Monthly Archives: April 2019

Buspirone Shortage In Healthcaristan SSR

(Epistemic status: Unsure on details. Some post-publication edits 5/1 to make this less strident.)


There is a national shortage of buspirone.

Buspirone is a 5HT-1 agonist used to control anxiety. Unlike most psychiatric drugs, it’s in a class of its own – there are no other sole 5HT-1 agonists on the market. It’s not a very strong medication, but it’s safe, it’s non-addictive, it’s off-patent, and it works well for a subset of patients. Some of them have been on it for years.

Now there’s a national shortage. My patients can’t get it, or have to go hunting from pharmacy to pharmacy until they find one that has it. I’ve told people find a source to stockpile a supply so they don’t run out. It feels like we’re living in the Soviet Union.

How did this happen? The New York Times writes:

The main reason for the buspirone shortage appears to be interrupted production at a Mylan Pharmaceuticals plant in Morgantown, W.Va., which produced about a third of the country’s supply of the drug. The Food and Drug Administration had said the facility was dirty and that the company failed to follow quality control procedures.

So the FDA shut down a major buspirone factory. But government agencies – ones that are a lot less nice than the FDA – shut down methamphetamine factories all the time without creating methamphetamine shortages. Why is the buspirone market so vulnerable? The Times again:

Rock bottom prices for some generic drugs are also contributing to the crisis. Consolidation among wholesalers has led to the creation of three buying consortium behemoths that purchase 90 percent of the generic pharmaceutical products in the United States, said Adam Fein, a consultant and chief executive of Drug Channels Institute. These “monster” buyers have squeezed manufacturers on prices, and “some of those generic manufacturers are deciding the profit is so low they can’t make money, and they’re exiting the category,” Dr. Fein said.

Is this really how economics works? There’s a medicine that millions of people desperately need? But nobody will produce it because they can’t make a profit? Huh? Isn’t the usual solution to just raise the price? And people will buy it at the higher price, because they need it so badly? And then you will make more profit, and can keep on making the medication? Isn’t “nobody will supply this product, it’s too cheap” just the economics version of “nobody goes there anymore, it’s too crowded”?

Sure, generic drug manufacturing is pretty consolidated. Most individual generic drugs are now manufactured only by one or two companies. If one of those few companies gets greedy (like Martin Shkreli did with Daraprim), they can increase prices by orders of magnitude without a lot of competitors to push back. And if one of those few companies suffers a shock (like the FDA closing the buspirone factory), it makes sense that there might not be enough competitors to pick up the slack.

But how come this is only happening in pharmaceuticals? How come (in capitalist countries) there are almost never meat shortages, bread shortages, laptop shortages, or chair shortages? Is there something unusual about the pharmaceutical landscape that predisposes it to this sort of thing?

I am not an expert in this area and may be getting some of it wrong. But from Berndt, Conti, and Murphy (2017) and a Berndt, Conti, and Murphy (2018), I gather that a big part of the story is the Generic Drug User Fee Amendments (GDUFA) of 2012 and 2017.

The story goes something like this. The FDA demanded that generic drug manufacturers pass FDA inspection before setting up shop. But the FDA didn’t have enough inspectors to review manufacturers in a timely manner. So companies kept asking the FDA for permission to enter the generics market, and the FDA kept telling them there was a several year waiting period. In 2012, Congress recognized the problem. Politicians, FDA officials, and industry leaders agreed on a new policy where generic drug manufacturing companies would pay the FDA lots of money (about $300 million last time anyone checked), and the FDA would use that money to hire inspectors so they could clear their backlog of applications.

The good news is, the FDA hired lots more inspectors and they are now pretty good at responding to generic drug applications in a timely way. The bad news is that the fees to the companies were designed in a way that subtly encouraged monopolies in generic drug markets. I don’t understand all the specifics, but there seem to be two main problems.

