Enter Covid-19, the disease caused by the new coronavirus, SARS-CoV-2. There are no proven therapies to treat an infection of that virus. One of the most promising candidates is Gilead Sciences ’ remdesivir, which has shown some potential in treating other coronaviruses, such as SARS and MERS. Due to its use in treating Ebola infections, remdesivir is known to be generally safe.
With the FDA’s two-hurdle requirement in place, however, remdesivir must be proven effective in clinical trials before it can be sold for widespread use in Covid-19. These clinical trials take time, and they still won’t answer many of the questions doctors have about the drug’s use against Covid-19 now. What’s the right dose to get remdesivir into the lungs? How early should it be given? If it is given to sick patients already on ventilators in hospitals, will it help? All these things will depend on the individual patient.
This is from Charles L. Hooper and David R. Henderson, “FDA Shouldn’t Keep Safe Drugs Off the Market,” Wall Street Journal, March 25 (March 26 print edition.)
Read the whole thing if you subscribe. I will post the whole thing in 30 days, the amount of time I’m contractually obligated to wait.
READER COMMENTS
Alan Goldhammer
Mar 26 2020 at 7:52pm
David – I don’t have a subscription and cannnot read your piece. However, from my experience I will make the following points.
FDA’s treatment IND regulations allow any company to sell their drug at a price that recovers research costs. You will have to ask companies why they don’t take advantage of this. Remdesivir was used to treat very few patients during the Ebola outbreak, hence you cannot say that this is a safe drug. The full safety profile of most drugs is not known at the time of approval. This is why companies spend a lot of money and have dedicated personnel who staff pharmaco-epidemiology groups. Drug prescribing information is constantly being updated as reports of new safety issues come in. I’ll give you one good example; Pfizer spent a lot of money getting travafloxicin, a broad spectrum antibiotic, approved. As I recall they went above and beyond studying it for many more medical conditions than they needed to. When it came on the market, it exhibited such bad liver toxicity that it had to be withdrawn. Of course the Cox-2 inhibitors are also a well known study of a drug looking good until it started getting used by lots of people.
I was part of the HHS AIDS Drug Parallel Task Force back in 1991. this group addressed the exact same issue you write about. We came up with reasonable guidelines for setting up a separate track outside clinical trials for HIV-positive patients to get experimental drugs provided this did not compromise the ongoing clinical trials.
Most community physicians are not skilled in conducting clinical trials. Data from the use of the drug in such settings is usually of low quality and not suitable to support a marketing approval. Clinical trials have defined protocols that have to be followed and there is a lot of monitoring and data collection. the questions you pose in the quoted paragraph are not going to be answered in community use settings. Even more important, such settings are never going to collect good safety data that might inform the proper use of the drug.
From a legal liability perspective, companies rely on FDA approval as a shield. this would not be the case if a drug is sold without approval. Would patients receiving the drug have to waive liability, would they have to sign an informed consent form, would the MD who prescribes the drug be liable?
Finally, experimental drugs are not manufactured at scale. It is quite likely, as it was with experimental HIV medicines, that not all patients who want access can get them. this is the case with remdesivir, which is not an easy drug to make to begin with.
The minute good clinical data comes out, the drug will be approved. I remember back when Novartis was testing Gleevec for a rare leukemia. the Data Safety Monitoring Board shut down the trial because everyone responded to the drug. Novartis CEO, Daniel Vasella, made a corporate decision to make the drug available to all who needed it for free while it was pending review at the FDA (much to the dismay of his production team who thought they would have 4 months or so to get it scaled up). I hope Gilead would follow this step.
Charley Hooper
Mar 26 2020 at 9:13pm
Any time a drug is given or sold to patients during the FDA review process, generally only bad things can happen. Because the drug is still under scrutiny by the FDA, anything untoward that happens to patients (such as an untimely death) may delay or scuttle the review and therefore puts that product’s launch at risk.
Regarding safety not being know at the time of approval, I agree. But that calls into question the whole process of the FDA approving drugs as “safe.” Safety lies on a spectrum and it changes as information improves as a drug is used.
