
As far as I know, neither political party has ever advocated “death panels”, that is, panels of experts that would deny Medicare coverage to FDA-approved treatments that don’t pass a cost/benefit test. But maybe they should.
Biogen’s new drug to treat Alzheimer’s was recently evaluated by the FDA. Here is Rachel Sachs in Health Affairs:
In March 2019, both trials were stopped halfway through, when “a planned interim analysis met prespecified futility criteria.” But Biogen went further into the data, analyzing the results and arguing that the high-dose patient population in the second trial had in fact experienced a statistically significant clinical benefit when measured against the placebo (a 22 percent reduction of clinical decline as measured on a chosen clinical scale), albeit a benefit absent from the high-dose population in the other trial or the low-dose population in either trial. On this basis, Biogen submitted an application to the agency for approval in July 2020.
In November 2020, the FDA convened a meeting of the pertinent advisory committee to review the evidence surrounding aducanumab. The advisory committee was highly skeptical of the post-hoc rationale used to provide evidence of the drug’s effectiveness, particularly when viewed against the backdrop of decades of failures of drugs which similarly targeted amyloid plaques.
I am no expert here, but in general I am extremely suspicious of claims of “statistical significance” that rely on post-hoc rationalizations. The committee was also unimpressed:
The FDA’s own advisory committee voted overwhelmingly that the drug’s clinical trials did not demonstrate evidence of effectiveness (with ten members opposed, none in favor, and one abstention).
Nonetheless, the FDA recently approved the new drug, which will have a $56,000 list price. Given that more than 6 million Americans suffer from Alzheimer’s (a number that is rising rapidly), this decision will have major consequences for health care costs:
The FDA’s approval of Aduhelm is likely to have health policy ramifications far beyond the FDA itself, though. As noted above, the potential financial implications for our health care system are staggering. Medicare Part B spends approximately $37 billion annually on prescription drugs program-wide, and Part D spends approximately $90 billion annually. Given the large patient population and high price tag, it is not difficult to imagine aducanumab alone commanding somewhere between those totals every year—for a drug with extremely limited evidence of clinical efficacy.
I believe that all American should be free to use Aduhelm (aka aducanumab) as long as taxpayers don’t have to pay for it. The more difficult question is whether it makes sense for taxpayers to pay vast sums of money to subsidize drugs that government experts view as having little or no benefit (relative to cost.) I’m skeptical of that idea, although I’m willing to listen to arguments for the other side.
PS. Three members of the FDA advisory panel have now resigned in protest:
A third member of a key Food and Drug Administration advisory panel has resigned over the agency’s controversial decision to approve Biogen’s new Alzheimer’s drug, Aduhelm, CNBC has learned.
Dr. Aaron Kesselheim, a professor of medicine at Harvard Medical School, said the agency’s decision on Biogen “was probably the worst drug approval decision in recent U.S. history,” according to his resignation letter obtained by CNBC.
PPS. I recently appeared on the Bob Murphy Show.
READER COMMENTS
Jon Murphy
Jun 12 2021 at 6:41pm
I think I may be missing something here because I don’t quite understand your point. Is Medicare obligated to pay for any FDA-approved treatment?
Matthias
Jun 12 2021 at 8:08pm
Well, those death panels would be exactly those people making the choices about which drugs to pay for.
Scott Sumner
Jun 12 2021 at 10:10pm
I recall that a few years back Congress set up a committee to make cost saving suggestions for Medicare, but that more recently Congress dismantled the committee. Congress seems very leery of anything that might be labelled a death panel, but perhaps someone else can provide more detail.
Certainly the author of the article I linked to assumed that Medicare would pay for the treatment.
Jon Murphy
Jun 12 2021 at 10:20pm
Ah ok. I understand now
Mark Brady
Jun 12 2021 at 7:44pm
How about limiting or removing patent protection for the drug?
Scott Sumner
Jun 12 2021 at 10:15pm
I’d prefer we do something like the FDA saying, “Approved, but not eligible for public subsidy.”
Floccina
Jun 12 2021 at 8:07pm
I agree 100%. It would be so great to have an Alzheimer’s drug that works but sadly this is not it. Medicare needs to have those “death Panel” which should really be called, refuse to pay for stuff that does work panels.
Floccina
Jun 12 2021 at 8:11pm
rather it should have read: “death Panel” which should really be called, refuse to pay for stuff that doesn’t work panels.
Scott Sumner
Jun 12 2021 at 10:12pm
I agree, which is why I used scare quotes for “death panels”. They would be cost/benefit evaluation panels.
