As part of its fact-finding process, the Institute of Medicine (IOM) Committee on the Public Health Effectiveness of the FDA 510(k) Clearance Process provided an opportunity for public comments on topics relevant to its task.
Medical Professional Association
A representative of the American Academy of Orthopaedic Surgeons (AAOS), which has more than 17,000 board-certified orthopedic surgeons as members, emphasized the role of surgeons in promoting patient safety through the responsible use of implantable medical devices, reporting of adverse events, and postimplantation reviews of patient outcomes. Failures occur and can have devastating effects on patients, but they rarely happen as the result of a single factor that could have been readily identified through premarket studies. AAOS has confidence in the 510(k) process, its rigorous review, and the use of standards, he said. When there is adherence to well-defined procedures, reviewers are empowered to follow the science to its logical conclusions, and there is compliance with internal protocols, the 510(k) process is a reliable and predictable pathway.
AAOS strongly believes that the current 510(k) process, combined with Food and Drug Administration (FDA) surveillance programs, provides the most favorable balance between benefits and risks. That balance is achieved through the 510(k) process’s inherent flexibility, which maximizes the benefits of early access to new technology while minimizing the risks associated with innovation. Products cleared through the 510(k) process harness
incremental improvements in technology as a result of small changes in the iterative design process, increase treatment options for individual patients and their disease phenotypes, and contribute to clinical quality improvement through the use of performance information on improved devices. Thus, the process promotes innovation in support of public health.
Incremental improvements in technology translate into substantial changes for patients, he said. In total knee arthoplasty, for example, there were failures from polyethylene wear due to oxidative degeneration. A Harvard innovation added an antioxidant that stabilizes free radicals and reduces the potential for oxidation. The change was cleared through the 510(k) process and has contributed to the extended life of many total knee implants. For patients, that means fewer revision surgeries and improved quality of life.
With flexibility and regulation comes responsibility. FDA must continually evaluate its performance in assessing risks and benefits and in determining safety and effectiveness. The question should be not whether but how changes should be implemented to maintain FDA’s performance while expanding it to accommodate greater volumes of 510(k) submissions. Administrative changes are necessary to ensure that decision-making authority resides with the most qualified people and to prevent interference in the review process that is not grounded in scientific inquiry.
Medical Device Manufacturers
The Medical Device Manufacturers Association (MDMA) represents more than 200 primarily small to middle-size medical device companies. The companies drive innovation and develop the technologies that improve patient care in the long term and reduce the cost of care. A representative of MDMA said that over 100,000 devices have been cleared via the 510(k) process since 1976, and there have been relatively few adverse events. Although some patients have had suboptimal outcomes, which are not to be overlooked, there is a lack of qualitative and quantitative data to demonstrate a systemic failure. There will always be outliers, but in the absence of hard evidence that suggests a systemic failure, he urged the committee to be cautious in suggesting broad sweeping changes in the system as a whole. He suggested that the FDA recall database would be one repository that could be part of a systematic review. If the data suggest systemic failures or isolated pockets of products that need additional scrutiny, MDMA is willing to address them and to make sure that the appropriate special controls or remedies are in place. But it is necessary for the data to demonstrate that
any proposed recommendations would address whatever underlying issues are identified.
The MDMA representative urged the committee to be sensitive, in formulating its recommendations, to the economic and resource realities of the current innovation environment and to recognize that there is a risk–benefit approach to device development. Some 80 percent of device manufacturers have fewer than 50 employees. If regulatory requirements to bring a device to market cost, for example, $100 million but the potential product market itself is $50 million, that will stem the tide of innovation. User fees are not an answer, he said, and he cautioned against making recommendations that rely on industry funding.
He also noted that it is important to look at products in their total life cycle; review should not be a binary event that ends when products receive clearance. Unique device identification and monitoring systems will allow each device to be tracked over the total life cycle of the product. Databases will contain information on how products perform in the marketplace.
A representative of another industry association, the Advanced Medical Technology Association (AdvaMed), said that the organization believes that the 510(k) clearance process is well designed to assess the safety and effectiveness of low-risk and moderate-risk medical devices whose risks are well understood from experience with similar devices. Although the basic structure of the 510(k) process is sound, there is always room for improvement. AdvaMed has been engaged with FDA and other key stakeholders on ways to improve the clarity and consistency of the process.