How does this agree with your earlier statement that:
We either get information from widespread use or we don’t.
Regarding using the FDA approval as a shield, I don’t believe that’s the case. When Vioxx had problems, aggrieved patients and their attorneys didn’t complain to the FDA, they went straight to Merck. Plus, drugs would still be “FDA approved.” Only that approval would be for safety, not safety and efficacy. So if there is a shield, it will still be present.
Regarding experimental drugs not being manufactured at scale, that’s because the companies know that they need to produce only enough to satisfy the clinical trials and QC/QA. However, if companies could sell drugs that were approved as safe, the drugs would no longer be “experimental” and the companies could plan ahead for sufficient manufacturing.
Drugs are certainly not approved the minute good clinical data comes out. If a company needs to run two Phase 3 clinical trials for approval and the first one shows good results, the company will still need to run a second trial and wait for the FDA review process. This typically takes years.
Alan Goldhammer
Mar 27 2020 at 10:52am
It’s getting too tiresome for me to rebut points. I’m already overloaded doing a daily news digest for fellow retirees from the pharma industry which includes several from your former employer who served in legal, drug safety and clinical development capacities. You are wrong about the Vioxx story in terms of when the information started coming in. Merck has successfully defended itself in virtually all liability cases using the FDA defense (I know several of the lawyers who were involved in this as well as other industry lawsuits).
You are totally wrong about the lack of quick approvals. If trials are done well and the data is good, drugs are approved quickly. Merck had their first HIV protease inhibitor approved in less than three months with one well designed Phase 3 trial. There are other good examples as well. Crappy designed trials and week data result in FDA limbo as they should.
Your statement on manufacturing is off base as well. Say a company does scale up and then the drug is shown not to be effective, and there is ample case history here in the antiviral drug space. There would be a significant sunk cost that is lost and would have to be written off.
In the information I’ve been tracking there are about 60 good FDA approved or unapproved clinical trialed drugs that safety data is available that may be as good or better than remdesivir. What is needed are better approaches to get things into controlled trials quickly so that we can see the data. Maybe remdesivir or the hydroxychloroquine/azithromycin drugs work; I would rather have 60 compounds in trial than one or two. It’s not at all clear for serious ICU patients that these drugs will work as they have other underlying medical conditions (cytokine storm which is being addressed by trials of several MAbs against the IL-6 receptor).
I’ll conclude by noting that there is now observational data coming out (and more work is being done) that show the anti-hypertensive angiotensin II blockers appear to reduce disease severity. Perhaps we should write Rx scrips for losartan and the other ARB drugs for everyone. Keep blood pressure low and reduce SARS-CoV-2 infections.
There will be a lot of lessons learned following the break in the pandemic and I’m working on a paper that addresses key issues. The sales of unapproved drugs is not one of the solutions.
Charley Hooper
Mar 27 2020 at 5:59pm
If you are correct, then why did Merck pay $4.85 billion to settle 27,000 lawsuits by patients and their families? That doesn’t sound like success to me.
What did I say to make you think that I disagree with this statement? I said:
Part of this planning process would be to consider the probability that the drug would fail somewhere along the line.
Dylan
Mar 28 2020 at 10:18am
Charley – I’d appreciate a little more of your expectations of how drug development would look if the FDA took your proposal and approved drugs based just on a Phase I trial. Would you include any sort of post marketing data collection process as a condition of early approval? Would you expect that we would still get large, double-blinded trials for most medications? If so, do you think recruiting for those types of trials would be more difficult in a world where the alternative is that the person can just buy the experimental treatment directly? If you don’t think we’ll get double, blinded trials, do you think other methods are sufficient for getting reasonably strong scientific evidence of efficacy and safety?
Phil H
Mar 26 2020 at 10:58pm
Alex Tabarrok has given some good arguments that the FDA’s processes aren’t great at the moment. But still: evidence that there is some problem with the FDA’s current processes is NOT evidence that a different process would work better. In particular, the process “let’s use this drug because it seems like it should work” has a very strong track record of being a complete disaster.
