News & Events
Scott Gottlieb, MD - Northwestern University Medical Center
This text contains Dr. Gottlieb's prepared remarks. It should be used with the understanding that some material may have been added or deleted during actual delivery.
2006 Clinical Research Educational Conference
Northwestern University Medical Center
Scott Gottlieb, MD
Deputy Commissioner for Medical and Scientific Affairs
May 18 , 2006
Thank you for having me here today.
We are entering a once-in-human-history opportunity for progress in the life sciences and their application to human health. We have seen life expectancy come close to doubling in the last century, from the mid-forties to the long life expectancies that await many of us here today. And many are convinced that when our present history is written years from now, one of the most significant pieces in that history will be the ability of all of us for the first time to comprehend the origin of diseases, to influence human health, and to develop on a scientific basis the means of preventing and even curing disease.
A lot of this progress owes to recent advances in genomics and proteomics, and the onset of a stage where we can develop therapies for disease that are based not on trial and error, but on rigorous deductive methods.
Over the last decade or so, this promise and the reality of basic biomedical research has been so exciting, so captivating that it has received the lion's share of attention from the public.
But the fact is that without a robust clinical research enterprise, much of what we learn through basic research cannot be applied effectively and efficiently to advancing the public health. And improving the public's health is what biomedical and health sciences research is ultimately all about.
The good news is that we are seeing far more emphasis on translational research, on taking the results of science directly to patient care.
But even while we have strengthened the bonds between clinical research and the practice of medicine, we still need to do more to also strengthen the bonds between basic science and clinical research, and to take advantage of modern tools and approaches to collecting and communicating clinical information.
Not enough has been done, or is being done, to modernize our approaches to clinical research, to take advantage of new scientific tools and new discoveries that can improve the way that we evaluate new medical products.
Some of this is because the tools that we use in the drug development process itself are in many cases outdated, whether it is a clinical trial endpoint or the way the trials themselves are being designed and executed.
Even while we have poured billions into the science of discovering new drugs, we have invested comparatively little in the science of how we go about actually testing and evaluating those drugs for safety and effectiveness. This is a problem that we are addressing at FDA, under the leadership of FDA’s Chief Operating Officer Dr. Janet Woodcock and others, through our Critical Path initiative, something I want to get back to in a few moments.
But even while we work to improve the actual scientific tools being used in the drug development process -- whether it is the validation of PET imaging as a better marker than plain film x-rays for measuring tumor shrinkage in trials of lymphoma drugs, or the creation and corroboration of a new assay for measuring drug side effects like liver and heart toxicity -- there is another place where we all need to work together: to develop better scientific tools and approaches for doing clinical research itself.
The fact is we are failing to modernize our approaches to clinical research, not just in how we collect clinical information but how we communicate it.
We – all of us – are dealing with four problems
First, in many cases, are not taking advantage of modern tools of information technology to improve the way we conduct and monitor clinical trials.
Second, in many cases, we are not taking advantage of these same tools to improve the way that we collaborate.
Third, we are not evolving the way that we structure clinical trials to leverage new insights into biology and genomic and proteomic predictors of peoples’ propensity to develop disease or respond to a particular treatment.
Finally, in many cases, we are not incorporating new information tools to improve the way that we communicate and share the results of clinical research.
We cannot afford to ignore these challenges. With the cost of developing new medical products continuing to increase, by some estimates it now costs more than $1 billion to develop a single, novel new drug, and with more and more potential breakthroughs taking longer to reach patients who need them, we need to take new steps to make the clinical research process more efficient and effective.
We simply must look at how we can leverage new technologies, new collaborations, and new scientific opportunities to improve the clinical research enterprise, if all of that basic bioscience being developed on the front end is going to be translated into safe and effective medical solutions for patients. Otherwise, clinical research itself may well become the bottleneck in the future to taking advantage science to improve health.
I want to talk about each of the four critical areas I outlined where we need to take new steps to improve the clinical research process, and tell you about some of the things that FDA is doing to address them by making better use of modern tools and approaches.
And I want to tell you where I believe we are going to need your help.
