Medical Device Daily Washington Editor

Given all the commotion regarding the 510(k) program at the Center for Devices and Radiological Health, it seems that changes are surely in the offing even before the Institute of Medicine (IOM; Washington) completes its review of the program. Medical Device Daily interviewed Donna-Bea Tillman, PhD, director of the Office of Device Evaluation, regarding the program last Friday, and Tillman let it be known that at the very least, more clinical data will be required of 510(k)s even before any official action is taken.

Tillman reinforced the message that a new day has arrived at CDRH, saying that "just because we did it a certain way in the past doesn't mean we're going to continue to do things the same way" where product approvals are concerned.

MDD asked Tillman whether it was safe to assume that most of the concern regarding the 510(k) process is over implantable devices. "No, I don't think that's true," Tillman said. She said that she and others at CDRH are "in the process of taking a look at the program," most likely a reference to the effort under the stewardship of acting CDRH director Jeffrey Shuren, MD, to analyze the 510(k) process.

This effort was announced by associate FDA director Joshua Sharfstein, MD, at this year's annual conference held by the Advanced Medical Technology Association (AdvaMed; Washington). Sharfstein said at the event that some of the standards for device clearances are poorly defined, and said that Shuren has "a very clear mandate to move forward on clear priorities" (MDD, Oct. 16, 2009).

Tillman acknowledged that there were "obviously some questions raised in the press and in Congress due to some problems with devices," adding that "it's time for us to take a good hard look at the program." Tillman also said, "we're planning a public meeting for next spring" to solicit input on the issue, but the agency has set no firm date as yet.

MDD asked Tillman what she sees as the biggest problem with the 510(k) process, and she pointed first to indications for a device's use.

"One of the things we're looking at is when a new indication becomes a new intended use," Tillman said, adding that her aim in this regard is to "better clarify what that is has always been a challenge" to regulate. "Another challenge ... is that science is not static," she said, so "managing and running a program like the 510(k) in a world where science is constantly evolving ... is difficult." She added that in her mind, technological change is "the biggest challenge" faced by CDRH in its administration of the 510(k) program.

As to which categories of devices she thinks are most problematic from a regulatory perspective, Tillman seemed most concerned about cutting-edge technology. She said the agency has "issues with things like combinations of products [and] issues with smart devices and smart technology" as well. She added that she sees a substantial demand on the center's resources engendered by devices associated with minimally invasive procedures, adding that the push toward personalized medicine is keeping the Office of In Vitro Diagnostics quite busy as well. "There are a lot of technological drivers that we have to deal with," she summarized.

Regarding the phrase "substantial equivalence," Tillman said the phrase "has a very precise legal definition," which includes the requirement that the new iteration retains the same intended use as the predicate. She noted that a cleared device can gain clearance for a different indication, but the potential for drift from intended use is a pressing issue among reviewers. "Trying to better understand when a new indication becomes a new intended use is something we definitely need to take a look at," she said.

As for technological differences, Tillman pointed out that threshold is set at the point at which those technological changes raise new types of questions about safety and effectiveness, but that the precise standard for when such changes creates a new device has not always been clear. "There hasn't perhaps been much clarity in the past and we're trying to build a little more clarity going forward" on this issue, she said.

Regarding whether technological drift has driven some of the feedback CDRH has received from outside the agency about cleared devices, Tillman said, "there has been some concern along those lines," adding that since the inception of the program, "more than 100,000 devices have gone to market under that program."

Tillman hinted that managing the "predicate tree" is proving to be a large task. "I think there is a question of how do you run a program that is predicated on this idea of substantial equivalence?" given the enormous base of potential predicates, she said.

Tillman noted that special 510(k)s are not an area of concern. She pointed out that specials "can only be submitted by a manufacturer who is making a modification to their own device" and observed that "I don't see special 510(k)s as being part of the problem" of predicate creep because they are "mostly submitted for fairly incremental design changes." Tillman said she believes that roughly 25%-30% of all 510(k)s arrive through this route.

The abbreviated 510(k), which employs conformity to a reference standard to introduce changes to a device, is "a program we are trying to better define," Tillman said. She noted that most 510(k)s "include a reference to a recognized standard," a fact which she said means that "the abbreviated 510(k) bucket is an odd bucket" because many 510(k) applications arrive with reference standards that are nonetheless not filed as abbreviated applications. "We are trying to get a little more clarity around that" mechanism, she said, describing the issue as largely one of nomenclature.

"The 510(k)s that people are seeing as more troubling are in fact the traditional 510(k)s because those are the 510(k)s where there may be a new indication for use or a shift in technology," Tillman remarked, such as a change of energy source in an ablation device. "We've always found a certain percentage of 510(k)s as not substantially equivalent," she said, and those applications become PMAs. However, she declined to address how CDRH might opt to clarify the line between a PMA and a 510(k).

As for immediate changes, Tillman said, "the one change that I've made right away – I've communicated this to my staff and to the device industry – that we need to make sure that decisions are based on a good legal framework and good science."

"If we have made decisions in the past that in looking back that we have some questions about or maybe if there's new data that makes us realize that ... maybe we would make that decision differently today," CDRH will use that data to inform decisions going forward, Tillman said. She advised that "just because we did it a certain way in the past doesn't mean we're going to continue to do things the same way" if the data suggest a different standard of proof for a device is in order.

As an example, Tillman said, "the agency just recently issued a series of 522 orders about dynamic stabilization systems."

"We had a number of devices – these dynamic stabilization systems for use in the spine – that we had found substantially equivalent to the rigid stabilization systems over the years," Tillman said, "and we realized that these are really fundamentally very different," a perspective fed to some extend by adverse events reports. She said that the agency has come to believe that the bench data from rigid systems "was really not going to be sufficient to characterize the dynamic ones. So we started asking the companies to provide clinical data in their 510(k)s."

"But then we also had this group of devices that we had cleared without any clinical data and I think we had some questions about those, and that was really the impetus for issuing those 522 orders," Tillman said.

She characterized this as "an example of a situation where in the past we had cleared these devices without clinical data, but we had recognized, based on looking at adverse event data, that this was not the best thing to do moving forward, so that isn't what we're going to be doing."

"I think people may see us making more of those kinds of decisions that it's not going to be enough for a company to just come in and say" that lack of clinical data was okay in times gone by. "In fact we are going to be asking" for clinical data more often in the future, she said.

Mark McCarty, 703-268-5690;

mark.mccarty@ahcmedia.com