In the 19th century, before they invented clinical trials, they literally sold snake oil. It’s a metaphor now, it was a reality then. We do not want to go back to that.
David Henderson
Mar 26 2020 at 11:03pm
You wrote:
This is from an NPR item in 2013:
Phil H
Mar 27 2020 at 4:51am
Thanks, David. It’s an amusing historical anecdote, but TCM isn’t really what we want to base modern medicine on.
More importantly, snake oil wasn’t the substance itself, it was the practice of making claims about medical efficacy that are not supported by medical science. The process of determining which medical treatments work and which don’t is incredibly difficult, painstaking work. Regulatory backsliding is unlikely to help. If remdesivir works, that would be great. But have you calculated the cost if it doesn’t? How many lives will be lost because doctors use the wrong drug? How much time and money will be wasted setting up the supply chains for the wrong drug? Until some level of clarity on these questions is found, betting on untested remedies is not much better than closing our eyes and making a wish.
Jon Murphy
Mar 27 2020 at 8:42am
Aside from the fact this is not historically accurate (“snake oil” was not as widespread a problem as one would think for the simple reason that people are not foolish), the FDA doesn’t currently regulate “snake oil.” Walk down any nutrition aisle and you’ll see rows and rows of supplements, meds, energy bars, etc etc., all of which have a sign on them “These statements have not been evaluated by the FDA.”
Phil H
Mar 27 2020 at 11:16am
Jon – that’s correct, and the fact that the FDA hasn’t approved them means they can’t make specific kinds of health claims, like for example, “this product is a treatment for covid-19”. You can sell anything. The FDA tells you whether or not you can claim that it cures an illness.
Many “snake oils” back in the day were sold with explicit claims that they would cure/treat named illnesses. That absolutely does fall within the remit of what the FDA does today.
Jon Murphy
Mar 27 2020 at 11:28am
The question is: “why”? Why is it that people are smart enough to be able to tell if Emergen-C actually prevents a cold or not (and thus no need for FDA regulation) but these same people, when under the care and guidance of a medical expert, suddenly become unable to evaluate the very same claims, and thus need the FDA?
Dylan
Mar 27 2020 at 12:49pm
People aren’t smart enough to be able to tell that, which is why those products sell despite lots of evidence that they don’t work.
The problem is, no one can tell without mountains of evidence, whether a drug will be beneficial and won’t cause more harm then good. I think people that haven’t worked in this industry are just really unaware of how little we know about drug discovery. Start with 100 really good ideas, that have scientific supporting evidence for why this pathway should be effective in some particular disease. You’re going to lose most of them before they even get to clinical trials. Either you can’t actually make a drug that has the properties you want, or it has toxicology issues, or fails in animal models, or a bunch of other things. You might get 10 of those original 100 into human trials. But, then your work is just really starting, because, again the vast majority of those 10 drugs won’t do what we expect them to do. They will have dangerous side effects maybe, but even worse, most of them, even the ones that showed some kind of efficacy in an animal model, just don’t work when we look at them in the clinic in a controlled setting.
The people that have sunk hundreds of millions of dollars into proving they work, are often able to squint hard and think they see that it might work in some subpopulation, and convince investors to pour millions of dollars more into trying to prove that, even though those efforts almost always fail.
In the end, out of those 100 ideas that made it into preclinical trials, we’ll be lucky to see one that is actually a working drug. And that is after the best efforts of huge teams of experts. So, it isn’t that we just distrust the ability of the general public to know what works, it is that we distrust the ability of well-meaning experts to know what drugs work unless they’ve done a really large, double-blinded controlled trial…and even then, that only gets us to be a little more certain. There’s still plenty of FDA drugs on the market that don’t really work, but have been approved because there just isn’t anything else.
Jon Murphy
Mar 27 2020 at 1:08pm
Which leads back to the question what is the FDA trying to do?