First, on improving the conduct and monitoring of the clinical trials themselves:
As clinical trials continue to evolve, in particular becoming increasingly large, decentralized and global, the FDA’s approach to bioresearch monitoring and human subject protection must also evolve and modernize.
Our regulations were written and structured and put in place many years ago, and since then clinical trials have changed a lot.
We need to make sure we are continuing to provide appropriate, risk-based safeguards, while also making sure that FDA is not an obstacle to progress in the design and conduct of clinical research.
There are new trial methods and designs, new methods of data collection, and new arrangements between sponsors and various contractors, among investigators, among institutions and among institutional review boards (IRBs). There is the rise of free-standing, for-profit study centers.
Many of these are positive developments. But these changes, the development of the clinical trials enterprise, challenge us to also adapt our approaches to how we oversee the clinical trials process. We need to modernize our approaches to accommodate and leverage this new landscape for clinical research.
So we have undertaken a new initiative called BIMO to do just that. I’ll give you some examples of some of the questions we’ll be working to answer.
For one thing, less frequently are clinical trials conducted at a single site, or a few sites. Often trials will involve 30 or 40 different sites, or more. Is it possible, reasonable or appropriate for a single principal investigator to oversee all those sites? How should FDA ensure that there is adequate and appropriate supervision at all sites? And, if an individual other than the PI is performing and integral and essential role in the trial, should that individual not be responsible and accountable to the FDA? We will be asking these questions, and also developing guidance on what we view as adequate supervision by the principal investigator to help address this latter question.
Another question we will be asking is what does it mean to keep a record electronically, where results like ECGs might be generated and fed directly into a central electronic database? In these cases, who is responsible for looking at these results to make sure that there is no medically urgent information contained on it? And what does it mean to retain more than one record in that case? These and other questions become increasingly urgent as we all move more fully into the electronic era.
Other things we might consider are more real time inspections of clinical trials sites rather than post-hoc audits, and how we can better capitalize on inspections we know industry is already doing, to better leverage these resources.
We will also be providing common standards and regulatory requirements for electronic data handling. And we will be developing guidance to better delineate the proper role of IRBs, and specifically on what adverse event information we think should be reported to IRBs from clinical trials.
Now for the second issue: collaboration.
To improve collaboration across the clinical research enterprise, we need to take new steps to make more effective use of information technology in order to create easier, less expensive ways of collecting and sharing clinical data.
Collaboration in medicine, and clinical research in particular, is increasingly done across electronic networks, facilitated by advances in medical computing and the growing adoption of electronic health records. To help capitalize on these advances, at FDA we are working on clinical research data standards and standards to help facilitate the exchange of data.
We are also working to harmonize these activities with the growing use of electronic health records. And we are taking new steps to standardize the input of a lot of the clinical trial information, such as the case report forms, to make collecting and sharing and reporting information easier and more effective.
In addition, insufficient attention has been paid to incorporating clinical research findings into clinical practice. Integrated health systems and practice-based networks also offer rich potential for the kinds of population-based research in "real world settings" that form the foundation of evidence-based medicine.
We need to more quickly adopt strategies to take advantage of these opportunities, including promoting mixed academic and non-academic clinical trials, and creating incentives that encourage collaboration in clinical research among academia, industry, private health providers, and managed care. Most of this does not involve FDA regulation or things that we can do at FDA to help facilitate the exchange of information, but sometimes it does.
We will have more to say on the subject of population-based data soon, including guidance we hope to be able to develop that will outline good principles for making rigorous use of the growing collection of population-based data being accumulated in large databases, especially among payers.
With the growth of large population-based databases, there is an increasing temptation to do retrospective studies that are sometimes underpowered and poor quality and not as useful as possible. We hope to be able to issue guidance on what we believe the proper and rigorous use of epidemiological databases is, especially for finding important safety signals.
On the third point I outlined at the outset, we need to take advantage of new discoveries in basic science such as genomics and proteomics to modernize the way that we approach clinical research and the way that we design clinical trials.
In recent years, we have seen a lot of advances in our ability to use tools such as gene and protein markers to help predict not just who is at higher risk to suffer from certain health events, but also, in some cases, who is most likely to benefit from a certain medical intervention.
