Patients are being harmed because FDA doesn’t commit sufficient resources to enforce federal requirements for hospitals and manufacturers to report adverse outcomes caused by medical devices, according to Larry Pilot, one of the original authors of FDA’s 510(k) device clearance process.
There is no need to revamp the 510(k) process, Pilot said. The agency should instead enforce existing laws to make device manufacturers and user facilities report adverse outcomes.
“It’s a good system we have now, but FDA needs to allocate more resources to enforcement, especially for user facilities,” Pilot said. “That’s why FDA, with its authority, and this opportunity to put the emphasis on reporting to the user facility required by law, and then by specific regulation, should be going for money penalties for user facilities that do not abide by the reporting requirement.
“How has FDA been managing what these user facilities are doing? What has the agency been doing?
“Because people like Amy [Reed] are being hurt by this. If she knew that it was, instead of one in 10,000, a one in 300 chance, she’d probably say no. This is an embarrassment to the agency.
“Going back to the responsibility of the external community, the users and manufacturers—I expect the FDA to have done a better job.”
Pilot received the FDA’s Award of Merit in 1977 for his contributions to the development of the agency’s medical device programs. He is a contributor to the Competitive Enterprise Institute in Washington, D.C.
Pilot spoke with Matthew Ong, a reporter with The Cancer Letter.
Matthew Ong: Patient advocates and the IOM report in 2011 say that the 510(k) process does not adequately protect the public from harm. What did you have in mind when you designed the 510(k) in the 70s?
LP: For that regulation, I would never say that I wrote it, because it’s a team of people; I was one of the supervisors, of course. What its intent was—and how it’s been altered through legislation, and sometimes altered not always for the benefit of the public by the bureaucracy, over the years and decades.
First, I disagree with the recommendation of the IOM, and I did publish a piece for the Competitive Enterprise Institute titled “Stifling Medical Device Innovation,” and I explained my position in that document.
Initially, the premarket notification process—the 510(k)—was implemented in order to identify to the FDA what was to be entered into commercial distribution in 90 days. It was a notification process, and the objective of the notification process was to enable the FDA to look at what the intended use of the device was and whether or not it fit into one of the three categories.
There was no requirement of feedback to the notifier, the submitter, unless the agency believes it was useful to convey some kind of message back to them. And there was no order attached to that, which is where we now move from 1976 to 1990, where there are extensive changes to the law, one of which stipulated that the 510(k) notification process now was to require the issuance of an order from the FDA in order to commercially distribute the device.
Before that, and I represented clients who submitted notifications to the agency, got no response whatsoever, and marketed devices in the 90th or 91st day. I did have clients who got notification, and for which the agency was maintaining that they had to go through premarket approval or something else, and for which the client to disagree and discard that recommendation, including for devices that were in Class III.
That was because of the particular nature of the transitional process, because devices which were in Class III—say, a heart valve—had not been manufactured by any other manufacturer; that new entity could market after the 90-day notification.
The company can market unless the device did not fall within the description, or the agency believed that it was not reasonably safe or effective. Because there was a three-year period during which those manufacturers of heart valves marketing on the day of the enactment, they could continue marketing, but by the end of that three-year time, they must have submitted an application for which then the agency would decide whether or not they could continue on the basis of the evidence that they provided to support the reasonable assurance of safety and effectiveness. If not, the agency may believe that it was imperative to reclassify what had been classified.
Then we move into back into the 510(k) process. The idea was to submit information to the agency to enable them to know who was marketing what and from where. That moved along fairly well from 1979 up to 1990, although there were some hearings and some events that suggested that perhaps there should be a better process, which led to, of course, 1990 amendments. That is the order process.
Now, with that process, the agency had much more leverage now over who sent in what, but in addition to requiring this issuance of an order, the requirement for standard was broadened into a special controls category, where the special control could be identified by the agency and for which those who were marketing Class II devices—there was no standard, and that was one of the impediments to the success of that venture into standards. Then the agency could identify what the special controls ought to be for which the agency would then review against what would be an order if they agreed.
From my perspective, both from the inside, the start of this program, through all the years that I have been in practice, CDRH has plenty of authority to stop what enters into the marketplace or to accelerate a remedial activity, including possible withdrawal of the device from the marketplace.
Over the years, it accomplished more feedback approaches to safety and effectiveness than were available for the drug side, where there is no requirement to report any adverse outcomes unless it were subject to a New Drug Approval.
