This Update will provide a summary of the proceedings at the U.S. Food and Drug Administration (FDA) Public Workshop entitled Strengthening the Center for Devices and Radiological Health’s 510(k) Review Process held on February 18, 2010, in Gaithersburg, Maryland.
I. Agency Introductory Remarks
Jeff Shuren, M.D., J.D., Director, Center for Devices and Radiological Health (CDRH), noted that the agency is extending the time period to provide written comments to the Docket on this topic until March 19, 2010.
He explained that the 510(k) program was created over 30 years ago with two goals: 1) the availability of safe and effective devices; and 2) to foster innovation. The program has undergone modifications over the years. Recently, there have been concerns regarding whether or not it achieves those goals. The agency faces a number of challenges today.
Therefore, CDRH is conducting two assessments to determine whether changes should be made to the 510(k) program:
- An Institute of Medicine (IOM) independent review to recommend administrative, regulatory and statutory changes. There will be a public meeting to discuss this on March 1, 2010; however, the final report is not expected until the end of March 2011.
- An Internal Working Group under Donna-Bea Tillman, Ph.D., Director of the Office of Device Evaluation (ODE), to identify actions the Center can take to improve consistency in the short term under the FDA’s existing authority. The Working Group report is due at the end of May 2010, and will be released for public comment soon after. CDRH intends to begin implementing changes by the end of September 2010.
Dr. Shuren stressed that today the agency wants to hear from outside parties. The Working Group has been brainstorming on this topic, but will be informed by the comments made today and the comments submitted to the Docket. The FDA is open to all ideas, including suggestions for statutory changes.
Dr. Tillman then provided a brief introduction to generally describe device classifications and the 510(k) program. She noted that the agency has the most difficulty in: 1) determining whether a device that has different technological characteristics raises new questions of safety and effectiveness; and 2) when to require a 510(k) for a change or modification to a device. She also noted that CDRH looks to the 1976 Congressional Record regarding how to interpret substantial equivalence.
Dr. Tillman asserted that technology has rapidly advanced since 1976. The FDA has facilitated this process and fosters the ability of medical devices to evolve. This is important, but the FDA also has to protect the public health. She noted that the FDA has cleared approximately 120,000 510(k)s since 1976. We are here today, she said, because technology will continue to change, the challenges will continue, and the FDA wants cutting-edge solutions. The agency intends to challenge existing paradigms; for example substantial equivalence may or may not be the right requirement. The FDA needs to figure out how to align medical device regulation with the future of medical device technology. The agency plans to continually reassess the regulatory framework so that the 510(k) program can be an agile program. CDRH does not have all the answers, and needs the perspective of patients, doctors, the medical device industry, and academia.
II. Issues for Discussion
Four FDA participants then presented the issues and questions for discussion at the meeting in the following four areas, described in more detail below (see also Attachment 1 for copies of the FDA presentations):
- Issues Related to Predicate Devices
- Issues Related to New Technologies and Scientific Evidence
- Issues Related to Practices CDRH has Adopted in Response to a High Volume of 510(k) Submissions
- Issues Related to Postmarket Surveillance and New Information about Marketed Devices
A. Issues Related to Predicate Devices
Christy Foreman, Deputy Director for Engineering and Science Review, ODE, noted that when devices are found substantially equivalent, this is a determination that the device is substantially equivalent to the original device type that was established based upon technology that existed in 1976. The FDA faces challenges in that today’s devices are not identical to those upon which the regulations and product codes were based. She presented the following specific challenges related to predicate devices:
- Manufacturers may lack sufficient information to select an appropriate predicate, as the 510(k) database and 510(k) summaries can lack critical details, requesting 510(k)s through the Freedom of Information Act (FOIA) can be a slow process, and a company’s current promotional material may not reflect what was actually cleared. The FDA is considering the idea of requesting redacted versions of 510(k)s at the time of clearance. In addition, the FDA is considering publishing information regarding the creation and assignment of product codes.
- “Old” predicate devices set the bar for comparison for today’s devices. In addition, there is an issue of how to deal with predicate devices with sub-par performance compared to the other devices in the class, as these are frequently chosen by sponsors as the comparator.
