[Editor’s Note 1: This is the first of a two-part Essenvia Insights story on the interplay between MedTech regulation and medical device innovation. Read Part 2 here.]
[Editor’s Note 2: Three regulatory experts interviewed for this two-part series – Carrie Kuehn, Steve Silverman, and Jay Vaishnav – will be panelists at a July 31 Essenvia roundtable webinar, “Shaping the Future of MedTech: Insights at the Crossroads of Regulation and Innovation.” Register for this free event now!]
Dr. Jekyll and Mr. Hyde. Superman and Lex Luthor. The Road Runner and Wile E. Coyote. All are yins to their yangs, so-called “good” versus so-called “evil” – but necessary evil, nonetheless. After all, the coyote just wants to have a delicious meal, much to our amusement; Luthor once saved the planet and billions of lives as a comic book US president; and Hyde has such incredible strength that he can tear through furniture, surely an asset to someone, somewhere.
Similarly, anyone with a nose can catch the familiar whiff of frenemy when inhaling a deep metaphorical breath of medical device innovation (yin) and MedTech regulation (yang). Sure, manufacturers understand at their core that rules are in place to ensure that only safe and effective products reach market. Does that mean companies are particularly happy if they run into regulatory hurdles, potentially stifling or delaying the development of devices that patients could be putting to good use today (and losing a potential boatload of money to boot)? Absolutely not.
But innovation and regulation in the medical devices industry represent an essential and never-ending balancing act and dance that must play out. But do MedTech manufacturers generally see regulations as obstacles to innovation?
“I don’t believe the right metaphor is ‘obstacles.’ I believe a better metaphor is ‘checkpoints,’” said Steve Silverman, a former director of the Office of Compliance within the US Food and Drug Administration’s (FDA) Center for Devices and Radiological Health (CDRH).
“That is to say, at the end of the day, the responsibility of the FDA and other regulators is to make sure that there is this push for innovation, which is great and critical for patient care, but by the same token, if innovation occurs without any kind of oversight, there’s a risk that parts will enter the market that aren’t safe and aren’t effective for patients, and that’s not a good result, either,” Silverman told Essenvia Insights in an interview. “So there has to be some kind of give and take that allows for some basic elements of safety and effectiveness to be to be maintained.”
Silverman, who’s now head of The Silverman Group consulting firm, pointed out that what he calls regulatory checkpoints aren’t developed to give any particular manufacturer a leg-up (or -down), but at the same time, regulators are sensitive to making sure innovative products indeed have a path to market.
For example, Silverman said the FDA “isn’t going to institute a policy that works to the benefit of a specific company. At the same time, individual companies can and do come to the FDA and say, ‘We understand what your expectations are, we understand what the what the basic rules are, where the guardrails are, and we think we have a good way of operating that allows us to be efficient while meeting these basic requirements and staying within the guardrails.’”
He further explained: “When this happens, the FDA definitely has this conversation with other companies and often says, ‘Yeah, fine, cool. Go for it.’ Or it says, ‘Well, what you want to do is not exactly right. But you know, if you made some changes, it would be OK.’ So those kinds of those kinds of company-specific conversations definitely happen.”
Silverman noted that while he “stands by” his comments, he nevertheless recognizes that innovators believe regulators can have an adverse effect on the pace of innovation. “There’s no question in my mind that if you talk to innovators that they will tell you that regulatory requirements that exist make it difficult for them to stay in business. But what I have seen from the FDA is a recognition of the need to be creative and collaborative and work alongside industry to try to solve some of these problems,” he said.
“I haven’t seen with the FDA – in the in the digital space, for example – the kind of stereotypical regulator model of, ‘No, you can’t do anything because we don’t understand it and we’re only willing to allow things that we I understand.’ I just I don’t see that. And I haven’t heard device companies credibly making that argument – in fact, I haven’t heard device makers trying to make that argument and I don’t think they credibly could.”
Using Change Control Plans to Balance Regulation, Innovation
Silverman points to the FDA’s 2020 launch of the Digital Health Center of Excellence (DHCoE) as a step toward finding ground where regulatory checkpoints are met but innovation is not hampered.
“I believe the primary – if not exclusive – motivation for the agency developing its DHCoE was to create a thought space that can take on questions like this, which is, how can the FDA adapt its regulations to allow for innovation while still maintaining basic regulatory benchmarks like safety and effectiveness,” he said. “So I wouldn’t characterize the atmosphere as a conflict between the FDA and industry, or the FDA serving as an anchor around the necks of industry innovators. A better way of characterizing it is as a push by the agency and stakeholders to find ways for sensible regulatory frameworks that provide basic services to consumers and to healthcare providers, while also both removing obstacles that unnecessarily hamper innovation and actually promote innovations.”
Predetermined Change Control Plans (PCCP), pushed last year by the FDA and fellow regulators Health Canada and the UK’s Medicines and Healthcare products Regulatory Agency (MHRA), is an example of how the DHCoE encourages innovation while walking the regulatory tightrope, Silverman said. PCCPs, sent to regulators along with marketing submissions, allow manufacturers to set modification limits for machine learning-enabled device software functions (ML-DSF) so companies won’t have to submit new applications should predetermined modifications occur and stay within preset boundaries.