For all devices—whether they were approved or cleared or never cleared but in commercial distribution—federal reporting regulations requires not only manufacturers, but subsequent to the 1990 amendments, user facilities to report incidents related to death or serious injury. Those are the controls or the expansion of controls that gave to the agency much more ammunition to correct what the agency believed was necessary to provide a reasonable assurance of safety and effectiveness.
My take on this is—relating to how the 510(k) process was implemented, first through the statute and then ultimate regulation—I disagree with the agency’s position and in particular, the IOM on the characterization of the classification process as risk-based.
It is not risk-based, and I have said this many times, because the criteria that applied to each one of the three classes was in the context of reasonable assurance of safety and effectiveness. A subset of safety is risk, yes, but this low, medium and high risk is fabrication.
The Europeans have a risk-based system, and what they rely on to clear or approve a device on is safety. They leave it up to the consumer and user or the intermediary practitioner the evaluation of effectiveness, whether the device is effective for its represented, intended use. For instance, a pacemaker needs to be safe—it’s not going to shock people inadvertently, it’s not going to fail when it’s needed. Now the medical community will decide the effectiveness of that device.
That is a risk-based system. But to me, that’s a lower standard that what we have here now, which is safety, which includes risk, and effectiveness, which goes beyond risk.
The risk-based system, which the FDA continues to advocate and incorporate into their websites—that’s a fabrication, many years after the three-tier classification was created. It became more intense at some point in time where I began to notice, and that was probably back around 2009, when Jeffrey Shuren testified before a Congressional committee.
I received an advance copy of that and I communicated with the staff person on the committee and I said, “This is not right. You should have somebody questioning him—what does he mean by risk? How is that incorporated into what the statute expresses?”
The statutory language, which didn’t change, is quite clear that it is about safety and effectiveness, whereas functionally, if people want to look at risk, fine, but it’s inappropriate for the agency to use risk as a surrogate for the explanation, when safety and effectiveness are the criteria.
For drugs, it’s absolute safety and effectiveness, whereas in the statute applicable to devices, it’s reasonable assurance. That was a distinction that was made during the legislative process, because the objective was to avoid pigeonholing this industry into the criteria that applied to drugs.
You are saying that the 510(k) process is not responsible for clearing devices that are potentially harmful. Who is at fault, then, for the deaths and harm to women by power morcellators?
LP: I know from what I’ve read that the manufacturer, the subsidiary of J&J, may have had information to suggest that there was a nexus between the use of the morcellator and possible sarcoma upstaging. If the subsidiary of J&J had that, the subsidiary was responsible for looking into that and saying, “Look, if this is true, is this going to be a reportable event?”
A good judgment could indicate, yes. I’ve represented companies who have been faced with these kinds of questions: “We just got this report, our device has been approved by the FDA, within the first five week, this has happened, what should we do?”
If you can’t figure it out, and this is advice I’ve given clients, “Well, let’s go to the FDA, don’t be afraid to go to them, because maybe they know something from their experience that we don’t know.” So we enlarge the scope to include the agency.
As for Brigham & Women’s Hospital, they were obligated, in my view, to notify the agency. If it’s true that the J&J subsidiary had enough information to be able to submit within 30 days, and if it’s true that the user facility had the information, and they didn’t report, both of them probably could be successfully prosecuted for failing to comply with the requirements of law.
If you go back further and you take the classification process and the review of the morcellator—how is it that a surgical instrument, which is different from the kind of surgical instrument that had existed in 1976 or 1979 when it was classified, is that substantially equivalent to what power morcellators are now?
That’s a judgment that the manufacturer, the sponsor, and the agency have the opportunity to resolve and make. Now what kind of an improvement in the legislative or the regulatory process can be made to that kind of judgment?
Somebody there should’ve said, “You know, this one, when you scratch it, it doesn’t give off any real perfume, and there might be a little smell to it.”
FDA needs to recognize that we’ve got special controls here, maybe that the agency should ask them about some of the clinical data, or require agreement on some kind of a study to evaluate whether grinding up pieces of stuff inside the abdomen that might be malignant, and say, let’s get some more information on that.
Part of the agency’s responsibility, at least when I was there, was, if we thought there was a signal that we should give to somebody, whether it was in a 510(k) or compliance context, we give them the signal.
Those are the kinds of judgment exercises that weren’t available in 1979, but could’ve been used in conversation with the applicant of the 510(k) and said, “Have you thought about this, do you have anything on that?” instead of, “Look, you haven’t done this, and we want to ask 1,000 questions.”
