David R. Henderson  

Ending the Requirement to Show Drug Efficacy

PRINT
The Many Faces of Means-Testin... Better Market-Oriented Proposa...

AdobeStock_85205479.jpg

In my January 14 post, "Why Are Drug Prices So High?", I referenced Charley Hooper's and my proposal that Congress go back to pre-1962 law and have the FDA be a certifier of only safety, not efficacy. In the longer piece that I linked to, we also suggested having drugs automatically legal in the United States if they are approved by a similar regulatory agency in any of a few developed countries.

That post led to much discussion--58 comments so far, which is close to a record for me. Fortunately most of it was civil and much of it was thoughtful. I particularly appreciated the comments of Jim Glass, who backed much of what I wrote. I also appreciated the comments of AlanG, who was critical of parts of my post and parts of Charley's and my short study.

Rather than reply in comments, it makes sense to reply in a new post.

We had written:

Who would want to take a drug that has not been shown, to the FDA's satisfaction, to be effective? Almost everyone.

AlanG answered:
Saying that almost everyone would not [DRH note: Alan agreed in a comment that he didn't mean to say write "not"] want to take a drug that has not been proven effective is an astounding statement. If this only refers to drugs where there is a documented history of use maybe I can buy it. But what evidence is there? I well remember the case of laetrile and its use as an oncology agents. It never worked. Just because there is a journal article or blogs doesn't mean a drug works.

His response makes sense only if we had written that everyone would want to take any drug that has not been shown, to the FDA's satisfaction, to be effective. Of course, we didn't write that. We wrote "a," not "any." And we went on to establish that many people in fact do take such drugs. Alan writes that "Just because there is a journal article or blogs doesn't mean a drug works." That's true and it certainly doesn't contradict anything we wrote. Moreover, Alan himself admitted that he has taken drugs that the FDA has not deemed effective. In a later comment, he wrote: "Yes, I've taken one or two drugs off label but I consider myself an [sic] super-educated consumer and I always read the full prescribing information of any new drug that I am considering."

We also wrote:

Market success would establish efficacy.

That was an overstatement and, quite rightly, commenter Stuart Buck jumped on it. The longer a drug is on the market, the more likely it is to be efficacious, but the term "establish" is too strong.

Stuart Buck reminded us that no drug is ever shown to be completely safe but then he went too far. He wrote:

This is a common suggestion, but it fails for the basic reason that no drug is "safe." There is simply no such thing. All drugs have side effects, and some are particularly dangerous. Cancer drugs are the best example. Chemotherapy was originally developed from mustard gas used in World War I, and some of the "best" chemo drugs we have today are incredibly toxic (google "adriamycin" and "heart failure"). The ONLY reason we would ever give a chemo drug to someone is NOT that it's safe, but that there is rigorous evidence that there might be a benefit large enough to outweigh the toxic and sometimes fatal side effects. Saying that we should approve chemo drugs for safety without efficacy is impossible nonsense.

Clearly, safety is contextual. But his last sentence, even though correct, can't be used to argue reasonably for an FDA requirement that efficacy be established. If it were so used, many anti-cancer drugs would not be used in their current uses: they have been shown to be efficacious for a particular cancer, but not for others. I wasn't clear whether Buck was claiming that doctors should not be allowed to prescribe and/or administer those drugs for uses where they have not been shown, to the FDA's satisfaction, to be efficacious. If he doesn't disagree with us, then our differences are smaller than he appears to think.

Late in the comment thread, the issue of reciprocity came up. Charley and I had discussed that in our piece, although it was not an idea I highlighted in the post. I think AlanG misunderstood our proposal. He wrote:

the idea of reciprocity of approvals has been around for over 20 years. My former employer was advocating for this back in 1992. While the idea was interesting back then, it would not have much of an impact as regulatory filings are usually done simultaneously and companies want a US approval as quickly as possible because of the lack of price controls here relative to other countries (though this may change because of Trump's announcement that he wants to negotiate pricing for Medicare and Medicaid).

This does not establish at all that the proposal would not have much impact. Even if drug companies file simultaneously, as he says, if one agency approves it well before another, then, under our proposal, it's legal here.


