Most likely the approval process which included running several studies cost them $100 million. It is not that they gave $100 million to the FDA. Running studies are very expensive, you have to pay doctors, administrators, supporting staff etc.
I'm a doctor and have been peripherally involved in several large clinical trials. I could believe that a multi year study involving many rounds of expensive investigations (eg screening MRIs in, say, the recent aducanumab study) could cost tens of millions of dollars, as this paper found (https://pubmed.ncbi.nlm.nih.gov/30264133/) - "Trials designed with placebo or active drug comparators had an estimated mean cost of $35.1 million (95% CI, $25.4 million-$44.8 million"
However Scott seems to be:
a) referring specifically to an 'FDA application fee', rather than the cost of running a trial and
b) talking about a relatively straightforward trial with short term clinical endpoints
so I'm going to assume he either misplaced a decimal point or is just mistaken. Happy to be corrected!
I believe he's referring to the cost of the trials combined with the fees.
Though I think I disagree with scott for a seperate reason.
The FDA tends to do a "he who pays for the trial gets the use patent" thing.
Before they ran a pair of full clinical trials to prove it worked as a sleeping medication a bunch of doctors might have had a hunch it had that effect due to patient reports but the evidence base was poor.
Just like patents encourage people to reveal their designs, drug patents and clinical use patents incentivise companies to pay to run the big clinical trials that would otherwise never get run which generate all the nice neat tables of effects, side effects and contraindications.
When you know that a drug is good for [use] becuase you've got the summary in front of you from a bunch of clinical trials it's hard to imagine a world where you don't know that information or only have a hunch about that.
It is kinda stupid for the FDA to not just routinely inspect substances approved in Europe and Japan for one's with a reasonably acceptable evidence base rather than waiting for companies to jump through symbolic hoops.
If something is expensive to create, we have a separate IP system for establishing temporary monopolies, which at least in theory can be tuned to produce economically efficient outcomes (e.g. protection that eventually expires by default). FDA approval has no such characteristic. There are problems with the IP system we have as well, but they are probably made worse by trying to also use the FDA to incentivize companies to do studies.
She sounds like a deeply pragmatic woman whose intensely committed to getting shit done.
She’s interested in saving babies and found allies to accomplish that goal. Shitting on the FDA would make things worse.
She could’ve written a libertarian blog or become a policy wonk but she focused on visible outcomes.
I think there’s a lesson to be learned here about the mentality of implementors versus wonks. Maybe she’s too short term and institutionally minded for this crowd but it’s a valuable perspective.
I think it depends on the level at which you want to get things done.
If you're a medical researcher, then you just have to accept the rules as the reality at which you operate in, regardless of how arbitrary and counterproductive they are. If the FDA mandated that double blind trials literally require everyone involved to be blindfolded, the only question to ask would be "what brand of sleep mask has FDA approval?"
If you're trying to operate at a political level and reform the system, then questioning the rules is exactly the kind of thing you should be doing.
It's a "game player" vs "game designer" distinction - they're two very different ways to approach the world, and while they can overlap you can't really be both at the same time.
Also, it was a speech she gave after winning an award for the work. Using a platform like that to drag anyone who was involved in the thing you're getting the award for is just ... not a thing that's done outside the confines of the internet.
It's not a normal thing to do, but it's been done. My favorite example is John Taylor Gatto's "The Psychopathic Schoolteacher," which was his acceptance speech for New York's Teacher of the Year award. https://www.naturalchild.org/articles/guest/john_gatto.html
From what I understand, companies live in fear of retribution from the FDA. Mary Ruwart writes about this in her book "Death by Regulation" providing a first hand account from her time inside the pharma industry. I suspect one of the reasons the AstraZeneca vaccine wasn't approved in January was retribution for a mistake AstraZeneca made in failing to immediately notify the FDA about a pause in the Phase III trial.
For academic researchers it's a different story, but they also have to work with the FDA too (for instance with trial design) if they want whatever they are working on to have a chance of getting approved. The FDA also provides funding to researchers, as we saw in Scott's last post, and grants compassionate use applications, etc etc.
Finally, researchers can be part of FDA advisory committees. I'm not sure if that is paid, but it is prestigious.
So I'm pretty sure fear of retribution is a factor when researchers are talking publicly about the FDA.
The EMA approved AstraZeneca vaccine but when the reports about blood clots came in, half of European national regulators suspended its use. EMA quickly reviewed the case and maintained that benefits overweight risks and suspension was not justified. The EU even sued AstraZeneca for missing deliveries at the same time deciding not to use AZ vaccine. They donated the remaining AZ vaccines to third-world countries but they come with short expiry dates and a lot of bad reputation that are increasing anti-vax sentiments.
For what it is worth, it seems that a stronger EMA would have been more beneficial for the EU. Similarly, without strong FDA I suspect that doctors and general people would have more doubts about the safety and efficacy of their medicines.
Re the EMA: its becuase the EU is a lot looser than the US. The EMA is more advice to state regulators.
From someone who used to work with the EMA they were saying brexit probably had a bad effect on its effectiveness as the UK beaurcrats played a non-trivial role in getting the various state regulators to keep calm.
EMA takes longer to approve drugs than the FDA, on average (see https://www.pmda.go.jp/files/000204615.pdf, slide 6). This could be a result of understaffing though rather than then being more strict. They have a different model than the FDA where they contract out a lot of the reviewing to private firms which some people say is better.
From talking to someone who works in drug trials, its also sometimes like the student who did the assignment and handed it in 2 days after it was assigned.... a month later watching one of their classmates who left it till the night before swearing about how it's "totally unfair" and an impossible deadline.
Sometimes they know that behind a story of delay is someone who kept screwing up and taking months to do a few days worth of tasks.
> I have no idea how you can write "they usually carry out their mandate well" in one place and then, in your main post, just go ahead and repeat your original belief—backed by an example you know is wrong—that the FDA does stupid and destructive things on practically a daily basis.
Which is a really bad thing for Kevin to be confused about, because if he’d read the next paragraph of Scott’s post, he’d read all about how Scott reached the conclusion because he anticipated this criticism and basically held the readers’ hand through the thought process.
> My problem is: doing anything in medicine is illegal until you clear a giant hurdle. To clear the hurdle, you have to pay millions (sometimes billions) of dollars, fill in thousands of pages of forms, conduct a bunch of studies that can be sabotaged for reasons like “this drug is too good so it would be unethical to have a control group”, and wait approximately ten years. You have to clear this hurdle to do anything, even the most obviously correct actions. Everything starts out illegal, and then a tiny set of possible actions is exempted from the general illegality. The easiest name for this hurdle is “the FDA”, since they’re the agency charged with enforcing it.
Meaning either Kevin drum conveniently stopped attentively reading at exactly the point where Scott’s post would provide the most fodder for his pro-FDA position, or he specifically cherrypicked that quote despite knowing that Scott fully addresses the concern immediately after the quoted section. (Note, my quotation from Scott is followed up by him reiterating the same point with specific examples for about 8 paragraphs, so there’s no way Drum could have had a microstroke or something and missed that specific paragraph.) In conclusion, Kevin is either a deliberately deceitful huckster or just a straight up moron and this piece has provided all the evidence I need to conclude that I don’t actually have to read anything else he has written in the past or will write in the future.
But I have to say, the most baffling part of Kevin's piece is when he "defends" the FDA by insisting that we start the clock of when we knew Omegaven was effective from when the company that makes it applied for authorization. Like “the FDA doesn’t actively go out and do it’s job proactively seeking new, useful medicines to help through the approval process so Americans can get prompt access to useful new drugs. Instead, they sit on their hands, even while getting a constant flow of emergency use requests, until someone pays the millions of dollars to make the adequate sacrifice to Moloch to get through their Kafkaesque approval process before they’ll even begin thinking about whether maybe they should form an exploratory committee to consider whether to process the paperwork which will begin the approval process” is some kind of ringing endorsement of the FDA.
> In conclusion, Kevin is either a deliberately deceitful huckster or just a straight up moron and this piece has provided all the evidence I need to conclude that I don’t actually have to read anything else he has written in the past or will write in the future.
Gotta say I oppose this mentality, except insofar as you don't *have* to read anything. If writing one dumb thing is enough to banish someone's posts to the shadow realm forever, you either have to apply that heuristic very selectively or end up with an incredibly cautious stable of writers.
This isn't a problem of dumbness. It's a problem of either fantastically implausibly bad reading comprehension or literally trying to actively deceive his readers. You can rule thinkers out when they're actively trying to make their readers dumber. But in the spirit of glasnost, if there's anything of his that you think would make me rule him back in, I'll read it and I promise to give it a fair hearing.
I... have exactly the same thing I said above to say to that. I didn't think he was dumb, just that I think one bad take shouldn't rule someone out forevermore
I've never read the guy, so I have no recommendations. I just think the approach you're putting forth here is mistaken in principle
Again, if this was a bad take in the sense that I think it was an honest mistake, I would 100% agree with you. (I'm a huge fan of Robin Hanson, despite my thinking that a lot of the points raised by his critics are valid, for exactly this reason.) But if the quote mining that Kevin did really was a result of a good-faith mistake of the kind which we should forgive... I'm sorry but that requires some deeply implausible assumptions. And even if we are to assume incompetence and not malice, we still need to assume *a lot* of incompetence to get what Kevin actually wrote. If we have that level of incompetence, then I wouldn't rule out Kevin on the grounds that he's actively trying to make me dumber, I'd just do it on the grounds that I have more reading material than I have reading time and I can do much better than someone that bad at basic reading comprehension.
Lots of people get hung up on one thing they read in some essay or post or article or book, and it so colors how they read the rest that they miss a lot of important detail. I imagine that's what happened to Drum in this article.
I think sticking someone with a "burn notice" is justifiable for deceit (as opposed to being honestly wrong). Fact-checking a fraud can get quite difficult, and cross-correlation of deceit is much higher than ignorance (if someone knows very little number theory, they may still know a lot about biology; if someone lies about number theory, they're probably also willing to lie about biology).