For Class II devices that might be on the fringe in the context of the application of that definition, if there is suspicion or belief, it has to be reasonable and supportable on the part of the agency, that some additional special controls—clinical studies—need to be undertaken before we can clear this device for the market.
Don’t you think they should’ve asked for something a little bit more in terms of evidence of safety when the morcellator was cleared over 20 years ago?
Registry trials do work to a certain extent, but not through a formal statutory requirement. The agency has the flexibility—special controls for the Class II devices—to track devices. The agency should judge the need and undertake that function.
I believe there were some people who were concerned about the power morcellator years ago. Why not tack on to that postmarket surveillance, and say, “Let’s work together and evaluate over time the performance of these devices?”
From my perspective, and having been one of the few people to start the process and review its performance on an experimental basis up until 1976 and in the real world after that, there’s more than enough authority for the agency to do what it needs to be done if the agency is administered by competent individuals who have a knowledgeable staff understand what the provisions of law and regulation are, and discharge their responsibility correctly.
Patient advocates are saying that individual practitioners—not only manufacturers and user facilities—should be required to report adverse outcomes. What do you think?
LP: If the agency cannot manage the present system for information supplied by user facilities and manufacturers, how can you expect a multiplier of 100,000 physicians all across the country? No.
If the agency would manage, in cooperation with the health care community and the industry, the responsibility it has over the medical device reporting system, you don’t need individual practitioners. But individual practitioners can always be encouraged to voluntarily report to the agency, and I think that’s a good thing.
Mandatory reporting? Can you visualize what millions of reports a year might be like? Impossible.
And FDA imposed on itself an unreasonable burden at the time they were beginning to apply the medical reporting regulations, because they took some very strict interpretations of what an adverse event was.
I had one client, maybe they’d send in 10 or 20 a month but they were conscientious about those few. The response they got from the agency’s counsel was, “Oh no, you have to send all those in.” But that would be a thousand a month, so they go from sending in 120 that are good reports, because they’re substantive, to junk. What are you going to do with 1,200 reports? And that’s just one company, so it ballooned up to the hundreds of thousands. At one point they had 200,000 reports coming in.
No, individual practitioner mandatory reporting would just complicate things for the agency enormously. It’s better to filter it through the user facility or manufacturer—report to them if you don’t want to report to the government, because the conscientious reporting is required by law.
Litigation is usually the best tool for regulating an activity. If you look at other incidents, even during the time I was at the agency, the plaintiffs bar had a much greater influence on determining what would be in the marketplace than the FDA. When I preach and lectured on this subject, I’d say, “You ought to be more worried about the plaintiffs bar, than you do the FDA.”
However, it should be the agency that is the monitor of that and the facilitator for the patient. I haven’t witnessed much improvement since Jeffrey Shuren has been there, or since that IOM report was issued. It’s unfortunate, all the missed opportunities, and that goes to the management of the organization.
How, then, can FDA more reliably pick up signals that patients are being harmed, when voluntary reporting hasn’t worked in the case of power morcellators? The agency said that it is unaware of the dangers until Hooman Noorchashm and Amy Reed filed a Medical Device Report in December 2013.
LP: It’s a good system we have now, but FDA needs to allocate more resources to enforcement, especially for user facilities.
The scientists evaluate the reasonable assurance of safety and effectiveness. The enforcement side of the agency is to detect whether a violation or possible violations have occurred.
And there’s a crossover when you have the medical device reporting problem, if it exists, where the hospital or manufacturer did not report these events as they should have.
The reporting is necessary as a signal, an early warning, for which then the scientists have to get involved, and the medical, technical people in the agency to evaluate, “Is it one in 300, or is it one in 10,000?” to establish what is a reasonable approach to the availability of the product in the marketplace, and the judgment that is exercised by the user community as through the advisory committee process.
That’s why FDA, with its authority and this opportunity to put the emphasis on reporting to the user facility required by law and then by specific regulation. The agency should be going for money penalties for user facilities that do not abide by the reporting requirement.
I know for almost 10 years, many user facilities have not been keeping records and did not have the required written procedures for evaluating whether a complaint justified an MDR, and FDA didn’t know that? They haven’t looked at any of this?
How has FDA been managing what these user facilities are doing? What has the agency been doing?
Because people like Amy are being hurt by this. If she knew that it was, instead of one in 10,000, a one in 300 chance, she’d probably say no. This is an embarrassment to the agency.
Going back to the responsibility of the external community, the users and manufacturers—I expect the FDA to have done a better job.
You’re saying that FDA needs to do more surveillance of user facilities, but how can this be done when the agency has limited resources?