Absolutely, 100% agreed. All the more reason for the FDA to step out. Increase that N.
Dylan
Mar 27 2020 at 3:46pm
Make sure that the products on the market actually have evidence to support what they claim to do, and that they won’t make you ill or kill you in the process?
I think that depends on how you think sufficient evidence is generated. Many of those drugs that don’t work, look like they work in poorly designed or underpowered trials. Without the FDA as gatekeeper I see a couple of potential options:
A new gatekeeper steps in like the insurance companies, and mandates the same/similar types of trials that are currently run. Best case scenario is they create a centralized organization that takes care of this, and they agree to reimburse based upon “approval” by this centralized agency. If it isn’t centralized, then you run into the problem where a drug company either has to run different trials to get on the formulary for different insurance companies, or you get into a lowest common denominator situation, where they only run trials that are weak and not conclusive.
We get a drug industry that looks like the supplements industry, with any trials that are run, run for the purpose of marketing materials and not for increasing scientific knowledge. I think we would be very unlikely to benefit from new treatments like gene therapy in that world, and instead would get a lot more expensive therapies that don’t do anything, like we already see with stem cells in less regulated markets.
Mark Z
Mar 27 2020 at 5:35pm
Dylan,
This all seems to assume that consumers, physicians, and hospitals have no interest or ability to discern in the efficacy of drugs, and the only reason we take drugs is because the FDA tells all of us they work. Absent gate-keepers, they’re may be more ineffective or even unsafe drugs that reach the market. But more drugs would be tried in general and at a more rapid pace. The gatekeeper imposes a level of risk aversion on society, certain assumptions about the risk of inefficacy or unsafe drugs relative to the benefits of more experimentation. I don’t see any reason to believe that the threshold they’ve happened to impose on us is the ‘right’ one.
Dylan
Mar 27 2020 at 8:17pm
This is right, I don’t think they have the ability to discern the efficacy of drugs. Unfortunately, “real world evidence” in the absence of controlled trials, is not very informative. There’s ways to make RWE work better and give you some information, and as Alan mentions in some post, the FDA has been moving in concert with industry to collect and analyze that data in a way that is scientifically meaningful…but absent that, you’re really just looking at anecdotal data on a large scale.
I agree that there is no reason to assume that the FDA has struck the right balance of safety vs. efficacy, I’d think that was more of a problem if there seemed to be a lot of drugs that worked, but had bad toxicology profiles that kept them off the market. But, at least for the companies I followed closely back in the day, that rarely seemed to be the issue. The much bigger problem is that the vast, vast majority of drugs aren’t very efficacious. Honestly, until you can figure out a way to make the hit rate better, which means a much better understanding of biology than we currently have, there’s not much I think you can do to increase the number of effective drugs on the market. We can vastly increase the number of ineffective drugs on the market, but that doesn’t seem like an improvement to me.
Dylan
Mar 28 2020 at 10:30am
Jon – Another question came to me. If the proposal is to be able to market drugs after a PhI safety trial, but before we get any evidence of efficacy. What would you say about the types of claims the company would be allowed to do for marketing purposes? Would this now come under the jurisdiction of the FTC and false advertising? Does the FTC now take over the role of the FDA in determining when we have enough evidence that something works? I could imagine a world where you were allowed to say for a new drug on the market that it was being investigated for such and such indication, but with a big black box warning saying something to the effect that 90% of drugs that have gotten to this stage have been found to be ineffective at treating their intended disease, and this also hasn’t been found to be safe to take in large trials, and so it might make you really sick and/or kill you.
David Henderson
Mar 27 2020 at 9:26am
You wrote:
What’s TCM? When I looked it up, it looked like Turner Classic Movies. But that’s probably not what you had in mind. What did you mean?
