Why does one new medicine cost billions?

One widely cited estimate puts the cost of bringing a single new drug to approval at about $2.6 billion.

That number sounds absurd. A pill costs pennies to manufacture. So where does billions of dollars go?

The answer is the single most important fact about this industry, and almost nobody outside it knows it.

You are paying for the failures

The money was not spent on the drug that worked. It was mostly spent on all the ones that did not.

Think about how a drug programme actually runs. A team picks a target. Chemists make thousands of candidate molecules. Most are useless. A few look promising in cells. Fewer still work in animals. A handful reach human testing.

Then human trials begin, in three stages, each larger and more expensive than the last. Phase 1 is a few dozen healthy volunteers, checking safety. Phase 2 is a few hundred patients, checking whether it does anything. Phase 3 is thousands of patients across many hospitals, and this is where the real money goes — often hundreds of millions of dollars for one trial.

And most candidates fail. They fail at every stage, including the last one, after nearly all the money has been spent.

That $2.6 billion figure is not the cost of a success. It is the cost of a success plus the share of failures it has to carry. When economists call it a capitalized cost, that is part of what they mean.

Why failing late is so expensive

Failing early is cheap. A molecule that does not work in a dish cost you a few weeks.

Failing in Phase 3 is a catastrophe. You have spent a decade and most of the budget, enrolled thousands of patients, and only then discovered the drug does not help enough — or that the target you picked was not really driving the disease.

Which is why the decisions made at the very beginning, long before any money is spent, matter so much. Choosing the wrong target is a mistake you may not find out about for twelve years.

Where this connects to everything else

This is the economic reason the evidence questions in Issues 2 and 4 are not academic.

If picking a target well could move the failure rate even slightly, it would be worth more than almost any efficiency anywhere else in the process. A programme that does not start is far cheaper than one that fails in Phase 3.

It is also why cheap, open, public evidence work is worth doing at all. Checking what the data already says costs almost nothing compared to finding out the hard way.

A caution about the number itself

Treat $2.6 billion as an estimate with real argument behind it, not a fact.

It comes from a specific 2016 study using confidential company data, and other researchers have published lower figures using different methods and assumptions. The disagreement is about what to count: failures, the value of money tied up for a decade, public research the company did not pay for.

Quoting the number without knowing that is exactly the sort of thing this newsletter is trying to talk you out of.

Try this

1. Go to ClinicalTrials.gov: https://clinicaltrials.gov
2. Search a condition, then filter by Phase 3 and status Terminated or Withdrawn.
3. Open a few and read the reason given for stopping.

These are real programmes that spent real years and real money and produced no medicine. There are a lot of them. That list is the thing the price of a drug is actually paying for.

One question to take away

If most of the cost is failure, is the way to make medicines cheaper to run trials more cheaply — or to be better at choosing what to put into a trial at all?

---

Source: DiMasi, Grabowski & Hansen, "Innovation in the pharmaceutical industry: New estimates of R&D costs," Journal of Health Economics, 2016. doi:10.1016/j.jhealeco.2016.01.012.