Comments and Sharing






COMMENTS (32 to date)
AntiSchiff writes:

Dr. Henderson,

If the choice were between abolishing the FDA and keeping it the way it is, I'd favor abolishing it.

That said, your ideas for reform would be very helpful and I support them 100%.

I would prefer, however, that all drugs be legal to purchase without FDA approval or a prescription, with the exception that antibiotics should require a prescription, since unnecessary or otherwise improper use entails negative externalities.

Toby writes:
Stuart Buck jumped on it. The longer a drug is on the market, the more likely it is to be efficacious, but the term "establish" is too strong.

I'm going to be contrarian here and point out that the most effective drugs, the drugs that eradicate a disease, will not be on the market because there is no demand for it anymore.

This statement is not necessarily true therefore: try to figure out a way to test this statement by measuring efficacy and time on the market given this fact!

Sol writes:

"it would not have much of an impact as regulatory filings are usually done simultaneously and companies want a US approval as quickly as possible" -- isn't part of the point here that if reciprocity is established, in many cases companies would probably focus on getting approval the place it was easiest / quickest to do?

Jon Murphy writes:

@Toby

"I'm going to be contrarian here and point out that the most effective drugs, the drugs that eradicate a disease, will not be on the market because there is no demand for it anymore."

I'm afraid I don't follow. Why would a drug which effectively cures or prevents a disease no longer be on the market? They eradicate the disease in the body or prevent it from taking hold in a body but the disease itself still exists. For example, firms still make polio and smallpox vaccines even though those diseases have been "eradicated" because those viruses still exist and people need to be inoculated against them/treated for them if they contract it.

Tom DeMeo writes:

The big question with this strategy is what happens to liability.

The FDA, for all its flaws, is a mechanism for public absorption of liability. Take this away, and take away the safeguards, and that liability hasn't magically disappeared. It remains to be seen how responsible everyone in the system will be, and whether the reallocation of this liability and responsibility will ultimately result in more effective freedom for patients.

What these discussions never consider is how many markets have intrinsically problematic liabilities that often are best managed by a democratically accepted system of acceptance in exchange for regulatory restrictions.

It may very well be possible to create a legal framework that manages to balance responsibilities and liabilities better than the current regulatory framework, but just assuming this will happen is naive.

David R. Henderson writes:

@AntiSchiff,
Thanks. I would like to abolish the FDA also. I’m proposing solutions, though, that have some substantial chance of being implemented soon and that have a great chance of making things substantially better. But it’s good of you to remind other readers and me to never forget the ideal.
Re antibiotics, you’re right that there’s an externality argument here. I’m not sure how effective our current system is, though, at restricting usage. Also, I think cattle raisers give huge amounts of antibiotics to their herds. I don’t know enough to say more.
@Toby,
Jon Murphy has answered you well.
@Tom DeMeo,
What these discussions never consider is how many markets have intrinsically problematic liabilities that often are best managed by a democratically accepted system of acceptance in exchange for regulatory restrictions.
Actually, not that you should know this because it was buried among the 58 comments, but I did consider this in response to AlanG. If companies want FDA certification as a safe harbor, they can do so. Those who want to risk liability can do so. I trust most drug companies to figure out which way to go.

dwb writes:

I think people are using "proof of safety" in entirely the wrong sense here. It is not clear to me whether people understand what this actually means.

"Safety" and "efficacy" are not black-and-white. Nor are they "contextual." Safety and efficacy are probabilistic concepts (partly due to natural genetic diversity in how people respond).

It is literally true that "all drugs have side effects," but this is a meaningless tautology and I think grossly misunderstands what safety studies are about.


"Safety" for example is typically measured as LD50: the dose needed to kill 50% of a test sample. We are talking about immediate death here.

Carrots have an LD50. As does Vitamin b-12, table sugar, salt, and caffeine. The LD50 of hydrogen cyanide is about 1 milligram/kg of body weight. Almonds contain a substance metabolized into cyanide, the LD50 of bitter almonds is about 3 grams/kg.

The LD50 for methotrexate (chemo drug, I looked it up) is 135mg/kg. Advil, 636 mg/kg. When you cook meat to 165 degrees, you are not killing every single pathogenic bacteria, or all the botulism. You are just killing "enough" that the remaining dosage of salmonella and pathogenic e coli is harmless (I cook mine only to 135 degrees and eat it red FYI). Bad bacteria are still there, but maybe dormant. Botulism is safe in small quantities too. Under the skin, it makes actresses look good.

