• Government & Policymakers
• Francis Collins, M.D., Ph.D.
• Janet Woodcock, M.D.
• Gregory Downing, D.O., Ph.D.
• Michael O. Leavitt

Dr. Janet Woodcock has dedicated nearly 20 years of service to the U.S. Food and Drug Administration (FDA). She is currently the Director of the Center for Drug Evaluation and Research (CDER), the drug approval arm of the FDA, a position she has held since March 2008. In her previous role as Deputy Commissioner, Dr. Woodcock oversaw the Office of the Chief Medical Officer. Prior to the creation of this new office in January 2007, Dr. Woodcock served as the Deputy Commissioner for Operations. Dr. Woodcock also served a previous term as Director of CDER from 1994 to 2003.

Throughout her tenure at FDA, Dr. Woodcock has tackled some of the toughest issues facing the agency. She has worked with her colleagues in conjunction with industry on a "critical path" initiative designed to improve the process of drug development. She has also spearheaded the agency's efforts to improve the way drugs are prescribed by using genomics to understand how and why individuals respond differently to the course of therapy.

In this interview posted in March 2007, Dr. Woodcock discusses some of the early steps that the FDA and the pharmaceutical industry have taken to advance pharmacogenomics, the study of how variations in the human genome affect the response to medications. She also addresses the pivotal role that the FDA plays in this growing field, as well as her thoughts on the future of personalized medicine.

The Building Blocks

Q: Drawing mainly from your lab's emphasis on diabetes, how are scientists today utilizing the information from the HGP and the HapMap to find the causes of complex diseases and reach toward treatment?

A: This is an enormously exciting time for unraveling those mysteries. My lab has worked on the problem of type 2 diabetes for more than ten years. We made some steady but slow progress over that decade. Now, we are exhilarated to engineer a leap forward that is unlike anything we have dreamed of being able to do.

Q: How has the Critical Path Initiative — the FDA's vision originally outlined in 2004 for improving the process of product development in the drug industry — progressed?

A: We released the critical path opportunities list and report here in March 2006, and it's been very well received. It has 76 different projects that we feel, if completed, would really enhance drug, device, and biologic product development in specific areas. We're working in a number of consortium arrangements and partnerships around the country to get some of these projects done, and I think the various industry groups and patient groups are really on board with the initiative and are making significant efforts to get collaborations going. We have some protocols that we are beginning to implement in specific areas as well. So, in general, I think that the initiative has gotten further in two years than one would have thought it might.

Q: What have you learned in the year or so since the FDA published the Pharmacogenomic Data Submission Guidance, a document whose purpose was to help integrate pharmacogenomics into drug development and also to encourage scientific progress in the field?

A: We've had very disparate types of submissions, so while you couldn't put them all into a database and say they're all analyzable together, they do show the power of genomics in a very broad range of situations and development. We are also learning about the difficulties in analyzing the data.

Q: Are standardized and advanced information technology systems critical to the adoption of personalized medicine?

A: Yes they are, and at FDA we have done a lot to advance standards in this area. For the Agency, many levels of IT transformation are needed. Ultimately, we need to have much better scientific computing capacity. But right now, in general, we're just trying to transform our basic IT infrastructure to automate our business processes.

What most people think about, though, is that FDA has all this clinical trial data and all this genomic and other information. We need to start building electronic databases to handle this so that we can do analysis that will really help us advance the field. There's a lot that has to be done: for example, you have to be able to phenotype the people as subjects, you have to understand their own personal characteristics very clearly, and you have to be able to match that with their genetic characteristics. You just can't do that on little bits of paper.

The Role of the FDA

Q: How can the FDA serve as a knowledge center for this type of information?

A: We're trying to set up a consortium that has the proper intellectual property protections so that it can do a lot of these activities. An example is the digital electrocardiogram (ECG) warehouse we've set up. We had to establish a standard for digital ECG submission. And then we created a database. We have over 200,000 ECGs in our database now. What we need to do now is set up other arrangements that will allow analysis and research to be done on that database. We could do that in other areas as long as we have the suitable confidentiality protection.