- Incremental design changes that are either submitted in 510(k)s or documented to the file can be challenging to deal with, as well as “Device Creep.” Sponsors often incorrectly interpret “could significantly affect safety and effectiveness” as “does significantly affect.” In addition, equivalence studies allow for the subject device to simply be non-inferior to the predicate, which can allow device performance to deteriorate.
- The issue of split and multiple predicates, and when it is appropriate to have more than one predicate, is also a challenge, especially when the predicates cross regulations and review divisions. Using the intended use from one device and the technology from another creates problems in assessing substantial equivalence, especially in trying to follow the 510(k) flow chart. An example where the use of multiple predicates is appropriate is combining the measurement of multiple physiologic parameters (each with a predicate) into a single machine. An example where it is not appropriate is the introduction of a new technology (high-intensity focused ultrasound vs. radiofrequency energy) for the treatment of uterine fibroids.
- There is a lack of clarity regarding the definitions and distinction between Indications for Use and intended use that leads to program inconsistencies.
B. Issues Related to New Technologies and Scientific Evidence
Arleen Pinkos, Reviewer, Office of In Vitro Diagnostic Device Evaluation and Safety (OIVD), noted that there are two main challenges in the area of new technologies: the type and extent of information needed to make decisions, and the decision-making itself.
She gave an example of a surgical robotic system that evolved from simple endoscopes through the 510(k) process. Challenges that this example illustrates include: what Indications for Use are appropriate, determining when different types of safety and effectiveness questions are raised, the effect of incremental changes on the patient, and the evidence needed to show substantial equivalence. She noted that human clinical trials are most often needed for devices when the outcome can’t be predicted, the changes might alter the clinical effectiveness, or introduction of a new device might lead to a new clinical practice. She gave as examples of types of technologies that may require clinical studies devices including nanotechnology or devices with new software algorithms. Other questions in this area include when it is necessary to demonstrate clinical utility, how to interpret clinical studies, how to extrapolate test data to the real-world, how to deal with poorly-designed studies after the fact, and how to deal with a lack of appropriate predicates.
Finally, she indicated that additional challenges include whether to require studies to address the potential for off-label use (as opposed to addressing it with a limitation in the label), the FDA’s limited authority to require manufacturing information in 510(k)s or to conduct pre-clearance inspections, and how to best use the de novo down classification process.
C. Issues Related to Practices CDRH Has Adopted in Response to a High Volume of 510(k) Submissions
Barbara Zimmerman, Deputy Director for Premarket Program Management, ODE, presented issues related to practices CDRH has adopted in response to a high volume of 510(k) submissions, specifically bundling, use of standards, and third party review.
Challenges related to bundling include the review timeline, bundling across Review Divisions and regulations, and the identification of bundled products after clearance. Challenges related to the use of standards include the lack of verification of stated conformity to standards, standards that do not contain objective pass/fail criteria, inconsistencies in the use of the consensus database, use of Form 3654 (which is not a declaration of conformity), and how to appropriately declare conformity. Finally, the main challenge related to third party review is that the agency continues to receive poor quality submissions from the Accredited Parties (AP), likely due to the lack of device-specific guidance documents and the fact that APs do not have access to previous decisions and reviews.
D. Issues Related to Postmarket Surveillance and New Information about Marketed Devices
Heather Rosecrans, Director, 510(k) Staff, ODE, noted that the major issue related to 510(k) devices in the postmarket period is that the FDA’s postmarket controls for these devices are very limited, as is the agency’s ability to modify or rescind 510(k) clearances.
She described a number of challenges in this area. The FDA’s authority to mandate postmarket surveillance studies after a substantial equivalence determination is limited to Section 522 studies (there are no “conditions of approval” for 510(k)s), and it is difficult to take action on any information collected in a postmarket study. The FDA has limited authority to rescind 510(k) clearance, and there are concerns about the “domino effect” of a recision for a predicate. She noted that the FDA proposed a recision regulation in 2001, but it hasn’t been finalized yet. How to use postmarket information collected in future 510(k) reviews for a certain device type, and how to deal with the “level playing field” issue, are not well-defined. Final printed labeling is not required prior to a substantial equivalence determination for a 510(k), and the FDA does not routinely review labeling for 510(k)-exempt devices. Finally, the FDA does not require the transfer of ownership of a 510(k) be reported to the FDA; therefore, it is difficult for the FDA to investigate adverse events, assign inspections and take appropriate enforcement actions where there is a new 510(k) holder. When changes in ownership are reported voluntarily, it is difficult for the FDA to verify this information.