(The main exception I'd make is that high-stakes deceit doesn't generalise especially well to low-stakes deceit. Most people have a threshold of stakes where they feel comfortable lying; if someone lies to avoid going to jail, they're probably willing to lie to avoid dying - or to avoid going to jail for something else - but they won't *necessarily* lie for money or attention.)
I think you hit the nail on the head, especially with regard to Kevin Drum. This is not the first time I have read a blogger responding to Kevin about some claim that comes down to "Ok, KD is either an amazingly bad reader, deceitful, or so deep in motivated reasoning that he doesn't even know if what he says makes sense." This has been going on for years and years, at least going back to 2009.
I am honestly a little confused about why so many people call Kevin Drum their favorite blogger, unless that is a euphemism for something.
"I want a world where if there is a $250 version of a drug, and a $10 version of the same drug, people are allowed to choose the $10 version even though the manufacturer did not give the FDA $100 million dollars."
Everybody wants that world. So how come generics aren't more widely prescribed? "Sure, I could give you a script for the fancy new drug the pharma rep tried selling me on, or I could write a prescription for the fifty-year old version that is now out of patent so there are three cheaper versions on the market". Why doesn't this happen?
But the thing is, we are living in an imperfect world, and *if* you get the world where:
- everybody from NBC News to random bloggers know that the babies are much less likely to die on fish-oil based fluids
- FDA doesn’t hold up approval for an additional five months arguing about the wording of the warning label
- ketamine eventually gets approved for use as a depression treatment
- if everyone knows COVID vaccines are safe and effective, then the government will have approved COVID vaccines as safe and effective
That is *also* the world where Aduhelm gets approved, because the drug company doesn't want to wrangle with the FDA over the warning label and pressure from patient advocacy groups who are grasping at straws works and everyone from NBC to random bloggers are saying it's a breakthrough miracle drug and the government relies on the pharma lobbyists telling them it's safe and effective.
It's the world where Dr. Ogu's Natural Herbal Remedies for epilepsy to AIDS get approved for the same reasons.
It's the world we used to have, before in the USA the 1906 Pure Food and Drug Act was passed, where the role of the FDA was limited to "there isn't any actual poison like arsenic in this item (anymore)" enforcement.
If it contains terra alba, barites, talc, chrome yellow, or other mineral substance or poisonous color or flavor, or other ingredient deleterious or detrimental to health, or any vinous, malt or spirituous liquor or compound or narcotic drug."
I do like the bit about poisonous substances in confectionary, don't you? Ah, the wonderful world where you could merrily put poison in sweets before the meddling and interfering hand of government intervened to retard you!
The pendulum swings wildly from extreme to extreme, and yes there should be a moderate mid-point, but when you have evil fools selling Miracle Mineral cures for autism, and parents so desperate for a 'cure' that they will torture their children, and people going on Amazon to give positive reviews of the product, I find it hard to say we're over-regulated.
Indeed, we face a complicated tradeoff of bad things, and the argument is largely about what the right point on the tradeoff would be. There is no practical point in the space of possible tradeoffs where we don't sometimes get approved medicines that have terrible side-effects, or drugs that are only marginally better than a sugar pill. There is also no practical point where we don't sometimes delay the availability of lifesaving or genuinely helpful drugs. (We can avoid the first by banning all new drugs, and the second by abolishing the FDA, but short of that, both errors will be made sometimes.)
ISTM that one fairly easy improvement we could make would be to use other countries approvals/experience to allow drugs in the US. Something like "if it's approved in the UK/EU/Japan/etc., and has been on the market a couple years there, it's approved in the US as well unless the FDA makes an explicit case to keep it out." Or "you can buy/prescribe things that have gone through safety but not efficacy trials under these conditions...."
I expect there are more ways to improve the tradeoff we're making, but I don't have great confidence about that--this is a place where all choices have painful costs associated with them. OTOH, Scott is speaking from experience as a practicing psychiatrist and has run into several cases in his practice where FDA policies seem to be making things worse. Probably we should pay attention to that, even if we're not sure exactly where the optimal tradeoff point is.
"painful costs". Yes, but who gets to make the decision about what painful costs are to be borne? That's the problem: it's not deciding between risks, it's *who* decides. If you think regulation is a good thing, then you think that the government is better at deciding who should bear what risks. Do they have a better incentive or better information than the individual who is making the decision for their own risk?
Well said — people think so strangely about this topic. "Once upon a time, candy very rarely had bad stuff in it, therefore there is no such thing as over-regulation" really seemed like a good argument to at least two people in this thread, apparently?!
Deisach contributed several specific examples with evidence, historical context with a link to learn more, and several well formed arguments. That’s great!
If we’re discussing the FDA, the historical context of why the FDA was established and what law established it is interesting. As is the example of the sort of bad thing that might be encouraged by the FDA’s abolition.
It wasn’t exactly a powerful and fully sourced peer reviewed journal article on the topic. But it’s perfectly good! And it started a few other conversations
> and what arguments?
short and unsatisfying arguments can still be arguments
I don't agree. I don't see any well-formed arguments; all I see is "yeah but with NO regulation, you could sell poison", and this is so obvious and well-covered that it doesn't really need examples and evidence — or mentioning. The question is, as always, how *much* regulation we need *now*; the situation with confectionery in 1906 is essentially irrelevant.
Well indeed, WC, that is a question we must all ask ourselves at some point in our lives. "Who am I, where do I come from, where am I going?" as the British Army patrol asked Kevin McAleer.
1. I am Thirteen Ways Of Looking At A WC
2. I am ragweed pollen on the breeze, and the solar UV index is moderately high.
3. I am he as you are he as you are me/And we are all together
An orbiting pendulum looks like it's simply swinging from any given point, and yet in actuality (and in spite of appearances), at no point in its orbit does the bob actually approach the happy point it is orbiting.
I'd be more sympathetic to Scott's position if we lived in a world where miracle drugs like imatinib were being developed every year*. But drug development is very, very, very hard. Your pre-test probability should be that most novel pharmaceutical interventions will not significantly outperform current standard-of-care + placebo. Given that landscape, the mandate and ethos of the FDA makes quite a bit of sense. The public (and the American healthcare system) is much more at risk from being victimized by ineffective and/or harmful treatments than having genuinely transformative medications delayed or withheld.
*(I'll grant Scott's point that the particular mandate/ethos was a very bad fit for development of COVID vaccines, which are the rare intervention with an enormous effect size)
The asymmetry I'm assuming is that for a given indication, a vast majority of pharmaceutical compounds won't work better than placebo. So if you have a marketplace flooded with Placebo Plus products with no meaningful regulatory restraint or oversight, you dilute the number of people getting effective treatment, you increase financial toxicity to the healthcare system, and you are exposing people to a wide range of tail-risk side effects.
Robin Hanson's work indicates that a whole lot of medicine is no better than a placebo. But the effect of this is just to waste money. If lots of people buy placebos that do nothing, and people with rare conditions (like those infants or the guy that lost his intestine) that can actually be improved by drugs get those drugs, it sounds like a net plus.
Leaving aside the ethics of constructing a society that caters to a small cohort of the information elite, only a tiny fraction of people have the time, expertise, and disposable income to evaluate the evidence base of candidate pharmaceuticals and conduct autologous n of 1 trials - and that fraction would turn to '0' if you deregulated the space and allowed shoddy, biased, and outright fraudulent research results to proliferate.
If you want evidence of that, you need to look no further than 2020. Many of the people who would describe themselves as 'diligent about [biomedical literacy and research]' went head-over-heels for a whole variety of placebo-equivalent antiviral candidates. There was so much clinical and biomedical data - much of it shoddy, unreliable, and non-replicable - that people's empirical apparatuses were swamped.
Imagine that, but for every indication and disease. How would a person with a life-threatening condition decide which drugs to try, in what combination, and in what order? How would insurers decide what was reimbursable? How would doctors be able to decide what was safe and effective to prescribe patients? How would the system bear the costs of this flood of products? Aducanumab alone had the potential to bankrupt Medicare, and that's just one shoddy monoclonal antibody.
Actually I think doctors all over the world were doign the same thing. Nobody knew what would help with covid, so they threw anything that might help at it--zinc, HQC, azithromycin, ivermectin, vitamin D. I think this always happens with a new disease. Eventually, people found some stuff that seems to work and switched over.
I go back and forth on this. It's definitely hard with the FDA involved. But it's weird how often drugs approved for one indication randomly help with another (eg fluvoxamine, an antidepressant that seems to also treat COVID - or ketamine, an anaesthetic that seems to also cure depression - or any of the substances mentioned at https://slatestarcodex.com/2014/07/17/psychotropic-base-rates-the-argument-from-antibiotics/). I'm not really sure what to make of this.
I'd love to see good data on the success rates for drug repurposing (and by success, I mean statistically significant differences in clinically meaningful outcomes in large double blind RCTs). I suspect the number is quite low, and that number is likely an underestimate given how much of the denominator would be hidden / unpublished.
I agree that many drugs have off-target activity that we don't always anticipate or understand. But I'm much more skeptical that the off-target activity is clinically meaningful, or if it is clinically meaningful, that it would outperform current standard-of-care.
Derek Lowe has had faint praise for drug repurposing, that may be worth reading. Here's one, though he mentions it in passing in a few other posts that I don't have the attention to find at the moment.
This is interesting, thanks for the link! It mostly aligns with my priors but he does bring up a few good examples (like thalidomide) where very high-signal off target effects were effectively repurposed within the same drug class.
I mean Covid drug repurposing didn’t seem to go great? A lot of drugs were identified and given to a lot of people. Azithromycin hydroxychloroquine remdesivir and a bunch a others, and i guess ivermectin although I can’t say if that one works or not.
combined OR on all-cause mortality for hydroxychloroquine is 1.11 (95% CI: 1.02, 1.20; I² = 0%; 26 trials; 10,012 patients) and for chloroquine 1.77 (95%CI: 0.15, 21.13, I² = 0%; 4 trials; 307 patients).
