On 19 January 2011, the FDA unveiled its much-awaited proposal for changes to the 510(k) process. Medical device industry representatives were generally pleased with the modest scope of the proposal. In a press release, AdvaMed president Stephen J. Ubl called the plan “clearly a good first step that will address some of the major problems with the program.” Mark Leahey, President and CEO of the Medical Device Manufacturers Association (MDMA) commended “FDA leadership and staff for working with various stakeholders in an open and constructive process to develop consensus proposals that improve patient access to safe and effective therapies.” The NY Times quoted Medtronic CEO William A. Hawkins as stating “This is a more balanced approach that sends a positive signal to industry that the F.D.A. is engaged, listening and concerned.”
Patient safety advocates were underwhelmed. Public Citizen’s Sidney Wolfe stated “the FDA is not being forceful enough about improving the safety and effectiveness of new devices.” Dr. Rita Redberg (UCSF, Editor of Archives of Internal Medicine) told theHeart.org that she is “worried about the implanted high-risk devices that are getting cleared without any data. There are lots of examples of stents, inferior vena cava filters, heart valves, that are [cleared through 510(k) but] not low risk.”
But is the FDA’s modest proposal really a good outcome for the device industry?
The 25 action items in the FDA plan
New guidance documents
The first 8 actions call for the FDA to draft new guidance documents to help industry comply with regulations. Of these 8 guidance documents, the 510(k) paradigm guidance appears to affect the broadest number of companies – clarifying the need for clinical data in 510(k)’s, and clarifying the appropriate use of multiple predicates. The guidance on “de novo” 510(k)’s should also be useful for a relatively small number of companies. While providing clarification of FDA’s thinking, guidance documents usually codify existing processes rather than improving them. We shouldn’t expect these actions to result in significant changes to the 510(k) process for most devices.
FDA skill set improvements
The next 5 action items in the proposal are (mostly internal) improvements to the FDA skill set, including “assessing [CDRH] staffing needs” and “enhancing training.” While “leveraging external experts” sounds appealing, historically FDA has found this challenging. For example, finding unconflicted experts has often been problematic for PMA panels. Again, these action items are unlikely to result in dramatic changes for most 510(k)’s.
Changes to the 510(k) process
The next 11 action items are proposals for actual changes to the 510(k) process. Let’s take these one-by-one. To me, four of the 11 programmatic and regulatory changes are significant.
First, the FDA proposes an “Assurance Case” Pilot Program, primarily aimed at improving medical device safety. In fact, FDA is already running a pilot via their recent guidance on infusion pumps. Typically, assurance cases are a formal way to organize multiple pieces of evidence (from activities of risk management, software development, testing, and possibly other sources) to demonstrate that software is safe. Software, QA and RA staff should read the detailed description of the Assurance Case Framework in the Carnegie-Mellon article Towards an Assurance Case Practice for Medical Devices. If adopted, the Assurance Case framework is likely to make more work for industry (and another set of documents for FDA reviewers to question), but is unlikely to result in dramatic changes to the 510(k) process. Given that Assurance Case documentation is totally new for both industry and FDA, it may take a while to evaluate even the pilot program.
Second, the FDA proposes to develop a new adverse event database. This is important. I don’t know anyone who thinks the current medical device surveillance system works well. Definitions of reportable adverse events vary widely among companies. The real question is whether this proposal can address the major shortcoming of the current system: while we might be able to know the numerator (the number of reported adverse events for each device) we don’t know the denominator (the number of times each device was used). With the current system we cannot determine the rate at which a device experiences adverse events. Will this change with the new system?
Third, the FDA will implement a “unique device identification (UDI) system, to improve adverse event reporting. Like the new adverse event database, this is a helpful but incomplete idea. Hospitals and physicians are supposed to file device-related adverse event reports, and a UDI system will help correlate adverse events to specific products. But how will FDA deal with hospital and physician non-compliance on reporting?
Fourth, the FDA will perform a program assessment of the IDE process. Good idea. Clinical trial design is challenging, and relating clinical data to clearances is not always straightforward. We’ll have to wait until the assessment is complete to know if any substantive changes are recommended.
