• Decrease font size
  • Return font size to normal
  • Increase font size
U.S. Department of Health and Human Services

News & Events

  • Print
  • Share
  • E-mail

Scott Gottlieb, MD - MIT Center for Biomedical Innovation

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.

Speech before
Massachusetts Institute of Technology's Center for Biomedical Innovation

Boston, Massachusetts

Remarks by
Scott Gottlieb, MD

Deputy Commissioner for Medical and Scientific Affairs
Food and Drug Administration

August 17, 2006


There has been a lot of discussion today on facilitating the development of innovative treatments, and facilitating pre-market access to these new medicines. But I want to focus my remarks on another increasingly important part of the drug development continuum, and that is how we evaluate new medicines after they are approved. Our ability to monitor drugs after we approve them, and to learn about and inform people of emerging information about new safety concerns, as well as new benefits, is becoming central to discovery efforts.

The principal reason for this is a scientific one. As we see many more very novel molecules and therapeutic approaches aimed at more unmet medical needs, continuing to learn about these new applications after approval is taking on added importance. Right now, when we look across the investigational new drug applications that we have in house at FDA, which are probably the earliest consolidated measure of the future of biopharmaceutical innovation, we are seeing an increased intensity of work around those applications. Drug developers are trying to break more new ground with more highly novel applications. This holds out a lot of promise for the future if it is reasonable to believe that new approaches, in the form of different kinds of molecules, might be able to help us conquer some of these vexing medical challenges that are not being fully addressed by the current crop of molecules that we have.

But more novelty also brings with it more uncertainty, especially when it comes to evaluating the safety of new drugs. Often times, the kinds of side effects that eventually emerge with new drugs are rare and cannot be predicted or unearthed with any reasonably-sized clinical trial. Often times, when it comes to very novel molecules whose mechanism is unfamiliar to us, it is sometimes going to be harder to predict a priori what kinds of rare side effects may eventually emerge.

This added complexity is not a reason to delay access to otherwise effective new medicines for those patients who can safely benefit from them. But it is an argument for why we need to remain vigilant after those medicines are approved. After all, there is no such thing as a completely safe medicine. All drugs carry risks and benefits. While the opportunities for health benefits have gone up with today's more diverse and often more complex medical treatments, there are also opportunities for things to go wrong. People are taking more complex medicines that sometimes have more subtle or complicated side effects. More people are taking more drugs, increasing the odds that two drugs could have a bad interaction, or may be used incorrectly. Our learning about the best ways to deploy effective new medicines does not stop at the time of FDA approval.

In short: With all the advances we have made in the last decade, also comes the need to do more to monitor for safety when it comes to using new medicines.

This starts with all of us understanding that just because we learn something about the safety of a drug after its approval does not mean the system failed, or that someone was wrong, or that the drug itself should not have been approved in the first place. Now it is true that we need to be uncovering these things sooner, and we are working hard to develop approaches to do that. But lifelong learning, and continual inquiry, is the cultural hallmark of medical practice. I worry that instead; however, we are pursuing quite the opposite, a culture of blame that focuses too much on assigning fault when medical problems are surfaced, rather than trying to learn from those problems, so that we can develop ways to mitigate and hopefully avoid them. Some people say this culture of fault finding is one reason too many medical errors in the hospitals go unreported. Across medicine, we need to focus less on finger pointing and more on working collaboratively to find facts sooner.

We also need to embrace the opportunity for our systems for unearthing side effects to also help us learn about new benefits from medicines. It is a two way street. I think it is fair to say that there are dozens of medicines approved over the last decade where the initial indication that the drug was approved for is dwarfed in terms of public health utility by the subsequent things we learned about those medicines, and the indications that they were subsequently approved to treat. Certainly this is true of the TNF inhibitors for example, or even common drug classes like ACE inhibitors and Statins.

At FDA, a lot of the steps we are going to be taking to develop better approaches to monitoring medicines after we approve them involve improvements in how we use new information tools. Right now, unfortunately, even as medical technology has progressed, our systems for learning about new medicines after they are approved have too often remained old school. At FDA, to address this gap, we are taking new steps to make systems work better, and to build ways to assure better patient safety by taking advantage of modern information technology tools and better reporting of information to consumers and healthcare providers.

Before I tell you about some of these new steps we are taking, I want to discuss just a few of the improvements we have already made on our existing systems. Earlier this year, we adopted the Systematized Nomenclature of Medicine or SNOMED, as the standard computerized medical vocabulary system to be used to electronically code important terms in the “Highlights” section of prescription drug labeling. This new standard will allow healthcare professionals to electronically access and share critical health and treatment information more easily and efficiently.

SNOMED is being used with structured product labeling in the new physician labeling design announced in January 2006, which is providing accurate, up-to-date drug information in an accessible, computer readable format. FDA also issued regulations this past year that require drug manufacturers, for the first time, to submit to FDA prescribing and product information such as the package insert or label in the SPL format. Using embedded computer tags, the prescribing and product information in the SPL format can be electronically managed, allowing a user to search for specific information. These tags can instruct computers to read specific sections of a drug label such as indications, dosage and administration, and warnings.