Patient safety is the number 1 priority of the medical device technology industry, but any regulatory requirement should balance FDA’s dual mission of protecting the public health and facilitating innovations that benefit patients. Since the 510(k) program was created in 1976, it has been based on risk: if a new device presents a risk that is greater than that posed by the predicate device, FDA could find that the newer device is not substantially equivalent, classify it into class III, and require submission of a premarket approval (PMA) application.
Risk assessments apply to both a new device’s intended use and its technology. The premarket notification program has worked extremely well for more than 30 years, she said. It has permitted FDA to review, on the average, about 3,500 submissions a year in a reasonably timely fashion and to ensure that products that go to market are safe and effective.
The agency has the legal authority to request as much information, including clinical data, as is necessary to make a premarket notification determination. And FDA alone makes the final decision of whether a medical device can be marketed in the United States.
The safety record for 510(k) devices has been strong, and FDA’s substantial postmarket controls have contributed to ensuring that both pre-
market notification and PMA devices meet their clearance and approval specifications and are made in quality systems that require the manufacture of reproducibly safe and effective products.
Attempts to paint the 510(k) process as cursory or fast-track are inaccurate and do not serve the interests of patients, the AdvaMed representative noted. Critics who persist in mischaracterizing the process do not take into account the years that it can take for a manufacturer to compile the data needed for a 510(k) submission. It is common for a 510(k) submission to contain hundreds or thousands of pages of documentation based on bench testing, animal testing, nonclinical tests, tests demonstrating conformity to standards, and whatever other requirements FDA may have.
About a year ago, AdvaMed sent a letter to FDA Commissioner Hamburg and then met with her to discuss three specific recommendations to improve the 510(k) process: prompt resolution of the regulatory status of pre-amendment class III devices, identification of class II devices that had cleared the 510(k) process and might need special requirements, and development of an internal committee to improve the Center for Devices and Radiological Health (CDRH) review process.
In March 2010, AdvaMed submitted comments to FDA in response to a series of agency questions asking for stakeholder feedback on ways to strengthen the 510(k) process. Included in the comments were recommendations to improve 510(k) summaries and the de novo process.
AdvaMed supports FDA’s current risk-based approach to medical device regulation as embodied in the 510(k) process, which makes safe and effective products and treatments available without unnecessary delays.
A representative of the Medical Imaging and Technology Alliance (MITA) stressed the overall good safety record of imaging devices compared with other environmental conditions in the health-related field. In 2007, for example, about 30 million magnetic resonance imaging (MRI) scans were performed; in 2006, over 68 million were performed. However, from fiscal year (FY) 2005 to FY 2009, only 890 medical device reports (MDRs) from all causes, most of which did not result in patient injury, resulted from millions of examinations. In comparison, the number of hospital-acquired infections is estimated at 1.7 million per year, which lead to about 99,000 deaths, and medication-related injury is estimated at 1.5 million per year. The risk of radiotherapy-related errors with serious medical consequences was estimated at 50–100 per million courses of treatment. Overall, the safety of imaging devices is quite favorable.
The 510(k) application has several key components, including the device description; intended-use statement; predicate-device comparison; declaration of conformity to performance standards and mechanical and safety standards; general clinical safety and effectiveness; clinical data, if applicable; device hazard analysis; software description; cleaning, disinfec-
tion, and sterilization, if applicable; and labeling and promotional material. Clearly, the 510(k) process does not consist of simply a signed statement that identifies a predicate device. It is data-driven and rigorous.
Safety is a key part of the 510(k) process in terms of standards and regulations. Devices cleared through the 510(k) process require conformity to international product-safety standards. IEC 60601-1 contains general requirements for safety, but there are also standards specifically for safety (for example, 60601-1-3 for radiation safety, 60601-2-33 for MRI, and 60601-2-37 for diagnostic ultrasonography). Safety is part of the 510(k) process in terms of device features (for example, display of fluoroscopic radiation time for fluoroscopy, acoustic output display for ultrasonography devices, and automatic exposure controls and audible signals to indicate duration and termination of exposure to x-rays).
Access to imaging devices is an important part of American health care, the representative from MITA said, and lack of access to imaging devices because of unavailability poses a public-health risk.
