Medical Device Link.

Originally Published MDDI September 2001

VERBATIM

From FDA reviewer to an industry VP:

C.R. Bard's Susan Alpert discusses her transition from director of FDA's Office of Device Evaluation to her new role in the private sector.

It's not news today, but on the morning that MD&DI spoke with Susan Alpert, vice president for regulatory sciences at C.R. Bard Inc. (Murray Hill, NJ), the purchase of the company by Tyco International Ltd. (Pembroke, Bermuda) was hot off the press. For an interview intended to focus on Alpert's observations about private industry after a lengthy career at FDA, including six years as director of the Office of Device Evaluation (ODE), that news provided a timely twist.

Asked how her experiences in private industry have differed from those in government service, Alpert identified the recent takeover announcement as a key example of the differences between the two. "Government doesn't change that way," she noted.

In this excerpted interview with Steven Halasey, editor in chief of MX, Alpert talks about her involvement in some of the major changes in the regulation of healthcare manufacturing over the past decade, and about how she views her new role in industry.

Q. You had a long career at FDA before coming to prominence as director of the ODE. Is that right?

A. That's correct. I joined FDA as a pediatric infectious-disease fellow and reviewing medical officer in the division of anti-infective drug products at the Center for Drug Evaluation and Research (CDER). I spent six years working at CDER, advancing from being a frontline reviewer to supervising in the areas of infectious diseases and anti-infective drugs, as well as dermatology drugs. So I simultaneously supervised a group of anti- infective medical officers and a group reviewing dermatology drugs.

While at CDER, I also got involved in a number of other things, such as the center's committee for advanced scientific education. Education has always been an interest of mine, so I really enjoyed my six years at CDER.

Then I moved to the Center for Devices and Radiological Health (CDRH), first joining as the senior medical officer in ODE, and then competing for and getting the position of director of the office.

Q. At that time, industry was especially concerned about the influence of the so-called drug model on the device center. To what extent did you feel a need to change the methods that ODE used to evaluate devices?

A. At FDA, the basic issue that arises for every product—whether it's a drug, a biologic, or a device—is how to evaluate its safety and effectiveness. And there are some components of the review processes that those different areas have in common.

Generally speaking, devices are designed to accomplish a certain task or set of tasks, such as cutting, coagulating, or stretching. Those kinds of activities can be evaluated without clinical trials because the manufacturer is not claiming the device will diagnose or cure anything, merely that it will perform a certain kind of activity during treatment. So not all devices need clinical trials.

In fact, the vast majority of devices never see a clinical trial because so much can be determined simply from the design of the device or by means of bench or animal testing. A lot can be deduced about the potential outcomes for patients if the design of the product is reviewed; bench testing is used to determine the reliability, reproducibility, and performance of the device; and animal testing is conducted to show that an operator can make the product do what it's supposed to do.

But the very sophisticated types of medical devices that began to be seen in the early 1990s raised a different set of questions. Because they are more responsible for the safety of the patient and the ultimate outcome of the treatment, those types of devices are the ones that really raise the need for clinical testing. And of course there are lots of varieties of clinical testing: small uncontrolled trials; large but simple trials; controlled, uncontrolled, and historically controlled trials; randomized and nonrandomized trials. There are a lot of potential models.

What happened in the early 1990s was that CDRH began to look at how medical devices were being evaluated. As a result, it was concluded that products were becoming so sophisticated that it was no longer sufficient to review them only in the light of bench and animal testing. The performance claims that were being made were really for patient outcomes, and that opened up the need for additional types of information.

Q. Was that the conclusion that was formalized by the so-called Temple Report in 1992?

A. Actually it began before that, with what device manufacturers were asking for in their claim structures. They were actually claiming cures and outcomes, and the device center didn't have the expertise or experience to evaluate those claims. That's what the Temple Report pointed out—the fact that the center didn't have a model for dealing with medical testing. And that's not a drug model, a biologics model, or a device model—it's just a good clinical testing model.

The evaluation that was done by the group led by Bob Temple resulted in changes within CDRH to ensure that if clinical testing was appropriate—that is, when a manufacturer made claims for diagnosis, mitigation, outcome, or cure, or when the product was essentially performing a procedure—then clinical testing would be done.

Q. How did those changes alter the emphasis of the device reviews conducted while you were director of ODE?

A. In the period after [former CDRH director] Bruce Burlington and I joined the device center, randomized concurrently controlled trials—which are the most common type of trials for drugs— remained uncommon for devices. Clinical testing, clinical study, and clinical trials increased significantly, but a lot of that had to do with the claims associated with new technologies that were being developed and our need to identify the safety issues and benefits of those technologies. And those new clinical trials were rarely the size and the kinds of trials that are routinely done in the drug center.

