I forgot to post this article 30 days after it appeared in the Wall Street Journal. Thanks to Brian Schwartz for reminding me.
Congress’s mandate that medications be proved effective is unnecessary and delays potential cures.
By Charles L. Hooper and David R. Henderson
March 25, 2020 6:28 pm ET
The federal government requires pharmaceutical companies to prove that their drugs are both safe and effective before putting them on the market. Before 1962, companies needed to prove only safety. While there is some appeal to this two-hurdle approach, evidence suggests that there is only a slight benefit and a tremendous cost. With the Covid-19 pandemic sweeping the world, there has never been a better time to revoke the Food and Drug Administration’s efficacy requirement.
The thinking behind the FDA’s two-hurdle approach is that ineffective drugs cost money and impose an opportunity cost: While the ineffective drug is being used, potentially effective ones are not. But researchers who have studied the 1962 law, known as the Kefauver-Harris Amendments, have concluded that before it was enacted, ineffective drugs were a small percentage of the market and therefore not much of a problem.
Further, the Kefauver-Harris Amendments dramatically increased the time and cost of getting new drugs approved. Evidence provided by University of Chicago economist Sam Peltzman suggests that the number of new drugs approved dropped by 60% in the decade following the law change. We have fewer ineffective drugs, but also far fewer effective ones than we could have had.
That’s not all that’s wrong with the Kefauver-Harris Amendments.
Before 1962, patients and doctors needed to try drugs to see if they worked for a particular patient at a particular time. It’s called trial and error and, despite the change in law, is still the standard medical approach.
To understand how little the FDA’s two-hurdle approach contributes, consider Merck’s Keytruda (pembrolizumab), which has recently become one of the hottest drugs on the market. Keytruda got some good press after Jimmy Carter, who was suffering from melanoma, was diagnosed as cancer-free after taking it. Mr. Carter was lucky because in one clinical trial Keytruda destroyed or reduced the tumors in only 34% of patients. Keytruda—which was the fourth biggest-selling drug globally in 2018 and brought in worldwide revenues of $11.1 billion last year—is far from a sure thing.
Does the FDA, which approved Keytruda as safe and effective for melanoma, know which patients will benefit from it? No. The potential benefits described on the drug’s label are what happened to strangers in another time and place. Patients still have to try Keytruda to know if it will help them. An FDA approval for efficacy is largely duplicative.
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.
And then, even if remdesivir is approved by the FDA, patients and doctors will still resort to trial and error, because the success rate won’t be 100%.
Instead, the FDA should allow patients and doctors to try remdesivir today. We know it’s safe, and there are no other proven therapies. With more widespread usage, the medical community will have a tool to fight back against Covid-19. Controlled clinical trials can continue largely as they normally would. Though access to an experimental therapy won’t be an incentive to join a clinical trial, many others remain—such as excellent free care and a sense of pride in helping society and other patients. This two-pronged approach would not only potentially save thousands of lives but also provide much more information.
If you were diagnosed with Covid-19 today, would you choose to take nothing? Or would you want a safe drug that has shown some potential against the virus? We suspect the great majority of Covid-19 patients would choose the latter option. And they would get it today, not months from now, when it would be of no use to them.
This same process plays out every day with other conditions that can be more dangerous than Covid-19. While patients and doctors continue to resort to trial and error out of necessity, the government insists that they initiate their experiments only after the FDA has completed its. This is a critically flawed approach to medicine.
Even as the pandemic robs Americans of their jobs, health and lives, it could reward us with a silver lining if we use the opportunity to take a hard look at our drug-approval process. Revoke the FDA’s drug-efficacy requirement.
Mr. Hooper is president of Objective Insights Inc., a company that consults for life-science companies. Mr. Henderson, a research fellow with the Hoover Institution, was senior health economist with President Reagan’s Council of Economic Advisers.
READER COMMENTS
Thomas Hutcheson
May 25 2020 at 10:56am
How much difference would it make vis a vis just requiring providers to give prescribers the information on effectiveness. I think the problem is that it is really difficult to know whether a new drug is cost effective, whether that information is generated up front or over time with use.
Alan Goldhammer
May 25 2020 at 7:31pm
Why not bring laetrile back on the market. It was also a safe drug.
Garrett
May 26 2020 at 6:38pm
Was it? Quick google search says it may cause cyanide poisoning
Grant Gould
May 26 2020 at 7:22am
I’m a bit confused by this analysis. How can you determine whether a drug is safe or not without reference to the conditions it cures? After all, chemotherapy drugs eg are extraordinarily unsafe by any effectiveness-independent measure, it’s only that they are effective against cancer that lets us call their nastier effects “safe”.