There is no such thing as "proof" or "certification" of safety. The FDA can only "certify safety" as a threshold question: A particular drug is "safe," at a given dosage, for a given (average) person. But even then, it's probabilistic.

The FDA has not "proven" or "certified" Advil "safe," they have (essentially) certified the safety studies and approved the threshold dosage warning. As in: "do not take more than X of Advil over a 24 hour period."

This is an important nuance. I do not see abolishing the FDA, there will always be a need for a certified, credible, study that establishes the threshold / guideline for safe dosage - at a minimum for appropriate warning labeling. I don't think anyone should care how or who does the study, as long as it's replicable.

Safety is only 1/3 of the cost of drugs, because it's much easier to establish a baseline for immediate death. You do not need a long term study to see that 50% of your rats died. The most expensive part might be replicating the study on animals most similar in biology to humans (which could mean chimps), while getting the Legal Dept to fend off the likes of PETA because you are killing animals.

Efficacy is far more difficult to prove. Efficacy is also a probabilistic concept. Not every drug works for everyone. It is also a conditional probability, conditional on a non-lethal dose. When a study says a drug (at a given dosage) works for 40% of patients, that does not mean it is "effective" or "ineffective." It might actually mean that it works for 40% of patients ( but we don't know which phenotype of human - maybe only people with O+ blood type). It can be extremely expensive to figure out which types of humans, and you might need a huge long term study since "effectiveness" may only happen over a long time (think of the cancer survival rate after 5 years). Moreover, it could actually be true that a drug only works for 20% of the population, depending on the disease, something you have to weigh when your clinical trial only has 247 people. You might need a heck of a lot more than 247 people in your study, making it extra expensive.

It is worthwhile for the FDA to "certify safety." in the sense of certifying the safe dosage. It might even make sense to have pharmacists, doctors, or nurses monitor dosage to ensure the patient is not getting poisoned (botulism in plastic surgery comes to mind).

In economic terms, people will pay for this kind of "insurance" anyway.

But for efficacy, you need often a much larger, more diverse sample to prove it, and study the patients over a longer period. This seems better suited to the market, where the sample size can be massive.

It might be true people are taking an ineffective drug, but so what? As long as people are taking a safe dosage, even of a chemo drug, so what?


pyroseed13 writes:

I like these proposals. My understanding is that when the FDA tests for "efficacy," they are essentially testing for how many people would have to be given the drug in order for an effect to appear in the data. But all that shows is that drug might be ineffective for most people, not everyone. There could theoretically be people within the sample who would benefit from drug. It would be wrong to deny those people access. Of course, insurance companies may only cover those drugs that have been shown to be efficacious for a large number of people, but for consumers who want to pay out of pocket, they would benefit.

Dylan writes:

Here's a post from a chemist working for a pharmaceutical company where he discusses some of the issues around FDA reform, and why he's not generally a fan. Short version, biology is hard, and we don't know how to prove drugs are effective any other way besides doing long and expensive randomly controlled trials. And even then, we get it wrong too often.

http://blogs.sciencemag.org/pipeline/archives/2017/01/23/i-do-hate-to-tell-you-this-but

Dylan writes:

@ DWB

"But for efficacy, you need often a much larger, more diverse sample to prove it, and study the patients over a longer period. This seems better suited to the market, where the sample size can be massive."

But you need more than a massive sample size, you need a control arm as well. How is the market going to provide that? And if it does, via some mechanism like insurance companies working with pharma companies to sponsor clinical trials...doesn't the end result look pretty similar to the FDA?

John Thacker writes:

One of the strongest cases for reciprocity is with travel vaccinations for diseases endemic to another country. In those cases, the market size in the US is simply not large enough to support applying to the FDA, but drugs are well-tested outside the US. This results in some pretty strange cases where people from the US go to Canada or elsewhere in order to get vaccinated. The US didn't get a cholera vaccine approved until June 2016, despite cholera vaccine being on WHO's list of essential medicines. There are plenty of diseases where even if a vaccination is available in the US, it is less effective or with more inconvenient dosing than that found elsewhere.

