Assuaging both agency and industry concerns requires a precarious balancing act in the 510(k) overhaul.
FDA’s massive revamp of its battered 510(k) market-entry program for medical devices has emerged as a struggle between two primary competing goals: industry’s need for better predictability in the system, and the public’s need for restored confidence in the process at the agency.
The two aren’t mutually exclusive. But they are seen as a delicate balancing act—an act that became dangerously out of balance in recent years. From industry’s perspective, the regulators are surreptitiously changing the rules by introducing testing creep and consequent unpredictability, while from CDRH’s viewpoint, industry has been getting away with too much predicate creep—slipping complex new technologies through as substantially equivalent to 34-year-old technologies. It’s a polarization of positioning that is built around a common recognition: a revamp is overdue.
Predictability was the recurring theme from both companies and consultants who stood before CDRH officials during a recent public meeting to offer their perspectives on revamping the program.
“The current 510(k) process is beginning to hobble device development and innovation, and there is a perception that the public’s confidence in the system has been shaken,” Coombs Medical Device Consulting principal Craig Coombs told the audience. “FDA can fix both of these and thereby strengthen the process if they focus on what often works, such as establishing predictable and reasonable testing regimens for the 510(k) process today, and shift resources away from those aspects that do not contribute to public health or confidence.” Coombs said that strengthening the predictability and reasonableness of the testing process would “improve innovation, maintain public health, and regain public confidence.”
A CDRH task force is assessing the pros and cons of the 510(k) program and will soon offer recommendations on how to change it, center director Jeffrey Shuren said. The task force is operating in parallel with an Institute of Medicine review that will not prepare its report until March 2011. In the meantime, Shuren said that any changes to the 510(k) process will be open for public comment sometime in June before they are fully implemented in September.
|Smith says that clearly established precedents are being abandoned, which hurts innovation and patient care.|
Hogan & Hartson partner John Smith couldn’t have agreed more on the increasing unpredictability and considerable delays occurring at CDRH. He said at the meeting that longstanding processes have been altered without official announcement or explanation and that there is an increasing reluctance to accept combination or split predicates. Smith also stated that clearly established precedents are being abandoned and more nonsubstantial equivalence decisions are being rendered, among other complaints. All of this translates to increased review times and “a cumulative negative impact on innovation and patient care,” he said.
Coombs raised a few eyebrows when he advocated scrapping the current reliance on predicate devices altogether. “We need to move away from the precedent-justified system to a more truly risk-based system,” he said. Coombs acknowledged that legislation would be necessary for such a change, but he said that in the meantime, the agency needs to maintain consistent guidance on how industry develops a substantial equivalence justification.
“Since last summer, new unwritten rules for substantial equivalence justifications have emerged and caused tremendous delays in submissions and review times,” he said. These changes include an insistence on a single predicate, a rejection of the reasonable use of split predicates, and resistance to even small changes in indications for use. “Changing such a significant portion of the 510(k) process without providing prior guidance harms device development and does not strengthen the 510(k) process,” Coombs said.
CDRH should consider allowing device clearances without the need for a predicate when there is no acceptable predicate available, Medical Imaging & Technology Alliance’s Richard Eaton told the audience.
“It is becoming increasingly difficult for a device developed today to be compared with a device that is 30 years old,” he said, advocating use of CDRH’s harmonized STED (summary technical document). STED could be an approach for such clearances “because it is risk based and involves conformance to central principles to demonstrate product safety,” he said.
MED Institute regulatory scientist Daniel Dillon also said the agency should get more use out of STED in the 510(k) process. “This document requires the assessment of known and anticipated risks that may be associated with the device, and an assessment of the clinical data that are available and pertinent to a new or modified device,” he said.
The agency should consider increasing the use of special controls, according to Dillon. “We believe FDA should reopen the classification regulations for those Class II devices that present the most risk and then promulgate special controls for those devices.”
He said FDA should also increase use of the de novo classification process for new low-risk devices that have no predicate. “Any device with a new intended use and whose technology is not high risk should be reclassified via the de novo process. We believe these changes can be easily made within the existing regulatory framework.”
|According to Eaton, CDRH should consider clearing devices without a predicate when there is no comparable one on the market.|
Whatever changes the agency decides to make, FDA needs to address what ExploraMed general counsel Ed Bright termed as “testing creep.” He described this as a result of FDA’s lack of a firm commitment on testing requirements to obtain a 510(k) determination. Although it may seem reasonable to bump up a study requirement from 100 patients to 200 or one-year follow-up data to a two-year follow-up, such changes in requirements would cost millions of dollars that most small companies may have trouble raising in the current economic environment, Bright said.