Q: How did the FDA change its traditional approach to become more proactive with respect to pharmacogenomics?

A: I think there was a core of people who felt that this was the right thing to do. It was actually stimulated by the industry saying, we're really concerned about doing this new science unless we understand how the regulators are going to interpret it.

Q: How has the FDA worked with other government agencies to promote personalized medicine?

A: Well, a good example is the Interagency Oncology Task Force (IOTF), between the FDA and the National Cancer Institute (NCI). IOTF has done a huge amount of work on different topics, many of which relate to personalized cancer therapy. We formed a spin-off called the Oncology Biomarker Qualification Initiative, among the NCI, the FDA, and the Centers for Medicare and Medicaid Services (CMS). We hope to key up a whole series of projects where we evaluate biomarkers to be used in concert with cancer therapy to individualize or target the therapy, or to stop the cancer process.

The Future

Q: What will the future of clinical trials look like in the age of personalized medicine?

A: The actual fact of personalized medicine means that the classic way we do the randomized control trial now, with a population-based endpoint, is going to have to be changed. We need to retain the features of clinical trials — such as randomization and the ability to make randomized comparisons — but also introduce methodologies and statistics that allow us to analyze individual responses, and then adapt trials more quickly as we learn the information. We're going to need to modify trials to be able to ask and answer a lot of questions, and also adjust to different individual characteristics. Once we learn enough about how to do it, it will seem the natural thing to do.

Q: What can the FDA do to further encourage the adoption of personalized medicine?

A: For drugs in development we're trying to create a Guidance on co-development of a drug and a diagnostic together. For drugs that are approved, we're cooperating with different groups to try to get some proof of concept trials done. There are several reasons why molecular diagnostic tests aren't widely used right now. One reason, of course, is lack of familiarity, but the second is that people are skeptical that the tests will actually improve therapy.

Q: Since public opinion often vacillates between demanding greater scrutiny of drug safety and desiring faster development of lifesaving treatments, how do you think personalized medicine will affect the public expectations of the FDA?

A: One size does not fit all when it comes to drugs. This means that people who don't benefit from therapy get harmed and then think that therapy should be taken off the market. And then people who needed that therapy but have no way to get it become very resentful. My theory for a number of years has been that we have to use science to get out of this box. No drug is all bad or all good — it's about who's going to have a positive response to a drug and who will have a suboptimal response. Personalized medicine, in my view, is the key. But we need to find better predictive safety biomarkers and other ways of predicting adverse events. Also, and equally important, we need to have markers of positive response.

Q: How do you think personalized medicine might evolve over the next several years?

A: We're going to see more targeted cancer therapies. I think cancer is leading the way because the cancer genome is mutated, and you can study the mutations and their cellular consequences. We'll see an increasing number of targeted therapies looking at either cellular expression, or gene sequencing. At the same time, we're going to see genetic tests that stratify the risk of recurrence or risk of aggressiveness of the tumor.

Q: And then what?

A: Then, I think, there would be parallel scientific activities going on with pharmacogenomics —trials where people are going to be looking at dose modification based on newly available genetic testing for drug metabolism and enzymes. At the same time, there's a whole other area that isn't widely explored yet — the genetic basis of drug response.

Q: How will scientists explore that?

A: That's going to come more from "fishing expeditions," as I call it. In all kinds of common treatment, researchers are going to have everyone's genomes and they're going to match them up with their responses in clinical trials. And then we're going to have this large amount of data that we don't know how to interpret, and we're going to have to do a lot of validation trials to determine what the really important variables are that modify drug response for good or ill. I think we'll see this emerging, and we'll have to develop consortia or other mechanisms whereby we can do this kind of validation work.

Q: And ultimately?

A: I think all of this will happen. Five years from now we're going to look back and think we were extremely naïve and uninformed at this point. We're going to uncover this tremendous amount of new knowledge. Everyone says, and I partly agree, that the diffusion of this knowledge into clinical practice and getting the diagnostic tests available in a reliable manner are going to be real challenges for us to work on, and ultimately overcome.

View of the Expert posted March 2007