The Good News on the FDA and ANDAs

Yesterday, I pointed out that generic drug prices are falling. So what accounts for the small number of large price increases in the generic drug market? It’s a combination of market shenanigans, supply shocks, and FDA delay.

The markets where price increases have been large tend to be relatively small. Daraprim, for example, is only prescribed some 8-12 thousand times per year in the United States. The small size of these markets is no accident. Keep in mind that whatever one may think of Shkreli, he did show a kind of entrepreneurial genius in scouring the universe of drugs in the United States to select one where monopoly power could be so effectively exploited. Shkreli found a market where 1) the total size of the market was low so there wasn’t much competition but 2) the drug treated a serious illness and 3) there wasn’t a good substitute so the value of the drug to the small number of patients was very high.

In addition, Shkreli knew that he had at least a 3-4 year window of opportunity to exploit monopoly power. To compete with Daraprim a competitor would have to submit an Abbreviated New Drug Applications (ANDA) to the FDA. Despite the name, Abbreviated, it costs at least five million dollars to go through the process and right now there is a backlog of nearly 3,000 ANDAs at the FDA’s Office of Generic Drugs. In recent years, it has taken 3- 4 years to get a generic drug approved. The cost is too high and the delay too long.

(I am focusing on the standard route to market entry and ignoring the possibility of importation or compounding which I discussed earlier. I’m also ignoring that Daraprim is unusual in that it was approved in 1953 before the current FDA system of safety and efficacy trials, and the FDA is being absurdly cagey about whether they would allow a simple ANDA for Daraprim. I may write about that in a future post– see here for a related case.)

So what’s the good news? In 2012 Congress passed the Generic Drug User Fee Act (GDUFA). Modeled after the very succesful PDUFA, the act earmarks fees paid by generic drug manufacturers to the FDA’s Office of Generic Drugs. As a result of those fees, the FDA has hired more reviewers and they are rapidly reducing the backlog. That’s the first piece of good news.

A second piece of good news is that FDA delay isn’t the only cause of the backlog. Another cause of the ANDA backlog was an unexpected increase in the number of ANDAs. I would have been much more worried if the number of ANDAs had decreased. Despite new user fees and some increase in regulation the increase in submissions is evidence that the US generic market is competitive, vibrant, and profitable.

The generic drug market in the United States has been very successful. We are constantly told, for example, that US pharmaceutical prices are the highest in the world and that is true for patented drugs but generic drug prices in the US are among the lowest in the developed world and most prescriptions are of generics.

We can address the price hiccups in the generic market by opening up to more world suppliers, speeding up the ANDA process and keeping costs of entry low. Overall, however, we shouldn’t let the price hiccups detract attention from the fact that the generic drug market is competitive, vibrant and thriving and we want to keep it that way.


FDA requires a bioequivalence study. To do this study, the generic company needs some of the marketed drug to compare to. But for special drugs, the supplier may restrict the sale of drug to only those with a valid prescription. This creates a catch22 of sorts. If the FDA mandates the drug be made available for clinical trials, that removes a hurdle.

Which drugs are sold directly from the drug maker without any intermediary?

No Federal or State law allows the kind of restriction on drug sales you describe.

After all, a drug distributor may need to analyze drugs it's reselling to verify they are not defective or counterfeit. A hospital may need to analyze drugs to make sure they are not contaminated and thus causing problems for patients.

Actually, the FDA requires a REMS for many products. This type of plan to reduce off-label or inappropriate often puts restrictions on how the distributor (intermediary) can sell the drug. To run a BE study, you may need a decent bulk purchase of the drug. Brand manufacturers interested in discouraging the competition from becoming approved can exploit this, and in fact require their distributors to not sell bulk supply even when there is not a REMS. Laws preventing importation of drugs from other countries also contribute to this. Are either the FDA restrictions, or the use of a contract by the drug maker, so inconceivable?

Daraprim is only one example (and I don't believe this is due to an FDA-required REMS), but it is easy to find articles and studies on this strategy.