Safety presumably is a relative measure of the drug’s side effects against its benefits, and so if you don’t know it’s effective as to its alleged benefits you can hardly define what it means for a drug to be safe.
If you’re going to drop the effectiveness requirement, why not just drop the safety requirement as well and be done with it? Require testing and disclosure and let doctors and patients do the tradeoffs past that. It would be far simpler than trying to make some objective standard of “safety” without regard to whether your counterfactual is headaches or hepatitis.
David Henderson
May 26 2020 at 8:51am
You ask:
That’s a good question. Interestingly, though, the FDA did just that for 24 years, between 1938 and 1962. So somehow they figured it out.
You write:
I agree.
Tom DeMeo
May 26 2020 at 12:40pm
“How can you determine whether a drug is safe or not without reference to the conditions it cures?
That’s a good question. Interestingly, though, the FDA did just that for 24 years, between 1938 and 1962. So somehow they figured it out.”
We often read posts by Prof Henderson where he cites a rhetorical flaw in an argument. “So somehow they figured it out” is such a flaw.
The position cited here reduces the role of the FDA. With that comes a different liability profile for selling such drugs, and a shift in burden in the role of doctors.
What if the FDA made no finding on Remdesivir? It is entirely possible that no patients would get to take it because the change in liabilities would crush any viable business model for producing drugs with liability risks.
Charley Hooper
May 26 2020 at 1:59pm
While liability is an important issue, the FDA doesn’t offer much protection today, even with approvals based on safety and efficacy.
Consider Merck’s Vioxx, which was approved by the FDA as a safe and effective drug. When a subsequent clinical trial questioned Vioxx’s safety, Merck was nearly buried in lawsuits.
David Henderson
May 26 2020 at 4:40pm
You write:
I think that’s right, although I would be interested in what my co-author thinks.
You write:
It’s hard to believe that liability “would crush any viable business modeler producing drugs with liability risks.” There are literally thousands of products sold, and probably millions, for which there is no government agency insisting on a showing of efficacy. Yet those products get produced.
But consider the worst case: drug companies need FDA approval to ward off liability suits. As Charley points out above, they might get sued even if there is FDA approval. But if drug companies needed FDA approval of efficacy in order to be willing to produce, then they would seek it. Our proposal would have zero effect. I would predict, though, that companies would produce some drugs without going to the FDA to seek approval for efficacy.
And my prediction is based on current practice. Look at the large number of drugs that doctors prescribe for ailments where the FDA has not approved their efficacy in those uses. Many cancer drugs, for example, are approved for cancer A and doctors get creative and try it on cancer B, an off-label use, and voila, it works on some patients.
Charley Hooper
May 26 2020 at 4:53pm
Regarding the liability question, that’s not my area of expertise. However, my understanding is that doctors and drug companies are regularly sued even with whatever cover an FDA approval offers.
Dylan
May 26 2020 at 8:53am
Thanks David for posting this. Nice to see some discussion of how you expect us to find out which drugs work, although personally I want a lot more detail than just trial and error.
I think it is interesting that you suggest controlled trials can continue on largely as they normally would. What’s the incentive for a drug company to run clinical trials? Would you favor not having Medicare/Medicaid pay for drugs until they’ve undergone clinical testing showing efficacy?
I also think it is funny that you suggest in one sentence that you don’t think the incentive to join a clinical trial would be reduced much, and then in the next paragraph you ask who would choose to take nothing vs. an experimental drug that has shown some potential against Covid-19? In a clinical trial, you might get the experimental drug, but you’ve got just as good of a chance of getting a placebo. If your choices are 50/50 chance in the clinical trial of getting the drug, or 100% by not being in the trial. Which do you think most people would choose?
It’s also lucky that you chose to highlight remdesivir, which has shown some efficacy in clinical trials, vs. hydroxychloroquine, which had a lot of anecdotal support, but whose benefit has been completely evaporated when looking at the evidence, and instead caused active harm.
Charley Hooper
May 26 2020 at 2:22pm
Even with clinical trials, we still need to resort to trial and error. That’s what we rely on today. If you want something better than trial and error, we’ll need to come up with a better system.
Clinical trials provide information. If the expected value of that information is greater than the expected cost of the clinical trial, and the company has no more enticing prospects (not all investments will be funded), then the drug company will initiate that trial.