In a recent response to inquiry from Rep. Mike Fitzpatrick (R-Pa.), FDA wrote, “We have general focused our enforcement resources on manufacturers—who are required under law to investigate any MDR-reportable complaint they receive—and not on user facilities. We have found that encouraging more reporting—and more complete reporting—by user facilities is a good use of our limited resources in this area.”
LP: What? That surprises me, because I believe too much of the agency’s resources go to manufacturers and are wasteful. Again, this is a topic I know very well, from inside and outside through many different experiences.
When I was at the agency, I lectured on this subject and on inspections, I said, “No inspection should last more than two days, I don’t care about the size of the company.”
Within two days, a good inspector should be able to determine—and we’re talking about the application of Good Manufacturing Practices, which nearly every manufacturer must comply with, almost totally with limited exceptions—if something is a little bit odd, or jump to the company’s records.
If it’s more than two days and if it’s a good inspection, then you’re on track at the very least an eventual warning letter.
Since my time, I’ve read warning letters where the agency has been in the facility for a month, sometimes with two inspectors, and then a warning letter goes out sometimes six months later. What is the importance of something six months later, after you’ve spent all the resources and time?
At the very least, take some of those huge resources you now have that we didn’t have at the time, and apply them to evaluating the user facility community, whether it’s hospitals for stay or ambulatory facilities.
Go out there and develop a statistical base for what you’re doing to evaluate: one, how many user facilities have written procedures, two, how are they following their procedures, and three, in following their procedures, are they making reports based on what is in the regulation itself as the criteria and their own procedures in implementing that.
If Brigham would’ve been inspected at any time in the past few years—let’s assume that they did have procedures but people sort of after awhile forgot about them—they would have picked that up in the inspection.
So I’m astounded by that kind of response, because it’s, “Oh, we put more of our resources here, but on this very important function, what we do is kind of encourage people.”
Encourage? I could never write a letter like that to Congress. And this is their associate commissioner for legislation. That letter must have passed through Shuren’s office, he had to have seen that letter that says, “Oh, we have a voluntary system for the user facility community, but with the industry, we have a mandatory system.” You’ve got to be kidding me.
This kind of judgment or failure to exercise judgment has been harmful to the agency and more importantly, to the public. This is our world—FDA’s world—and we need to at least have some minimal understanding of surveillance, or when it’s important to step up surveillance.
If Brigham at Harvard isn’t reporting, what about other groups and other user facilities?
In summary, what do you think is the root cause, then, for what happened to patients and families who were affected by power morcellators? And what can we learn from this case study?
LP: It is a failure of FDA to properly administer the flexible authority and the required authority that it has. The root cause here is the failure on the part of the community to get the information to the FDA.
It’s not a fault of the statute, or even the implemented regulations. It’s the failure of the agency to responsibly evaluate the performance in the marketplace.
Now, if people are going to be dishonest on either side, whether it’s in the agency or within the industry, then they should be punished, and we’re going to make them as an example that is administratively and judicially fair.
It is a disappointment to me, as someone who has been in this arena from its inception in 1970 to the present, and seeing it from both sides, because I’d rather have a more congenial relationship between the agency and the health care community and the industry.
It’s much more effective, instead of the “us vs. them” mentality the agency has now.
Let’s circle back to your relationship with FDA. How did you become involved in creating FDA medical device legislation in the 1970s?
LP: My interest in the topic on medical device legislation began 1969 when I joined the government and then shifted over to FDA. Medical device regulatory processes were minimal then—it literally hadn’t been looked at since 1938, when the term device was identified.
So then we moved forward to do what it took to get to the 1976 amendments, and that was an almost six-year process. To implement the regulations, of which we were able to accomplish most by 1979, and that’s when I left the agency.
Another fellow, David Link and I were the two people who were assigned the responsibility in 1970 to plan for and interact with those who would ultimately write and approve the legislation. We created the Office of Medical Devices, which ultimately became the Bureau of Medical Devices, and which is now the Center for Devices and Radiological Health.
I was responsible, as a surrogate to Dave, who was the director, on all matters relating to regulatory issues, but in that context of implementing the regulations, Dave and I both were intimately involved in everything across the board, including the 510(k) process. I still practice law.
You wrote the existing three classes for medical devices with David Link?
LP: That concept was introduced—not exactly three classes—as a result of an intradepartmental task force for Ted Cooper was the chairman, and he was the director of the Heart and Lung Institute at the time. But this task force was directed upon the request of President Richard Nixon and his consumer message in 1969, where he suggested that the agency take a look at its authority over the medical device industry as a complement to the regulation of drugs.