BC
Jun 13 2021 at 11:38am
Health care inherently involves life and death decisions. In private health care, those decisions are made by individuals, in government health care by some government panel. To debate whether a particular panel is or is not properly labeled the Death Panel is a distraction. Someway somewhere, Medicare must implement a Death Panel.
Garrett
Jun 12 2021 at 9:37pm
In Bob Murphy’s article on Market Monetarism, he writes:
Given your recent posts about the COVID recession, would you agree that he is mischaracterizing your view?
Scott Sumner
Jun 12 2021 at 10:14pm
No, I think that’s a correct characterization of my views for 99% of recessions, albeit not for the Covid recession. But I did not anticipate something like the Covid recession when I formulated the policy proposal.
BC
Jun 13 2021 at 3:03am
“all American should be free to use Aduhelm (aka aducanumab) as long as taxpayers don’t have to pay for it.”
Important point to emphasize. Otherwise, we could end up with a Charlie Gard or Alfie Evans situation. So, there needs to be some sort of Right to Try, even if not approved for government subsidy. A panel that denies Right to Try would be aptly labeled a Death Panel, a panel deciding on behalf of others that it was time for them to give up. Of course, a Right to Try would also mean accepting that the poor would not have “equal access” to the same drugs as the rich, which some people seem loathe to do. (“Equal access” is in scare quotes because it’s defined to mean something unequal: that some people should be eligible to spend Other People’s Money on things that others have to spend their own money on.) So, demand for “Equal Access” seems to lead directly to Death Panels, or at least Rationing Panels.
The question of what Medicare covers is more complicated. Of course, taxpayers should not have to subsidize questionable treatments. On the other hand, Medicare is mandatory. So, people that wanted to pay for experimental or questionable treatments, either directly or through high-premium plans that covered such treatments, might not be able to because too much of their money was spent on Medicare instead, perhaps covering other treatments that they didn’t even want. With mandatory government monopoly, it’s hard to distinguish sometimes between Other People’s Money and one’s own money. Maybe the King Solomon solution is to allow a government panel to determine cost-benefit but allow individuals to opt out of Medicare.
Demands for “Equal Access” and/or mandatory participation in government programs lead directly to either Death Panels or the alternative of runaway spending. Europe tends to get the former (Charlie Gard and Alfie Evans). The US tends to get the latter (“Congress seems very leery of anything that might be labelled a death panel”).
Phil H
Jun 13 2021 at 4:48am
The argument for deregulating medicine, as I understand it, goes something like this:
In place of the current systems of regulation, new mechanisms would appear to help patients and doctors understand the relative utility of various drugs and treatments. Essentially you’d have one or more private FDA issuing their seals of approval to good drugs and turning them down for bad drugs, and both doctors and patients would come to trust the private certifiers, even if they can’t trust the claims of the drug manufacturers.
But it seems fairly plain that private certifiers would be under exactly the same kinds of pressures as the FDA apparently succumbed to here.
So on a level of efficiency (understood as: delivering effective drugs), I can’t see why I should believe deregulated medicine will work better than the current system. The FDA is imperfect, but what’s the rationale for thinking a privatised system would be less imperfect?
(This is not directly addressed to Scott, who didn’t make this argument in his post. It’s just a general comment/question!)
john hare
Jun 13 2021 at 6:54am
Because it wouldn’t be illegal for people to try cures and procedures that haven’t run the gauntlet of FDA approval. Obviously there would be some bad outcomes, and more obviously, there would be many good outcomes. The logical and numerical balance favors it. Emotionally there are many that don’t want to be responsible for making decisions that may have bad outcomes regardless of the balance. And by default they don’t want others to have that freedom either.
dylan
Jun 13 2021 at 8:36am
Under the current system, ~9 out of 10 experimental drugs aren’t found to be effective or are considered too unsafe to use. Under a system where you don’t have to prove efficacy, what makes you think that ratio would improve? I tend to think that it would be much worse.
Jon Murphy
Jun 13 2021 at 9:14am
I assume you’re being hyperbolic here. In order for the ratio to be worse, it’d have to be every drug would have to be ineffective or unsafe. It would be an odd decision for a firm to produce all useless goods. Another firm could come along and produce something only marginally effective and safe and steal 100% of their business.
So no, the ratio could not get worse.
(By the way, the current system has an approval rate of around 90%, not a rejection rate).
