell, bioethicists that oppose human challenge trials mostly take absolutist stands and pontificate. I've said it before: if bioethicists did cost-benifit analysis, they'd find out that the world would be a better place without bioethicists.
* To be fair: lots of people like these arbitrary parameters that are based solely on personal preferences, such as Blackstone's ratio. But different people express wildly different values for their ratio (e.g. Ben Franklin and Blackstone disagree by an order of magnitude --- and others differ by yet another order of magnitude), so I don't think that such a parameter can be meaningfully used in a cost-benefit analysis; if a personal-preference parameter can vary by several orders of magnitude, it is just a fudge factor to ensure that you get your desired result.
You could do that, but for an OTC product like a vaccine, where the public's access does not run through a learned intermediary prescribing physician, who can explain the nuances of the scientific uncertainty and risk-benefit calculation to lay patients in terms they can understand to make sure there's informed consent and make an independent decision, it would necessary to provide some kind product liability protection for manufacturers. The normal prescription drug or vaccine liability protections wouldn't apply in the situation you're describing, and manufacturers would be at risk of major mass tort liability that could easily swamp the entire profit of the product even if it the end of the day it turns out the lawsuits are bogus.
Yep. Don't trust the vaccine? Don't take it! Just don't try to stop other people from taking it if they want to!
The upside of a pan variant vaccine seems pretty worth a challenge trial, so long as the participants know that the risks, even if low, aren’t zero. Hell, going to the moon had far more risk for arguably far less upside.
It seems like we're in a position where we let ethicists dictate what is acceptable to society, rather than society laying the foundation for ethicists.
Agree, but its not clear ther current approach is supported with a sound cost benefit analysis.
It also comes down to the bioethics framework. It's like the trolley problem--do you pull the lever and kill one person to save five? The current framework is very conservative about pulling such a lever.
Is there a link to that framework?
I don't deny the complexity of the problem along multiple dimensions, but I've struggled to find math behind the various covid approvals or requests for additional data. If you're going to ask for a trial to be extended several months to gather additional data (or just telegraph it will not be approved without it) you should be transparent and rigorous about your assumptions. What risks that data would reduce and the cost in lives/dollars of the delay laid out in a comprehensive manner.
This I don't know. I'm just going based on what I learned in my research ethics courses. I don't know how the math is actually calculated, and how precise of a science weighting these costs/benefits is.
I’m aware of Fast Grants but I don’t think the CEO of a payments firm was the “scientific expert” the commentariat had in mind the other day.
Yes, I know this was written and scheduled before then and is sheer coincidence.
If folks want to read a scientist I’d recommend Eric Topol’s Substack.
What I’m interested in reading in this area is the intersection of science and policy. How does regulatory policy shape what’s possible with regard to progress with new Covid vaccines, and what are the levers we could push to change policy, speed innovation, AND still produce a safe and efficacious product? This collaborative post is a nice push in that direction, thank you.
Of course, but now and again bringing in another voice to corroborate yours is sensible.
Following this suggestion, I subscribed to Topol and read the last two columns. He's obviously a very smart and technically well-informed guy. However . . .
In his May 15 post "The Covid Capitulation" (https://erictopol.substack.com/p/the-covid-capitulation) he describes how we're in the midst of a dangerous new COVID wave and we're capitulating to it. Almost two months later, one can say that we're not in the midst of a new wave, even though of course we'd like to see deaths go down much more (https://www.worldometers.info/coronavirus/country/us/). So no explosion from the BA.2 variants.
His most recent post on June 27 "The BA.5 Story" (https://erictopol.substack.com/p/the-ba5-story) paints a worrisome picture of how bad a BA.5 wave will be. Maybe! I hope not, but unlike him I'm no expert. But even still I'd tell the experts to be careful about predicting the next dangerous wave because if it doesn't happen, your future predictions of the same will be heavily discounted.