The FDA's model works particularly poorly with rare diseases, and especially with rare diseases that are common outside the country. A somewhat similar situation happened during the Zika scare, where Puerto Rico was denied access to possible countermeasures used by its neighbors because they are regulated by the FDA. (The FDA would, no doubt, have changed its mind quickly due to mainland pressure if Zika spread in the continental 48.)

Jeff writes:

I've worked in Biotech for 15 years, and wanted to weigh in.

Reciprocity is a fine idea, and would save literally hundreds of thousands of dollars and hundreds of hours of work on each drug, as there are subtle and aggravating differences between EU and FDA approval, despite harmonization attempts.

Removing the efficacy bar is fraught with problems, and I think the proposal stems from a poor understanding of what the FDA does. The approval bar is actually a judgement by the FDA on whether a drug has a positive Risk/Benefit ratio. Risk is quantified by a safety database, and Benefit is established by showing evidence of effectiveness... this is quantified differently based on the disease and endpoint, with leeway provided in certain cases depending on unmet medical need (See Sarepta).

Removing half the Risk/Benefit equation from the FDA mandate, would mean you only have systematic, controlled data on Risk. The vast majority of individual doctors, and especially patients, are extremely poor at estimating Risk, and and even worse at determining benefit themselves. They simply aren't educated enough to understand the complexity of medical study data. Frankly even experts in the field and at the FDA have trouble with this on many drugs, shown by the frequency of drug withdrawals. Many studies have shown that actual drug usage is more dependent on what is new, what is in fashion, and what is being marketed, rather than what actually works best or has the best Risk/Benefit ratio (see some of the comparative efficacy of anti-epileptic drugs work done by the NIH).

With some drastic investment and work, we could shift the system to an observational post-approval setting establishment of effectiveness. For this to even have a shot at working, you would need to ban virtually all drug marketing, and severely improve medical publication and communication standards. Vaccines actually sort of use this model (for various reasons), with mixed results. The usage setting (outside of a controlled trial) in this model is inherently less controlled, so there is more noise around both safety and efficacy signals. This means exposing drastically more patients to see treatment deltas, and this is done in a setting where you can't really control physician training, patient education, patient consent, and the data quality.

In this post-approval work you would invariably harm/kill a lot more people, treat a lot more people with useless drugs, and have a lower quality of scientific information by which to evaluate therapy choices under your model.

AlanG writes:

Thanks for posting a new set of comments. let me see if I can parse my early comments so that they are clearer. As others have already noted, safety is never assured upon approval and we are constantly surprised when new adverse events arise. If you really argue that that some modicum of safety is required under the proposal it has to be defined up front. Simple pharm/tox studies on animals or Phase 1 studies on humans won't do this. They will also fail to catch important adverse events that might show up when you study a population the size of a clinical trial. So even if you do go through Phase 1, the likelihood of an unsafe drug being given to lots of patients is quite high. Now one way around this is to require all such drugs be given under informed consent such that the patient acknowledges that efficacy has not been determined nor has a full safety profile been established. this would also have to be done to waive any product liability so that the company is protected as they could not use an FDA defense in court.

Say under the above scenario you increase the patient size to what is seen in Phase 2 trials (several hundred patients), why not try to measure preliminary efficacy so that we don't end up selling the newest laetrile? If you do sell drugs at this point or earlier, is it expected that insurers will reimburse or would the patient pay full price out of pocket? If the latter then pricing would have to be very low and once a base line price is established it will be hard to move. AlanG Pharma sells his drug for $20 a month (routine co pay) but then finds out it cures disease X; how easy do you think it would be for AlanG to raise the price of the drug? Sure, I could raise it somewhat but probably not to the level of a Hep C antiviral ($70K for the regimen).

The other problem is one that we dealt with in the early days of HIV research. Patients were clamoring for access to any experimental drug that was being tested. I was part of a Public Health Service Work Group along with FDA, NIH, patient advocates and other industry reps. We looked at this issue with an eye to expanding access but not compromising the clinical trial process. We came up with a plan that was acceptable in that it allowed experimental access but if companies found that there were too many patients taking experimental drugs such that clinical trials could not be conducted the program would be shut down. Fortunately it never came to that.

