Sumatriptan tablets science commentary

Why are drugs so expensive?

The reason is not what you might think.
by T. J. Nelson


L ots of people are worried about high drug prices these days. Some of the newer antibody drugs cost hundreds of thousands of dollars per patient per year. But too many people are arguing in a vacuum. Without understanding the issue, their solutions could make the problem worse.

Of course some of these people are politicians, and making problems worse is their job. But there is no mystery why drugs are expensive. In this article I will explain why by walking you through a cost analysis of a typical drug.

I helped develop a drug for a major neurological disease which we're currently testing. I have no financial interest in it, and as a pure research guy I started out knowing nothing about how it's done. I assumed it was done simply and logically. But ... it isn't.

Pharma is one of the most regulated industries we have. People are extremely risk-averse when it comes to their health; the government's piles of regulations reflect that. The regulations do their job, but they also raise the price enormously.

The latest example is Turing Pharmaceuticals, which just raised the price of Daraprim (pyrimethamine), a drug for protozoal infections, from $13.50 to $750 a pill. Daraprim is generic, and sells for four cents a pill in India. That's 18,750 times cheaper.

Companies do this for a variety of reasons, the most common one probably being that their profits are declining because they stopped innovating and cut back on research to save money. Sometimes it's because they have a monopoly on their product. They get away with it because insurance often pays. Look for more of this to happen under Obamacare.

Why not just buy them abroad where they're cheaper? Unfortunately, it's a federal crime (21 USC 331) to import drugs from other countries. The FDA calls any drug that has not gone through their system an unapproved drug, and considers it a public health threat. They will seize the drug, and fine, disbar, or prosecute anyone who sells it.[1] Pharma supports this because imports would give foreigners who don't submit to the government's crushing paperwork an unfair advantage.

In a sense Turing is only taking advantage of the system. Once a drug goes off patent, any drug company can make it. But the FDA grants the first generic for a drug a 180 day exclusivity period. It also requires a company to demonstrate comparable pharmacokinetics as the original. Complying with the FDA's rules for manufacturing a drug requires a huge investment.

In a free market, Turing would have simply priced itself out of business. Indeed, that now seems to be happening thanks to the compounding pharmacy loophole. Suddenly, it seems, compounders are no longer evil. (See here for a typical example of the media's wrath against compounders.)

The bright spot is that these high prices do subsidize the discovery of new drugs. Here's a brief description of how that process works.

Step 1: Basic Research

Research is the area I know best. Basic research is what I enjoy doing more than eating. It is, of course, expensive, and the nature of science is it doesn't always lead where you want it to. Not all topics are good for research at any given time, but managers want us to cut corners and study things they can patent and sell quickly.

Quick research does not guarantee quick results, or even accurate ones, so when the economy turns south managers always cut R&D first. Increasingly the trend is to leave the basic research to academia. Industry scoops up the results ... they patent it ... and the money pours in!

As you might imagine, this sometimes causes friction between industry and academia, with industry making fatuous claims about the reliability of academic research and academia condemning industry for being profit-oriented and unimaginative. Both attacks are unwarranted.

Step 2: Hit the Target, Win the Kewpie Doll

The next step is to use all that knowledge and wisdom provided by the researchers to confabulate a theory about how to cure the disease. For example, if you believed cholesterol caused heart disease you might theorize that blocking cholesterol synthesis will reduce plasma cholesterol and thereby prevent it. The target in this case would be the enzyme that synthesizes cholesterol.

These theories are almost always wrong, because they're usually invented by people who need to raise funds or by people whose goal in life is to be famous and win the Nobel prize for curing something. But you have to have a plausible theory that's easy to understand, or the investors or whoever is controlling the funds will not pay for the next steps, and your colleagues will ignore your idea.

Eli Lilly just lost almost a hundred million bucks on evacetrapib, a drug that inhibits an enzyme called cholesterol ester transfer protein, or CETP, which converts LDL to HDL. Curing heart disease would be immensely rewarding as well as profitable, but this setback proves once again that we still don't understand cardiovascular diseases very well.