Which leads us to the next question: even if only 10% of drugs are safe and effective, does it make sense to sit on them? Take, for example, the long process by which the FDA approved emergency use for the COVID-19 vaccines. It took them approximately 10 months. Given how insanely effective the vaccines are, that process implies hundreds of thousands of excess deaths. When we take into account the pause ordered by the FDA and subsequent reduction in public confidence of the vaccines, we’re talking probable another few thousand deaths. And that’s just in the US alone! So, does it make sense to delay for the sake of test review?
dylan
Jun 13 2021 at 11:26am
Jon,
I believe you misunderstood my statement. The article you linked to is the number of drugs submitted to the FDA that are approved or rejected. The number I quoted is the number of drugs that enter human trials that eventually make it to approval (the number is a little better now, recent studies estimate closer to 14%) Note, if you start with preclinical testing, the number is far worse, only 1 in 1000 drugs that are investigated ever make it to human trials.
And yes, those numbers could be worse. I’ll quote again from my favorite writer on the subject.
As for the sitting on safe and effective drugs, of course it doesn’t make sense to sit on those. It makes sense to sit on the others though. If we had a cheap and quick way to tell the difference, then we should absolutely do that. Without that information though, how would you tell which of the hundred vaccine candidates to take? They are not all going to be effective and some will be unsafe. As you are no doubt aware, two decades of research has gone into mRNA vaccines, and until this coronovirus, they had pretty much come up empty. Vaccines have one of the highest success rates in the clinic, but that still means only 1 in 3 that start human trials will be effective.
dylan
Jun 13 2021 at 11:37am
Have a full response with links, but seems to have got caught in moderation for the moment. A quick response to say that I think you misunderstood me, the article you linked to is the percentage of drugs that are submitted to the FDA for approval (after running Ph III trials) that are approved by the agency. My reference was to the number of drugs that start human trials that are eventually approved. Over the last few years this number has gone up and is now 13.8% according to one study, 1 in 1o is just easier shorthand.
Hopefully my earlier response will not stay in moderation for too long.
Jon Murphy
Jun 13 2021 at 2:50pm
Ah yes, I did misunderstand.
I’ll wait until your response is released (I think anything with 2+ links is automatically flagged)
Phil H
Jun 13 2021 at 8:38am
Thanks, John. My feeling is that your response doesn’t answer my question. There is a freedom argument against regulation, but I was specifically asking about efficiency. I agree that more people would try more treatments, but given the number of people who try to use Chinese medicine, homeopathy, and faith healing, I think this will simply give us a tidal wave of more bad treatments, with very few new discoveries. That means that the net result will be a drop in the average efficacy of all treatments; in fact, a large drop.
BC
Jun 13 2021 at 11:25am
“The FDA is imperfect, but what’s the rationale for thinking a privatised system would be less imperfect?”
Competition and selection. Some private certifiers might be too conservative in endorsing drugs, others too easy. Professional doctors can learn such reputations, look at multiple certifiers’ recommendations and read the underlying rationales, and make an informed decision with their patients. It’s like asking what’s the rationale for thinking that an open system of competitive blogs would be less imperfect than one government-run blog.
dylan
Jun 13 2021 at 11:32am
I think that would depend a lot on who is paying the certifiers. How well did “independent” credit rating agencies do leading up to the great recession?
Jon Murphy
Jun 13 2021 at 2:51pm
The credit agencies are actually evidence toward BC’s point. That industry is hugely regulated and monopsonistic. We had major big player problems.
dylan
Jun 13 2021 at 3:39pm
I might need the summary of big player argument. 100EUR is a bit out of my price range for an internet debate.
My suggestion was only meant to be that private organizations that serve quasi-regulatory purposes are under the same pressures to get captured as their public counterparts.
But, I’m also pretty sure that “hugely regulated and monopsonistic” applies at least as much to healthcare as it does finance. A clean slate approach might allow for better outcomes overall, and that’s worth discussing. However, I’ve been assuming proposals like abolishing the efficacy requirement from the FDA are made ceteris paribus and try to evaluate them as such.
Jon Murphy
Jun 13 2021 at 5:36pm
That’s essentially the “Big Player” argument in a nutshell. When there is one major buyer (or seller) of information (ie the Big Player), the oversight agencies (such as the credit agencies) tend to go along with whatever the big player wants to hear. With credit agencies, they did whatever the SEC wanted to hear. With health care, it’s whatever the FDA wants to hear. When there is competition, even if that competition is low quality, it increases the quality of information in the marketplace!
Dylan
Jun 14 2021 at 9:47am
Interesting. I’d like to see this fleshed out more, I believe the standard explanation is that the credit agencies bowed to what their customers, the banks, wanted to hear not so much the SEC, which seems fairly plausible to me. The contrarian in me is always open to hearing other explanations though.