The value of that information could be large. Imagine launching a drug for bladder cancer with the message that it might have some efficacy versus saying that only 20% of 150 patients in a clinical trial recurred within two years versus 60% on standard therapies. That information has tremendous value for patients and physicians.
If you treat people like lab rats—raw materials with no rights—then, yes, you may have a hard time getting enough patients for a COVID-19 trial. If you treat them as humans—ends in themselves—and make it worth their effort, then you shouldn’t have any problem.
Dylan
May 26 2020 at 4:39pm
Yes, but it is trial and error that is backed up by scientific evidence. It’s true that even after running a clinical trial, we don’t know exactly who a drug will benefit. Although we’re often at least part of the way there through biomarker analysis, like this recent readout of data that hopefully leads to us being able to prospectively identify the patients that will respond best to a specific immuno-oncology treatment. But clinical trials for efficacy at least give some guideposts, if a drug passes that, you at least have decent confidence that a drug in the average patient will do more harm than good. For drugs that haven’t proven efficacy, you’re much safer assuming the opposite.
You seemed to conveniently leave out the most important part of this, those patients and physicians are not the people paying for the drugs. I’m guessing that you would prefer that this was not the case, that patients were more responsible for shouldering at least a portion of the cost themselves, as opposed to having zero idea of what a drug even costs, if so, I’d be in agreement with you. But, that’s not the world we live in, the world we have is one where Medicare picks up the tab for roughly a third of all drug spending, and they have no ability to negotiate on pricing. Given that, what do you expect to happen when drugs are on the market with no evidence of efficacy? Will Medicare be required to buy these? Or, will they be allowed to require some form of clinical trials to show efficacy before they pay for anything? If so, do you think that private insurers might want to piggyback on that effort so that they don’t have to run their own clinical trial apparatus? Would that world look very different from what we have now?
Charley Hooper
May 26 2020 at 5:16pm
Maybe someday we’ll be able to tell which patients will benefit beforehand, based on biomarkers or something else, but we aren’t there yet. When we get there we can discuss the implications.
Regarding clinical trials as guideposts, I’m agreeing with you. I’m not anti-clinical trial. I think clinical trials can offer a lot of good information. But I want them to be initiated when there’s the promise of offering good information rather than something a drug company is forced to do when checking off the boxes of an FDA submission.
Regarding Medicare paying for drugs, there are many subtleties and moving parts involved. The drugs used in hospitals (Part A) are reimbursed differently from drugs prescribed in a hospital for at-home use (Part B) and differently from outpatient drugs prescribed by an office-based physician (Part D). Some Medicare patients are covered by private insurance. Depending on the setting, there are drugs that Medicare isn’t required to pay for. Frequently it’s the hospital or the health plan that decides which drugs to cover.
In other words, the linkage between “FDA approval” ==> “Medicare must pay for it” does not exist today.
Dylan
May 26 2020 at 6:56pm
No, we’re not there yet. We’re getting closer in some areas, which is what the whole “personalized medicine” revolution is supposed to be about, but I think by the time we get “there” that means we’re probably at a stage where rational drug discoery makes sense, and where the risk that we see in clinical trials today is not nearly as big of an issue. Right now the problem as I see it is, biology is really messy, and we’ve got very little idea what will work and what won’t before we try it in a whole bunch of people and compare it to some kind of control.
Glad to see we agree on the importance of clinical trials. I’m not at all wedded to the FDA model of drug efficacy testing, but before getting rid of it I’d like to have a pretty good handle on what the new incentive structure looks like to provide the market with that kind of data.
No offense intended, but this seems an awfully lot like dodging the question. Yes, the drug payment environment is complex (I’ve written a whole series of pieces on how we’re actually paying a far larger fraction of our healthcare budget on pharmaceuticals than it looks like, because the CMS data doesn’t include those really expensive hospital administered drugs in their drug category). However, just because Medicare doesn’t purchase all drugs on the market, doesn’t change the fact that they do pay for something like a third of all drug spending , and private insurers pay for most of the rest.
If the FDA isn’t responsible for validating efficacy any longer, I see a couple of possibilities. 1) You replicate this functionality within Medicare/Medicaid and/or large insurers pool together to do it. or 2) There’s some rule that Medicare has to cover drugs that pass PhI safety, maybe at some kind of reduced rate. The first is definitely more palatable to me, but doesn’t seem to change much from the status quo? I guess the difference is that some companies may choose to sell experimental medicines to people who pay out of pocket, either because they are rich or they don’t have any other options, or both. That makes the world a bit better from a freedom perspective, but otherwise doesn’t seem to make many people better off, since those drugs are likely to do more harm than good, and we don’t even get the benefit of the knowledge that you get if it is part of a clinical trial (i.e. all those companies offering unproven stem cell therapies in less regulated markets aren’t carefully collecting data).