You write:
We discuss this in the op/ed. I want to point out, though, that if you mean everything you said above, you would have to ban a practice that a huge percent of doctors engage in and have engaged in for decades: prescribing for off-label use. An example: Doctors wondered if Prozac could be used to help women with pre-menstrual syndrome. They started prescribing it and in many cases it worked. Eli Lilly was not allowed to tell doctors that. It became so widely prescribed, though, that Eli Lilly decided it would make sense to bear the cost of showing efficiency and so a year or two later, the FDA approved it for PMS. With your model, you would have prevented doctors (before the FDA approved it for PMS) from prescribing it for PMS because “betting on untested remedies is not much better than closing our eyes and making a wish,” right?
Alan Goldhammer
Mar 27 2020 at 11:00am
“An example: Doctors wondered if Prozac could be used to help women with pre-menstrual syndrome. They started prescribing it and in many cases it worked. Eli Lilly was not allowed to tell doctors that. It became so widely prescribed, though, that Eli Lilly decided it would make sense to bear the cost of showing efficiency and so a year or two later, the FDA approved it for PMS. ”
That was then and this is now. FDA is more amenable to well done observational trials that can show new benefits for approved drugs. I spent three years working to get a project off the ground at PhRMA that improved the chances of using observational medical data for both safety and efficacy. We spent two years working on the business plan following inception in late 2005. I presented the plan to industry CEOs in 2008 and got $20M to get this project started. The group lives on at: https://ohdsi.org/ and a bunch of company and academic researchers have been spending all week in a virtual study group to see how this can be brought to the current pandemic.
Christophe Biocca
Mar 27 2020 at 11:01am
TCM: Traditional Chinese Medicine (which is where snake-oil originated).
Off-label use does make the current situation a bit ridiculous: if remdesivir had gone through FDA approval for ebola (which, if it was effective, would pretty much happen regardless of side-effects, considering the lethality of the disease), then there would be nothing stopping doctors from prescribing it for a completely different virus, without waiting for evidence of covid-specific efficacy.
Phil H
Mar 27 2020 at 11:24am
As Christophe says, TCM is traditional Chinese medicine, a venerable but non-scientific tradition.
Prozac for PMS? Lord give me strength. Perhaps Merck would like to sell me a pill for the disappointment of a bad movie? Perhaps Novartis would like to medicate me if I’m not doing quite as well in exams as I think I should… doh! That one actually happened.
Yes, I would infinitely prefer that Prozac not be pushed at doctors for another questionable syndrome, at least *until we’re sure it actually works*. Because one thing has been proven again and again and again and again and again in medical history: doctors’ observations are not good enough. The placebo effect, cohort effects, and reversion to the mean create incredibly powerful impressions, and doctors *cannot* distinguish between drugs that work and drugs that don’t on the basis of their observations.
Econlog has had a rash of posts lately about how important it is not to let ideology trump technical knowledge. This is an example of just that. Of course everyone wants a good treatment for covid. But if you don’t test them properly, we won’t get good treatments. We’ll get a rash of nonsense placebos.
nobody.really
Mar 27 2020 at 1:26pm
As Henderson might tell you, if you want to avoid bad movies, stick with TCM.
Mark Z
Mar 27 2020 at 6:00pm
Even from a standpoint that believes the FDA – up to this point – was perfectly efficient and optimized, when society is faced with a more contagious, deadlier disease than usual, it is only rational to lower regulatory standards, and let drugs be sold based on less data than normally, because the cost of doing nothing has gone up. Many people are arguing that this would cause more ineffective drugs to be released. Sure it would. The point, I think, is that it’s worth the cost. The FDA, we should all remember, does not guarantee that approved drugs are safe and effective. Rather, it sets a threshold that reduces the probability that unsafe or ineffective drugs are released; it also reduces the probability of safe effective drugs being released. There’s a tradeoff, and some implicit cost-benefit analysis is made. When the hazards of having no treatment at all go up – as in the case of severe coronavirus cases – the tradeoff is shifted such that it is only rational to be less risk averse toward new drugs. For the sake of maintaining broad appeal, I won’t posit that this specifically warrants eliminating the gate altogether, but it definitely warrants lowering the threshold. There’s nothing ideological about that, it’s just a matter of how new risks necessarily tilt the cost-benefit analysis.