Going with 1.11
The pharmacies steadily dispensed an average of 383,685 prescriptions for hydroxychloroquine (HCQ) per month, and 1839 prescriptions for chloroquine (CQ) per month. However, from February to March, prescriptions for hydroxychloroquine jumped by 86.2%—from 367,346 to 683,999. Prescriptions for chloroquine jumped from 2346 to 6066—an increase of 158.6%. Among those patients given hydroxychloroquine, the number who were simultaneously prescribed azithromycin jumped from 8,885 in February, to 101,681 in March—a 1044% increase.
(I am now going to pull numbers out of my colon, and I don’t expect this to have really any relation to actual hcq deaths. I probably misinterpreted at least two of my sources somehow)
Assuming of 300k given hcq, the case/hospitalization and hospitalization/death ratio for the general population are the same as for that group (which there are no grounds at all to make), 300k* 600k deaths / 35M ~~= 5.1k deaths. 11% increase = 500 deaths. If it’s 75k like another source suggests, 125 ish. Which is quite a few, but not exactly at the scale of the vaccines being delayed. (But I would submit again that the fda process being very strict may have pushed the drug companies to have actually good development processes and those with to the top)
I believe the drugs that were useful were the steroids, which is much less an experiment than the others.
I think both. Huge amounts of time and treasure and genuine genius have gone into treatments that looked great until the clinical trial, where it turned out they didn't do any good. But also stuff that works can be impossible to get approval for because it can't be patented and producing the evidence that it works sufficient to get approval costs hundreds of millions of dollars.
> Given that landscape, the mandate and ethos of the FDA makes quite a bit of sense.
There seems to have been ample evidence to the contrary for the fish oil. It wasn't even a "new" drug. Something that has a proven safety record in other jurisdictions shouldn't have to overcome the same hurdles as a truly new drug.
The pendulum has never swung back to the old days you bring up. Even though we could rely on third-part certification like Underwriters Lab, fraud laws for false claims, and Hanson's would-have-been-banned stores.
I don't find it hard to say we're over-regulated. Here, I'll say it. "We're over-regulated". The counter to your "merrily put poison in sweets" claim is the requirement that manufacturers of industrial alcohol merrily put poison into their products. I've seen an estimate that, since prohibition, 10,000 people have been killed by this government requirement.
Essentially, your claim is that individual action in the marketplace is insufficient to solve problems of unhealthy foods. Perhaps that's the case, but it doesn't explain why the government has been encouraging people to eat unhealthy foods (too many carbs, because "eating fat gives you heart disease", when that's exactly backwards) for most of my lifetime.
So, do you suppose there could be some possible middle-ground between allowing people to put poison in candy and preventing people from using a treatment that's known to work and approved in many other countries?
This is like the dumb internet argument style where anyone who proposes less government is accused of wanting to move to Somalia, while anyone who proposes more government is accused of wanting to move to North Korea.
Is there a middle ground between having a forest fire burn down your entire forest and not having it burn down any part of the forest?
Part of the problem is that you're thinking this decision can be made by itself: that we have a knob which can control the amount of regulation of substances, and that we just tell our politicians where we want the knob to be set, and them and all the other politicians put in their setting, the average gets taken, and that's the setting.
Firstly, most people are dumbshits, as I'm sure you've noticed. Second, we never send our best and brightest to govern us. In fact, the best and brightest are too smart to want to be politicians. Third, once you give a politician power, they never want to give it up, so the knob only gets turned in the direction of more control, and only rarely less control. Fourth, political power is fungible. You may think you're giving a politician the power to keep us safe from unsafe substances, but what if they trade it to another politician (who has promised to eliminate unsafe substances from the market) in exchange for the power to benefit an evil corporation?
Your case would be more powerful if you offered examples of times, probably more than we know, the FDA put the kibosh on as yet to be known harmful therapies. I keep a quote handy I fall back on occasion: Thomas Chalmers, the famous medical researcher, once wrote, “One only has to review the graveyard of discarded therapies to discover how many patients have benefitted by being randomized to a control group.” Not quite apples to apples to your response above, but it's close to the spirit.
Also, I don't follow your two-track description of the FDA. The first, "stupid," and the second, THe FDA is a product of the charter organizing committees set forth for them: slow, deliberate, and overly cautious. The latter doesn't validate the former. If you didnt imply that, I misread your intent—apologies for the misinterpretation.
The most radical FDA reform proposal I've heard is that the FDA should use its current system to check if drugs are safe, but be laissez-faire about whether or not they're effective. The safety checking is (as far as I know) relatively quick and easy and not usually most of a delay (though I would have to double-check this before asserting it officially). I think this is kind of like what they do with vitamins/supplements right now, which are cheap, numerous, easy-to-get, often ineffective, but safe enough that if you dared someone to take a randomly chosen Whole Foods product I think only the most over-cautious people would have any qualms.
Overall, I like that idea, but implementation is going to have to resolve the issue of "safe" not being binary.
To pick and extreme example, chemotherapy drugs are literally poison, but they're net benefits to patients with certain forms of cancer because undertreated cancer is worse for the patient than the chemotherapy drugs. You're probably going to want some kind of backup process where some standard of evidence of benefit can be used to get approval despite known or likely hazards of the drug. The tricky part is to stop the acceptable level of risk for basic certification and the standard of proof of effectiveness for approval despite risk from drifting to the point you're effectively back to the current system.
Indeed. And sometimes how we think about risk is just weirdly broken or mis-calibrated (the reaction to the super rare AZ or J&J vaccine-associated blood clots, for example). Certain people get all weird about me taking medicine that ups my risk of blood clots and breast cancer, when realistically the risks are comparable to the average female baseline risk (which is higher than the normal rate for a male), and they also significantly reduce my risks of falling victim to other things, like prostate cancer, alongside the positive effects which I want. If I were to get certain surgeries, some doctors would force me off my medications for weeks leading up to it because of that fear of blood clots, despite the risk not actually being high (cis women on birth control are at significantly higher risk and generally doctors won't force you to stop that before surgery) and going off the medications being quite unpleasant in and of itself.
You can't really consider them in isolation though. We ought to accept a drug that prevents 10,000 fatal heart attacks a year even if it causes a dozen liver failure each year too. I mean, the FDA didn't accept the first beta blocker historically and Europe did but they shouldn't have accepted such a risky drug without proof of efficacy. Of course, an easy solution is to be open to safe drugs bit stringent with dangerous ones.
Tier 1: Drugs that have proven reasonably safe can be manufactured and sold, but can't be marketed with claims of curing disease, and won't be covered by insurance.
Tier 2 is drugs that are also proven effective. That's the only tier we have now.
Now watch as the cost of a drug is multiplied by a factor of ten the day it's moved from Tier 1 to Tier 2, because insurance is paying so they can charge $100 / bottle for what was formerly $10/bottle over the counter.
I feel like at this level, though, it wouldn’t resolve a bunch of the fundamental objections. Ketamine, to take one of your frequent examples, doesn’t seem like it meets this bar: you say it causes “acute distorted cognition for the next few hours” at clinical doses, and it has lots of potential harms at recreational doses. That doesn’t seem like a level of safe that we’d be happy to have it sold over the counter at Whole Foods, which presumably would move it into the “must show effectiveness” track and we’re back where we started.
That’s not to say that the FDA is the right answer, but lots of medications have some bad effects that trade off against their good effects, so just evaluating safety without an eye to effectiveness seems like it doesn’t actually result in lots of medications meeting the bar.
The FDA would be more useful and believable if they restricted themselves to an advisory role. Then you would have insurance companies saying "Sorry, we don't pay for this because the FDA advises against it." but people could still get the compound they believe helps them.
When I take a drug to treat my diabetes, or blood pressure, or some other chronic condition, or even cancer -- then yeah, I want to be pretty damn sure it's effective, not just safe. Leaving my conditions untreated is *not* safe.
Of course! So why do you trust the government to do what you should be doing? What would happen if they failed to ensure that your medicine is effective, like they failed us by telling us that fat is bad for you? What mechanism is in place to correct their failure? If a private company created medicine that was not effective, they would probably go out of business. How might the FDA "go out of business" given that they are reliant on politician goodwill, not customer goodwill?
> So why do you trust the government to do what you should be doing?
Whom should I trust instead ? Every snake-oil salesman who comes my way ?
Look at it this way. The last time you rode in an elevator, did you stop to consider the probability of it crashing to the ground, burying you deep in the elevator shaft ? Did you check the local reviews for the construction company who built the elevator before getting into it ?
If you trust "the government" -- or rather, the teams of engineers and scientists whom the government either employs, or forces to pass standardized test -- to maintain your elevator, then why not medicine ?
Who should you trust? Walgreen's. Kinney's. Rite Aid. CVS. They should be the ones testing for safety because unlike any government agency, they compete with each other and will HAPPILY test their competitor's product and gleefully point out anything that isn't safe.
I trust the elevator company because they will lose everything if any one of their elevator crashes to the ground because of incompetence on their part.
Why do you trust the government? Who is monitoring them? Not CVS. Not Walgreen's. Not Kinney's. Because none of them have any interest in catching the government's errors, because all that will do is cause everybody to stop buying from their local pharmacy.
If you ask me to predict the future, I will ask you what incentives people have. CVS, Walgreen's, Rite-Aid have good incentives, the FDA has bad incentives.
Basically that they should approve for safety and effectiveness, but not clinical utility. And have explicit tiers available for measured effectiveness (biomarkers / symptoms / survival etc).
Sure, but that opens the door to homeopathic-style treatments. Is it safe? Of course, there's not enough of anything in this to do anything good, bad or indifferent. Is it effective? Well, if you don't have anything really wrong with you and it triggers the placebo effect, sure. If you do have something really wrong with you, um, yeah.
But hey, at least it's a lot cheaper than Aduhelm at $56,000 per year! Which is something approved as safe but not effective:
"Why does this drug cost so much?
There are millions of people and their families who want a new Alzheimers treatment. Any Alzheimers treatment.
"We have established a price for Aduhelm that reflects the overall value this treatment brings to patients, caregivers and society -- and one that will enable continuous innovation," wrote Biogen's CEO Michel Vounatsos in a letter announcing the approval.