Scott Sumner
Jun 13 2021 at 12:05pm
You asked:
“The FDA is imperfect, but what’s the rationale for thinking a privatised system would be less imperfect?”
In general, competition and choice produces more innovation and better outcomes than government monopoly.
robc
Jun 14 2021 at 9:04am
This isnt really a reply to Scott, but couldnt figure where else to put it. I bring this up on a number of related issues, but no one ever tells me why the Kosher-certification model wouldnt work. I assume that is because it absolutely would work.
Organic, gluten-free, restaurant inspection, drugs, occupational licensing…all dont need gov reg, just apply the kosher model. Those that care can distinguish between the multitude of certifying orgs.
Scott Sumner
Jun 14 2021 at 12:30pm
I agree.
Phil H
Jun 17 2021 at 1:52am
No, this is not a good example at all. For this reason: you have no idea if the kosher system works. Nobody has any idea if the kosher system works, because when the kosher system goes wrong, there are zero consequences. The kosher system is a great example of how markets definitely can work: they deliver what people want very well. And kosher is just a random bunch of rules that people apply if and when they want it. The fact that people want it *defines whether kosher works or not*.
Medicine is nothing like this. Medicine either works or it doesn’t, and there are objective criteria that determine that. The question for a market is: will the market that delivers what people like eventually line up with what makes them well? It’s possible that it will, but the empirics to date don’t look good.
robc
Jun 17 2021 at 10:36am
The goals for medicine is no more objective than for kosher.
In general, the former is health of the body and the latter is health of the soul ( I presume). But really, everyones goals are different so the kosher system makes the most sense.
Thomas Lee Hutcheson
Jun 13 2021 at 6:59am
There are several points at which decisions could be improved. FDA could apply more rigorous (or at least some kind of) cost benefit analysis, Medicare could apply a cost benefit analysis of what it will pay for (which could lead to some bargaining over price) or doctors could prescribe with some notion of cost-benefit. My guess is that Medicare is the best point.
Alan Goldhammer
Jun 13 2021 at 7:09pm
In all my years working in pharma regulatory afffairs, this is perhaps the worst approval I’ve seen. FDA has required a post-market study to show efficacy but Biogen has NINE years to do this trial and in the meantime they can count the money coming in. What would be interesting is, if at the end of the day the drug did not work, could the company be sued for fraud. Lilly must be apoplectic as they were the pioneers of research into amyloid plaque and Alzheimer’s. They never found a drug that worked despite spending a huge amount of money doing so. They should have taken the route Biogen did and would have had a drug 12 years ago.
the Medicare decision is complicated as this drug is infused and not covered under Part D of medicare but rather Part B (physician administered drug). Medicare routinely reimburses for drugs and I’m not aware of a case where they have not. Anyway, HERE is a good primer on how Part B works.
Ken P
Jun 13 2021 at 8:54pm
You correctly state that the price is a “list price”. But analysts expect the actual drug cost to between $10k and 20k. This is an important difference in price. Insurance companies/Medicare don’t pay list price. Also, I believe the treatment consists of 12 injections over time which is important. If it actually works, the drug would be a big cost savings because Alzheimers patients often require in-home healthcare or nursing home care as well as other costs.
IMO, cost/payment is a separate question. Lots of people die waiting for a drug that is already developed to be approved. It sounds like this drug passes the do no harm test and approval might draw in some competition that with more efficacious products. It’s not the rapid improvement in progress that an improved regulatory approach would bring but it beats what we have.
Ken P
Jun 14 2021 at 9:24am
*I meant cost/payment is a separate question from approval.
Also – Thanks for heads up on interview with Bob. I will check that out.
dylan
Jun 14 2021 at 9:40am
Ken,
Your first paragraph I absolutely agree with, both that the list price isn’t what will actually be paid, and that an effective Alzheimer’s drug could charge a pretty high price and still end up saving the system a lot of money in the longer run.
I don’t think I understand your 2nd paragraph though, particularly this:
In what way will this lead to more competition with more efficacious products? The problem with Alzheimer’s is we have nothing that works (including most likely this drug). We don’t even have great ideas on what to try. The two leading contenders, the amyloid and tau hypothesis have so far been shot down in the clinic fairly convincingly, although there are still people trying both approaches. Other MOAs are still in pretty early and speculative development. This isn’t for lack of trying! Billions have been poured into R&D in the space (Lilly alone has probably spent that much). Everyone knows that if you can find something that works just a teeny bit, you’ll make billions (and Biogen is out to prove that you don’t even need a drug that works a teeny bit). This approval seems like a good way to encourage others to try to bring to market other drugs that they can’t prove work (but where if you squint real hard you can maybe see a way in which they could work). I don’t see how it promotes competition for new, effective drugs.