We need to leverage this new scientific information to improve the way we structure and conduct clinical research. For example, early on, when you are conducting a clinical trial to investigate a new medical product, you want to do studies in people who have a lot of the events you are interested in studying, and if possible people who stand a better chance of realizing more of the benefits, and fewer of the side effects of a drug you are giving them.
The genomic revolution and the proteomic revolutions are already making these things possible. But we will need to work together to enable these public health opportunities to inform how we work as clinical researchers, and how we work as regulators. This is the most immediate impact that genomics and proteomics will enable – the ability to identify people at greatest risk of bad outcomes and to identify people capable of responding to a treatment, both related to a growing mastery of mechanistic data rather than continued reliance on the highly empiric approach that we use today.
There are dramatic examples of how making better use of this science in the clinical research process can enhance efficiency, and improve health outcomes.
In oncology, the classic example of where these kinds of tools can improve clinical research is the drug Herceptin. This drug, for the treatment of breast cancer, was evaluated in clinical trials in people with the HER2 receptor; the receptor toward which Herceptin is directed. This allowed researchers to avoid testing the drug in patients who would be likely to experience benefits but could still experience the drugs potentially serious cardiac adverse effects.
Also if the original trials had included the roughly two-thirds of people who were not HER2 positive, it would have been harder to show that drug worked at all, requiring a larger, longer clinical study than the one FDA needed ultimately to prove that the drug was highly effective. By focusing on people with factors like HER2, who are pre-selected because of their ability to respond well to an intervention, you make the effect size larger which means you can study a smaller population and expose fewer patients to experimental treatments.
Another example in oncology is with the drug Tarceva. While there are ongoing debates about what measurement of EGFR correlate best with response to that drug, at a more basic level, it is widely appreciated that measurement of the amount of EGFR receptors on cancer cells correlate with response to the drug and this is reflected in labeling.
There are also many indicators of likely events already in use, for example people with high lipids, recent infarction, diabetes, and other risk factors are more likely to have a heard attack then others and such selection criteria are regularly incorporated into trials. There will surely be new genomic and proteomic predictors of risk that can be used similarly.
Enriching trials based on these kinds of markers, is something that FDA is actively encouraging sponsor to do. Particular promise for this kind of work clearly rests with oncology, because what is a tumor but a genetic abnormality. But the principles apply to almost any therapeutic space.
We have already shown we are willing to look at results from enriched clinical trials, especially when there is a strong response in an unmet medical need, in something other than the entire potential population, and to allow results generated from at-risk populations, that are pre-selected based on good science, as the basis for labeling. The extent to which labeling would focus on the particular group studied or be generalized depends on the particular case, but clearly the outcome data even in a subset is of great importance.
For example, the very first showing that the hypertension drugs known as ACE inhibitors could prevent death in people with heart failure was done in highly selected population, people with advanced and severe heart failure. In fact, this first outcome study that showed the drug improved survival was in the CONSENSUS study; a study with only 253 people, but the magnitude of the effect was very large, and so we wrote that result into label right away.
Given the magnitude of the potential for benefit, we did not wait and say we would not know how the drug would work in people who were less sick.
There are a lot of other very interesting opportunities to extend these principles, and a lot of promise from advances in genomics and proteomics that has not yet been fully realized. This will also help address another critical challenge, the agonizingly slow pace and low levels of recruitment for clinical trials. A recent survey by the American Society for Clinical Oncology for example reported that fewer than four percent of eligible cancer patients participate in clinical trials.
To help facilitate these approaches to clinical trial design, we are in the process of developing a guidance document specifically on enrichment in clinical trial design and how sponsors can take advantage of these opportunities.
There are a number of issues where we believe, through better guidance to product developers and researchers we can help facilitate advances in enrichment designs. For example, how much data do you need on the “off” subset, the people who do not have a high level of whatever marker you are using to enrich your trial population.
Some of our best opportunities for developing novel treatments for unmet medical needs, and doing so in a way that is efficient and does not make us expose people to experimental drugs unnecessarily, all rests with our ability to develop better, well validated approaches to structuring clinical trials. This is a major part of our Critical Path initiative – to develop better science on more adaptive clinical trial designs that allow us to incorporate what we understand about patients and about therapies into how we design clinical trials.