With this information, physicians will be able to quickly search and access specific information they need before prescribing a treatment, resulting in fewer prescribing errors and better informed decision making. In addition, having the labels submitted to FDA in SPL will improve the FDA drug labeling review process, so that FDA can provide immediate access to the most recent information about medications to doctors and patients.

This new labeling format will also be integrated into FDA's other e-Health efforts through a variety of ongoing initiatives. As prescription information is updated in this new format it will be used to provide medication information for DailyMed -- an interagency online health information clearinghouse, sponsored by the National Library of Medicine.

The DailyMedis making up-to-date information about FDA-regulated products widely available on the Internet at no cost. Third party vendors are already creating tools and platforms that leverage this source of highly reliable drug information in ways that we could not have imagined. This is significant yet underappreciated opportunity offered by this sort of electronic information trove. Private entities are figuring out even better ways to push this information to doctors and patients in formats that more easily integrate into the way that people work so that they have the facts they need right at the point of care.

During the past year FDA also issued three final guidance documents to help develop new ways and improve methods to assess and monitor the risks associated with drugs and biological products in clinical development and general use. The final guidances describe additional safety testing, monitoring, and interventions that may be helpful in selected circumstances and address pre-market risk assessment; the development, implementation, and evaluation of risk minimization action plans, called RiskMAPs, and good pharmacovigilance practices and assessment of reported adverse events.

Finally, to help implement these and other new steps, under the leadership of Dr. Steven Galson, the head of FDA's drug center, FDA announced some additional improvements in the way that our post market safety programs are managed inside our drug center.

Now, this gets me to how we continue to improve our approaches to safety. Let me be clear. I do not think that continuing to make improvements in our approaches to monitoring after medicines are approved is a question of adding new hammers to FDA's toolbox of existing authorities. Nor is it a question of simply finding more ways to restrict the personal medical decisions that doctors and patients make in order to try and reduce the chances that people might use medicines, even after making their own personal balancing decisions between risks and benefits, in ways that expose them to more side effects or fewer benefits from a medicine. Risk management approaches that put burdens on providers need to be used with a lot of care. Especially as medicines become more personalized and more targeted to narrow groups of patients, doctors and patients are going to need more flexibility, not less, in order to make individual decisions guided by good information.

Rather, I believe that improving on our current approaches is going to require us to work together to continue to make sure that that agency has the resources and tools to more effectively use its existing authorities. In particular, there is a lot of improvement we can make in the way we collect information and the way we analyze it, but this is going to require new efforts to improve the information systems we use to evaluate newly approved medicines. Now that we are well into the Information Age, now that modern information technology has had a fundamental impact on so many parts of our lives, there is no good reason we cannot have a fundamentally better system for helping to protect consumers and assure the safe and effective use of new drugs.

So I want to focus the remainder of my remarks today on some of the new steps we are going to be taking to improve our information tools for collecting and evaluating new medicines after they are approved by FDA.

Information technology provides remarkable new opportunities to accomplish all of these goals and more. There has been growing recognition of this fact in government, especially at the Department of Health and Human Services, under the leadership of Secretary Leavitt. But I do not think there has been as much recognition of what all of this means for the areas of greatest concern for FDA's mission: assuring the safety and effectiveness of medical products, and helping to protect and advance the health of the public. Whether it is developing and filing information from clinical trials with the FDA electronically, using sophisticated tools to help us analyze this information, or monitoring patient data in real time for signals that approved treatments may be causing potentially serious side effects -- all of this can have a fundamental impact on our mission.

First we need to improve the information we rely upon, and our tools and resources for gathering that information in a timely and reliable way. This starts with having access to robust, rigorously collected epidemiological databases that can help us track potential rare adverse events.

To these ends, we recently signed collaborations with a number of databases, and plan more in the future, including collaboration with the VA Medical System. As part of our commitments under the reauthorization of PDUFA, we also hope to get specific resources for working on guidances that will help establish the best approaches for using this kind of epidemiological data effectively. Prospectively collected data sets like these offer a lot of opportunity for routine monitoring and for signal spotting, but they also have certain limitations. We hope to use these guidances to establish the best approaches for using these data sets in a rigorous fashion that maximizes their promise and recognizes their limitations.

The fact still remains, for example, that sometimes to answer very difficult safety questions, such as whether a naturally occurring event is occurring more frequently as a result of a medical product, we will need the benefits of a placebo controlled trial to tease apart these very subtle safety problems. We need to clearly lay out what epidemiological data can, and cannot tell us. We need to identify the cases, for example, where more rigorous information is needed -- where epidemiological data alone does not tell the whole story.

We also need to do more to actively collect information, and not just rely on passive databases that report information sometimes months and maybe years after actual events occurred. To these ends, we are working on developing scientific efforts aimed at facilitating more active monitoring systems, similar to the MedSun initiative we have for medical devices, which actively solicits information from selected hospitals on adverse events with certain medical devices. Today the MIT Center for Biomedical Innovation is announcing one of these efforts.