First, if you manufacture a drug, the FDA will charge you a fee, but the fee doesn’t scale linearly with how much of the drug you produce. So suppose Martin Shkreli owns a very big Daraprim factory. The FDA might charge him $1 million per year to fund their inspectors. Suppose you are a small businessman who is angry at Martin Shkreli’s fee hike, and you want to open a competing Daraprim factory in your small town, using your small amount of personal savings. Probably your factory will be much smaller than Martin Shkreli’s. But the FDA will still charge you the same $1 million per year. At worst this means you make no profit; even at best, Shkreli’s economy of scale gives him a big advantage over you. So you may decide not to enter the market at all. From the second paper:

President of the Pharma & Biopharma Outsourcing Association, Gil Roth, remarked, ‘We have a single generic client that we do a short run of production for. Why are we charged the same as a Teva facility that pumps out a billion tablets?’ Another commented, ‘At least a flat tax is based on a percentage, either of revenue or profit. This is a flat fee, which makes it a regressive tax on smaller businesses, both contract manufacturers and small generics companies’

I think the fee might even be per factory, which encourages companies to concentrate all their manufacturing at a single site – like the Mylan one that just got shut down, thus affecting the whole country’s buspirone supply.

Second, traditional economics suggests that if some company has a monopoly on a product that people really need (like a medication), they will charge very high prices. But many generic drugs are produced by only one company each – and Shkrelis aside, most of them charge affordable prices. Why? Berndt et al argue it is because of the possibility of competition: if Shkreli raises his prices too high, some other company can move in and undercut him. But FDA licensing procedures make this undercutting harder than it could be: it will take months to years, and thousands to millions of dollars, for the other company to move in (at which point Shkreli can just say “Haha, no” and lower his prices again, meaning the undercutter would lose all the money they put in).

Historically, the system has worked anyway – because lots of companies are sitting on pre-existing FDA approval to make certain drugs. If a company had ever made a drug in the past, they had FDA permission to make it again whenever they wanted. So if Shkreli raises prices on Daraprim, some other company that made Daraprim ten years ago can set up a new factory tomorrow and undercut him. This helped prevent would-be Shkrelis in most markets, and provided a safety valve for shocks like the one creating the buspirone shortage today.

But GDUFA weakened this system by mandating that any company with FDA approval to manufacture a drug pay yearly inspection fees to the FDA, whether or not they were actively manufacturing it. That turned FDA approval for drugs you weren’t actively manufacturing into a liability; you were paying fees, but not making a profit. Companies started voluntarily cancelling their FDA approvals for older drugs so they wouldn’t have to pay the fees. That meant monopolists lost a lot of their potential competition. And that cleared the way for people like Shkreli to hike prices.

You get more of what you subsidize and less of what you tax. Unfortunately, the FDA is inadvertently taxing companies for being in the generic drug business. And it’s taxing them more if they’re not a monopolist with economies of scale. That means we get fewer companies in the generics industry, and more monopolists.

So my very tentative guess as to why buspirone is more plagued by shortages than bread or chairs is because number one, the need for FDA approval makes it hard for new companies to enter the buspirone industry, and number two, the FDA’s fee structure favors large-scale monopolies over small-scale competitors.


The price of insulin is much too high. Vox argues that this is because of the “lax regulatory environment” and the “free market approach”, and that if we could just become socialist like all of the cool countries, everything would be fine.

Insulin is off-patent. It was discovered almost a hundred years ago. But somehow, all the insulin sold in the US is brand-name. This is shocking and obviously the root of the problem. What’s going on? Vox links NEJM’s Why Is There No Generic Insulin?, but summarizes it by saying it’s “because companies have made those incremental improvements to insulin products, which has allowed them to keep their formulations under patent” and because “older insulin formulations have fallen out of fashion.”

I am not diabetic. But if I were, I don’t think I would worry that much which kinds of insulin were vs. weren’t fashionable. What’s really going on?

Here my source is partly the NEJM paper above, but also Health Affairs’ Biologics Are Natural Monopolies. Both agree that the key point is insulin’s nature as a “biologic”. It’s not a simple molecule you can make with a chemistry set. It’s a complex peptide hormone of about seven hundred atoms, arranged in a series of helices and threads and tentacles. The only way to manufacture it is to genetically engineer some microorganism to make it for you.

The FDA usually requires generic manufacturers to prove that their drug is identical to the brand name drug they’re copying. But genetic engineering is hard, microorganisms are uncooperative, and insulin is too complicated to say with certainty that any one insulin molecule is “identical” to any other. So the FDA has lowered their standards for biologics to require proof that a generic biological is “similar”.