She provided the following key points from the FDA’s presentations today:
- Increasing transparency is essential.
- Achieving consistency is critical.
- Developing clear definitions, guidance, and additional authorities may be required.
- Utilizing rational application of guiding principles is vital.
III. Comments from Open Public Sessions, Round-Table Discussion
See the Meeting Package (Attachment 2) for the list of registered speakers from the Open Public Sessions and the Round-Table Discussion. In addition, Hogan & Hartson LLP’s presentation, given by John J. Smith, M.D., J.D. is included as Attachment 3. A brief summary of the discussion is presented below.
The speakers almost uniformly supported the 510(k) program, and the concept of substantial equivalence, and believed that the program is rigorous and effective in its current form. Of course, like any program, it was noted by a number of speakers that it would be possible to strengthen the 510(k) process. Several of the speakers questioned what specific problem had been identified to trigger the FDA’s recent investigations of the 510(k) program, and whether there was data to support that there is a problem. Once the problem has been clearly defined, solutions can be identified. A number of speakers criticized the recent press coverage of the 510(k) program as sensationalized and inaccurate. A dissenting opinion was voiced by Dr. Diana Zuckerman, President, National Research Center for Women & Families, who believes the 510(k) program has major issues and therefore requires major changes.
Many of the speakers noted that in the past few years the process has become less predictable, data requirements are often out of proportion with the devices under review, and review times have become longer. However, this appears to be an implementation problem, rather than a problem with the 510(k) process itself. Several speakers stated that the process could be improved by the FDA management taking a bigger role and providing more training for reviewers. In addition, it was stated that the agency should open the lines of communication with sponsors. Device-specific guidance documents, additional Special Controls, and additional use of consensus standards would also be helpful to industry in understanding the requirements. Several speakers suggested harmonizing with risk-based systems or standardized submissions from other countries, for example the European Union.
Dr. Tillman noted during the Round-Table Discussion that the FDA understands that industry wants consistency, and that it would be ideal to have a device-specific guidance or standard for every device type. However, she said this is not possible given the variety of devices the Center regulates. Therefore, the FDA is asking for concrete ideas on how to achieve consistency, and achieve the balance between being flexible and predictable. Other Round-Table participants suggested more management involvement in reviews, training of the FDA staff on new technologies (by academia, for example), and having reviewers concentrate in particular types of devices.
Speakers also emphasized the need for both predictability and transparency in the 510(k) program. They pointed out that the FDA must encourage innovation, and not make data requirements so burdensome (particularly by changing them mid-stream) such that the small companies that tend to be the innovators cannot comply with them. Companies are willing to do what the FDA asks in terms of testing, but what is expected must be clearly defined and adhered to. In addition, the agency should be certain any “new information” relied on by the FDA is credible and universally accepted (vs. anecdotal) before changing requirements. Certain speakers believed that any new requirements for a device type should be retroactively applied to products of that type on the market. Several speakers defended the use of split predicates as essential to the 510(k) process, and denounced what they viewed as the FDA’s new “conservatism.” In addition, several speakers noted that “device creep” has been presented negatively by the FDA, when in fact this represents innovation and is what the 510(k) program was meant to do. Several manufacturers also rejected the notion that device safety and effectiveness may have deteriorated through the 510(k) program, when manufacturers are always striving to make safer and more effective devices than competitors, not simply “non-inferior” ones to predicates. Many speakers supported wider and more predictable use of the de novo down classification process for appropriate “new” technologies.
V. Concluding Remarks
Dr. Shuren reiterated the timelines for the assessments of the 510(k) program as outlined in his opening remarks.