Yesterday I pointed out that, according to Austin Frakt, one in five prescriptions is off-label. Moreover, it's the older, generic drugs that are most often used off-label. See, e.g., That 20% of drugs are used for unapproved purposes may be music to Tabarrok's ears, but what about the risks to patients, patients who are mostly unaware that they are taking drugs for purposes that have not been approved. As Frakt pointed out in his Upshot column, off-label use can easily be prevented (it is in other countries), but we choose not to in large part as a way of offsetting the relatively high price of approved drugs in the U.S. Think about that: 20% of legally prescribed drug use is essentially unregulated. But we pay a high price for the deceit, not only in terms of the risk to patients but in the free riding on other drug companies' investments.

Be clear about what you're saying:

1) Peer reviewed publications and science doesn't matter, approval for the uses in other countries doesn't matter, only FDA approval matters for safety.
2) Requiring drugs to go through additional level of approvals to bring the on-label uses on side is exactly the same thing as trying to get pre-1953 drugs to go through the modern regulatory approval. The FDA will (based on what it does with pre-1953 approved drugs taken through the modern process) grant several years of exclusivity to drugs that go through that process, which means much higher prices-- perhaps on a Shrekli level, since he has used that very tactic before.
3) "Off-label use can easily be prevented?" Really? In the same way that we're easily able to prevent people from getting inappropriate painkiller prescriptions (and then easily able to prevent them from satiating their need with heroin after we shut off their painkiller prescriptions?) You're proposing that we can somehow force doctors to avoid reading peer-reviewed journals and see what drugs are prescribed for in other countries? That we can easily tell them "Yes, this will cure your patient, improve their quality of life, and possibly keep them alive, but don't let them have it?"
4) It's difficult for you to complain about the high price of on-patent drugs and simultaneously complain about generics "free riding on other drug companies' investments." At the very least, it's incoherent to do so without calling for relaxation of FDA regulatory and approval policy to bring costs down-- simply ordering drug companies to lower their prices without lowering their compliance (and R&D costs) costs is a pain worse to those companies than any free riding. Indeed, for "free riding" to be a problem, it must result in less R&D investment by the pharmaceutical companies than without it-- which means that you're directly arguing for higher prices in order to incentivize R&D.

Rayward, you are essentially proposing Shrekli and Valeant as the proper models for the industry, whether you like it or not. I must strongly reject that, as it would be terrible (and expensive) for patients and kill far, far more people in the name of avoiding "risk."


As usual, John Thacker nails it.

There is strong merit to everything you say. But in an America where flagrant abuse of marketing off-label usages occurs for groups known to suffer from bad side effects, causing many (statistical) deaths and much suffering, there is very good reason to be concerned.

Check out the following for a well-researched investigative journalism piece (soooo rare these days) on J&J's essentially unpunished criminal behaviour with respect to Risparadol:

10 part series. There's a lot of outrageous lies and abuse to uncover, and he barely scratches the surface in this series.

Isn't the gift of the case for allowing off label uses is that the safe part of safe and effective is ultimately more important than the effective part?

There may be rare cases where off label drugs are safe for the on label users but not for the off label users, but usually, an off label drug is tested on control subjects in the initial approval process and with people who are benefiting from a similar compound in a different way, so the truly unsafe off label uses are likely to be few.

Maybe quite a few are "uncomfortable" (such side effects can be tolerated if the benefit associated with the on label use are great) and not proven to be effective. But, so long as that is a pretty small subset of all off label drug use, it should be manageable, especially in the age of the Internet where it is much easier to aggregate user experiences (for outright illegal psycho-active drugs, the website "Erowid" manages a quite effective system with volunteer labor and donations) and get the word out to physicians about problematic off label uses quickly to minimize harm. Perhaps in the cases where off label uses are actively promoted commercially, there should be a tort duty to actively monitor the off label use promoted for ill effects.

Indeed, I think that there is a lot to be said for having a two stage FDA approval process. First, you prove that a drug is safe and may due to with reciprocal approvals from other jurisdictions, and it is legal to prescribe and sell it. Then, you prove that a drug is effective, and it may be (must be?) covered by government funded or government subsidized health financing systems (e.g. Medicare, Medicaid, VA Health, Tricare, and Obamacare subsidy approved plans). The financial incentive to conduct the whole approval process is still there and the FDA protecting the public fisc in the process, but people are allowed to try experimental drugs that are proven to be safe sooner without having to be in an experimental trial.