Finally, on the topic of communication:
When it comes to how we share and communicate the results and findings of clinical research -- I believe there is a lot more we can do to improve the transparency around the information that gets generated from clinical research, as well as the tools and approaches we use to make this information available to patients and doctors.
We need to make more of this information available at the time and the place where people are confronted with a medical decision.
Right now, when you are a doctor practicing medicine, bottom-line information is not easily accessible when a medical question comes to mind.
Even when this information is accessible, the specific facts that are relevant to a doctors’ particular medical specialty is not always easy to pick out from the reams of general medical information that is available.
A cardiologist for example is inundated with warnings and writings about psychiatry drugs while surgeons get deluged with data on diabetes. Moreover, matching up a patient’s particular characteristics which the subtle signals and data nuggets that might alert doctors to a rare problem is not as easy as it should be.
We need to use the same information technology that is becoming commonplace in clinical research to improve clinical practice, and the way that results of research can supplied to physicians.
We need, for example, a system where the same IT tools are able to keep track of this medical information, and readily integrate updates, and then seamlessly alert doctors to new information, and the most relevant facts for each individual medical encounter. This is the system everyone envisions. It is true that the bottleneck to these kinds of systems have been the lack of information technology tools, and the widespread adoption of electronic medical records that make archiving and retrieving this kind of data possible.
But the problem has not only been about hardware and software, but also about the absence of the right data in the right formats.
This is where FDA can help. We are fundamentally in the business of turning raw medical data into useful knowledge about medical products. For many years, we did not do as good of a job as we could have of sharing this knowledge widely. In some cases, we did not have the resources, but in others we just did not think we had to. This is another place where information technology tools are also helping.
With advances in the way that doctors and patients retrieve medical information, at FDA we have been able to fundamentally re-think the way we provide it. And we are doing just that. Let me tell you how.
Starting last November, healthcare professionals, patients, and other consumers have been able to access high-quality, up-to-date information on medications via the Internet free of charge through the Daily Med system, which receives, distributes and displays medication information that has been developed by Federal agencies. This is a monumental change in the way medical information is made available that would not be possible without the leadership and hard work of the professional staff at FDA’s Drug Center and especially my colleagues Dr. Janet Woodcock and Dr. Steven Galson.
A key element of the Daily Med is the electronic product label, known as the package insert or the physician label. This is the onion-skin piece of paper folded up inside the tops of your pill bottles. Until now, the content and format of these paper labels was not conducive to easy accessibility, and they were not easy to search to find the most pertinent information. It was hard to even find them on the web. Often, getting an electronic copy of the label meant downloading a PDF file.
So we recently released a new rule that will fundamentally re-make this package insert, or the physician label, making the information that is provided more readable, more accessible, and more amenable for electronic dissemination.
This is an important step toward creating an electronic environment for drug safety and drug effectiveness information that will make it easier for doctors and patients to get up-to-date information right at the point of care.
There are many other steps we are taking.
We recently required drug makers to start submitting label changes electronically in a standard format that, after verification by the FDA, will make updating labels with new information seamless, and automatic. Imagine a world where the most relevant information in the product labels, for example new updates or the most common side effects or the prescribing information, can be easily pulled off the web or better yet, pushed to a doctor on her PDA.
You can be instantly updated each time new safety warnings are added or new information about a common drug’s benefits. Doctors could receive special alerts for the drugs they are most likely to prescribe.
In addition to these things, FDA is taking other steps to help build the systems, infrastructure, data standards, and strategies to support and sustain the management of that information well into the future. We are modernizing our system for reporting drug safety information, called Med Watch, into a new system that we creatively dubbed, “MedWatch Plus.” We will be expanding on-line reporting, increasing interoperability, and moving toward a consolidated system for reporting all adverse events across the agency.
Another program we will be launching is a new, regular report on drug safety that is generated from the Med Watch surveillance data that we receive. It will be similar to the Morbidity and Mortality Weekly Report that is published by the C.D.C. This will help better inform doctors, more regularly, about what we are finding.
FDA already reports on the data in Med Watch, but right now this information comes out quarterly. By making this information available more regularly, our hope is more doctors will be prompted to report more of what they are finding. We should be able to release more bottom line medical data more quickly while protecting commercial confidential information.