An extreme analogy: if we faced an epidemic of a highly contagious disease that killed most of the people it infected within a day, I think even the most adamant FDA enthusiasts would admit that it would be rational to try pretty much everything with a remote chance of working that isn’t known to cause spontaneous combustion. The relative cost of unsafe or ineffective drugs drops radically compared with the benefit of effective treatments. The current situation should push us in that direction, not necessarily as far as the hypothetical usually fatal disease, but unless one believed that the approval process in the status quo ante was already too lenient, I think its indisputable that it moves the cost-benefit analysis in the direction of less risk aversion regarding drug safety.
Phil H
Mar 27 2020 at 9:27pm
Hi, Mark.
Thank you! You are giving exactly the kind of arguments I was suggesting. I don’t think your arguments come out correct, but I agree that that kind of thinking is exactly the right way to go about this.
“FD…threshold…reduces the probability of safe effective drugs being released.”
I don’t see much evidence of this. Is there some stock of effective drugs that aren’t available in the USA? There’s a couple; and there’s some evidence that the FDA delays releases, but it’s all pretty marginal.
“even the most adamant FDA enthusiasts would admit that it would be rational to try pretty much everything with a remote chance of working”
I think this misunderstands what FDA approval does. You can *try* any drug any time. (It just takes money.) FDA approval says that a drug maker may sell a drug to doctors with the explicit message that “this drug effectively treats condition X”. Off-label prescription is legal, remember (and quite common). I agree with you that it’s good to try lots of things. I don’t think it’s a very good idea to crowd the marketplace with a lot of explicit marketing, if we know in advance that 99% of that marketing will be incorrect (not malicious, just incorrect). And I would suggest that we do know that, because drug companies generally do try to make health claims for their drugs when permitted to do so, even when those health claims are not well-supported by the evidence.
Mark Z
Mar 28 2020 at 9:36am
“I don’t see much evidence of this. Is there some stock of effective drugs that aren’t available in the USA? There’s a couple; and there’s some evidence that the FDA delays releases, but it’s all pretty marginal.”
That’s not really my main argument here. The sample sizes for clinical trials are surprisingly small – often a few dozen patients – and unrepresentative of the relevant patient population due to the difficulty of recruiting. There is a non-negligible chance of just bad luck leading to a drug seeming significantly less safe or less effective than it., and consequently hitting a dead end. There’s an even higher chance that the drug’s effectiveness will be understated by the composition of the trial population, because drug efficacy often varies greatly with genetics as well as environmental factors. Getting a drug to market increases sample size, however, allowing one to better infer efficacy. In clinical trials, the reality is N is just too low to consistently infer efficacy with the degree of confidence that it seems most people attribute to the process.
“I think this misunderstands what FDA approval does. You can *try* any drug any time. (It just takes money.) FDA approval says that a drug maker may sell a drug to doctors with the explicit message that “this drug effectively treats condition X”.”
I imagine a drug would probably have to be marketed as treating the disease in question to be worth producing and used widely enough to get enough data to tell if it works.
Speaking generally: you seem to view the FDA as useful mainly as a means of cost control, and absent the restrictions, people would waste vast sums of money on useless medication. As far as over-the-counter drugs go, I don’t think this is a serious problem. People wasting a little money on useless tablets isn’t much of a concern for the state, IMO, and the cost isn’t that big. Where the costs are significant are with prescription drugs, especially treatment for chronic conditions (cancer probably most f all). The thing is, for these kinds of illnesses, drugs are heavily subsidized. Mostly the patient and doctor and pharmacist are wasting someone else’s money on a nearly useless drug. So with a stringent approval process, our systems follows a model of subsidizing demand on the one side and restricting supply on the other. It would make more sense to reduce both the restriction on supply and the subsidization of demand, so we have more of an incentive to care about costs.