There is nothing wrong with the rest of the 11 programmatic and regulatory changes, but none appear to be game changers:
- The FDA will use “notice to industry letters” as a standard practice to inform industry when CDRH has changed its regulatory expectations.
- The FDA will conduct additional analysis to determine why devices that cite more than five predicates (in the 510(k)) have a greater rate of MDR’s.
- The FDA will create an SOP to regularly update the list of device-types eligible for third-party review.
- The FDA will create a set of SOP’s covering their generation of guidance documents and regulations.
- The FDA will issue a proposed regulation on the transfer of 510(k) ownership.
- The FDA will issue proposed regulations clarifying device-labeling requirements, and will develop an on-line database of device labeling.
- The FDA proposes non-proprietary photographs of devices in a public database.
Deferral of issues, pending outcome of IOM study
Finally, the last action item is for the FDA to ask the Institute of Medicine (IOM) to study 7 contentious issues, listed below. (The IOM is already in the process of evaluating the 510(k) process, and has released several interim reports which can be accessed here).
- Under what circumstances can the FDA rescind a 510(k)?
- When can FDA require post-market studies as a condition of 510(k) clearance?
- Should FDA establish a Class IIB where more clinical data, manufacturing data or post-market surveillance is required for 510(k) clearance?
- Under what conditions should some devices no longer be allowable as predicate devices?
- How can the regulatory terms “indication for use” and “intended use” be consolidated into the single term “intended use”?
- Should companies be required to keep one physical unit of each submitted device, for the FDA to access upon request?
- Should FDA ask congress for power to consider off-label use when reviewing 510(k) applications?
The IOM recommendations are expected in the summer of 2011, and will likely become the basis for another round of device company hand-wringing. (Device companies are concerned that they have no representation in the Institute of Medicine, and therefore no influence on the study.) In the meantime, changes in these areas are deferred and the 510(k) process goes on as before.
Is the FDA’s action plan really a good outcome for industry?
It appears that industry has gotten what it wants, and the current 510(K) program is more or less intact. But is maintaining the status quo really in industry’s long term best interest? I’m not so sure.
First, the FDA’s relatively limited proposal has fostered the impression that a self-interested device industry has steamrolled the agency, putting profits ahead of patients. Bloomberg’s headline was “Device Companies Dodge Controversial Changes at FDA.” The Wall Street Journal wrote that the “U.S. backs down on … medical devices.”In the court of public opinion, device industry credibility has been weakened.
Second, the device industry’s real issue – patient access to the latest and best technology – is not even on the table. At best, transcatheter aortic valves will be available in the US five years after Europe, and two years after a compelling pivotal trial. Cardiovascular optical coherence tomography, a new tool for assessing stent parameters and apposition, was available in Europe and Japan years ahead of its US 510(k) clearance. More than anything, industry and patients innovative products to get to market faster.
To put the timeline issue on the table, the medical device industry needs to build credibility. To build credibility, the device industry needs to get out in front of some very real product safety and surveillance issues. The FDA’s proposals are a response to very real post-market device performance issues, including repeated recalls of infusion devices, automated external defibrillators (AED’s), orthopedic implants, and others. Call me naïve, but can we really expect FDA to consider expediting pre-market clearance without improved post-market product safety and stronger post-market surveillance?
Industry has opposed many FDA proposals, but has not offered constructive alternatives. In October, AdvaMed said it did not support the use of “assurance case” frameworks in 510(k) submissions, but offered no alternative methods of improving product safety. In the same month, the Medical Device Manufacturer’s Association (MDMA) said it “opposes the adoption of an ‘assurance case’ framework for 510(k) submissions” and that FDA should rely on current regulations. Industry has known for a long time that the MDR reporting system is broken, yet there has been no movement from industry to document best practices for reporting adverse events. Is there really no room for improvement?
Industry reacted to draft FDA proposals with tremendous energy. In addition to detailed AdvaMed and MDMA position papers, the industry marshaled 15 senators to sign a joint letter to the FDA, asking the agency to “proceed cautiously” and expressing concern “that many of the recommendations in the report could result in a more burdensome and time-consuming approval process.” If the industry applied the same energy proactively, to address real concerns about patient safety and product reliability, could the industry build enough credibility to start the discussion about clearance and approval timelines?