“A PCCP allows the FDA to say, ‘For certain kinds of digital products where we know that there is going to be innovation, we will work with device firms in advance so when they – for example – get a product approved, as part of their approval, they will detail for us certain types of changes that they expect to make to the product over time,’” Silverman explained. “A PCCP also lets the FDA say, ‘We will essentially acknowledge that the company is going to do this and create a pathway so they can make those changes as expected over time. But they don’t have to come back to us for approval in advance if they’re operating according to the playbook that [the agency] originally reviewed and blessed.’”
The PCCP concept is solid, Silverman said, because innovation can happen so quickly that regulators don’t want to keep ever-evolving products and medical technology out of the hands of caregivers, users, and patients.
“If manufacturers have to get regulatory approval every time they want to innovate a product, then it’s going to cause problems for them. The FDA recognized that and created PCCP,” he said. “Sure, it’s imperfect. There are a lot of questions. It’s not widely used right now. It’s not widely understood. It has to be improved over time. But it’s still a good idea, and it will be improved over time. And then the FDA will be able to leverage it over time.”
Silverman went on: “But PCCP doesn’t solve all the problems that device makers can have. There are other questions that the FDA hasn’t answered and that device makers haven’t answered. An example of that is generative artificial intelligence (GenAI). Nobody knows how GenAI works. Nobody understands the actual technological process by which this technology can learn on its own and get better over time. The FDA doesn’t want to prevent GenAI. But by the same token, how do you regulate something and understand it when nobody understands it? So the FDA and device makers are still working with and struggling with that question, and that’s just an ongoing push.”
In the very simplest of terms, GenAI creates new content and data by analyzing currently available datasets. It can be used to support a variety of somewhat mundane regulatory activities, including creating new product submissions, identifying production and quality trends, pinpointing device safety risks, developing adverse event reports, and drafting responses to regulators.
In this month’s voiCEOver blog entry, Essenvia CEO Soumya Mahapatra warned that “MedTech companies must fold GenAI into RA departments or risk being left behind and losing their competitive edge in the global marketplace. …Manufacturers that streamline and modernize regulatory operations by using GenAI can ensure better efficiency and process harmonization across the lifecycle of the products they make.”
GenAI is Quite Regulatable, Experts Say
But is GenAI ultimately regulatable? Former FDA compliance director Silverman says yes.
“There are lots of parts of GenAI that are absolutely regulatable today,” Silverman said. “Let’s say ACME device company uses generative AI to power its diagnostic device, and there are going to be parts of this diagnostic device that are genuinely confusing and that need special treatment over time. That is absolutely true. But at the end of the day, the FDA can still walk into ACME’s factory and say, ‘OK, show me how you’re dealing with suppliers, show me how you’re dealing with complaints. Do you have a robust complaint handling process? Are you meeting adverse event reporting requirements? Are you properly packaging this product?’
“So there are lots of elements that the FDA can evaluate using its current methodology that provides assurances of safety and quality for device makers that don’t slow down innovation that the agency can deploy successfully today,” he added. “So it’s not the case that, because there is this critical component of these devices, the FDA can’t do anything or just says no, until it’s hot enough that it can do something. It’s just about recognizing that there are these critical new technological innovations that need to be wrestled with over time, and in the meantime, there are parts of these devices that are capable of being evaluated using the agency’s traditional techniques.”
Jay Vaishnav, a longtime MedTech regulatory expert, a fellow of the Regulatory Affairs Professionals Society (RAPS), and coeditor of the book “From X-Rays to AI: Navigating US Regulations in Radiological Health,” agrees with Silverman, pointing out that “nothing is unregulatable.”
“Sure, regulating GenAI will be difficult,” she told Essenvia Insights. “But the FDA has made long strides in the last few years in its regulation of software and AI. That said, the whole paradigm that the agency uses wasn’t built with GenAI in mind because it didn’t exist at the time. So it’s definitely been challenging.”
Clinical Trials and Regulators
During her career in the MedTech industry, Vaishnav has seen projects scrapped because the innovation juice wasn’t worth the regulatory squeeze. “Even just the idea sometimes doesn’t get off the table,” she said. “Someone will say, ‘There's this idea we have for a medical device … but wait, you know, it uses a drug, or it uses AI, or there’s no predicate, and the FDA. …’ and the idea stops then and there.”
Vaishnav, who is also a seven-year veteran of the FDA, says requirements around clinical trials can be the kiss of death for some products, particularly for smaller manufacturers.
“Any time a device needs a big clinical trial to make it to market, that’s a substantial burden on a company. When you switch from just requiring bench data, performance data, or retrospective testing – anything that requires a prospective clinical trial is difficult and expensive,” she said. “I wouldn’t tie that burden to a specific regulation, but some devices raise FDA eyebrows more than others, and prompt them to request large amounts of clinical data. Most Class II devices don't require clinical data. But for the minority that do, it’s much more challenging to get them to market.”
AND THE CONVERSATION CONTINUES!
Mark your calendar now for a July 31 Essenvia roundtable webinar, “Shaping the Future of MedTech: Insights at the Crossroads of Regulation and Innovation.” This can’t-miss panel discussion with regulatory experts Carrie Kuehn and Jay Vaishnav, former FDA medical device compliance office director Steve Silverman, and Essenvia regulatory affairs VP Dhriti Roy will dissect and expand on points made in this two-part Insights story. Register for this free event now!