That’s what the task force was set up for, and they completed their work in 1970 and issued what is known as the Cooper Committee Report, which laid the foundation for what ultimately became the Medical Device Amendments of 1976.
And the concept of this tiering, that ultimately became Class I, II and III, was recognized among this industry or this practice of medicine that the breadth and depth was clearly much different than anything that existed in the drug arena—because we’re looking at Band-Aids all the way to heart valves, so to speak.
That’s how the legislation developed into the necessity to classify devices, that is devices that were in commercial distribution at the time of the enactment on May 28, 1976—we’re getting up to 40 years now.
Dave and I thought we’d get a head start on the process in the early 70s and Commissioner Charles Edwards said, “Go do it!” so we organized two test panels, one in the orthopedic arena, and one in the circulatory system—cardiovascular, thoracic—to evaluate how a committee consisting of physicians, scientists, engineers, industry, consumer, could work together to develop a plan to evaluate distinctions based on what were adequate controls to provide reasonable assurance of safety and effectiveness.
So obviously, for a tongue depressor, the floor would be Device Registration and Listing, Good Manufacturing Practices, etc. But for those devices for which something more was necessary to provide that reasonable assurance of safety and effectiveness, the concept of compliance with standards would be the Class II devices—Class I being general controls, Class II being standard.
It was contemplated at the time that standards would be rather broad—maybe electrical safety, sterility—some broad categories applicable to the particular device or group of devices for which safety and effectiveness would be reasonably compliant.
And then the third group, for those devices that met the definition of Class III—life-sustaining, life-supporting or for which there was not a reasonable risk of illness or injury associated with the use.
With those two test panels, we were able to agree on a questionnaire system to be used by the individuals who were on the advisory committee, and then subsequently, after their work was done, to publish in the Federal Register a notice to promulgate a rule by regulation.
That’s how the process began before 1976. So all the 10 or 12 advisory committees that we had had looked at approximately 1,800 different classes of devices to place them into one of these applicable slots.
But when the amendments were enacted on May 28, we had enough background to hold a formal, public hearing to have the panel agree on recommendations, which then were published in the Federal Register as a proposal for the public to comment, and for the agency—then it was the Bureau of Medical Devices as part of the FDA—to evaluate those comments and publish a final rule with explanatory text. That became the official classification process for devices.
It took from 1976 to the mid-80s before all those proposed regulations were completed and finalized. That formed the basis for the formal classification system.
If devices were developed after that time and for which they couldn’t slot into one of those three categories, because they might be completely novel—say, an artificial pancreas, what would you do with an artificial pancreas? Or if there was a pre-enactment device before 1976 that had been in lawful commercial distribution, but not known to the agency or the committee at the time, then those would ultimately have to be classified, because they were pre-enactment.
On a separate issue, the Supreme Court ruling in Medtronic v. Lohr (1996) generally held that FDA approval of a Class III medical device preempted state product liability law. What was the rationale of that ruling, and why does it sharply limit the types of recourse a patient might have for injuries?
LP: Before Medtronic v. Lohr, there was another case involving the 510(k) process. The immunization from products liability is with reference to the premarket approval process.
That’s what the intent was back in 1976, so that when FDA approves a device, manufacturers wanted some protection from frivolous lawsuits. The companies said, “If you are going to approve the PMA device, we want some insulation here, because this judgment-making process, while it’s mutual, FDA is giving us the green light that there’s a reasonable assurance of safety and effectiveness. We don’t want any frivolous lawsuits for the PMA devices.”
The Class III approved devices are entitled to this exemption, but it does not apply to Class I or II devices. The Supreme Court has ruled on its application, and it would take a Congressional amendment to the statute to modify or eliminate the exemption.
This exemption from liability suits would be difficult to accomplish. If there was to be some fraud associated with the Approval process, this could be the basis for some further litigation as to application of the exemption. At present, there could be other avenues for judicial relief that may be possible, but I am not aware of any.
I really respect what Hooman and his wife, as well as what some of the others have done, but I disagree with his penchant for a legislative change.
FDA should use present resources to provide greater surveillance over how user facilities are complying with the MDR regulation. If either the manufacturer or user facility failed to comply with the regulation, they could be subject to civil and/or criminal penalties as well as lawsuits from plaintiffs for negligence per se.
It’s within the authority of the agency to detect violations, but better, to prevent by trying to understand what’s going on the in the marketplace before waiting for an incident.