“Will be”… I would correct to “already have been”
We’ve been listening to the experts doom-monger since day one. None of those terrible predictions have come to fruition.
Arguably, that’s because we listened to them before careening off the cliff in early 2020.
But since then adherence to their recommendations has fallen to basically nothing, especially in the aftermath of the wide availability of vaccines.
And still, the world has not ended. Just a steady trickle of deaths which was always unavoidable.
I dunno, a million dead in US alone seems pretty dramatic to me
What are we comparing it to?
Because there's no way by which we stop that. Maybe, maybe, if we do absolutely every goddamned thing right from a public policy perspective, starting with Trump, we reduce the death toll to 600,000. Near-universal, bi-partisan masking and limited public interactions through early-to-mid 2021, a blind sprint to vaccinate the elderly and vulnerable followed by everyone else, vaccination rates hitting Icelandic levels (mid-80's), prompt FDA streamlining to roll out variant boosters...
But there is just no world in which our technological toolkit is sufficient to *stop* COVID, regardless of policy. And viewed in that context, all of the agonizing coming from the public health people has been worse than counterproductive. They've been mostly wrong on everything except the initial 2020 call to panic, and it's impossible to know how correct that was because we followed it. Basically every time they've made a call since, it's been incorrect.
China's plan to pretend we do have the toolkit to stop COVID, a.k.a. to ride Dynamic Zero COVID down in flames, doesn't seem great either. Even Australia's admittedly brilliant transition from zero COVID to "vaccinate everything that moves and get on with it" wasn't within reach for the United States. We have actual borders and there was no legal framework to enforce Australian-style lockdowns every time something slipped across them.
Seriously ? Mortality rates from COVID in us are far worse than in most peer countries. They could have been much worse still without initial lockdowns and without the (admittedly belated) mask mandates. Hundreds of thousands of lives in either direction are nothing to scoff at, in my book.
This specific CEO is also known as one of the best people at execution in tech. He’s not making a scientific case as much as he is making a case about removing self imposed obstacles that impede Getting Shit Done.
A good system generally needs a productive tension between the “getting shit done” people and the worry warts. Throw one side out of a system and expect the “get it done” people to look past risks that turn out to matter, maybe a lot. The worry warts will probably overlook the costs of inaction and crossing every t and dotting every i. That a GSD wants to remove obstacles is interesting and I’d like to hear from people who value the obstacles what he is leaving out.
He’s also the founder of the Arc Institute at Stanford. Not that this necessarily increases his knowledge, but his wife is a former CZI Biohub investigator (the virus hunters from The Premonition) and Stanford biochemist. I’m admittedly biased as I’ve followed Patrick Collison closely for a while and am involved in various efforts that he’s supported, but he’s quite clearly got more expertise here than reducing him to “payments CEO” implies.
Are nanoparticle-based vaccines safe though? I've seen a lot of hype around nanoparticle delivery for a number of drug classes (with good reason!) but there are serious questions about how the body clears them. Which again, is why we run large clinical trials in the first place, to tease out these effects before they're shown to be harmful.
Meta-comment here: I'm familiar with Patrick Collison and the good work he's done, but I'll restate a comment I've made previously: it would be beneficial to have some guest posts written by some actual scientists. It's fine for Mr. Collison to tout this work, but I'd rather read someone poking holes in the argument and then showing that the benefits outweigh the risks, rather than ignoring the risks entirely.
I agree that more guest posts (and best of all, a back and forth between those on the opposite sides) would be valuable.
I have a meta question though - what does it mean to be an “actual scientist”? I think there is a defensible case that Patrick Collison does not meet a reasonable bar, but let’s actually look at what he’s done (as opposed to his lack of a degree) - he has spent the past two years working on Fast Grants, reviewing literature, funding and following clinical trials (disclosure: I worked with an org that helped recruit participants for one of them), founded a scientific institute at Stanford, reads science textbooks, sometimes tracks down the raw data of studies and replicates the analysis, etc. He doesn’t have a degree but I would give him >50% odds at being able to pass at least undergraduate science exams with fairly minimal additional study. The bar for “actual scientist” should probably be based on “has some verifiable scientific knowledge” and “has actually practiced science”. He seems to meet both bars in a binary sense, the question is really about how much of each is required, unless you insist on credentialism.