"There may be rare cases where off label drugs are safe for the on label users but not for the off label users"

If you have time to review the series, you will find that J&J used a variety of grey marketing mechanisms to promote off-label usage prescriptions from groups where negative effects were known by the company, but suppressed (in fact, worse, they manipulate the results).

Namely, higher cardiovascular death for elderly and that the drug caused some percentage of pubescent boys to develop genuine and large breasts. The company hid both results and promoted the off-label use anyways by paying lots of money to "experts" to promote off-label usage.

I would have more confidence in the current system if J&J people who were responsible, whether by paying fines/suits or jail time, and if the cost had not been fully passed on to sharedholders. Check the articles. It's not some tiny isolated small case. The violation represented tens of billions of dollars of sales, more than tenfold what they had to pay in settlements.

I think that there is some conflating of concepts. Under current law, any doctor can prescribe a drug for an "off label" use. Only the drug company is prohibited from promoting that off-label use.

The FDA approves that a drug is safe an effective for a given disease: i.e. US govt certifies that taking the labeled drug at the labeled dose for the labeled disease is safe and effective for treating the labeled disease. With FDA approval comes a period of exclusivity (to compensate for R&D) and opens up government payment for that use.

Off label use by doctors makes sense because they can adapt the latest science to treat a disease. They are responsible for the patient's health, and therefore assume the risk for off label usage. The government is not certifying.

Drug companies cannot promote an off label use. They can spread lot of literature, and let others spread the world. Like a PAC under our campaign finance laws. The rationale, from what I understand, is to prevent the appearance of government imprimatur, and thereby protect the public. The problem is that this is less efficient way to spread information and that no one is responsible for its accuracy, and we have a plethora of bad doctors and snake oil salesmen: so much for public benefit. I'd rather some system that e.g. a drug company can use to promote the latest science, without FDA sanction.

J&J WAS promoting off-label use, for groups that they KNEW would experience negative effects, negative effects which they covered up prior to promoting the off-label use. Yet, no individual was held responsible, shareholders footed the entire bill of the lawsuits.

Nathan: that's partly my point. From a consumer protection perspective, we should be more concerned about side effects than whether J&J promoted a drug, yet the primary offense is promoting off label use.

Excellent comment, and I would also add: I kinda question just how well you can really prevent off-label prescriptions. Doctors can simply falsify diagnoses if they want to prescribe a particular medication in order to treat some condition it hasn't been officially approved for. In some cases, this will be as simple as filling in some codes on a form.

+1 great comment John.

However, this topic has really brought out the crazies on the site. They are getting more numerous and vociferous by the day.

That wasn't very clear.

I hope this is clearer.

Drug makers do R&D to prove to the FDA and other regulators that a drug is safe within parameters and marginally effective in treating conditions A, B,...

The FDA approves the sale of drugs with labeling describing what their research has proved to the criteria of FDA law and regulation.

Doctors get to prescribe any drug using the data from the drug makers that is approved by the FDA, AND ANY OTHER RESEARCH DATA AVAILABLE TO THEM WHETHER APPROVED BY FDA OR NOT.

The principle is the doctor is deciding treatment based on objective data that is not influenced by profit. If the We the People NIH research suggests a set of drugs can treat a medical condition, that is knowledge unbiased by profit. The doctor is putting himself on the line by prescribing a drug for a patient, but if he has informed the patient, the decision on treatment is the patient's.

The history of drug companies pushing drugs to profit from use that is harmful is rampant, and the restrictions on drug companies even distributing reports of say NIH funded research is driven by repeated abuse by drug company sales and marketing.

After all, no one creates detailed laws and punishments for things no one has ever done. Ie, there are no laws prohibiting the murder of Martians.

Ah yes good old Alex here to sing the praises of criminal monopolists. What next a post praising pedophiles for exploiting advantages in the dating market?

You gotta love those economists .... ;)

The problem is that what he did was not criminal. It was entirely legal. It was legal because the FDA made it possible for him to profit so greatly by legally restricting entry to the market.

Denouncing the FDA's overbearing approval process =/= praising greedy monopolists. In fact, it's the opposite relationship.

"he did show a kind of entrepreneurial genius..."

It's a lower and lower bar to be a John Galt super hero these days. What happened to inventing things?

Why invent new things when the government protects the market for old things?