The nonprofit Institute for Clinical and Economic Review, which scrutinizes drug prices, argues that because the drug has not been proven to be effective, a more reasonable and cost-effective price would be much lower -- maybe as low as $3,000. That's far from the $56,000 Biogen will charge.
The impossible problem.
More than 120,000 Americans die from Alzheimer's each year, making it among the leading causes of death in the US.
But treatment of the disease has been stuck for years, and the FDA had not approved a new drug to treat this disease since 2003."
The bargain is not "how many thalidomides is one omegaven worth", it's "how many aduhelms is one omegaven worth".
Omniscience is not available to anyone, even someone who works for the government. It is only through trial and experimentation that we know anything. The FDA's remit in verifying effectiveness is very problematic. It should only be testing for safety. It is up to other parties to determine effectiveness.
In the US, you can buy homeopathic remedies and herbal medicines and such over the counter. And you can go to a chiropracter or accupucturist or naturopath for your cancer. So maybe this would be a terrible world, but it woudl be the terrible world we already live in, at least in the US.
Still don't have an answer to the question of "how many patients does this (lack of approval for the fish oil formula) impact every year." I am betting that it is under 1k and probably under a hundred, not the tens of thousands originally given.
FDA-employed researcher who is a partner of a family member, enormously paraphrased: "the article correctly described the FDA's difficulty with too fast/too slow on drugs, but noted that the 'F' part of FDA doesn't have a similar issue, mainly because it doesn't approve foods before they can be sold as food; if something appears to be making people ill, it’s super fast and super easy to get authorization and budget to study/ban it, but for foods which no one has thought to test yet and which are only on the radar on priors like the one I'm currently working on, it took a year to set up protocols and it will be at minimum a year before any results of anything are to be looked at".
I disagree with your assessment that you're "basically correct" when you're off by a factor of 2-4 in the detail that serves as the crux of the point. You're right that this is an example of the FDA being overly restrictive on a drug that ought to be widely available, but there is an absolutely massive difference between an organization that delays new drug releases by 5 years compared to 20 years.
I agree I came on way too strong originally, but I think there's some leeway in when you date from. Omegaven was invented in 1991, there was a WSJ article about it being a miracle drug in 2006 ( https://www.wsj.com/articles/SB116338867216321332 ), there were studies confirming it in 2011, it was well-known in the libertarian blogosphere by 2014, and it got approved in 2018. Which of those is the starting point for saying the FDA was "delaying"? In a default-yes medical regimen, I think people would have started using it around the WSJ article in 2006, so maybe 12 years earlier.
One of the questions I was thinking about when reading your initial article was: Assuming we can only work with the single lever of FDA restrictiveness, how many Omegavens is one Thalidomide worth? The number of years the FDA delayed Omegaven strongly influences my answer to that question (as does the rate at which the FDA encounters "Omegavens" and "Thalidomides" in the wild). That's what I mean when i say that the particular detail you were most incorrect on is an essential one, and being off by a factor of 2 is a large one.
Very good point, I apologize for implying otherwise.
It seems to me, however, that Scott is arguing for a standard wherein the FDA would have gotten Thalidomide wrong. In this case, it would have been by adopting a European approved drug, but it probably would have also approved it in its own channels by only looking for first-order safety effects as opposed to inter-generational ones (perhaps he believes we ought to have stricter safety standards for researching drugs given to pregnant women?). I do suspect that on net this tradeoff would be worth it; that the cost of Thalidomides would be more than made up for by easier drug access and accelerated research.
The history as I have heard it is that prior to the thalidomide scandal, the FDA checked only for safety; after it, it started checking for effectiveness as well, ("Kefauver-Harris Drug Amendments"), even though the safety check caught thalidomide.
The first COVID vaccines passed the safety check in, IIRC, May 2020. The rest of the delay was effectiveness.
How many people died between May 2020 and whatever month in 2021 you choose to mark for the "general availability" of the vaccine, and why shouldn't we lay those deaths at the feet of the FDA? Somebody's got to go to jail.
I think overall it would still be worth it if some thalidomides got through, but see my response to Brad above - I don't think there's any need to compromise safety here.
Any clock should probably start with when it was approved for use in EU. Off label applications are allowed in US even when they haven't been "proven" to be effective. Simple reciprocity could have avoided the entire fiasco.
But Scott could have written the exact same paragraph, with exactly the same amount of outrage, only replacing '20' with '10.' This means he's basically right and just needs to make a minor correction.
Consider two different ways someone could be off by a factor of 2-4.
Example 1: I claim we should do my economic plan because it has a positive NPV of $1B compared to the next best alternative plan with a positive NPV of $0.9B. If it turns out my economic plan actually only had $250-$500M of value, my main point is refuted and I should retract.
Example 2: Imagine I'm going around screaming, we need to stop this serial killer because they've already killed 20 people! And it turns out they had, in fact, only killed 5 to 10. If I'd falsely attributed 10-15 deaths to the serial killer, I should correct that and discontinue saying the false number once I learn the truth. But the main point of stopping the serial killer is very much still in place!
Scott usually operates within a utilitarian-ish framework wherein certain things trade off against others, and therefore are generally examples of type 1. I generally operate within this framework as well, and when doing so being off by a factor of 2 *may* lead to significantly different conclusions!
Yes, Omegaven getting delayed is evidence against the FDA no matter what the duration of the delay is. However, the degree of evidence is proportional to the length of that delay (among some other factors like the magnitude of impact that delaying has). So when you're off by a factor of 2, you've doubled the evidence of your example! That should be a big deal.
That evidence against then has to trade off against evidence for: the FDA didn't approve thalidomide in the US largely because they deemed European approval insufficient. Its rules provide some protection, the question is whether the cost of that protection is too high. If you fudge the evidence by a factor of 2, that can significantly shift the answer.
>"I want a world where if Europeans have been giving fish oil based nutrient fluids to babies for years, and everybody from NBC News to random bloggers know that the babies are much less likely to die on fish-oil based fluids, and all the studies that have been done confirm it, you are allowed to give those babies the fish oil based fluids without waiting another five years for an FDA approval."
Isnt the problem in this case that it was not as clear cut as you are trying to make out, based on https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7671012/ , it seems like there was actual opposition from within the medical community, not the FDA, and that along with a lack of support from the drug company was the reasons things took so long. Ie Gura calls out
`The fact that we demonstrated that IV soybean oil was detrimental, but IV fish oil was not, created controversy in the medical community. Papers and commentaries were published that suggested that Omegaven caused growth failure, bleeding, and progression of hepatic fibrosis. These criticisms led to us to do more translational and clinical studies.`
This being said it looks we got a pretty positive paper in Nature in 2011, https://www.nature.com/articles/jp2010182 , after which point it seems like education more that the FDA was the problem. Do we have any recorded cases of doctors who wanted to use Omegaven having any difficulty in acquiring it after this point?
Yes, "lack of support from the drug company" was relevant *in the context of the current system*, where without drug-company-level resources you can't push anything through the difficult FDA approval process. If you make it cost millions of dollars to do something, there will always be someone who resists paying.
There was (bad) opposition from within the medical community, but I don't think it would have passed a simple cost-benefit analysis, and a quick check at Europe, where this had been used for a long time, would have defused it.
In terms of "did doctors have trouble acquiring it" - doctors at hospitals that were willing to do an IND (very complicated FDA related process) did not, but very few hospitals were willing to do that. If you can call making people go through a very complicated process to vote "voter suppression", I think it's fair to call this "drug suppression". Of course you can always just do the complicated thing, but many people won't.
People buy snake oil cures and use homeopathic remedies for cancer now, in the world with the FDA behaving as Scott describes.
Typo thread: "the Silenor brand [paid] a $100 million FDA application fee for approval as a sleeping pill"
Addressing the claim rather than the typo, is this actually true?
Does it really cost $100 million to submit an application for a sleeping pill?
This seems ... implausible.
Most likely the approval process which included running several studies cost them $100 million. It is not that they gave $100 million to the FDA. Running studies are very expensive, you have to pay doctors, administrators, supporting staff etc.
I'm a doctor and have been peripherally involved in several large clinical trials. I could believe that a multi year study involving many rounds of expensive investigations (eg screening MRIs in, say, the recent aducanumab study) could cost tens of millions of dollars, as this paper found (https://pubmed.ncbi.nlm.nih.gov/30264133/) - "Trials designed with placebo or active drug comparators had an estimated mean cost of $35.1 million (95% CI, $25.4 million-$44.8 million"
However Scott seems to be:
a) referring specifically to an 'FDA application fee', rather than the cost of running a trial and
b) talking about a relatively straightforward trial with short term clinical endpoints
so I'm going to assume he either misplaced a decimal point or is just mistaken. Happy to be corrected!
I believe he's referring to the cost of the trials combined with the fees.
Though I think I disagree with scott for a seperate reason.
The FDA tends to do a "he who pays for the trial gets the use patent" thing.
Before they ran a pair of full clinical trials to prove it worked as a sleeping medication a bunch of doctors might have had a hunch it had that effect due to patient reports but the evidence base was poor.
Just like patents encourage people to reveal their designs, drug patents and clinical use patents incentivise companies to pay to run the big clinical trials that would otherwise never get run which generate all the nice neat tables of effects, side effects and contraindications.
When you know that a drug is good for [use] becuase you've got the summary in front of you from a bunch of clinical trials it's hard to imagine a world where you don't know that information or only have a hunch about that.
It is kinda stupid for the FDA to not just routinely inspect substances approved in Europe and Japan for one's with a reasonably acceptable evidence base rather than waiting for companies to jump through symbolic hoops.
If something is expensive to create, we have a separate IP system for establishing temporary monopolies, which at least in theory can be tuned to produce economically efficient outcomes (e.g. protection that eventually expires by default). FDA approval has no such characteristic. There are problems with the IP system we have as well, but they are probably made worse by trying to also use the FDA to incentivize companies to do studies.