This latter point is very important in that you might have a large number of "experimental" drugs being taken such that a definitive clinical trial for any one drug could not be done. The argument is that there would be evidence coming from this widespread use that would suggest what is good and what is not. What exactly are we going to use for this evidence? I don't think (or at least hope) that it would be Internet forums. It would have to be some well designed trial.

Finally, who is going to monitor the use of all these experimental drugs? What would be the responsibility of reporting adverse events which are sure to happen? Would doctors incur any liability if the prescribed such drugs? I just think that there are a lot of legal and scientific issues that would not make this a proposal that a pharmaceutical company would take advantage. I assure that AlanG Pharma won't be selling any pharmaceuticals that are not FDA approved (but of course I will continue the lucrative dietary supplements business!! Hey, maybe this is the workaround for the proposal, just consider all these products dietary supplements)>

AlanG writes:

Jeff writes:

With some drastic investment and work, we could shift the system to an observational post-approval setting establishment of effectiveness. For this to even have a shot at working, you would need to ban virtually all drug marketing, and severely improve medical publication and communication standards. Vaccines actually sort of use this model (for various reasons), with mixed results. The usage setting (outside of a controlled trial) in this model is inherently less controlled, so there is more noise around both safety and efficacy signals.

Quite right. One of the initiatives that I directed was an Observational Medical Outcomes project that could do this for both safety (the primary reason for the project) and also efficacy. It was predicated on having access to large databases of deidentified patient information that could be aggregated and analyzed. I secured the initial $20M of funding from pharma companies and the project was funded for a couple of more years before it was transitioned off to the Reagan/Udall foundation as a public/private partnership.

During the research phase a lot was learned about how to aggregate data (not as simple as one might think) and new search algorithms were developed. For intrepid readers the link to the project is here: along with a link to the original OMOP project that I directed.

Personally, I think achieving David's vision will be more difficult than doing the clinical trial for FDA.

[Broken HTML fixed in two places. Please proofread your comments when previewing. You have to close quotes and close links, not just start them. Please see our FAQ page for detailed instructions on using the quote and link buttons.--Econlib Ed.]

Larry L. Wasem writes:

Dr. Henderson:

In 1976, I was a legislative assistant for Congressman Steven D. Symms (R-Idaho). Another of his chief aides, Paula Hawks, and I worked on a very simple piece of legislation for the Congressman we called the "Medical Freedom of Choice Act". A very simple bill, it merely repealed the 1962 Amendments to the Food, Drug and Cosmetic Act. At one time in 1977, we had something like 150 House co-sponsors. Unfortunately, Rep. Paul Rogers (D-Florida) chaired the subcommittee that the legislation would have to be heard in and Mr. Rogers had been instrumental in getting the 1962 Amendments adopted. He would never allow a hearing. I liked Mr. Rogers but he told me one day that we could have a hearing when we had 218 co-sponsors (1/2 of the House). It was good legislation then, it is even more important now. One of the more disappointing aspects of the entire process for me, was reading the testimony of the pharmaceutical industry during the hearings for the 1962 Amendments; for the most part all the dire predicitions came true. However, when we approached the big pharmaceutical companys in 1976, the had all changed their minds. Competition had been decimated and they were now fine with the burdensome regulatory system. Crony capitalism at its worst. I hope your idea can get some traction. It is a good one.

dwb writes:

@Jeff,

"Removing half the Risk/Benefit equation from the FDA mandate, would mean you only have systematic, controlled data on Risk. "

Why are we assuming that if the FDA does not force drug companies to perform efficacy data, it won't exist? Or that doctors will not publish?

I do understand risk, and I think you are vastly over-complicating it.

A family member of mine has extremely high familial cholesterol and triglycerides (Hypertriglyceridemia). Her doctor did not rely on FDA risk analysis, the condition is extremely rare. Her doctor had to experiment with several combinations of drugs over a few years until something brought down her numbers. One drug alone did not do it. One of those is insulin - even though she is not diabetic. It is a secondary effect, insulin is not normally used specifically to help lower triglycerides. If she had to wait for a clinical trial to prove efficacy, it would never happen. Trials involving cocktails are exponentially more expensive to prove, and it will never be cost effective for uncommon diseases.