Sometimes people try to get around this. They have a drug they think works, and maybe it works in animals, most of the time, but they don't really know why. There's strong pressure to weave a complicated story out of nothing, by cherry-picking their data and doing bad experiments. Thankfully it's rare; but it's fear and ambition, not financial greed, that makes people do this.

Step 3: Gimme drugs

Next the chemists and biochemists come in and try to design a drug that has the desired effect. Nowadays, a lot of this happens on the computer, partly because it's cheaper and partly because there are so many bureaucratic obstacles to doing it the old way. We now have multipage forms to fill out before we can get animals or use radioisotopes. We even have forms to fill out before we can use chemicals.

Once you get a molecule, you test it against the target and in cells. If it still works you send it out for toxicity testing.

This step requires utter ruthlessness. If there's the slightest thing wrong with a molecule—it's too hard to make or it doesn't dissolve well enough, or the dose that causes toxicity is too close to the therapeutic dose—the drug is dropped. It's just not worth the risk. That means that about 999 out of every 1000 molecules are discarded.

If you still have a drug at this point, you patent it and the 20-year clock starts ticking. Patenting is expensive too. Patent lawyers cost 500 bucks an hour. They spend weeks on each patent. The drug has to be patented in the USA, then Europe, then Japan. (We don't usually bother with the rest of Asia because they don't adequately protect intellectual property over there.) The patent clerks always have some dumb objection, which means another week or so with the lawyers.

It doesn't help if the guy in charge has no idea about what claims he wants to protect. The lawyers pretend to get frustrated, but in fact they are as happy as pigs in mud when that happens.

In the USA, the clock keeps ticking even while the FDA is reviewing it. They're not in bed with industry as some people think, but they try hard not to turn them into enemies, either.

Step 4: Preclinical and clinical trials

I will skip over this part, because it's incredibly boring. Suffice it to say it takes the better part of a decade, and we have forms up the wazoo. Every step has to be approved: Institutional Animal Care and Use Committees, Institutional Review Boards, Investigational New Drug approvals, Certificates of Analysis, and on and on until we finally get to the New Drug Application, which is the government's way of deforesting the continental United States.

We had a company manufacture some drug for us. It cost a quarter of a million bucks for a couple thousand doses and they sent us a pile of forms about one foot high. I still have that useless pile of paper in my file cabinet. Its only purpose is to cover somebody's butt if there's a problem.

That manufacturing cost was 127 bucks a dose. The patient needed 18 doses. (To save you from draining your calculator, that works out to $2,286 per patient.) And that cost didn't include our rent, salaries, electricity, and the cost of all those little sticky Sign Here tabs that we use.

Most of this cost comes from government regulations. But we're only getting started.


If you want an honest discussion of big pharma costs, written for the layman, read Drug Truths: Dispelling the Myths About Pharma R&D by John L. LaMattina. LaMattina was a big R&D honcho at Pfizer, and he is impeccably honest about its strengths and ... as they say in industry ... their challenges.

An even better one is Leading Pharmaceutical Innovation by Gassmann, Reepmeyer and von Zedtwitz. They wrote that in 2006 big pharma spent $43 billion on R&D, but only 29 new molecular entities made it through the FDA. Get your calculator back out: that works out to $1,482,758,620.69 per drug.

Those big companies didn't spend that 43 billion for fun. Only 11% of new drugs survive clinical testing. Only one in a thousand molecules even make it that far. Most—often the best ones—are thrown away because they have some risk for some small portion of the population, or some other reason. Even if a drug survives, only 3 in 10 pay back their development costs.

Here's how the cost breaks down[2]:

Budget item Percent of total cost
R&D, licensing 20-40%
Manufacturing 15-30%
Administrative costs 5-15%
Marketing and distribution 20-30%
Margin 20-35%

That 35% margin might look high, but when you consider that 89% attrition rate, it doesn't look so good at all; and the margin is 0% if the drug bombs out. (Of course, marketing costs are lower too. But the R&D cost is the same.)

Patents are not the problem

Some people claim that weakening patent protection would make drugs cheaper. One commentator quoted MIT economist Benjamin Roin as saying that pharmaceutical companies refuse to develop drug candidates that lack strong patent protection.