New World Regulatory Solutions scientific affairs director Glenn Neuman complained that his group is seeing too many inconsistencies at the review level. He said a better approach would be to establish and publish clear requirements for 510(k) submissions and “hold everyone’s feet to the fire. Then we would all know what we are doing and it wouldn’t be like this black box approach that it is now. Standardizing and continually updating requirements will best protect public health and provide a level playing field.”
Neuman said device predicates should represent the accepted standard of practice, and if there is an old predicate that doesn’t meet current standards, then it shouldn’t be on the market anymore. He said FDA now is strongly recommending specific predicates for submissions. “We had a predicate rejected in an [investigational device exemption submission] in exchange for a better-known predicate by FDA,” he explained. “Just because it’s more widely used doesn’t mean it’s more accurate. Predicates are a double-edged sword and maybe the agency should use the de novo process more.” He also said the terms “intended use” and “indications for use” are confusing and they should be consolidated into one term.
Hogan & Hartson’s Smith said that in moving forward, it is important to recognize the flexibility of the 510(k) process. “You have constantly evolving products that should not be held to druglike standards.” Also, he said, the agency has to recognize Congress’ intent to gauge data requirements and match those to the device at issue.
Earlier in the meeting, CDRH had a chance to vent. Reviewers face significant challenges with “incremental design changes and device creep,” Office of Device Evaluation (ODE) engineering and science deputy Christy Foreman told the audience. She said modifications to 510(k)-cleared devices are often not submitted to the agency unless they are changes that could significantly affect safety or effectiveness or there is a modification to the intended use.
“But these submissions for modifications are based on a firm’s determination regarding the effect on safety and effectiveness,” and not FDA’s determination, Foreman said. “This information is kept in internal firm files and not subject to FDA review except during inspections. Firms often interpret the regulation, which states ‘could significantly affect’ as ‘does it significantly affect?’” Based on the latter interpretation, if the manufacturer has data showing that a change doesn’t really affect the safety and effectiveness, then the agency is not given an opportunity to reach the same conclusion.
Foreman also raised challenges about the use of predicate devices, such as when companies choose a poorer-performing device as the predicate when better-performing devices are available. They choose the poorer-performing device to make it easier to obtain a substantial equivalence determination, she explained.
|Neuman says that standardizing 510(k) requirements is the best way to protect public health and provide a level playing field.|
“This construct may serve to inhibit device improvements for performance for the entire device class and I don’t think that serves the public health,” Foreman said. “We also run into cases where the original device may no longer be marketed due to subpar performance, yet this device serves as a valid predicate because at the time it was equivalent.”
ODE deputy director for premarket program management Barbara Zimmerman hinted at making changes to CDRH’s third-party review program because the reviews by accredited parties are fraught with problems. She noted that about 300 (8% of all 510(k)s submitted) third-party submissions are processed annually. The poor quality of these reviews requires additional FDA resources to evaluate and address problem issues.
Zimmerman said that problems from third-party reviews are typically caused by a lack of a device-specific guidance document to help accredited parties review the application when they receive it. Also, accredited parties do not have access to previous decisions or reviews of other similar device types, unlike FDA reviewers. “We need to think through what we can do to assist our third parties in dealing with these challenges,” she said.
ODE 510(k) staff director Heather Rosecrans described challenges from having limited authority to rescind a 510(k) clearance. “In the absence of a robust rescission authority, it is difficult for FDA to address problematic predicates,” she said at the meeting. A rescission regulation was proposed in 2001 but was never finalized.
Another challenge highlighted by Rosecrans results when FDA collects postmarket safety information on a certain device type, and then the agency asks for new information to address a safety signal seen in devices currently under review. This can create an inequality perception among those arguing for a level playing field, which Rosecrans believes more regulatory authority would address. She also said the agency has concerns about final printed labeling not being required before or after devices are cleared—the center often finds that manufacturers make changes to the draft labeling without FDA clearance when such changes could require a new 510(k), she explained.
“We believe increasing transparency is essential for all of our stakeholders and for the FDA staff,” Rosecrans said in summarizing FDA’s perspective. “Achieving consistency is critical…and developing clear definitions, guidance, and additional authorities may be required, which would all go through a public process.”
|Ulatowski said that devices are misbranded if devices if they are labeled for use with the SS1, pictured here. Image courtesy of STERIS CORP.|
A recent letter to endoscope manufacturers from CDRH compliance director Tim Ulatowski warned them of possible misbranding of reusable devices labeled for reprocessing by the Steris System 1 (SS1) processor.