What the doctor needed to know was safety- what the dosage is, what the signs of overdose are, what the signs of effectiveness are. Not the FDA approved risk-benefit.

I can certainly see the argument for a trained professional monitoring dosage and progress with lab results. Plus informed consent.

If, as you say, studies show people are buying drugs while ignoring the risk-benefit analysis, exactly what value does FDA approval really add?

I am struggling with what the harm is in allowing trained professionals to experiment with drugs proven safe. If they recklessly endanger patients by ignoring safety, we jail them, its an actual crime.

It is probably true that we might treat a lot more people with "useless drugs." We already do that: the demand for vitamins is massive and they are largely useless too. I don't see how we will kill more people as long as the safety baseline is established.


@Dylan
"You need more than a massive sample size, you need a control arm as well. How is the market going to provide that?"

This is a fair point. But, why should proof of efficacy be necessary for approval?

I think you are over-complicating the situation. Disease progress is pretty objective, and a doctor who is monitoring monitoring lab results can figure it out. Doctors publish results and go to conferences, that's how the family doctor knew what to try for a rare condition.


Thomas Sewell writes:

If we politically can't get rid of the FDA, why can't we at least make it somewhat market-based and responsive?

Turn it into a certifier for recognizing medical testing laboratories, similar to UL and the other Nationally Recognized Labs.

Then it can also certify non-U.S. agencies like the EU as recognized drug and medical device testers as part of the same process.

With some shared liability on the part of the testing labs to ensure quality standards, at the end of it all you'd have a market for independent medical testing.

As a bonus, as suggested, make the barrier to being able to use a drug the safety test (even if you require a big orange label on it that it's only tested for such and such a dose and only for safety) and make the barrier for the patent owner or licensee making the drug being able to advertise or promote a specific use the effectiveness test.

mariorossi writes:

Dr Henderson,

Do you believe we are currently overinvesting in clinical trials? The main aim of the proposed reforms seems to be reducing the demand for them.

It seems to me that clinical trials are a public good. The approval process is one possible mechanism to encourage their production. Patent protection is obviously another subsidy to such activity and basic research grants yet another.

I think that it would be an interesting arguement: instead of forcing companies to conduct studies, should the government directly provide grants to produce the studies?

I think clinical trials can be viewed as unpatentable intellectual property. Everyone gets the right to use the information generated by clinical trials for free. This is a potential source of undersupply. Should we change IP laws to require efficacy in order to obtain protection (e.g. the patent holder is the person that proves a treatment works rather than who comes up with a new drug)? Should we make clinical trials patentable (e.g. the inventor of a drug cannot market a compound for a specific use wihtout permission from the clinical trial patent holder)?

Michael Byrnes writes:

dwb wrote:

Why are we assuming that if the FDA does not force drug companies to perform efficacy data, it won't exist? Or that doctors will not publish?

This is a good point. And I think this is an area where the private sector can and will pick up some of the slack.

Efficacy data from large clinical trials is a big part of what drug companies rely on to justify the prces they set. To use a high profile example, consider Sovaldi, Gilead's $80,000 (at $1,000 per pill) cure for hepatitis C. They can get away with that price because the drug has been shown to have a 90% cure rate in large efficacy trials.

In an alternate, safety-only FDA world, Gilead either does the needed efficacy work anyway or they sell Sovaldi much more cheaply because they cannot justify the price based on the efficacy data.

It is dead wrong to assume that 3rd party payers are stupid sheep who happily pay whatever price a pharma company asks for, regardless of justification.

AlanG writes:

Michael Byrnes raises the critical point that I also tried to cover. "It is dead wrong to assume that 3rd party payers are stupid sheep who happily pay whatever price a pharma company asks for, regardless of justification."