You might argue, in the spirit of Robin Hanson, that even absent subsidization, social pressure will lead people to waste lots of money on ineffective treatment. That might be, but there is also plenty of evidence that subsidization increases healthcare consumption, so some of the need to restrict what healthcare people can spend money on is clearly a result of all the subsidies we give people to buy healthcare.
Dylan
Mar 28 2020 at 4:15pm
Mark Z, thanks for the post, that helps spell out some things that I think tend to be left unspoken, and makes it easier to follow the logic, which is always appreciated.
This seems certainly true, although my expectation is that it mostly goes the other way, in that drugs that aren’t really effective get on the market anyway, rather than we’re losing out on effective drugs. Look at the incentives, almost everyone in the process, from the companies developing therapies to the doctors, to the patient groups, are looking for any evidence they can find to suggest that some new drug works. There’s not a lot of incentives in the system to push back against that as is, and I’d imagine even less in a more open market.
I think that it is a good point that there are probably genetic markers differences we’re not yet aware of that make it so that a drug would work well in some people and not in others. And I can see an argument to be made that’s “hey, try this drug, it might work for you” might be better than not having access to the drug at all> But I’m not sure that leads to better overall outcomes.
I can’t speak for Phil, but that’s not my primary concern. My biggest issue is with continuing to generate the type of knowledge that gives some confidence that a given drug is effective, and also advancing the state of our scientific knowledge on human biology more generally, so that we continue to be able to invent better and better therapies. It’s not intuitively obvious that a less regulated approach has the right set of incentives to make that outcome likely, and I can see us instead settling into a suboptimal equilibrium that looks more like the supplements market than the prescription drug market.
That pessimism comes from years of watching the biopharma industry as mostly an outsider that still doesn’t understand a lot of things very well. However, one thing I learned early on when looking at clinical trials is, whenever possible you really want double-blinded controlled studies with a predefined endpoint(s). That last point is particularly important one that everyone knows, knows at least until they have a drug fail. Then, all of a sudden, those same experts start looking through the data until they find some subpopulation that looks like it shows a benefit, and convinces themselves that it is real, and then usually manages to find investors who are convinced it is real…but at the moment they can’t get it on the market without doing another trial that is setup to show whether that is a real outcome, or just an artifact of the data. Spoiler, turns out to almost always be the latter. What I wonder about is, in a world where drugs can get on the market without showing efficacy, if we somehow still got that first trial, would we ever get the second one? Or, would they just look at that data, and market to doctors that it looks like this works in this type of patient? Would we ever end up with rigorous data showing whether something works or doesn’t?
David Henderson
Mar 27 2020 at 6:05pm
Phil H.
You wrote:
Your lack of concern for those women is really quite breathtaking.
Phil H
Mar 27 2020 at 9:14pm
Nice deflect, and I probably am a really bad person. But all the ad hominem arguments in the world won’t change the point under discussion.
The discovery of effective drugs is hard; if you want to make a policy argument for something that the FDA should do/not do differently, you have to provide a *complete* policy argument. That means some level of assessment of what benefits and costs the current policy has; and some attempt to quantify the benefits *and costs* of the proposed new policy. Your post doesn’t examine the costs at all.
Obviously this is just a blog post, and no-one’s expecting a fully-argued paper. (For those, go and see Alex Tabarrok at Marginal Revolution.) But I will continue to post reminders that complete arguments are necessary, at least until Econlog gets fed up with me.
David Henderson
Mar 27 2020 at 9:23pm
I think you probably are a bad person.
I’m fed up with you. Others may not be.
David Henderson
Mar 27 2020 at 9:49pm
You wrote:
I don’t know you and so I don’t know whether you’re a bad person. You might be and you might not be. I’ve noticed on many internet discussions that I think less even of some of my friends, let alone strangers, as a result of those discussions. You’re a stranger to me. My guess is that you are like most of us: you have good and bad qualities.
But now to the issue. Our article was about why the FDA should drop efficacy requirements. You challenged us. I took your challenges to mean that the FDA should keep efficacy requirements. Then I gave a case where doctors found an off-label use for Prozac, dealing with women’s PMS, and then the use was so popular that Eli Lilly went to the FDA with efficacy studies that then allowed it to market Prozac for PMS.