Maybe? I suppose that's fair and I realize I'm going to open the door to a lot of nitpicking on who constitutes a scientist and who doesn't. I would argue that Patrick Collison is extremely scientifically literate (if you took a slice of only the most educated set of the populace, he'd probably still be in the 99th percentile). I'm wondering if he himself would consider himself a scientist; my guess is that he probably wouldn't (I'm not sure he "has actually practiced science"). But I don't know and it really wasn't the point I was trying to make. I'd like to see a guest post from Eric Topol or Angela Rasmussen or Derek Lowe or any number of top scientific bloggers who are very good at mass communication and to whom most of Yglesias's readership hasn't been exposed.
My overall frustration though, with this post and the several Yglesias COVID trial-related posts over the past few years is a lack of acknowledgment as to *why* clinical trials are the way they are. It's really not because of excess bureaucracy! It's because we've set up the burden of proof to be on the drugmaker to demonstrate both safety and efficacy before they are allowed to market their drug.
So yes, by all means the government should be investing billions of dollars to speed up manufacturing at risk. That's money well spent even taking into account the drugs/vaccines that will fail trials. And yes, it's inexcusable that the FDA hasn't given the green light to the second generation Omicron-specific vaccines. Where I object is statements like, "For many vaccine candidates, we could run human trials concurrent with primate trials (once basic safety data has been obtained)." The parenthetical there is doing a LOT of work, and especially so for new platforms. Safety data necessarily needs at least a somewhat large trial to have the statistical power to be meaningful, and you can't simulate time. That's enormously frustrating but all the models in the world can't replace real world data.
I agree with your first paragraph (which is why I tried to preface with the note about it being a meta-question), and apologies to you for getting this comment (I could have replied to a number of commenters asking the same question). It’s really just a thought that bubbled up while reading all the requests for “actual scientists” and the few trying to reduce Patrick Collison to a generic tech founder.
> the several Yglesias COVID trial-related posts over the past few years is a lack of acknowledgment as to *why* clinical trials are the way they are. It's really not because of excess bureaucracy! It's because we've set up the burden of proof to be on the drugmaker to demonstrate both safety and efficacy before they are allowed to market their drug.
Your second sentence is both true and not at odds with your first statement being false. The system we set up is *highly bureaucratic*, intentionally. The argument is that this system is not the optimal one! It’s true that the reason we have such a bureaucratic system is that we have set up a system that requires proof of safety and efficacy, *as determined by a bureaucracy*. You even admit that it’s inexcusable to have not approved the new Omicron vaccines! The entire reason for that is the bureaucratic process. Sure, real-world data is important, but people like Patrick Collison are arguing for *actually allowing the collection of real world data*. Ultimately, someone is going to decide to just route around this and go somewhere that allows them to actually move forward (although I guess the US being the primary consumer mitigates this to some extent, and is another example of the US being able to just kinda muddle through on its inherent advantages).
It’s impossible (to me, I guess) to look at what is happening, say, with monkeypox vaccines, where we literally own millions, but only thousands were available because the storage facility certification had lapsed (IIRC that it lapsed, not had never been certified), the FDA never got around to recertifying it, but then the EU certified it and the FDA announced they’d honor said EU certification, but never got around to *actually doing whatever bureaucratic paperwork to officially honor it*, and then claim that excess bureaucracy isn’t a problem. We used to have a Lyme disease vaccine! Even after the COVID vaccine trial results were announced, it took *weeks* for the FDA to get around to reviewing it. Why is an American university going to South Korea to trial a new platform? Why has Novavax been available outside the US for a year (even if one concedes the initial delay in the US was justified, those issues have long since been resolved, not to mention the real-world evidence we’ve seen from people who _actually received it_)? Why do I have to buy the good Japanese sunscreen from 3rd party sellers on Amazon that list it as a “watery essence”? The evidence is endless about this.