Another typo: "$100 million dollars" doesn't need both the "$" and the "dollars"
The NBC article is available on Archive.org: https://web.archive.org/web/20130612022529/http://rockcenter.nbcnews.com/_news/2013/06/07/18833434-drug-treatment-omegaven-that-could-save-infants-lives-not-yet-approved-by-fda
Thanks, I've added the correct link into my essay.
When I read what Gura said about the FDA, my immediate thought was, 'stockholm syndrome'.
But, another possibility is, how dare any one who is engaged in pharmacology research criticise the FDA, or even engage in faint praise?
I'm sure it does happen, but yeah, I had a similar thought.
She sounds like a deeply pragmatic woman whose intensely committed to getting shit done.
She’s interested in saving babies and found allies to accomplish that goal. Shitting on the FDA would make things worse.
She could’ve written a libertarian blog or become a policy wonk but she focused on visible outcomes.
I think there’s a lesson to be learned here about the mentality of implementors versus wonks. Maybe she’s too short term and institutionally minded for this crowd but it’s a valuable perspective.
I think it depends on the level at which you want to get things done.
If you're a medical researcher, then you just have to accept the rules as the reality at which you operate in, regardless of how arbitrary and counterproductive they are. If the FDA mandated that double blind trials literally require everyone involved to be blindfolded, the only question to ask would be "what brand of sleep mask has FDA approval?"
If you're trying to operate at a political level and reform the system, then questioning the rules is exactly the kind of thing you should be doing.
It's a "game player" vs "game designer" distinction - they're two very different ways to approach the world, and while they can overlap you can't really be both at the same time.
Also, it was a speech she gave after winning an award for the work. Using a platform like that to drag anyone who was involved in the thing you're getting the award for is just ... not a thing that's done outside the confines of the internet.
It's not a normal thing to do, but it's been done. My favorite example is John Taylor Gatto's "The Psychopathic Schoolteacher," which was his acceptance speech for New York's Teacher of the Year award. https://www.naturalchild.org/articles/guest/john_gatto.html
From what I understand, companies live in fear of retribution from the FDA. Mary Ruwart writes about this in her book "Death by Regulation" providing a first hand account from her time inside the pharma industry. I suspect one of the reasons the AstraZeneca vaccine wasn't approved in January was retribution for a mistake AstraZeneca made in failing to immediately notify the FDA about a pause in the Phase III trial.
For academic researchers it's a different story, but they also have to work with the FDA too (for instance with trial design) if they want whatever they are working on to have a chance of getting approved. The FDA also provides funding to researchers, as we saw in Scott's last post, and grants compassionate use applications, etc etc.
Finally, researchers can be part of FDA advisory committees. I'm not sure if that is paid, but it is prestigious.
So I'm pretty sure fear of retribution is a factor when researchers are talking publicly about the FDA.
The EMA approved AstraZeneca vaccine but when the reports about blood clots came in, half of European national regulators suspended its use. EMA quickly reviewed the case and maintained that benefits overweight risks and suspension was not justified. The EU even sued AstraZeneca for missing deliveries at the same time deciding not to use AZ vaccine. They donated the remaining AZ vaccines to third-world countries but they come with short expiry dates and a lot of bad reputation that are increasing anti-vax sentiments.
For what it is worth, it seems that a stronger EMA would have been more beneficial for the EU. Similarly, without strong FDA I suspect that doctors and general people would have more doubts about the safety and efficacy of their medicines.
Re the EMA: its becuase the EU is a lot looser than the US. The EMA is more advice to state regulators.
From someone who used to work with the EMA they were saying brexit probably had a bad effect on its effectiveness as the UK beaurcrats played a non-trivial role in getting the various state regulators to keep calm.
That's definitely true. EMA was based in London before. Many EU countries have lax standards and they also want to assert their independence.
EMA takes longer to approve drugs than the FDA, on average (see https://www.pmda.go.jp/files/000204615.pdf, slide 6). This could be a result of understaffing though rather than then being more strict. They have a different model than the FDA where they contract out a lot of the reviewing to private firms which some people say is better.
All the vaccines have a similar blood clotting risk.
https://twitter.com/moreisdifferent/status/1421475970062557184?s=20
I have written about the AZ ban in the EU over clotting. It was a mistake and lives were lost as a result:
https://moreisdifferent.substack.com/p/europes-reckless-precaution
This feels more like "fish don't notice the water" to me, tbh.
From talking to someone who works in drug trials, its also sometimes like the student who did the assignment and handed it in 2 days after it was assigned.... a month later watching one of their classmates who left it till the night before swearing about how it's "totally unfair" and an impossible deadline.
Sometimes they know that behind a story of delay is someone who kept screwing up and taking months to do a few days worth of tasks.
Kevin writes:
> I have no idea how you can write "they usually carry out their mandate well" in one place and then, in your main post, just go ahead and repeat your original belief—backed by an example you know is wrong—that the FDA does stupid and destructive things on practically a daily basis.
Which is a really bad thing for Kevin to be confused about, because if he’d read the next paragraph of Scott’s post, he’d read all about how Scott reached the conclusion because he anticipated this criticism and basically held the readers’ hand through the thought process.
> My problem is: doing anything in medicine is illegal until you clear a giant hurdle. To clear the hurdle, you have to pay millions (sometimes billions) of dollars, fill in thousands of pages of forms, conduct a bunch of studies that can be sabotaged for reasons like “this drug is too good so it would be unethical to have a control group”, and wait approximately ten years. You have to clear this hurdle to do anything, even the most obviously correct actions. Everything starts out illegal, and then a tiny set of possible actions is exempted from the general illegality. The easiest name for this hurdle is “the FDA”, since they’re the agency charged with enforcing it.
Meaning either Kevin drum conveniently stopped attentively reading at exactly the point where Scott’s post would provide the most fodder for his pro-FDA position, or he specifically cherrypicked that quote despite knowing that Scott fully addresses the concern immediately after the quoted section. (Note, my quotation from Scott is followed up by him reiterating the same point with specific examples for about 8 paragraphs, so there’s no way Drum could have had a microstroke or something and missed that specific paragraph.) In conclusion, Kevin is either a deliberately deceitful huckster or just a straight up moron and this piece has provided all the evidence I need to conclude that I don’t actually have to read anything else he has written in the past or will write in the future.
But I have to say, the most baffling part of Kevin's piece is when he "defends" the FDA by insisting that we start the clock of when we knew Omegaven was effective from when the company that makes it applied for authorization. Like “the FDA doesn’t actively go out and do it’s job proactively seeking new, useful medicines to help through the approval process so Americans can get prompt access to useful new drugs. Instead, they sit on their hands, even while getting a constant flow of emergency use requests, until someone pays the millions of dollars to make the adequate sacrifice to Moloch to get through their Kafkaesque approval process before they’ll even begin thinking about whether maybe they should form an exploratory committee to consider whether to process the paperwork which will begin the approval process” is some kind of ringing endorsement of the FDA.
> In conclusion, Kevin is either a deliberately deceitful huckster or just a straight up moron and this piece has provided all the evidence I need to conclude that I don’t actually have to read anything else he has written in the past or will write in the future.
Gotta say I oppose this mentality, except insofar as you don't *have* to read anything. If writing one dumb thing is enough to banish someone's posts to the shadow realm forever, you either have to apply that heuristic very selectively or end up with an incredibly cautious stable of writers.
https://www.lesswrong.com/posts/u8GMcpEN9Z6aQiCvp/rule-thinkers-in-not-out
As they say! You might be selecting against some of the very traits that make someone worth reading here, by pruning anyone who sticks their neck out
This isn't a problem of dumbness. It's a problem of either fantastically implausibly bad reading comprehension or literally trying to actively deceive his readers. You can rule thinkers out when they're actively trying to make their readers dumber. But in the spirit of glasnost, if there's anything of his that you think would make me rule him back in, I'll read it and I promise to give it a fair hearing.
I... have exactly the same thing I said above to say to that. I didn't think he was dumb, just that I think one bad take shouldn't rule someone out forevermore
I've never read the guy, so I have no recommendations. I just think the approach you're putting forth here is mistaken in principle
Again, if this was a bad take in the sense that I think it was an honest mistake, I would 100% agree with you. (I'm a huge fan of Robin Hanson, despite my thinking that a lot of the points raised by his critics are valid, for exactly this reason.) But if the quote mining that Kevin did really was a result of a good-faith mistake of the kind which we should forgive... I'm sorry but that requires some deeply implausible assumptions. And even if we are to assume incompetence and not malice, we still need to assume *a lot* of incompetence to get what Kevin actually wrote. If we have that level of incompetence, then I wouldn't rule out Kevin on the grounds that he's actively trying to make me dumber, I'd just do it on the grounds that I have more reading material than I have reading time and I can do much better than someone that bad at basic reading comprehension.
Lots of people get hung up on one thing they read in some essay or post or article or book, and it so colors how they read the rest that they miss a lot of important detail. I imagine that's what happened to Drum in this article.
I think sticking someone with a "burn notice" is justifiable for deceit (as opposed to being honestly wrong). Fact-checking a fraud can get quite difficult, and cross-correlation of deceit is much higher than ignorance (if someone knows very little number theory, they may still know a lot about biology; if someone lies about number theory, they're probably also willing to lie about biology).
(The main exception I'd make is that high-stakes deceit doesn't generalise especially well to low-stakes deceit. Most people have a threshold of stakes where they feel comfortable lying; if someone lies to avoid going to jail, they're probably willing to lie to avoid dying - or to avoid going to jail for something else - but they won't *necessarily* lie for money or attention.)
Especially since, as I was amused to note, there are people in the comments to Kevin's blog who say exactly the same thing about Scott.
I think you hit the nail on the head, especially with regard to Kevin Drum. This is not the first time I have read a blogger responding to Kevin about some claim that comes down to "Ok, KD is either an amazingly bad reader, deceitful, or so deep in motivated reasoning that he doesn't even know if what he says makes sense." This has been going on for years and years, at least going back to 2009.
I am honestly a little confused about why so many people call Kevin Drum their favorite blogger, unless that is a euphemism for something.