Alan Goldhammer
May 26 2020 at 9:26am
If you’re going to drop the effectiveness requirement, why not just drop the safety requirement as well and be done with it? Require testing and disclosure and let doctors and patients do the tradeoffs past that. It would be far simpler than trying to make some objective standard of “safety” without regard to whether your counterfactual is headaches or hepatitis.
These are always easy statements to make by anyone unfamiliar with what is in a new drug application submitted to a regulatory agency. When I was still working in regulatory affairs at PhRMA, I received a communication from a pediatrician in Illinois taking pretty much the same view and that he would be happy to review data on drug safety and efficacy in his spare time. This was laughable. When we were negotiating the renewal of the Prescription Drug User Fee Act back around 2006 we hired an economist to help us do some quantification of the FDA review process. FDA provided us with a lot of information so that we could do our analysis and they had their own economist double checking thing. IIRC, the average application required over 100 FTEs throughout the entire process from submission of the investigational new drug application to the meetings that are held during protocol development and on to the review of the final submission. One pediatrician in Illinois working on weekends cannot do this kind of work as I explained to him responding back.
Now David and Charlie argue that parts or maybe all of this review are wasteful and that we should allow ‘safe’ drugs on the market. What is meant by safety? As noted above, many chemotherapy drugs are not particularly safe. We know there are a number of other drugs that also pose risks to certain populations. As I noted in my post above, laetrile is safe but totally ineffective. Should it be allowed on the market as some people continue to believe it to work on certain kinds of cancer.
Most of the people who write about these types of approaches have little understanding of the regulatory process and the decision making that goes on. Furthermore, many physicians have an incomplete understanding of the drugs they prescribe (just look at the over prescription of opioids). It’s fine to say this should be a decision between the patient and his/her doctor but without the baseline of knowledge contained in the FDA approved prescribing information this is not going to be a particularly informed decision. Until, I see a proposal from the libertarian community that addresses the information deficit, I cannot take them seriously.
Charley Hooper
May 26 2020 at 2:37pm
Even with the exhaustive FDA review process, can the FDA tell which patients will have an adverse event? No. Can the FDA tell which patients will have a favorable outcome? No.
If the current system is so good, why do physicians and patients still need to rely on trial and error?
BC
May 27 2020 at 4:14am
“While there is some appeal to this two-hurdle approach, evidence suggests that there is only a slight benefit and a tremendous cost”
Ironic that the two-hurdle approach was adopted even though it does not pass its own two hurdles. It’s not safe because of the side effects of keeping effective drugs off the market. It’s also not effective because ineffective drugs were a small percentage of the market and therefore not much of a problem and also because patients and doctors need to go through a trial-and-error phase anyways.
Some people like to say that they’re not pro-regulation, only pro good regulation. If so, then they should demand that any regulation be proven safe and effective in random control trials before widespread adoption.
David Henderson
May 27 2020 at 9:59am
Well done, BC. You’ve actually anticipated much of the book on the FDA that Charley is working on and, in fact, the subtitle he is likely to use.
Charley Hooper
May 28 2020 at 1:30pm
Good points. I think a lot of regulation follows this pattern:
(1) Identify the risk of X happening, where X is something bad and the root emotions are fear and distrust.
(2) Without clearly showing whether X is actually happening, the probability that X will happen, or the how the costs of accepting X are greater than the costs of trying to prevent X, a new regulation is put in place. In other words, there’s a lack of numeracy and clear thinking.
(3) The intentions of the regulation to prevent X are used to judge the merits of the regulation rather than the actual mechanisms and outcomes of the regulation
(4) The new regulation becomes the default case and all challenges are waved off with mentions of how X might happen sans the regulation. In other words, there’s a lack of curiosity, inquiry, and clear thinking.
(5) The wisdom and motives of challengers are questioned. This helps solidify the belief that virtuous and smart people know that the regulation is a good idea.
(6) Serious challenges to this regulation must prove that X isn’t happening, X won’t happen, or that the costs of trying to prevent X are greater than the costs of X itself. Those who challenge existing regulations face a tougher challenge than those who advocate for new regulations.
Comments are closed.