I don't see anything in David's original post or in those from others about how pricing of these efficacy unproven drugs(because that is what they are)will be. Mr. Byrnes notes that Gilead is able to get a nice price for Sovaldi and there are countless other examples of high cost drugs as well. If insurers are not going to reimburse for non-efficacious drugs the price will have to be quite low. As most economists know, market entry at a low price makes it very difficult to raise the price by a quantum leap or two without sparking outrage. thus, we are probably looking at an OTC drug pricing model. This might work fine for solid oral dosage form drugs but I fail to see how it would be applicable to newer biotech products whose manufacturing is much higher. Instead of a patient facing a $20 co-pay they might have to pay 10 or 20 times as much.

mariorossi writes:

@Michael Brynes

I don't think anybody is arguing that efficacy studies would disappear. But removing efficacy studies from the FDA process removes a subsidy for them.

So you can either argue that studies would be conducted anyway (at which point I see no benefit to the change, it's like abolishing a 1$ minimum wage, not really binding), or that less studies would be conducted or that better studies would be conducted. The benefit of less studies seems very questionable to me, while we are all in favour of better studies.

While effcacy studies as a marketing tool are certainly useful, there are loads of other (effective) marketing tools. Those other marketing tools will become marginally more attractive.

Michael Byrnes writes:

@ mariorossi

It is not just marketing I am talking about.

In recent years we are seeing a trend, throughout much of the world, towards greater sophistication among payers. This is obviously a response to rising costs. Payers are universally concerned about getting value for their dollar (or other currency) and about the overall impact of paying for pharmaceuticals on their bottom line.

The wave of the future is going to be value-based pricing. Harder to do than it sounds, but we are already seeing steps in that direction. In the U.K., for example, manufacturers have to make representations about how well their drug will work, how many people will use it, what the overall cost impact will be - and if it does not live up to those benchmarks the company must rebate some of the cost of the drugs back to the NHS.

This trend won't be stopped by things that the regulatory agencies do.

Jeff writes:

@AlanG
I had actually run into some of the IMEDS work in the past, it is an interesting program, good work. Though after working with hundred of physicians on clinical trials, I think we are miles away from having even the basic infrastructure in place at site of care level needed to generate the databases effectively. I think only an industry wide push by the payors to standardize EMR will get this done. The first steps seem to be happening from what I've heard, but progress is not rapid unfortunately.

@Michael and dwb

Sorry, as soon as there is no requirement to conduct adequate well controlled clinical studies, pharma will immediately stop doing them. The nutraceutical, and homeopathic anti-cancer clinic industries provide ample evidence of what will happen. Industry will conduct poor controlled, highly biased 'trials' to generate marketing materials. Providers will use hype and their perceived position of expertise to charges thousands or even hundreds of thousands of dollars for procedures that do not work. Just google Robert O Young for the latter. Patients will pay themslves, or lobby insurance companies to pay so heavily that it gets covered at least for some people... patients do it because they are desperate and want hope and don't understand Risk/Benefit analyses in the complex field. Insurance will continue to cover the least amount possible to try and avoid public outcry and additional government interference, which is exactly what they do now.

If you still have the government require high quality trials, industry will just take longer to conduct them. Their incentive to move fast will be reduced because they are generating revenue now. The incentive can actually be to never complete the studies (particularly in non-competitive indications), since the results may be negative leading to drug withdrawal.

A 'pay for evidence' model can be tested in part by looking at NICE in the UK. The practical effect has been to spur a little bit of work, but mostly just to reduce available medicines and the rate of innovation as the government inevitably sets reimbursement levels low (since they are the biggest payor)

Michael Byrnes writes:

Jeff wrote:

The nutraceutical, and homeopathic anti-cancer clinic industries provide ample evidence of what will happen. Industry will conduct poor controlled, highly biased 'trials' to generate marketing materials.

This is a good articulation of the "payers are dumb sheep" view that I don't agree with.

Insurers are not going to pay $1,000 per pill, or anything close to that, for a drug supported by a few poorly controlled, highly biased trials.

Charley Hooper writes:

@Sol

isn't part of the point here that if reciprocity is established, in many cases companies would probably focus on getting approval the place it was easiest / quickest to do

That's exactly the point. Companies wouldn't need to file with a dozen different governments' regulatory agencies. Further, these regulatory agencies would then compete with each other to provide the best service.

This proposal turns monopolistic organizations into competitive ones.

Charley Hooper writes:

@Toby,

I'm going to be contrarian here and point out that the most effective drugs, the drugs that eradicate a disease, will not be on the market because there is no demand for it anymore.