"I want a world where if there is a $250 version of a drug, and a $10 version of the same drug, people are allowed to choose the $10 version even though the manufacturer did not give the FDA $100 million dollars."
Everybody wants that world. So how come generics aren't more widely prescribed? "Sure, I could give you a script for the fancy new drug the pharma rep tried selling me on, or I could write a prescription for the fifty-year old version that is now out of patent so there are three cheaper versions on the market". Why doesn't this happen?
But the thing is, we are living in an imperfect world, and *if* you get the world where:
- everybody from NBC News to random bloggers know that the babies are much less likely to die on fish-oil based fluids
- FDA doesn’t hold up approval for an additional five months arguing about the wording of the warning label
- ketamine eventually gets approved for use as a depression treatment
- if everyone knows COVID vaccines are safe and effective, then the government will have approved COVID vaccines as safe and effective
That is *also* the world where Aduhelm gets approved, because the drug company doesn't want to wrangle with the FDA over the warning label and pressure from patient advocacy groups who are grasping at straws works and everyone from NBC to random bloggers are saying it's a breakthrough miracle drug and the government relies on the pharma lobbyists telling them it's safe and effective.
It's the world where Dr. Ogu's Natural Herbal Remedies for epilepsy to AIDS get approved for the same reasons.
It's the world we used to have, before in the USA the 1906 Pure Food and Drug Act was passed, where the role of the FDA was limited to "there isn't any actual poison like arsenic in this item (anymore)" enforcement.
https://en.wikisource.org/wiki/Pure_Food_and_Drug_Act_of_1906
"In the case of confectionery:
If it contains terra alba, barites, talc, chrome yellow, or other mineral substance or poisonous color or flavor, or other ingredient deleterious or detrimental to health, or any vinous, malt or spirituous liquor or compound or narcotic drug."
I do like the bit about poisonous substances in confectionary, don't you? Ah, the wonderful world where you could merrily put poison in sweets before the meddling and interfering hand of government intervened to retard you!
The pendulum swings wildly from extreme to extreme, and yes there should be a moderate mid-point, but when you have evil fools selling Miracle Mineral cures for autism, and parents so desperate for a 'cure' that they will torture their children, and people going on Amazon to give positive reviews of the product, I find it hard to say we're over-regulated.
https://www.amazon.com/hz/reviews-render/lighthouse/B074Q254HB?filterByKeyword=mms&pageNumber=1
Bad things are created by government action. If you are unwilling to acknowledge that, you should be ignored.
Indeed, we face a complicated tradeoff of bad things, and the argument is largely about what the right point on the tradeoff would be. There is no practical point in the space of possible tradeoffs where we don't sometimes get approved medicines that have terrible side-effects, or drugs that are only marginally better than a sugar pill. There is also no practical point where we don't sometimes delay the availability of lifesaving or genuinely helpful drugs. (We can avoid the first by banning all new drugs, and the second by abolishing the FDA, but short of that, both errors will be made sometimes.)
ISTM that one fairly easy improvement we could make would be to use other countries approvals/experience to allow drugs in the US. Something like "if it's approved in the UK/EU/Japan/etc., and has been on the market a couple years there, it's approved in the US as well unless the FDA makes an explicit case to keep it out." Or "you can buy/prescribe things that have gone through safety but not efficacy trials under these conditions...."
I expect there are more ways to improve the tradeoff we're making, but I don't have great confidence about that--this is a place where all choices have painful costs associated with them. OTOH, Scott is speaking from experience as a practicing psychiatrist and has run into several cases in his practice where FDA policies seem to be making things worse. Probably we should pay attention to that, even if we're not sure exactly where the optimal tradeoff point is.
"painful costs". Yes, but who gets to make the decision about what painful costs are to be borne? That's the problem: it's not deciding between risks, it's *who* decides. If you think regulation is a good thing, then you think that the government is better at deciding who should bear what risks. Do they have a better incentive or better information than the individual who is making the decision for their own risk?
Well said — people think so strangely about this topic. "Once upon a time, candy very rarely had bad stuff in it, therefore there is no such thing as over-regulation" really seemed like a good argument to at least two people in this thread, apparently?!
Deisach contributed several specific examples with evidence, historical context with a link to learn more, and several well formed arguments. That’s great!
If we’re discussing the FDA, the historical context of why the FDA was established and what law established it is interesting. As is the example of the sort of bad thing that might be encouraged by the FDA’s abolition.
It wasn’t exactly a powerful and fully sourced peer reviewed journal article on the topic. But it’s perfectly good! And it started a few other conversations
> and what arguments?
short and unsatisfying arguments can still be arguments
I have to agree with you there, WC; things have never been the same since they took the alcohol out of gripe water:
https://www.medicinenet.com/why_is_gripe_water_banned/article.htm
I was reared on the old-fashioned formulation and look out I turned out! 😁
I don't agree. I don't see any well-formed arguments; all I see is "yeah but with NO regulation, you could sell poison", and this is so obvious and well-covered that it doesn't really need examples and evidence — or mentioning. The question is, as always, how *much* regulation we need *now*; the situation with confectionery in 1906 is essentially irrelevant.
Cool. Here’s a real well formed argument by me. https://astralcodexten.substack.com/p/contra-drum-on-the-fish-oil-story/comments#comment-2571594
Nobody’s really taken a look at it, sadly. Wanna take a crack at disagreeing with it?
"What are you contributing here? "
Well indeed, WC, that is a question we must all ask ourselves at some point in our lives. "Who am I, where do I come from, where am I going?" as the British Army patrol asked Kevin McAleer.
1. I am Thirteen Ways Of Looking At A WC
2. I am ragweed pollen on the breeze, and the solar UV index is moderately high.
3. I am he as you are he as you are me/And we are all together
4. I am the guttersnipe's guffaw at your somewhat unfortunate initials https://en.wiktionary.org/wiki/water_closet
13. Ceci n'est pas treize
6. I am the head of the spear in battle
Who can say, WC, who can say? To quote Philip Larkin:
"Ah, solving that question
Brings the priest and the doctor
In their long coats
Running over the fields."
An orbiting pendulum looks like it's simply swinging from any given point, and yet in actuality (and in spite of appearances), at no point in its orbit does the bob actually approach the happy point it is orbiting.
I'd be more sympathetic to Scott's position if we lived in a world where miracle drugs like imatinib were being developed every year*. But drug development is very, very, very hard. Your pre-test probability should be that most novel pharmaceutical interventions will not significantly outperform current standard-of-care + placebo. Given that landscape, the mandate and ethos of the FDA makes quite a bit of sense. The public (and the American healthcare system) is much more at risk from being victimized by ineffective and/or harmful treatments than having genuinely transformative medications delayed or withheld.
*(I'll grant Scott's point that the particular mandate/ethos was a very bad fit for development of COVID vaccines, which are the rare intervention with an enormous effect size)
The asymmetry I'm assuming is that for a given indication, a vast majority of pharmaceutical compounds won't work better than placebo. So if you have a marketplace flooded with Placebo Plus products with no meaningful regulatory restraint or oversight, you dilute the number of people getting effective treatment, you increase financial toxicity to the healthcare system, and you are exposing people to a wide range of tail-risk side effects.
Robin Hanson's work indicates that a whole lot of medicine is no better than a placebo. But the effect of this is just to waste money. If lots of people buy placebos that do nothing, and people with rare conditions (like those infants or the guy that lost his intestine) that can actually be improved by drugs get those drugs, it sounds like a net plus.
Leaving aside the ethics of constructing a society that caters to a small cohort of the information elite, only a tiny fraction of people have the time, expertise, and disposable income to evaluate the evidence base of candidate pharmaceuticals and conduct autologous n of 1 trials - and that fraction would turn to '0' if you deregulated the space and allowed shoddy, biased, and outright fraudulent research results to proliferate.
If you want evidence of that, you need to look no further than 2020. Many of the people who would describe themselves as 'diligent about [biomedical literacy and research]' went head-over-heels for a whole variety of placebo-equivalent antiviral candidates. There was so much clinical and biomedical data - much of it shoddy, unreliable, and non-replicable - that people's empirical apparatuses were swamped.
Imagine that, but for every indication and disease. How would a person with a life-threatening condition decide which drugs to try, in what combination, and in what order? How would insurers decide what was reimbursable? How would doctors be able to decide what was safe and effective to prescribe patients? How would the system bear the costs of this flood of products? Aducanumab alone had the potential to bankrupt Medicare, and that's just one shoddy monoclonal antibody.
Actually I think doctors all over the world were doign the same thing. Nobody knew what would help with covid, so they threw anything that might help at it--zinc, HQC, azithromycin, ivermectin, vitamin D. I think this always happens with a new disease. Eventually, people found some stuff that seems to work and switched over.
"But drug development is very, very, very hard"
I go back and forth on this. It's definitely hard with the FDA involved. But it's weird how often drugs approved for one indication randomly help with another (eg fluvoxamine, an antidepressant that seems to also treat COVID - or ketamine, an anaesthetic that seems to also cure depression - or any of the substances mentioned at https://slatestarcodex.com/2014/07/17/psychotropic-base-rates-the-argument-from-antibiotics/). I'm not really sure what to make of this.
I'd love to see good data on the success rates for drug repurposing (and by success, I mean statistically significant differences in clinically meaningful outcomes in large double blind RCTs). I suspect the number is quite low, and that number is likely an underestimate given how much of the denominator would be hidden / unpublished.
I agree that many drugs have off-target activity that we don't always anticipate or understand. But I'm much more skeptical that the off-target activity is clinically meaningful, or if it is clinically meaningful, that it would outperform current standard-of-care.
Derek Lowe has had faint praise for drug repurposing, that may be worth reading. Here's one, though he mentions it in passing in a few other posts that I don't have the attention to find at the moment.
https://blogs.sciencemag.org/pipeline/archives/2020/09/11/drug-repurposing-how-often-does-it-work
This is interesting, thanks for the link! It mostly aligns with my priors but he does bring up a few good examples (like thalidomide) where very high-signal off target effects were effectively repurposed within the same drug class.