But even proving biosimilarity is orders of magnitude tougher than anything that small molecules have to go through. From the Health Affairs article, slightly transposed for readability:

The typical [small molecule] generic drug takes firms 1-3 years, $1-$5 million, and no human clinical trials to introduce. [In contrast], the entire biosimilar development process has been projected to span 8-10 years and cost upwards of $100 million. Human clinical trials involving hundreds of patients can cost $20-40 million to simply confirm that the candidate biosimilar generally replicates the reference product’s short-term positive and negative clinical effects.

Brand-name insulin companies make a bad situation worse by patenting their manufacturing techniques, using different patents than the drug patent, which may still be in effect when a generics manufacturer is trying to copy their drug. For example, Sanofi has somehow managed to get 74 different patents on their Insulin Lantus, which this I-MAK report describes as a “patent thicket”. Many of these patents seem to be totally illegal, and exist only so that it would cost a generics company time and money to challenge them in court. Most generics companies look at the process of trying to prove their hideously complex molecule is “biosimilar” to Sanofi’s hideously complex molecule, without using any of the 74 different manufacturing processes Sanofi uses to make it, and decide against entering the market.

Then the FDA mandates that biosimilars have a different name than the product they are replacing (ordinary generic drugs may not use the trade name, but can use the same chemical name). This makes it harder to have prescriptions for one cover the other, and doctors may have too much inertia to switch to a new drug with a new name. This limits potential sales for these products.

So the reason companies aren’t making generic insulin is that the FDA approval process for generic insulin is very onerous, brand name companies have excessive and illegal patents that make the approval process even worse, and companies’ ability to sell what comes out the other end is limited.

I realize my political slant makes me blame poor regulatory choices for these sorts of things pretty often. And the Health Affairs article I’m drawing from makes a different argument than I do, arguing that their biological properties make insulins “natural monopolies” and that policy choices are only secondary to this. You should consider my biases before you necessarily take my words at face value.

But the NEJM article mentions that plenty of poorer countries do have biosimilar generic insulins, including such gleaming-high-tech bastions of cutting-edge pharmaceutical excellence as Peru. Which of the following do you think is true?:

1. Peru has better technology than the US, and so is able to make cheap biosimilar insulin using processes that our own scientists and engineers can’t manage.

2. Peru has a bigger market than the US, so there’s more money in creating generic insulin to sell to Peruvians than there is selling it to Americans.

3. Peru has a better regulatory environment than the US, and this is enough to make producing biosimilar insulins cheap and easy.

Extreme fringe libertarians have a certain way with words. For example, they call taxes “the government stealing money from you at gunpoint”. This is a little melodramatic, but words like “patent loopholes” and “onerous review processes” sound a little bloodless for something that probably kills thousands of diabetics each year. So I would like to take a page from the extreme libertarian lexicon and speculate that the problem with insulin costs is that the government will shoot anyone who tries to make cheap insulin.

And then Vox writes an article saying that the problem is “the free market” and we need more government intervention. Fine, whatever. I have despaired of anyone ever analyzing this topic in any greater depth than that. The drug situation is going to keep getting worse – for small molecules, for biosimilars, for whatever. People are going to keep blaming “the free market” and implementing more poorly-thought-out regulation. And so the cycle will keep going, ad infinitum. Vox will keep writing this article once a year or so, and I’ll keep telling them they are bad and wrong, and we’ll both get some clicks out of it. The system works!

I want to clarify that I’m not criticizing the current FDA administration. The FDA has recently done a great job trying to shift their processes marginally in the direction of approving more medications, approving more companies entering the generic market, promoting more competition, and generally doing everything right. Even the GDUFA was a step in the right direction, in that it was necessary in order to get generics approved at all. This is probably part of why drug prices are starting to drop (note that there’s a complicated debate over how true this is and what statistics to use, but I think even the skeptics agree the trend is positive, and they are rising less quickly than they have in the past). There are probably some small steps they could still make to improve things – I get the impression that having the government pay for FDA inspections using tax dollars instead of having the distortionary GDUFA system would help. And patent reform would be great. But a lot of this is concessions to political reality that are probably outside the FDA’s control.

The current trends are good, and further small fixes could be better, but they probably aren’t enough to make drugs affordable and consistently available to patients. If this is even possible, it’s going to require more dramatic changes – not just having good regulators who try to make the best of the current system, but reforming the system entirely. This is a tough order (and I’ll try to blog later about what might be involved). But it’s the only thing that I can imagine allowing us to eventually catch up to Peru.