In short the FDA decided that Prozac was efficacious for dealing with Prozac.
But that didn’t satisfy you. You said:
So there are two ways to interpret that. One is that you have no sympathy for those women. The other is that you think even proof of efficacy is not enough. I took the first interpretation. Was I wrong and what you’re really arguing is that efficacy requirements are not enough? But if the latter, what are you advocating? Are you advocating that the FDA approve only drugs that you, Phil H. approve?
As to whether EconLog is fed up with you, I don’t know. I know only my own thoughts. I’m close but I’m not quite fed up with you yet. Part of it will depend on how you respond.
Phil H
Mar 28 2020 at 1:19am
Nice to have admin privileges so that you can change hasty replies, huh? All that’s fine, just chill on the ad homs.
I am not a big fan of the medicalisation of everyday life. Men and women have survived the bane of PMS for thousands of years. It is far from clear to me that it is a “disease” that requires medical intervention. I do know that the vendors of Prozac are keen to make more money off it by “finding” (or inventing) new indications. Hence my dismissive snort at the notion of this particular new development. And of course, this is not my own thought; it is taken from people that I read who know much more about this situation than I do. Ben Goldacre’s Bad Pharma is maybe the locus classicus.
But you’re right that this is not really relevant to the correctness or otherwise of the FDA’s regulatory stance. Here is the kind of calculation we’d need to do to work that out.
First: Costs and benefits of the current stance. Costs: the patients who went untreated by an effective treatment in between the discovery of that treatment and the FDA’s approval. Benefits: the ineffective treatments prevented by the FDA’s stance. To actually do these calculations is far beyond me, but here’s a possible approach: we could operationalise the cost by looking at the difference between average annual off-label prescriptions of Prozac for PMS before approval and average annual prescriptions for the same indication after approval, and multiplying by the number of years. (Could convert $ using the price of Prozac.) We could operationalize the benefits by looking at other indications for Prozac requested by Lilley but turned down by the FDA (dollarized by estimated size of market?).
Second: Costs and benefits of the suggested stance. Benefits: faster discovery of drugs, operationalized as value of those drugs during what would have been the approval delay period. Costs: Use of ineffective drugs over the same period.
My contention, which I think is supported by what Dylan was saying above, is that the costs of the suggested stance would be very large, because drug discovery is very difficult. Therefore that factor tends to dominate, and in general, tight regulation is good.
I suppose there’s an argument for speed in the current crisis, but you’d have to be very careful in how you make it. Because rolling out ineffective treatments is a major cost.
Mark Z
Mar 28 2020 at 9:01am
“I am not a big fan of the medicalisation of everyday life. Men and women have survived the bane of PMS for thousands of years. ”
We’ve survived a lot of things that are very painful and debilitating that we can now deal with chemically, which I think is great. Feel free to forbear, but let the rest of us medicate away.
Regarding drug discovery being difficult: isn’t this why it’s useful to repurpose existing drugs? Given the neuropsychological etiology of the menstrual cycle-related mood swings, I don’t see how it’s implausible that an anti-depressant would be useful in that situation as well. Same neurotransmitters are relevant in both cases, no?
Charley Hooper
Mar 29 2020 at 1:17pm
Your suggestion to do a cost/benefit analysis is a good idea. University of Chicago economist Sam Peltzman did a similar analysis and came to the following conclusion:
Were some ineffective drugs kept off the market by the 1962 Amedments? Yes. Were effective drugs also kept off the market? Yes. There were simply fewer new drugs overall with no evidence that the proportion of good drugs to bad drugs increased. Peltzman said:
Peltzman said it was as if,
In conclusion, he said that consumers typically gain more from the benefits of new drugs than they suffer from the costs of ineffective drugs, which
Meaning the costs of the 1962 Amendments exceeded the benefits.
Comments are closed.