Ken P
Jun 14 2021 at 11:41pm
Dylan,
Yes, the amyloid plaque hypothesis is largely shot down but mostly due to lack of efficacy of the treatments that target it. This study seems to give a (perhaps last) breath to the amyloid approach.
It sounds like you are more knowledgeable on this space than I am. Have previous treatments successfully removed/slowed plaque or did they only have a mechanism of action that would be expected to reduce impact of plaque?
My point is really that we need to be getting more real world data/feedback from the end user and we need to iterate. It might be that it only works on a subset of patients with something in common. That could lead to better understanding of the therapy. Perhaps a different epitope on the amyloid protein should be targeted or a different binding affinity may make more sense. The current approach is like developing a car with only one or two designs (selected by execs who haven’t held a ratchet in their hands in decades) that actually make it to trial that utilizes a 50 ft long road. Then that car either makes it or doesn’t on a 1-mile track in a trial viewed by some regulator.
Dylan
Jun 15 2021 at 7:56am
Yes, previous drugs have reduced amyloid plaques in the brain. All of the drugs target a slightly different part of Aβ, and there are some theoretical reasons for thinking the part that Biogen targets could be more effective…but that’s not worth a whole lot without data of clinical efficacy.
The problem is, “real world evidence” isn’t worth all that much. Look at aducanumab, two Ph III trials that should have been identical, with showing pretty different results. If it is so difficult to figure out if something works under highly controlled circumstances, think how difficult it is when you have millions of patients getting it under wildly different protocols. Some people seem to think that with widespread use in the real world, it will just become obvious which drugs work and which ones don’t, unfortunately that’s not the case.
Ken P
Jun 15 2021 at 8:36pm
To me, paying and approving are separate things. Like David Henderson says, the FDA shouldn’t keep safe drugs off the market. I know you disagree with that. But I’m coming from the alternate pole as you. Mine is: “Our bodies. Our choice.
What we need is a randomized control trial on the effects of regulation itself. A location where Our bodies, our choice prevails. That alternate example is hopefully on the way somewhere in the world. Otherwise we will continue to live in this world where tens if not hundreds of thousands of people die every year unnecessarily for extra bureaucracy. A world where Covid tests are not allowed because people might make decisions on imperfect tests. A world where Fitbit and Apple leave extra sensors out of their watches to avoid regulatory scrutiny, and algorithms that could have detected Covid from watch feedback. Fortunately computers are in a different world where regulators don’t decide how much RAM is too much RAM or what is spreadsheet software is acceptable for you to use.
Dylan
Jun 16 2021 at 12:55pm
We’re not necessarily in disagreement. In principle, I also think people should be free to take whatever they want. If that’s the end of your argument, I can respect the principled stance despite whatever the outcomes of that policy change might be.
What gets me about the piece (and others like it) that you linked to, is the lack of acknowledgement of the difficulty of proving efficacy. Charley Hooper, in a recent exchange with me in the comments, talked about a drug working if we give it to a person and their tumor shrinks, as if showing causality was simple. He’s obviously an expert that has spent longer in this world than I have, so I know that he isn’t that naive. But, when explaining to people who don’t know the drug world very well I think it is easy to get a different impression.
As to your second point, there’s actually a lot of places on earth that are considered “non-regulated” drug markets where you have the freedom to try unproven drugs. This has led to things like “stem-cell tourism” clinics popping up in those countries offering unproven (and generally considered ineffective) stem cell treatments for a whole host of diseases.
Michael Rulle
Jun 15 2021 at 8:35am
That is all we need, another issue for two sides to hate each other about. I grant there are likely situations where refusal to subsidize will appear rational—-even moral. However, a death panel, which will have another name —-as it should—-will be a bureaucratic entity. It will be given “standards” of behavior. But it’s real purpose will be to not waste money. And like all bureaucratic entities, it will expand its mission and gradually become unaccountable. Like the EPA, for example.
Maybe we can put Ezekiel Emmanuel in charge—-assuming he is not yet 75.
Colin
Jun 18 2021 at 12:56pm
The major problem with removing FDA efficacy requirements and moving to a safety only, is that those two parameters should be measured together.
Chemo treatments are not “safe”, but are usually better than the alternative.
Comments are closed.