A representative of the Institute of Molecular Technologies said that the current 510(k) process has been a useful and effective tool for bringing technologies to the marketplace efficiently, and for provide public-health assurance that devices have been appropriately cleared for the marketplace. There will always be exceptions to any clearance process or review process, no matter how many types of controls are in place, and vigilance is always necessary. However, in the current regulatory system, many innovations that pose only low or moderate risk would fall under the PMA process if it were determined that there were no predicate devices. The representative suggested a risk-based approach to devices that do not clearly have predicate devices. There are two approaches. The first approach would be comparison with a generic device type already regulated through the 510(k) process. The generic device would have associated predicate devices, but not as specific devices that one would be required to compare the new devices with. The new devices would be considered not substantially equivalent (NSE) because of minor differences in intended use or in performance as described to FDA. The new devices would present low or moderate risk. The second approach to the generic device type would be nongeneric devices for which there is no predicate for comparison, but the risk posed is still low to moderate. Here, the representative recommended a risk-based approach to analyze what information would need to be brought forward in a 510(k) application for the technology.
One of the issues with the de novo process is that it requires filing of a 510(k) application with FDA even when it is known that no adequate predicate is available. A filer must wait for FDA to review the submission, perhaps 3–6 months or even longer; get an NSE letter; and then refile a submission to FDA de novo. The participant recommended cutting out
that process, going directly to a submission, and allowing a manufacturer to follow the risk-based approach based on the Global Harmonization Task Force (GHTF) model for assessing risk with the documentation that must be submitted for agency review on the basis of that risk. This would mean submitting the same kind of information that is provided in the de novo process but without having to go through the first step of getting the NSE turndown and then resubmitting information to the agency.
A representative of a startup ultrasonography company said that in recent years, much impressive ultrasonographic technology has reached clinical users through the 510(k) process. From 2003 through 2008, there were an average of 41 MDRs per year regarding ultrasonography, no class I recalls, and an average of six class II recalls per year. During that time, there were over 100 million ultrasonographic examinations per year, for which well over 100,000 devices were used. That, the participant said, is a good safety record for an important modality that has gone through much innovation. A key part of that success is the FDA guidance document that provided a framework for ultrasonography innovation. The ability to prepare and clear 510(k)s in a reasonably efficient manner has been an important part of medical device innovation. If a device manufacturer is not filing at least one 510(k) every 2 years, it may be falling behind the competition. More important, it means that clinicians have older tools and that changes that are taking place in technology—whether in electronics, software, or signal-processing materials—might not be working their way into the clinical environment as quickly as possible.
A representative of another device company emphasized the process of risk assessment. The international standard, ISO 14971, has a long list of questions regarding risks that manufacturers can consider when developing a new product—for example, What kind of energy does the device put out? What kind of mechanical forces are involved? What kind of human errors can reasonably be expected? What kind of patient interfaces might present issues? She recommended that risk analysis automatically be part of the 510(k) submission. She also drew attention to the European regulatory pathway of clinical evaluation for products that do not have to undergo clinical studies. That encompasses review of the clinical literature for the intended uses of a new product and comparing the new product with existing products.
A participant who is a professor, lawyer, and cofounder of a startup medical device company asked the committee to keep in mind three fundamental points when considering revisions of the 510(k) process. First, as one increases regulatory burdens or regulatory pathways, he said, one runs the risk of affecting patient autonomy and physician ability to practice medicine. Off-label use of products is well accepted, and the American Medical Association has a policy statement supporting it. Second, system and structure issues are separate from implementation and administrative issues. Third, there is a lack of systemic data on whether the system is working in a way that provides protection for patients and whether an inordinate number of unsafe products are being allowed onto the market. There is some information on the innovation side but very little on the safety side.
A psychologist who studies human error from a systems perspective spoke about how to reduce the likelihood of error relative to the safety and efficacy of medical devices. The 510(k) process contributes to the perpetuation of error in the use of medical devices, she said. Current data reveal trends and what kinds of errors occur, but data on why errors happen are needed. The 510(k) process focuses on changes in devices, but errors in the design of predicate devices are not resolved. The 510(k) process can be a good instrument, she said, but it needs to address the question of what factors contribute to error.
A representative of the National Research Center for Women & Families raised concerns about studies that are based on the Manufacturer and User Facility Device Experience (MAUDE) database. MAUDE is widely perceived to be underreporting problems, he said. He is conducting a study of class I recalls of devices that caused severe injury or death, and his results show that the vast majority are 510(k)-cleared products. The study is in review, and he will provide it to the committee.