Either we are or we aren't going to regulate drug prices; and since we aren't, we can't (and shouldn't) prevent Shrekli or anybody else from engaging in what is essentially arbitrage. Since we choose not to regulate the price, arbitragers such as Shrekli are providing a service by making the market more efficient, bringing the price up to the market price so that supply and demand will be in balance. Of course, we could choose to regulate neither the sale nor the price, which Tabarrok might prefer, or we could choose to regulate the price but not the sale, or we could choose to regulate the sale but not the price, which is current policy. To criticize arbitragers such as Shrekli is to ignore the choice we have made.

Private supply must meet private demand, no matter the social cost. It will make things better unerringly and always. It is not a matter of faith, rather it is something that is only to be believed.

The "social cost" of supply *not* meeting demand is higher, I would posit, than the "social cost" of supply meeting demand.

It can always be argued for a specific case. But when a monopolist excludes 99% of the market (making up numbers) in the profit-optimizing strategy, it is doubtful that in the specific case it will maximize social gains in any timeframe.

First, we're discussing competitive markets, not monopolist markets. A monopolist can only maintain control over a market through brute force. Discussing the balance of supply and demand in such a market is a meaningless endeavor. The market needs to be open to many suppliers willing to meet latent demand.

Second, excluding potential buyers is probably the worst way to earn a profit, even for a monopolist. In a competitive market, you can virtually guarantee a damaged reputation, lost market share, and eventually bankruptcy. In a monopolist market, excluding buyers only eats into potential revenue.

Can you explain where they are excluding buyers? I don't know specifics, but it seems that the insurance companies and Medicare/aid are generally absorbing the increased cost of this small-market drug.
Not disagreeing with the reputation part, though Shkreli at least doesn't seem to care about that.

Supply was meeting demand for forty straight years uninterrupted before Shkelli entered the picture. What is "more efficient" after Shkelli?

Shkelli exploited governmental regulatory rules to his financial gain. Nothing was more efficient. This is a classic case of the government interfering in markets with the best of intentions. The result is often unexpected and bad.

If there is no value creating purpose to his raising the price, why should the government let him? All you've got to do is have one of the agencies say its illegal and force him to sell at the old price. Who would be hurt by such action? This is a really simple solution that can be easily implemented, and the only individual in the entire world hurt by it is Martin Shkelli, whom we've already identified as an immoral parasite who produces nothing.

What don't you seem to understand about the fact that the government has no control over drug prices? The government has no authority to "let" him raise the price, so asking why the government should "let him" raise it is meaningless.

Ask the Venezuelans about the efficiency of price controls.

"This is a really simple solution that can be easily implemented"

It's an illegal solution under current law. Implementation means literally an act of Congress. Not so easily implemented.

I'm no fan of Shkreli, but its always amazing to me that the drug companies are portrayed as the villans. Doctors and hospitals are responsible for the lion's share of healthcare costs, and their pricing structures are entirely opaque. Only recently did we get some information, and it shows huge differences in pricing.

A lot of what goes on in the drug industry is advertising and drug pushing through sales reps. Most of the big companies don't even do much R&D anymore, they just buy start ups that do R&D. Big Pharma are basically just giant sales/marketing/legal entities. The sales aspects are dubious (they are more about getting people to take drugs no matter what as opposed to trying to help people find drugs that improve their health), and the legal aspect seems pretty zero sum. We all know patents are necessary to spur R&D, but its debatable how long the patent needs to be to do so. Most of the legal expense of trying to extent patents past their original dates is a zero sum game.

So you've got companies that don't really make drugs and spend a lot of their revenue on zero sum economic activity (aggressive sales techniques, lawyers). Most people don't mind the idea of drugs being expensive to pay for R&D, it's all the non-R&D practices of drug companies that irk people.

"We all know patents are necessary to spur R&D."

This is not a myth that is universally accepted.

Indeed on this topic there's an interesting book by the economists Boldrin and Levine called "Against Intellectual Monopoly":
which was also reviewed by Alex on this blog years ago:

A recent example of how the price of new drugs can be set very high thanks to the monopoly granted by patents is the one of Sofosbuvir (Sovaldi):
Unfortunately the solutions proposed by the report are rather cosmetics and ineffective: a missed opportunity.

Yes, a lot is spent on marketing. But R&D aint cheap, whether it is done in house or through purchasing a company. Clinical trials are even more expensive.