It was clear that premarket approval (PMA) products needed to be studied because there was no other way to anticipate their impact on patients or on the procedures in which they were being used. But over the years some 510(k) devices had undergone a radical transformation from very simple, user-operated products to very sophisticated, computer-operated devices. We needed to understand how such new devices operated, how they affected patient procedures, and what problems and benefits they carried with them. For a very small percentage of those 510(k) devices, clinical testing was required to answer those questions.

When I went to CDRH, many people thought it was anathema to require clinical testing for a 510(k) product. But since then, it's become better understood and accepted that sometimes the answers to the most important questions about a device can only be found in the clinic. Often, such testing is only needed to confirm that the device performs as its design and bench testing suggested it would. Then, the clinical testing generally involves a very reasonable number of patients in the hands of a few investigators. And the advantage of being able to obtain this kind of clinical data even for 510(k) products is that it then becomes possible to down-classify those products. People are much more comfortable with that model now than they were when I joined the center.

Q. Was this shift in emphasis also a big change for industry?

A. In some device areas, there was already a lot of clinical testing going on, but invisibly. What we did was to focus a light on the need for such information, to point out that no matter how wonderful a device might seem to be, its impact on patients still needs to be evaluated. And to do so, we recruited many more clinical reviewers, people with the expertise needed to understand clinical trials and evaluate clinical data.

The need for clinical testing is much more accepted today than it was a decade ago. In fact, many companies now plan their clinical testing not only to meet FDA requirements, but also to support applications for reimbursement coverage or to gather data that can increase sales.

Q. Was the device center cognizant that it was going through a process of developing a new approach to evaluating devices, or did it sort of back into those changes?

A. To most people within the center, I think it was very clear that we were changing purposefully and beginning to ask different kinds of questions about medical devices. But the center's changes came concurrently with technology advances for a vast number of devices and, in reality, many areas of industry had already recognized that clinical testing was required. In the cardiovascular device area, for instance, clinical testing was already the norm more than the exception when I joined the center.

Q. Industry has generally given you high marks for working to reduce the backlog of product submissions awaiting approval while you were director of ODE. Since joining the corporate world, how have you found industry's sense of timing to be different from FDA's?

A. Companies do their work and send their packages off to FDA, and then they wait, literally with bated breath, 30 days for an IDE, 90 days for a 510(k), or 180 days for a PMA application. And those are long times to wait after the company has already completed its work.

When I was at ODE, we recognized that one of the difficulties we faced was that our efforts came at the end of the line. It wasn't only the pace of the work we were doing, it was that we were doing it while companies waited, and waited, and waited. That's why we created a lot of different mechanisms for companies to interface earlier in the product development process. That recognition had a lot to do with the development of programs such as pre-IDE meetings, which made it possible for companies to get agency input earlier.

A similar program that I take great pride in is the modular review program for PMA devices, which permits the sponsor to submit pieces of the PMA application as they are completed, instead of waiting for everything to be finished. That way, the results of bench testing and animal testing can be reviewed while they are still fresh in the minds of the investigators, and when questions can be answered much more easily. That approach relieves a lot of pressure, because it enables the company to get some things finalized while other matters are still under way.

When I put on my company hat, I can really appreciate the advantages of such a program. The modular approach makes the PMA process a continuous series of events, each of which contributes to the overall progress of the product application. And it eliminates those long periods of time when the company is merely waiting for something to be completed, so that the entire application can be gathered together and submitted—maybe next year.

Q. You spoke recently at an ASQ meeting. Do you miss having a very prominent and visible public role, or are you still very active publicly?

A. I'm still active in some of the same areas, only from a different perspective. For example, I'm quite active in AdvaMed and other trade associations. I represent Bard on AdvaMed committees and I still work across industry, but now from within. My impact is different now, because it is only that of a single company.

Obviously device industry people know my background, so when I speak with trade association members about issues relating to FDA, or submissions, or product development, they know the kinds of insight I can provide. I hope that in those contexts I'm able to provide some insight into how regulators think.

I certainly can't speak for the agency now, and in any case, the device center has changed a lot in the past couple of years. But it can still be useful to point out some of the issues that FDA is likely to raise, or to suggest how a reviewer might see a problem.

Bringing that insight to the table keeps me visible in the device industry. But it is very different to comment on policy rather than to help develop policy. That's quite different.

Copyright ©2001 Medical Device & Diagnostic Industry