A medical device submission consultant provided the committee with a report in which he identifies two root causes of the current problems in the 510(k) system: lack of a process for determining what data are necessary to demonstrate device safety and effectiveness and the dysfunctional requirement to demonstrate substantial equivalence to a specific, legally marketed predicate device to determine, according to risk, whether a new medical device is in class I, II, or III. Focusing on the latter, he said that the 510(k) risk-classification process leads to many problems. For example, much of the agency’s review time is consumed in determining the adequacy of a company’s substantial-equivalence justification. In all but the simplest of 510(k) applications, that effort consumes 25 percent or more of reviewers’ time. Every minute spent on reviewing the justification is a minute not spent on reviewing the data that prove the safety and effectiveness of a new device.
Once a bad decision is set in a 510(k) predicate history, innovation suffers. Additional factors contribute to the substantial-equivalence issue. For example, what are the distinctions between intended use and indications for use? How do we properly define technology? The substantial-equivalence argument rests on parsing such terms. Industry cannot predict how a reviewer will interpret the terms, because they are not consistently defined or applied. That gives rise to poor or delayed decisions of substantial equivalence, which in turn delay innovation. The 510(k) substantial-equivalence process also creates public-confidence issues. The press has wrongly characterized the substantial-equivalence process as a shortcut for industry or an abbreviated review for the agency.
The solution to the 510(k) substantial-equivalence problem, he said, is for FDA to adopt a truly risk-based classification system that is blind to whether a device is innovative or “me-too.” The European and GHTF systems are good examples of well-tested processes for rational risk classification. The IOM committee may need to call for legislation that allows repair of the broken 510(k) risk-classification process. By implementing a modern risk-classification process that has flexibility for continuous improvement, we will increase the agency resources available to review the safety and effectiveness of devices, improve the predictability of the review process, improve public confidence in our work, and begin to restore the environment that fosters innovation for better public health.
The committee heard testimony from representatives of Truth in Medicine Incorporated, a patient-advocacy organization that focuses on educating the public about the potential risks posed by and complications of the implantation of synthetic surgical mesh into the human body. Mesh is used in hernia repair, bladder suspension, and treatment of pelvic-floor disorders. Statements were given by the organization’s president and founder, the executive director, and several individual members, all of whom shared their personal experiences with the device. Those participants attended to represent the thousands of others who have had similar adverse experiences with medical mesh products. The organization noted among its accomplishments its successful urging of FDA to issue a public-health notification warning of the serious risk poses by and complications of the transvaginal placement of synthetic surgical mesh. The warning was issued to health-care practitioners in October 2008.
Participants shared their clinical experiences of chronic pain and discomfort, infection, chronic inflammation, incontinence and urinary retention, disability, multiple surgical attempts to remove mesh and address complications, and other illness, which had brought some of them close
to death. Some described having endured over 20 operations in less than 10 years in attempts to remove mesh that had migrated or eroded. Others described the challenge of finding a doctor willing to perform further complicated and risky operations to continue to remove bits of mesh. They candidly described the toll that those health outcomes had taken on their lives, such as the inability to work, loss of employment or personal business, loss of health-insurance coverage, financial ruin, homelessness, and stress on personal relationships, including effects on intimacy with spouses and partners. It was pointed out that additional people had registered to provide comment at the workshop but were unable to attend because of health issues.
Participants explained how mesh systems were cleared by FDA for marketing through the 510(k) process. They expressed concerns that the 510(k) process allows unproven medical devices onto the market, inasmuch as clearance does not require proof of safety or efficacy of class I or class II devices. As a result, they said, an uninformed, unaware public is endangered daily by unsafe and unproven medical devices.
The organization specifically recommended that the committee consider the following changes in the 510(k) clearance process:
Educate the American public about the difference between premarket approval and premarket notification.
Make adverse-event reporting mandatory, with clear consequences for silence by doctors, hospitals, and medical device makers.
Create a specific guide for FDA and CDRH to make better use of their regulatory authority. The decision-making process for when and how to use FDA’s regulatory authority should not be left to the discretion of agency employees.
Include a mechanism which stops medical device makers from paying doctors to use products off-label to increase the sales of their products.
Participants called for expansion of informed consent, making it mandatory, for example, for a medical implant device package insert to be reviewed by the doctor with the patient 3–7 days before surgery (not on the day of surgery or when the package is opened in the operating room). It was also stressed that stakeholder involvement should be a critical component of the 510(k) process. One suggestion was for FDA to meet regularly with patients who have been adversely affected by the process. Participants also said that the 510(k) process should be more transparent.