R&D: A Merck in Progress

Gary Herman, head of the new experimental-medicine division, is teaching Merck's storied scientists to fail...faster
Sep 01, 2007


That was the collective vow sworn by Big Pharma last December following Pfizer's $1 billion–down–the–tubes withdrawal of the cholesterol compound they had touted as the most important drug of the decade. The question is, What's the right organizational construct to support innovation—or at least to stop Phase III failures?

Gary Herman, executive director of Merck's Department of Experimental Medicine. Herman spearheaded early development of Januvia and is now bringing his ideas to additional therapeutic areas.
The large-cap pharmas are all carrying out interesting experiments in how they do R&D, hoping to save time and money at no cost to quality. With Januvia, its novel type 2 diabetes drug, Merck has shown that it can be done. With a new CEO, who believed in bringing all the voices in the company together as one, and the implementation of an integrated new R&D model, Merck got Januvia to market in record-breaking time. Now, a mere six years after it was born, the drug stands to earn $762 million in sales.

That achievement prompted us to start Pharm Exec's "R&D Innovation" series by interviewing Gary Herman, the executive director of the firm's year-old Department of Experimental Medicine. Herman spearheaded the Januvia early-development program—and is now charged with working his magic across Merck's many therapeutic areas.

Lesson number one: Change is never easy. And the process of getting better products out faster has as much to do with corporate culture as it does with science. Merck's storied scientists must now work in new ways—collaboratively, and with a ready acceptance of failure and ability to move on...quickly.

When did you first realize that the old way of doing R&D was no longer working and needed to be fixed?

It's been an evolution. Back in 2001, I started work on our dipeptidyl peptidase IV inhibitor—which became Januvia. We were significantly behind in this area of type 2 diabetes. We began to think about how we could do things much faster to determine whether our compounds were working. We used single-dose studies in patients with diabetes, giving them a glucose challenge. We then measured all kinds of biomarkers to figure out what the optimal dose would be. We were able to shave more than a year off of our development timeline. With Januvia, there was a company-wide recognition that we could really do things much faster.

How long did it take to get Januvia from lab to market?

The compound was born in early 2002. It was on the market at the end of 2006.

Four years—that's less than half of the drug-development process.

People worked very hard. It was prioritized. We recognized we could use experimental medicine to make a much faster determination about whether a compound was going to work. In fact, most drugs that go into man are not going to work—only about 8 percent end up working.

That was when I started to develop an interest. Lots of synergies were going on in the company. In 2004, 2005, there was a decision to have some of us working full-time trying to develop and integrate the discipline into how we do drug development.

Are there pivotal points in the discovery process where the biomarker approach is most useful?

The sweet spot is early in clinical development: After you've gone into human beings, you try to determine as early as possible whether the medication is likely to have a benefit. You're getting some confirmation of the effect on the biology that things are happening in a desirable way, but it's not proving this is doing everything it's going to do as a medication.

lorem ipsum