All of these initiatives, as well as several others, together make up what we are calling Facts@FDA – a new electronic architecture for making timely, reliable information available to doctors and patients when they need, where they need, on demand. Taken together, these efforts comprise a big step into the future of healthcare information, a future when medical decisions can be guided by the most up-to-date and reliable information possible.
Finally, there is one more issue I want to discuss in closing, that also has to do with the way clinical information gets communicated to physicians and ultimately, to patients. That is the important role that the medical journals play in evaluating and promulgating the results of clinical research.
Medical journals -- both general and specialized -- have a vital part to play in shaping the views and behaviors of scientists, clinicians, and policymakers about medical practice and health policy issues of importance and concern. They can often serve as a bridge between these sometimes disconnected communities. They help put clinical and scientific findings into a common, practical focus.
I have served on the editorial staff of two major medical journals and consider myself a friend of these institutions. My esteem for the medical journals and their unique mission is why I have been deeply disappointed by some of the recent practices and pronouncements of some of our leading medical journals.
Editors of major journals have espoused greater transparency in the clinical research system, and that is a necessary end. But journals are not doing enough to adopt these same principles. Instead, the journals have enhanced the value of their franchises by maintaining very strict embargo policies, and long publishing cycles, that can bottle up clinical results for months and in rare cases years.
That can have public health consequences.
Now don’t get me wrong, the peer review process, although sometimes needlessly long, can serve medicine well by making sure that the medical results that get published and relied upon by millions of doctors have undergone close and expert scrutiny.
It is also true that FDA has its own peer review process where medical decisions undergo review by senior scientists, and sometimes are the subject of large regulatory briefings and meetings inside our drug center. Difficult medical questions get hotly debated by our career scientists and our medical staff.
In fact, the culture inside FDA is very similar to the culture inside any medical institution, including the medical journals, where we are required to reach scientific consensus around complex medical questions. No one person possesses all the necessary expertise and information, especially on complicated drug safety questions. It is for this reason that we have recently gone to new and great lengths to be more transparent about our deliberations, and to put in place processes that help resolve scientific disputes by allowing people to have individual scientific opinions strongly considered, especially diverging views.
But when it comes to the medical journals, my concern is this: The medical journals cast blame for the time it takes others, including FDA, to scrub medical information and put it through rigorous peer review before communicating final conclusions. Yet the journals themselves often take months to catalogue, review, format, and publish the articles they get submitted to them and are sometimes slow to update these articles when new findings emerge.
I think there should be a level playing field across the clinical research enterprise that engenders a culture of transparency around research results. And I think there should be a sense of immediacy shared across institutions involved in the research enterprise, FDA included, where important results are communicated more quickly, and important updates are made when medical information evolves.
A number of journals are starting to web-publish many more of their most important studies. This is an encouraging development. Clinicians and patients deserve nothing less than timely, complete information around the results on which they base their medical decisions. It is worth pointing out that some journals have also recently talked about developing more transparency around their peer review process as well, including more disclosure and transparency around their peer reviewers, who are often the same outstanding and principled experts FDA also relies on for public input at our advisory committee meetings.
To these ends, FDA has also worked hard to achieve greater transparency in its review process as well as around its own advisory committee process and especially the way that it discloses findings about the safety of medical products. We will continue to make these goals among our highest priorities.
But because we are all so dependent upon the work that the medical journals do, I also think the medical journal should be doing more to adopt these same principles in their own work, to foster a culture of trust, and to create a more level and transparent playing field for information. They should take a lesson from others who have made great strides in these areas, and commitments to these goals, sometimes at the journals’ own prompting – and that includes the FDA.
Private biomedical research has flourished under a largely self-governed arrangement, yielding immense benefits to society. But this arrangement relies upon a public trust that can sometimes be quite fragile.
At FDA, we are equally committed to these goals, and to adapting our own regulatory approaches in ways that accommodate good scientific ideas and new innovations in clinical research. Only by leveraging better scientific approaches in how we develop medical products, will we be able to see the benefits of all of the recent advances that have been made in basic science turned into practical medical solutions and advances in the public health. Thank you.