I mean Covid drug repurposing didn’t seem to go great? A lot of drugs were identified and given to a lot of people. Azithromycin hydroxychloroquine remdesivir and a bunch a others, and i guess ivermectin although I can’t say if that one works or not.
(And they didn’t work, and hcq specifically made things worse?)
https://www.nature.com/articles/s41467-021-22446-z
combined OR on all-cause mortality for hydroxychloroquine is 1.11 (95% CI: 1.02, 1.20; I² = 0%; 26 trials; 10,012 patients) and for chloroquine 1.77 (95%CI: 0.15, 21.13, I² = 0%; 4 trials; 307 patients).
Going with 1.11
The pharmacies steadily dispensed an average of 383,685 prescriptions for hydroxychloroquine (HCQ) per month, and 1839 prescriptions for chloroquine (CQ) per month. However, from February to March, prescriptions for hydroxychloroquine jumped by 86.2%—from 367,346 to 683,999. Prescriptions for chloroquine jumped from 2346 to 6066—an increase of 158.6%. Among those patients given hydroxychloroquine, the number who were simultaneously prescribed azithromycin jumped from 8,885 in February, to 101,681 in March—a 1044% increase.
https://www.contagionlive.com/view/hydroxychloroquine-chloroquine-prescriptions-covid-19-us
(I am now going to pull numbers out of my colon, and I don’t expect this to have really any relation to actual hcq deaths. I probably misinterpreted at least two of my sources somehow)
Assuming of 300k given hcq, the case/hospitalization and hospitalization/death ratio for the general population are the same as for that group (which there are no grounds at all to make), 300k* 600k deaths / 35M ~~= 5.1k deaths. 11% increase = 500 deaths. If it’s 75k like another source suggests, 125 ish. Which is quite a few, but not exactly at the scale of the vaccines being delayed. (But I would submit again that the fda process being very strict may have pushed the drug companies to have actually good development processes and those with to the top)
I believe the drugs that were useful were the steroids, which is much less an experiment than the others.
Is drug development really hard? Or is it really hard to develop a drug that has a no-kidding-sure-fire $1BB+ market?
I think both. Huge amounts of time and treasure and genuine genius have gone into treatments that looked great until the clinical trial, where it turned out they didn't do any good. But also stuff that works can be impossible to get approval for because it can't be patented and producing the evidence that it works sufficient to get approval costs hundreds of millions of dollars.
> Given that landscape, the mandate and ethos of the FDA makes quite a bit of sense.
There seems to have been ample evidence to the contrary for the fish oil. It wasn't even a "new" drug. Something that has a proven safety record in other jurisdictions shouldn't have to overcome the same hurdles as a truly new drug.
The pendulum has never swung back to the old days you bring up. Even though we could rely on third-part certification like Underwriters Lab, fraud laws for false claims, and Hanson's would-have-been-banned stores.
I don't find it hard to say we're over-regulated. Here, I'll say it. "We're over-regulated". The counter to your "merrily put poison in sweets" claim is the requirement that manufacturers of industrial alcohol merrily put poison into their products. I've seen an estimate that, since prohibition, 10,000 people have been killed by this government requirement.
Essentially, your claim is that individual action in the marketplace is insufficient to solve problems of unhealthy foods. Perhaps that's the case, but it doesn't explain why the government has been encouraging people to eat unhealthy foods (too many carbs, because "eating fat gives you heart disease", when that's exactly backwards) for most of my lifetime.
Indeed, the government has often *heavily subsidized* unhealthy foods!
So, do you suppose there could be some possible middle-ground between allowing people to put poison in candy and preventing people from using a treatment that's known to work and approved in many other countries?
This is like the dumb internet argument style where anyone who proposes less government is accused of wanting to move to Somalia, while anyone who proposes more government is accused of wanting to move to North Korea.
Is there a middle ground between having a forest fire burn down your entire forest and not having it burn down any part of the forest?
Part of the problem is that you're thinking this decision can be made by itself: that we have a knob which can control the amount of regulation of substances, and that we just tell our politicians where we want the knob to be set, and them and all the other politicians put in their setting, the average gets taken, and that's the setting.
Firstly, most people are dumbshits, as I'm sure you've noticed. Second, we never send our best and brightest to govern us. In fact, the best and brightest are too smart to want to be politicians. Third, once you give a politician power, they never want to give it up, so the knob only gets turned in the direction of more control, and only rarely less control. Fourth, political power is fungible. You may think you're giving a politician the power to keep us safe from unsafe substances, but what if they trade it to another politician (who has promised to eliminate unsafe substances from the market) in exchange for the power to benefit an evil corporation?
Scott
Your case would be more powerful if you offered examples of times, probably more than we know, the FDA put the kibosh on as yet to be known harmful therapies. I keep a quote handy I fall back on occasion: Thomas Chalmers, the famous medical researcher, once wrote, “One only has to review the graveyard of discarded therapies to discover how many patients have benefitted by being randomized to a control group.” Not quite apples to apples to your response above, but it's close to the spirit.
Also, I don't follow your two-track description of the FDA. The first, "stupid," and the second, THe FDA is a product of the charter organizing committees set forth for them: slow, deliberate, and overly cautious. The latter doesn't validate the former. If you didnt imply that, I misread your intent—apologies for the misinterpretation.
Brad
The most radical FDA reform proposal I've heard is that the FDA should use its current system to check if drugs are safe, but be laissez-faire about whether or not they're effective. The safety checking is (as far as I know) relatively quick and easy and not usually most of a delay (though I would have to double-check this before asserting it officially). I think this is kind of like what they do with vitamins/supplements right now, which are cheap, numerous, easy-to-get, often ineffective, but safe enough that if you dared someone to take a randomly chosen Whole Foods product I think only the most over-cautious people would have any qualms.
Overall, I like that idea, but implementation is going to have to resolve the issue of "safe" not being binary.
To pick and extreme example, chemotherapy drugs are literally poison, but they're net benefits to patients with certain forms of cancer because undertreated cancer is worse for the patient than the chemotherapy drugs. You're probably going to want some kind of backup process where some standard of evidence of benefit can be used to get approval despite known or likely hazards of the drug. The tricky part is to stop the acceptable level of risk for basic certification and the standard of proof of effectiveness for approval despite risk from drifting to the point you're effectively back to the current system.
Indeed. And sometimes how we think about risk is just weirdly broken or mis-calibrated (the reaction to the super rare AZ or J&J vaccine-associated blood clots, for example). Certain people get all weird about me taking medicine that ups my risk of blood clots and breast cancer, when realistically the risks are comparable to the average female baseline risk (which is higher than the normal rate for a male), and they also significantly reduce my risks of falling victim to other things, like prostate cancer, alongside the positive effects which I want. If I were to get certain surgeries, some doctors would force me off my medications for weeks leading up to it because of that fear of blood clots, despite the risk not actually being high (cis women on birth control are at significantly higher risk and generally doctors won't force you to stop that before surgery) and going off the medications being quite unpleasant in and of itself.
You can't really consider them in isolation though. We ought to accept a drug that prevents 10,000 fatal heart attacks a year even if it causes a dozen liver failure each year too. I mean, the FDA didn't accept the first beta blocker historically and Europe did but they shouldn't have accepted such a risky drug without proof of efficacy. Of course, an easy solution is to be open to safe drugs bit stringent with dangerous ones.
My version is two tier:
Tier 1: Drugs that have proven reasonably safe can be manufactured and sold, but can't be marketed with claims of curing disease, and won't be covered by insurance.
Tier 2 is drugs that are also proven effective. That's the only tier we have now.
Agreed this is a good first draft of a proposal.
:+1:
Now watch as the cost of a drug is multiplied by a factor of ten the day it's moved from Tier 1 to Tier 2, because insurance is paying so they can charge $100 / bottle for what was formerly $10/bottle over the counter.
I feel like at this level, though, it wouldn’t resolve a bunch of the fundamental objections. Ketamine, to take one of your frequent examples, doesn’t seem like it meets this bar: you say it causes “acute distorted cognition for the next few hours” at clinical doses, and it has lots of potential harms at recreational doses. That doesn’t seem like a level of safe that we’d be happy to have it sold over the counter at Whole Foods, which presumably would move it into the “must show effectiveness” track and we’re back where we started.
That’s not to say that the FDA is the right answer, but lots of medications have some bad effects that trade off against their good effects, so just evaluating safety without an eye to effectiveness seems like it doesn’t actually result in lots of medications meeting the bar.
Ketamine is safe, *if* taken at the recommended doses. It is also an addictive drug (US schedule 3)
Addictive drugs would certainly be left out of any such reform in the US.
Though I'll note that alcohol *is* sold freely at Whole Foods.
The FDA would be more useful and believable if they restricted themselves to an advisory role. Then you would have insurance companies saying "Sorry, we don't pay for this because the FDA advises against it." but people could still get the compound they believe helps them.
When I take a drug to treat my diabetes, or blood pressure, or some other chronic condition, or even cancer -- then yeah, I want to be pretty damn sure it's effective, not just safe. Leaving my conditions untreated is *not* safe.
Of course! So why do you trust the government to do what you should be doing? What would happen if they failed to ensure that your medicine is effective, like they failed us by telling us that fat is bad for you? What mechanism is in place to correct their failure? If a private company created medicine that was not effective, they would probably go out of business. How might the FDA "go out of business" given that they are reliant on politician goodwill, not customer goodwill?
> So why do you trust the government to do what you should be doing?
Whom should I trust instead ? Every snake-oil salesman who comes my way ?
Look at it this way. The last time you rode in an elevator, did you stop to consider the probability of it crashing to the ground, burying you deep in the elevator shaft ? Did you check the local reviews for the construction company who built the elevator before getting into it ?
If you trust "the government" -- or rather, the teams of engineers and scientists whom the government either employs, or forces to pass standardized test -- to maintain your elevator, then why not medicine ?
Who should you trust? Walgreen's. Kinney's. Rite Aid. CVS. They should be the ones testing for safety because unlike any government agency, they compete with each other and will HAPPILY test their competitor's product and gleefully point out anything that isn't safe.