It is an agreement to work with FDA on exploring the science around the best approaches for real-time drug safety monitoring. Under this collaboration, CBI will work to identify alternative methods for monitoring the safety and efficacy of new and well-established pharmaceutical products and medical devices. In addition to voluntary reports the FDA receives from patients and the healthcare community, the CBI model uses large healthcare datasets to recognize patterns that indicate unexpected efficacy or problems with safety.

Currently, techniques used to detect such adverse effects, because they are voluntary, result in underreporting of incidents. CBI's new approach to post marketing surveillance will be based on advanced computerized methods to automate the process of recognizing unusual patterns of infectious disease.

Such systems, developed by the researchers who will take part in this project, are now key tools used by public health departments and the CDC in the fight against bioterrorism and pandemic influenza. The group plans to leverage what we have learned from biosurveillance methods developed over the past eight years to help detect outbreaks and track influenza and SARS and applying them to the problem of post-marketing surveillance.

When it comes to prescription drugs, these active monitoring approaches need to be prospectively validated to better understand their utility in day-to-day signal detection and to assess the added value they have in identifying important safety concerns. We will be working with MIT as well as other scientific groups to help stimulate the expansion of this science.

Finally, our work is by no means done once we collect post market information. Raw data does not itself provide answers. We need to analyze this information, and try and draw conclusions about what it is telling us about the safety and effectiveness of medicines. This is the hardest part of our job, turning this raw data into rigorous medical information that physicians and patients can have confidence in and integrate into their own risk-benefit balancing.

Right now, one of the shortcomings of our approach is that we are not taking advantage of modern information tools to help us with this analysis. A lot of this work is still done by hand, by individual experts evaluating sometimes reams of information. Computers will never replace this human intervention, but we are sure that we can be making better use of sophisticated information tools to better support this work.

I will give you one example. Right now the system we use for collecting reports of adverse events, called our Adverse Event Reporting System or AERS, is largely a cataloguing and tracking system. When a potential adverse event is reported to FDA, AERS will help route the report to the right scientist at FDA, who will then go about evaluating the report to figure out what it means. But the system does not do enough to add a layer of analysis, for example, comparing the adverse event to all the other events we might be seeing, or scanning the entire medical literature to look for reports that might help put the event into sharper clinical context. We are working on developing a system that will add more analysis, which we are calling AERS 2.

Under the leadership of Dr. Janet Woodcock and others at FDA, we have made a lot of progress already to create a framework for improving these systems, so that we are using our computational tools not only to track our information, but better analyze it. Our first step was to form a bioinformatics board at FDA. This new board will help us work across our different science and medical centers to develop common platforms for functions like adverse event reporting, which are really activities that all of our centers engage in. By building common platforms that can then be customized for our different centers and our different activities, it will enable us to use our resources more efficiently to build more sophisticated applications that are supported by an improved infrastructure.

In the end, it is clear that all of this is going to require significant new investments in time and resources aimed at improving the FDA's IT backbone, which has grown outdated over the years. It will also require continued support by Congress. But we are committed to making these enhancements because they are critical to our ability to improve our post market safety system.

All of this is, of course, a two-way street. To make it worthwhile for health care professionals and organizations, we need to make better information available to patients and doctors who can use it to make more informed healthcare choices. So we are also taking steps to improve our tools for communicating this information, through or new label format and more electronic reporting of labeling information to start. We are also working on better vehicles for doing more regular reporting of the safety signal observations that we make in our AERS system, while continuing to protect confidential information.

For example, one of the things we are working on right now, and hope to be able to unveil soon, is a report -- similar in some ways to the CDC's Morbidity and Mortality Weekly Report. This effort, led by Dr. Paul Seligman and others at FDA, will allow us to make regular reporting from what we are observing in the AERS reports, to help doctors be aware of the signals we are seeing so that they will hopefully be stimulated to remain vigilant and file reports of what they might be observing as well with FDA.

We will keep working hard to find additional effective ways to protect patients from risks from medical products and to help doctors and patients understand the risks and benefits of the products we regulate.

Developing and sharing better information about the risks and benefits of treatments is of critical importance in helping people get more for their money in health care. Information technology has changed the every sector of our economy, whether it is sophisticated systems for relaying banking information across international markets, or the digital scanner at the local drug store that both speeds you through the check-out line and permits the store to adjust its inventories. IT is a big reason why industries have been able to make significant improvements in quality over the last decade and implement quality systems and just-in-time delivery. It is a big reason why the US economy has experienced huge productivity growth.

But IT has been slowly and sparingly adapted to healthcare. Healthcare has not realized the same benefits from these new tools. That needs to change and it is changing. I have no doubt that this more modern, more effective system of product surveillance and patient protection will come to pass. But to make it come to pass as soon as possible, we need to overcome some important obstacles and need to work on new approaches.

So I am delighted to be here at MIT, as they unveil today their own efforts to pursue some of those approaches, and at FDA we look forward to developing similar partnerships to advance our tools for learning about how new medicines can be more effectively used to improve public health.