Humer [Laughs] That is a good question. I think in general, not only the timeline, but also the risk of drug development, is underestimated. It takes us somewhere between 8 to 12 years to develop a new drug, and it may well be that only after 9 years of development you find out that the drug doesn't work or has unwanted side effects. The risk is often underestimated. This comes up plainly in, for instance, statements that because now we can test for SARS, we should have a treatment for SARS in a couple of months. It just cannot happen that fast. It is even difficult for experts to predict the timeline. In the early '90s I heard some of the world's experts in virology say we would have an AIDS vaccine in 5 years. We are 10 years on, and there is not an AIDS vaccine in sight because of the complexity of the virus.

American Way Of the good ideas that researchers generate, not just in your labs but any labs, how many of those ideas actually become drugs that go on the market?

Humer If there are 100,000 substances produced in the pharmaceutical labs, perhaps 1 will become a drug.

American Way It's like buying lottery tickets.

Humer [Laughs] Well, it's not quite lottery tickets. I think there's a little bit more sophistication behind it.

American Way There's also a terrific payoff when a drug works.

Humer But if you do the calculation properly, it also shows that only one out of four drugs that reaches the market creates a decent return for the company. Very often the drug that you have developed is not for the disease you started out thinking it was for. It may not cure, or be able to treat, as many patients as you thought. Not all drugs reach the so-called blockbuster status [where a drug achieves annual sales topping $1 billion].