I may not have been entirely fair in addressing your specific example of a statement that frustrates you, so I’ll try to be better here. You’re frustrated with statements like "For many vaccine candidates, we could run human trials concurrent with primate trials (once basic safety data has been obtained)." I am reading your specific frustration as the parenthetical obscuring that obtaining basic safety data is time-consuming. Ok, so let’s say we grant that, and even go so far as granting that it takes the majority of the time devoted to trials (this is actually the opposite of how I understand it, but I have low confidence that I am correct about that). Doesn’t that still mean that running the subsequent trials concurrently would save time?
I will admit to being a bit confused as to why we’d still need the primate trials at all if we’re running the human trials at the same time - do the primate trials allow us to study something adjacent or inform future vaccines?
I have very little to object to in either of these replies. I'll only say that I didn't mean to imply that excess bureaucracy was *not a problem*, simply that the safeguards built into clinical trials are there for good reasons (the old Chesterton's Fence analogy).
I place new versions of mRNA and other vaccines already demonstrated to work and be safe in the same category as seasonal flu vaccines. The FDA already has a framework to deal with this and I share the frustration that this process hasn't been expedited.
Where I believe we need to be more cautious is with new technologies and/or modes of action. Is our current clinical trial framework optimal for this? Almost certainly not. But I don't think it's as far off as the libertarian voices would have us believe.
I think that is almost certainly true. A generally safe bet is that the most extreme complaints are likely overstated in reality (or, at the very least, need to make the concessions that reality is not necessarily optimal). I personally believe that the FDA and CDC are far gone enough that we should basically just be collecting complaints, stack ranking them by how easy they are to fix (even if that means conceding to political considerations), and moving down it. I think that is a POV shared by many vocal people, and is a POV that leads to underrating the value of explaining how we got here (and why it’s no longer appropriate). I fall victim to that too.
Well, we know there are small but definable risks for myocarditis and Guillain-Barré Syndrome for the mRNA vaccines, for example. There are always risks and potential safety issues with any new therapy.
Huh? You're arguing that side effects from drugs or vaccines are unexpected? On the contrary, Guillain-Barré Syndrome is a well-known side effect from many vaccines and is something that clinical trials specifically look for. And again, the reason to run larger clinical trials is specifically to identify (and quantify) these safety risks. 1% is different from 0.1% or 0.0001%, etc.
More importantly, the human body is really, really, really complicated. A drug or vaccine that targets a protein or pathway of interest will almost certainly hit other pathways as well. Those of us in drug discovery know to look for these complications and do whatever we can to mitigate them.
The current mRNA vaccines are delivered via nanoparticles encasing the mRNA. I couldn’t tell from the level of description given here whether this is exactly the same as the new ones (just the new ones have chosen fractional protein targets instead of full proteins) or if there’s a different delivery method.
Thanks! I should've known that.
Of course, "nanoparticles" encompasses many different *types* of nanoparticles, all of which are going to behave a little differently. But still, having the mRNA vaccines delivered that way safely is a good precedent to set.
Nanoparticle-based vaccines cover a huge realm of different modalities whether of biological (e.g., self-assembling protein nanoparticles, virus-like particles, exosomes) or synthetic (e.g., ferritin) origin. Some of these groups have other assets in later stage clinical trials and so the modality risk has been addressed somewhat in those studies if we (and regulators) want to make that logical "bridge".
Again, it will vary substantially between groups and for different approaches, but generally the issue of regulatory hurdles are big one that prevent the rapid deployment of new assets from the same platform. Hopefully the FDA is loosening up to this, so that we can address safety on a modality-by-modality basis and move to more quickly deploy differentiated vaccines.