He reminded them about a December FDA notice in which the agency informed healthcare facilities that Steris Corp. had significantly modified the SS1 and that FDA had not approved or cleared this modified product. Ulatowski said that if devices are labeled for use with the SS1, then they are misbranded under section 502(f)(1) of the Federal Food, Drug, and Cosmetic Act because they fail to bear adequate directions for use.
The letter recommended the companies follow four steps:
1. Immediately review all labeling, including online information, for references to SS1 and revise it to comply with requirements.
2. Consider adding the notice: “The Steris System 1 (SS1) is not a legally marketed device,” and that the company’s labeling will be revised to identify reprocessing methods using legally marketed devices.
3. Take immediate action to validate at least one reprocessing method using legally marketed devices if the SS1 is the only method described for reprocessing your device.
4. Determine whether your labeling changes require a premarket approval (PMA) or 510(k) submission.
a. Relabeling a product approved in a PMA may require the submission of a PMA supplement.
b. Relabeling a reusable device cleared in a 510(k) premarket notification may require the submission of a new 510(k).
In the wake of a long legal battle with FDA, TMJ Implants Inc. (TMJI) CEO Robert W. Christensen announced to close sympathizers in April that he is filing for Chapter 7 bankruptcy.
|TMJI's CEO again took to his YouTube channel recently to discuss FDA's "biased, corrupt" overregulation of the medical device industry.|
What pushed him to this step, he told FDA Webview, was a March 30 letter from FDA associate chief counsel for litigation Tara Boland that offered him 10 months to pay $340,000 in civil money penalties for not filing 17 disputed MDRs. Christensen had asked for 10 years to pay, in light of his and the company’s financial condition. Boland attached an agreement form that demanded $170,000 immediately and $17,000 a month thereafter plus 7.5% interest on late payments.
Christensen says bias against his devices by CDRH dental devices chief Susan Runner was first revealed in 1999 when she ordered his two prostheses off the market pending reclassification into Class III. She then expedited a competitor product by TMJ Concepts Inc. while delaying action on his. According to Christensen, Runner said that TMJI’s devices were 1960s technology that should be modernized.
“It was Runner who told me a week before the first panel hearing in May 1999 that I shouldn’t come back to the panel as we would not prevail,” Christensen said in an e-mail to the author of this column. “Give me a break! I told her I would think about it overnight and get back to her. I did and said we are coming. TMJI’s presentation was a million times better than TMJ Concept’s. Dr. Skinner, the lone orthopedic surgeon on the committee, said ‘nice presentation’ to me right after it. He knew we were correct.” That panel meeting voted to recommend approval of both of Christensen’s two prostheses, but Runner and her staff allegedly did not want his biggest seller (a unique partial metal-on-bone joint) on the market, so it was deferred by Runner to a different panel meeting a year later.
At that second meeting, which Christensen says was “rigged,” the panel voted not to recommend the partial joint’s approval. After a 19-month delay during which the company lost $6 million, CDRH approved the first, low-sales full prosthesis (TMJ Concepts’ competing device had been approved in two months and had gained a substantial market lead). The following month, over staff objections, then Office of Device Evaluation director Bernard Statland approved TMJI’s partial joint.
Christensen’s troubles weren’t over with the staff-resisted return of his bestseller to the market. As he continued to complain internally about the injustice he believed he had suffered, including demands for HHS Inspector General and FDA Office of Internal Affairs investigations, other FDAers had their feathers ruffled and resented his refusal to “just get over it,” Christensen alleged.
Twelve months later, FDA struck again. This time it conducted a two-week inspection of Christensen’s MDR files that prompted a warning letter six months later, beginning TMJI’s mortal second battle with the agency over 17 incidents that FDA said should have been the subject of MDRs. Christensen disagreed, starting a hair-splitting war of words over whether disease-progression surrounding an implant is a reportable MDR “event” involving a “serious injury” that might be viewed as implicating his device.
Mindful what market damage his FDA competitor TMJ Concepts could wreak with those 17 MDRs, Christensen took the battle all the way to the 10th Circuit Court of Appeals, losing at every stage. Each review level was apparently convinced by the rubber-stamping below it in a “perfect” paper record.
Christensen told me that throughout the MDR civil money penalties ordeal, “I kept showing FDA and HHS I owned nothing but a poor company….I used our millions to bring the company through the approval process, never asking anyone else for help. It cost us millions, to the point I am now penniless.”
The protracted legal battle waged with FDA leaves Christensen jobless, in debt to the tune of hundreds of thousands of dollars, and lacking the wherewithal “to pay our home mortgage. It will be lost.”