Making Medicine — Ten years, and most of them fail

Thousands of candidate molecules go into the funnel. A handful reach human trials. Most of those fail, often late and expensively, because they turn out to be unsafe or simply no better than what already exists. The price of an approved drug is mostly the price of the failures that came before it.

• Trials run in phases: safety in a few dozen people, then whether it works, then a large comparison against current treatment. Each phase is bigger and costlier than the last.
• A generic is the same molecule once the patent expires. That is why an identical drug can sell for a small fraction of its original price a few years later.