I trust the elevator company because they will lose everything if any one of their elevator crashes to the ground because of incompetence on their part.
Why do you trust the government? Who is monitoring them? Not CVS. Not Walgreen's. Not Kinney's. Because none of them have any interest in catching the government's errors, because all that will do is cause everybody to stop buying from their local pharmacy.
If you ask me to predict the future, I will ask you what incentives people have. CVS, Walgreen's, Rite-Aid have good incentives, the FDA has bad incentives.
> I will ask you what incentives people have.
Interesting. Asking that question is exactly why most people I know trust the (Canadian) government more than private companies.
I saw a slightly less radical proposal along the same lines here:
https://www.niskanencenter.org/how-can-the-fda-foster-greater-resilience-in-the-medical-marketplace/
Basically that they should approve for safety and effectiveness, but not clinical utility. And have explicit tiers available for measured effectiveness (biomarkers / symptoms / survival etc).
Sure, but that opens the door to homeopathic-style treatments. Is it safe? Of course, there's not enough of anything in this to do anything good, bad or indifferent. Is it effective? Well, if you don't have anything really wrong with you and it triggers the placebo effect, sure. If you do have something really wrong with you, um, yeah.
But hey, at least it's a lot cheaper than Aduhelm at $56,000 per year! Which is something approved as safe but not effective:
"Why does this drug cost so much?
There are millions of people and their families who want a new Alzheimers treatment. Any Alzheimers treatment.
"We have established a price for Aduhelm that reflects the overall value this treatment brings to patients, caregivers and society -- and one that will enable continuous innovation," wrote Biogen's CEO Michel Vounatsos in a letter announcing the approval.
The nonprofit Institute for Clinical and Economic Review, which scrutinizes drug prices, argues that because the drug has not been proven to be effective, a more reasonable and cost-effective price would be much lower -- maybe as low as $3,000. That's far from the $56,000 Biogen will charge.
The impossible problem.
More than 120,000 Americans die from Alzheimer's each year, making it among the leading causes of death in the US.
But treatment of the disease has been stuck for years, and the FDA had not approved a new drug to treat this disease since 2003."
The bargain is not "how many thalidomides is one omegaven worth", it's "how many aduhelms is one omegaven worth".
Omniscience is not available to anyone, even someone who works for the government. It is only through trial and experimentation that we know anything. The FDA's remit in verifying effectiveness is very problematic. It should only be testing for safety. It is up to other parties to determine effectiveness.
In the US, you can buy homeopathic remedies and herbal medicines and such over the counter. And you can go to a chiropracter or accupucturist or naturopath for your cancer. So maybe this would be a terrible world, but it woudl be the terrible world we already live in, at least in the US.
Still don't have an answer to the question of "how many patients does this (lack of approval for the fish oil formula) impact every year." I am betting that it is under 1k and probably under a hundred, not the tens of thousands originally given.
Still have not been able to find this information, but I admit I was very possibly wrong about that.
FDA-employed researcher who is a partner of a family member, enormously paraphrased: "the article correctly described the FDA's difficulty with too fast/too slow on drugs, but noted that the 'F' part of FDA doesn't have a similar issue, mainly because it doesn't approve foods before they can be sold as food; if something appears to be making people ill, it’s super fast and super easy to get authorization and budget to study/ban it, but for foods which no one has thought to test yet and which are only on the radar on priors like the one I'm currently working on, it took a year to set up protocols and it will be at minimum a year before any results of anything are to be looked at".
I went through a similar analysis as yours on reddit when I first read Kevin's article (https://www.reddit.com/r/slatestarcodex/comments/ozz9je/interesting_criticism_of_the_recent_fda_posts/h83hnn8/).
I disagree with your assessment that you're "basically correct" when you're off by a factor of 2-4 in the detail that serves as the crux of the point. You're right that this is an example of the FDA being overly restrictive on a drug that ought to be widely available, but there is an absolutely massive difference between an organization that delays new drug releases by 5 years compared to 20 years.
I agree I came on way too strong originally, but I think there's some leeway in when you date from. Omegaven was invented in 1991, there was a WSJ article about it being a miracle drug in 2006 ( https://www.wsj.com/articles/SB116338867216321332 ), there were studies confirming it in 2011, it was well-known in the libertarian blogosphere by 2014, and it got approved in 2018. Which of those is the starting point for saying the FDA was "delaying"? In a default-yes medical regimen, I think people would have started using it around the WSJ article in 2006, so maybe 12 years earlier.
One of the questions I was thinking about when reading your initial article was: Assuming we can only work with the single lever of FDA restrictiveness, how many Omegavens is one Thalidomide worth? The number of years the FDA delayed Omegaven strongly influences my answer to that question (as does the rate at which the FDA encounters "Omegavens" and "Thalidomides" in the wild). That's what I mean when i say that the particular detail you were most incorrect on is an essential one, and being off by a factor of 2 is a large one.
To its credit, the FDA didn't approve thalidomide: https://en.wikipedia.org/wiki/Thalidomide_scandal#United_States
Very good point, I apologize for implying otherwise.
It seems to me, however, that Scott is arguing for a standard wherein the FDA would have gotten Thalidomide wrong. In this case, it would have been by adopting a European approved drug, but it probably would have also approved it in its own channels by only looking for first-order safety effects as opposed to inter-generational ones (perhaps he believes we ought to have stricter safety standards for researching drugs given to pregnant women?). I do suspect that on net this tradeoff would be worth it; that the cost of Thalidomides would be more than made up for by easier drug access and accelerated research.
The history as I have heard it is that prior to the thalidomide scandal, the FDA checked only for safety; after it, it started checking for effectiveness as well, ("Kefauver-Harris Drug Amendments"), even though the safety check caught thalidomide.
The first COVID vaccines passed the safety check in, IIRC, May 2020. The rest of the delay was effectiveness.
How many people died between May 2020 and whatever month in 2021 you choose to mark for the "general availability" of the vaccine, and why shouldn't we lay those deaths at the feet of the FDA? Somebody's got to go to jail.
I think overall it would still be worth it if some thalidomides got through, but see my response to Brad above - I don't think there's any need to compromise safety here.
> how many Omegavens is one Thalidomide worth?
That seems like a red herring. Omegaven had a proven safety record in Europe.
Any clock should probably start with when it was approved for use in EU. Off label applications are allowed in US even when they haven't been "proven" to be effective. Simple reciprocity could have avoided the entire fiasco.
This isn't journalism, it's blogging. You don't have to be exactly right the first time, you only need to be willing to correct yourself.
My point was that I believe Scott got it wrong the second time as well when he evaluated his initial take as basically correct.
I do believe Kevin was unnecessarily harsh, since the facts do support Scott's conclusion, but more weakly by a factor of 2-4.
But Scott could have written the exact same paragraph, with exactly the same amount of outrage, only replacing '20' with '10.' This means he's basically right and just needs to make a minor correction.
Consider two different ways someone could be off by a factor of 2-4.
Example 1: I claim we should do my economic plan because it has a positive NPV of $1B compared to the next best alternative plan with a positive NPV of $0.9B. If it turns out my economic plan actually only had $250-$500M of value, my main point is refuted and I should retract.
Example 2: Imagine I'm going around screaming, we need to stop this serial killer because they've already killed 20 people! And it turns out they had, in fact, only killed 5 to 10. If I'd falsely attributed 10-15 deaths to the serial killer, I should correct that and discontinue saying the false number once I learn the truth. But the main point of stopping the serial killer is very much still in place!
Scott's error was very much the later type.
Scott usually operates within a utilitarian-ish framework wherein certain things trade off against others, and therefore are generally examples of type 1. I generally operate within this framework as well, and when doing so being off by a factor of 2 *may* lead to significantly different conclusions!
Yes, Omegaven getting delayed is evidence against the FDA no matter what the duration of the delay is. However, the degree of evidence is proportional to the length of that delay (among some other factors like the magnitude of impact that delaying has). So when you're off by a factor of 2, you've doubled the evidence of your example! That should be a big deal.
That evidence against then has to trade off against evidence for: the FDA didn't approve thalidomide in the US largely because they deemed European approval insufficient. Its rules provide some protection, the question is whether the cost of that protection is too high. If you fudge the evidence by a factor of 2, that can significantly shift the answer.
>"I want a world where if Europeans have been giving fish oil based nutrient fluids to babies for years, and everybody from NBC News to random bloggers know that the babies are much less likely to die on fish-oil based fluids, and all the studies that have been done confirm it, you are allowed to give those babies the fish oil based fluids without waiting another five years for an FDA approval."
Isnt the problem in this case that it was not as clear cut as you are trying to make out, based on https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7671012/ , it seems like there was actual opposition from within the medical community, not the FDA, and that along with a lack of support from the drug company was the reasons things took so long. Ie Gura calls out
`The fact that we demonstrated that IV soybean oil was detrimental, but IV fish oil was not, created controversy in the medical community. Papers and commentaries were published that suggested that Omegaven caused growth failure, bleeding, and progression of hepatic fibrosis. These criticisms led to us to do more translational and clinical studies.`
This being said it looks we got a pretty positive paper in Nature in 2011, https://www.nature.com/articles/jp2010182 , after which point it seems like education more that the FDA was the problem. Do we have any recorded cases of doctors who wanted to use Omegaven having any difficulty in acquiring it after this point?
I feel like I addressed this pretty thoroughly.
Yes, "lack of support from the drug company" was relevant *in the context of the current system*, where without drug-company-level resources you can't push anything through the difficult FDA approval process. If you make it cost millions of dollars to do something, there will always be someone who resists paying.
There was (bad) opposition from within the medical community, but I don't think it would have passed a simple cost-benefit analysis, and a quick check at Europe, where this had been used for a long time, would have defused it.
In terms of "did doctors have trouble acquiring it" - doctors at hospitals that were willing to do an IND (very complicated FDA related process) did not, but very few hospitals were willing to do that. If you can call making people go through a very complicated process to vote "voter suppression", I think it's fair to call this "drug suppression". Of course you can always just do the complicated thing, but many people won't.