You don't need an MIT economist to tell you that. It's no secret that drugs lacking strong patent protection don't get developed. Them babies belong to the generic manufacturers, not big pharma, but the generics are, in a sense, parasitical: they exist to take drugs that have already been tested by somebody else and approved by the FDA but have gone off patent. When a big drug patent expires, insiders call it a patent cliff. It means tens of thousands of layoffs; too many cliffs at one time means bankruptcy.

If you want someone to market unpatentable drugs, you have to figure out who's going to pay the hundreds of millions of dollars to test them. Push this issue too hard and you'll wake up one day to find a new government agency doing it with your tax revenue.

Patent trolls are a big problem in the computer industry, but they're hardly even on the radar in pharma. Licensing costs are pretty cheap, compared to everything else, because there's a one-to-one correspondence between new chemical entities and patents (or, there's supposed to be). I've been told that a big company can and will grind a small one into the dust if they want a patent badly enough. More often they'll just buy the company and fire all the employees to get the IP. Even so, no one who understands the industry would ever claim that patents stifle innovation. In fact, quite the opposite is true—as Roin's conclusion proves.

Others think price controls are the answer. These people think drug manufacturing should be a philanthropy. But they would be the first ones to criticize the industry for not developing enough new drugs after price controls kill innovation. Or they would complain because the new ones are too much like the old ones—so-called copycat drugs.

Capitalism is the answer

It might sound paradoxical, but when government subsidizes something, demand goes up because people think it's free. With price controls one of two things will happen: the industry will get effectively nationalized, in which case the costs will be hidden in your taxes. Or the industry will give us the finger and move to China. If you think the package inserts are hard to read now, wait till some guy from Anhui province starts writing them.

Brian D. Smith, author of The Future of Pharma, summed it up nicely by quoting Sir Richard Sykes, former chairman of GlaxoSmithKline, who said:

“Today, regulatory authorities are totally risk-averse and it is almost impossible to bring some molecules to market.”

As much as I complain about overregulation, those regulatory authorities aren't living in a vacuum. They're doing what people want. If you want to know why drug prices are so high, close the door on your medicine cabinet and take a look.

Update (Oct. 06, 2015) See here for the story of how government regulations caused the price of colchicine to jump by 20-fold. People clamor for more regulations, then are shocked when they cause the price to jump.

But often government's solution is not to get rid of the harmful regulations, but to add new ones to fix the problems created by the old ones. It's an endless cycle that ends with so many rules no one can remember them all—and every rule adds more to the cost.

Turing is indisputably an example of greed. But the beauty of capitalism is that it turns greed into a force for good. It is the only system that can do this. So if you want lower prices, the answer is simple: bring back the free market.


[1]. The only exception is if your U.S. doctor certifies you began treatment overseas, and if your condition is serious enough, and if there is no unreasonable risk, and if the drug is unavailable here at any price, and it's not a controlled substance, you might be allowed to bring in a three month supply for personal use.

[2]. Swiss Health Care and Pharmaceutical Market, 2002, Interpharma, Basel. Cited in Leading Pharmaceutical Innovation by Gassmann, Reepmeyer and von Zedtwitz, p.2

See also:

Science commentary
Why You Should Feel Bad For Big Pharma

Science commentary
Why Do So Many Drugs Fail?

Bad Pharma: Fact or Myth?
The pharmaceutical industry is dying. When it's gone, some of the blame for its loss will be laid at the feet of the authors of all those pharma-bashing books.

Book reviews

Essential CNS Drug Development

Drug Truths: Dispelling the Myths About Pharma R&D by John L. LaMattina

The Future of Pharma: Evolutionary Threats and Opportunities by Brian D. Smith

Leading Pharmaceutical Innovation: Trends and Drivers for Growth in the Pharmaceutical Industry, 2nd ed

Guidebook for Drug Regulatory Submissions by Sandy Weinberg

Clinical Trial Methodology by Peace and Chen

Fundamentals of Clinical Trials Friedman et al

Principles and Practice of Clinical Trial Medicine

Name and address
oct 04, 2015; last updated oct 23, 2015