Y'all never heard the expression "criminal genius".

Moriarty weeps.

Well, it is rare to hear 'entrepreneurial genius' so blatantly connected with 'criminal genius,' especially as Prof. Tabarrok does not discuss Shkreli in criminal terms, merely in the much more mild fashion of 'whatever one may think of Shkreli.'

It will be an extraordinary day indeed when any member of the GMU econ dept. will associate entrepreneurial genius with criminality. One should note, to be fair, that Shkreli apparently committed no criminal acts in connection with becoming the owner of a profitable drug monopoly.

i wouldn't be sure of this statement
"Shkreli apparently committed no criminal acts in connection with becoming the owner of a profitable drug monopoly"


I would like your view on Biosimilars and how to create a successful market for a group of products which are not automatically considered interchangeable for the purposes of permitting substitution at pharmacists. In the absence of this it would seem difficult to generate the sort of competition that deliver the price reductions that we would hope for in the off patent space

We already have a real life example of this that the manufacturer is trying to obfuscate the information on. Genentech received approval for Avastin (bevacizumab) to treat certain cancers and it's antii angiogenic properties proved useful in controlling age-related macular degeneration. The company did not pursue FDA licensure for that compound but rather developed Lucentis (ranibizumab injection) which costs much more than that of compounded Avastin. Genentech argued that the only Lucentis should be used for treating the condition despite the fact that both compounds have the same mode of action. The issue was settled when NIH sponsored a comparative clinical trial that showed both were equally effective. Retina specialists can prescribe compounded Avastin off label saving a lot of money for patients and the health care system.

I'm curious about one thing - if generics in other developed countries cost more than in the U.S., the advantage to the U.S. health care system in importing more expensive drugs to reduce 'price hiccups' is what, exactly?

Because that is what these two lines seem to imply - 'We are constantly told, for example, that US pharmaceutical prices are the highest in the world and that is true for patented drugs but generic drug prices in the US are among the lowest in the developed world and most prescriptions are of generics.

We can address the price hiccups in the generic market by opening up to more world suppliers, speeding up the ANDA process and keeping costs of entry low.'

Though possibly, world suppliers means a country like India. Even more possibly after the Indians submit properly to American sponsored trade treaties which no one is allowed to read beforehand. Because I'm pretty sure that this sort of 'generic' import will not be allowed into the U.S. - 'According to the California Technology Assessment Forum, a panel of academic pharmacoeconomic experts, representatives of managed care organizations, and patient advocates, a 46% discount would bring the average price of treatment to about $40,000, at which price sofosbuvir-based treatment regimens represent a "high value" for both patients and healthcare systems. .... The company (Gilead) also announced that it would sell a name brand version of the product in India for $300 per course of treatment, approximately double a third party estimate of the minimum achievable cost of manufacture.'

Yep, the pharmaceutical market at work - even when charging 99% less in India than the U.S., Gilead is still charging twice as much as what a competitor feels would be a reasonable price to make a profit from based on the cost of manufacture.

Of course, regulating the sale of drugs but not the price of drugs provides opportunities for price gouging. Tabarrok's solution is the market solution: regulate neither the sale nor the price. That's fine, but it comes with the potential for an extremely high cost, so high that it's unlikely the U.S. will make that choice. Instead, we make the choice silently, by allowing with a wink and a nod a large and growing market in the sale of generic off-label drugs. Soon enough the unregulated off-label market will exceed the regulated market, or at least that's the hope of some. Then some catastrophe will occur, and there will be lots of hand-wringing, and critics of the FDA will argue that it confirms all of their criticism of the FDA. I'm no economist, as the saying goes, I only pointed out the size and ramifications of the off-label drug market. For those who wish I hadn't brought up this inconvenient information, take it up with Frakt, he's the economist who specializes in health care.

He's only proposing that patents not be offered for grandfathered drugs, not proposing that there be no approval process for new drugs.

Does anyone know why an ANDA is so expensive? It sounds about 10x higher than it should be.

It's not just the cost of the bioequivalence study(s) but also the manufacturing cost and associated regulatory costs of getting the facility inspected, etc.

What is "expensive." I am involved in NDA processes that costs hundreds of millions in clinical trials. $500 million is not at all surprising. $5 million for an ANDA is nothing.

Comments for this post are closed