While theoretically possible, this is highly unlikely in practice. It could only work for a contagious disease where everyone takes a 100% effective vaccine.

For cases such as cancer and cardiovascular disease that arise spontaneously, there will always be new patients who present with the condition. Layer in the fact that many people don't take drugs, most drugs don't work for everyone, and most drugs control a disease rather than cure it, and you can see that drugs don't go away.

Charley Hooper writes:

@Tom DeMeo,

I'm not an expert on legal liability, but I don't think FDA approval protects companies much, if at all. Just consider the avalanche of lawsuits against Merck when it removed Vioxx from the market.

Vioxx had been approved by the FDA and the FDA-approved label mentioned the very problem that Merck was later sued for. The only new information was that the risk was greater than estimated earlier and Merck conceded defeat by voluntarily withdrawing Vioxx from the market.

I think the only companies that have legal protection are vaccine manufacturers who enjoy the legal shield of the National Childhood Vaccine Injury Compensation Program.

Charley Hooper writes:

@Dylan,

But you need more than a massive sample size, you need a control arm as well. How is the market going to provide that? And if it does, via some mechanism like insurance companies working with pharma companies to sponsor clinical trials...doesn't the end result look pretty similar to the FDA?

There's one reason this is very different to the situation right now with the FDA: forced versus voluntary action.

The FDA tells drug companies what trials to run, even if they don't make any sense. One of my small clients spent years and millions of dollars running clinical trials that were pointless, given the drug they were developing. When one key person left the FDA, his replacements basically asked, "Why did he tell you to run those trials?"

Drug companies, working on their own, could design trials that made sense, answered a valid question, and were cost effective. That's a night and day difference.

Toby writes:

@JonMurphy

If the disease still exists then the disease has not been eradicated. Neither is the drug then the most effective drug.

Point is that the (future) demand for a drug can be (weakly) declining in its effectiveness. A statement that the time on the market will vouch for its effectiveness is therefore not necessarily true.

Similarly, we see very ineffective drugs being on the market for a very long time and are in fact gaining market share. Homeopathic remedies are an example.


Dylan writes:

@ Charley Hooper

The pharma companies are the ones that are actually designing the trials right now, just with guidance from the FDA. I don't doubt that the FDA has given some bad guidance over the years and made companies jump through hoops that they shouldn't have to jump through, but I've personally seen a lot more of the opposite. Companies routinely cherry-pick through or completely ignore the advice the agency gives, choose bad endpoints, under-power studies, misrepresent to investors what the agency has communicated (compounded by the fact that CRLs aren't publicly disclosed), often don't publish trials that failed so insurance companies and doctors only get a partial picture of the data package for the drug and generally have no idea that there is missing information, post-hoc subgroup analysis. The list goes on and on. And I (mostly) don't blame the companies or their management for this kind of behavior, but the incentives for management are perverse. I'd like to think that insurance companies could step-up and refuse to cover ineffective drugs, but since they cover drugs that are ineffective and really expensive now (see Exondys 51 for the most recent example) I remain pretty skeptical.

dwb writes:

@Jeff
"Providers will use hype and their perceived position of expertise to charges thousands or even hundreds of thousands of dollars for procedures that do not work. Just google Robert O Young for the latter."

No, that is just hyperbolic overwrought nonsense. We did not have an epidemic of snake oil salesman in the US prior to the 1962 requirement of efficacy. The 1962 law was passed in response to Thalidomide. Thalidomide was not even a crisis of *efficacy* testing, it was a crisis of poor or non-existent safety testing (which failed to uncover birth defects). Had people known the true risk of birth defects, with additional safety testing, I don't think pregnant women are taking a drug that causes birth defects in 50% of babies.

Thalidomide is still used as a treatment, by the way, just not on pregnant women.

Nor did the FDA (or a few felony charges) even stop Robert O Young.

Bad facts make bad law!

FDA will never stop the Bernie Madoffs and's of the world. Heck, getting charged with a few felonies did not stop Young

Stuart Buck writes:

Interesting comments from a lot of folks who actually know what they are talking about:

https://endpts.com/dear-president-trump-dont-destroy-the-fda-we-know-and-respect/

Comments for this entry have been closed
Return to top