FDA seeks new regulatory science framework for innovative tech

The FDA is taking stock of the regulatory science tools it will require for the next decade, as it looks to not just keep pace, but shepherd medical device innovation through a time of rapid development and the weaving of medical technologies into everyday life.

At a meeting at its headquarters co-sponsored by the Medical Device Innovation Consortium, agency leaders set the stage for the road ahead, calling it the start of a conversation that will help set the FDA’s agenda as it works to build out its capabilities.

“This year is the 50th anniversary of the Medical Device Amendments of 1976, and the technologies of today and those of the future weren’t contemplated 50 years ago,” said CDRH Director Michelle Tarver. “But our rubric, our foundation, still persists.”

“The technologies that we are seeing today — like autonomous AI systems, next-generation sequencing, bioelectronic medicine, next-generation implantables — all challenge us to consider new evidence generation pathways, novel measurement tools and creative models,” Tarver said. “The boundaries of medicine are expanding, and that challenges us to expand our scientific frameworks as well.”

Previously: Vision check with FDA CDRH Director Dr. Michelle Tarver

To start, the agency needs to tackle three assumptions about medical devices that won’t hold true in the coming years, said Kolaleh Eskandanian, a senior research scientist at MedStar Georgetown’s AI CoLab.

“First is that a device is fixed at authorization, and that with every change you have to submit a full FDA regulatory package,” said Eskandanian, who formerly served as chief innovation officer at Children’s National and as executive director of its Sheikh Zayed Institute for Pediatric Surgical Innovation.

“The second is the assumption that only a clinician should operate and/or monitor that device,” she said. “And the third is that only in a hospital environment can that device be operated.”

Additionally, the industry needs to pursue offerings capable of evolving alongside their patients, said Eskandanian, especially for children as they grow.

“We have to make sure that we have a good plan in place to come up with methodologies for growth-accommodating types of devices, so children don’t have to go through surgery throughout their life,” she said. “We have the example of predetermined change control plans for software as a medical device. Why can we have something similar for age bands?”

Previously: Kolaleh Eskandanian on why device developers shouldn’t pass over pediatrics

In the meantime, people’s conceptions of the industry will change as well.

“From the patient perspective, we’re at a place where when you say medical device, most patients think of a singular thing,” said Leslie Ritter, senior VP for advocacy and access at the National MS Society. “And if you think 10, 15, 20 years from now, it’s going to increasingly be an ecosystem working together, both for use clinically and for research.”

And while the healthcare system has largely relied on infrequent doctor visits to gauge where a patient is in the progression of their disease, the FDA and clinicians will need to have a framework for digesting the data gathered from wearables and daily monitoring, said Ritter, as well as make it clear what patients can expect when that type of device carries a regulatory green light from the agency.

But despite the promise of new technologies, near-term obstacles will complicate and slow their development, said Josh Makower, co-founder and director of the Stanford Mussallem Center for Biodesign.

“I think the reality of the medical device world at the moment is exceptionally challenging,” said Makower. “We’re in an environment where the availability of capital, especially early-stage capital, is at a nadir that I don’t think I’ve seen in my career.”

“Most of the capital is shifting to later-stage technologies, because it takes so long for returns on the investments,” he added. “There are multiple factors, but a piece of this is the extremely large sucking sound produced by AI, which is sucking a tremendous amount of capital out of the capital market to AI itself.”

Previously: Josh Makower warns of the ‘biggest challenge’ to medtech innovation: ‘It is dire.’

That is drawing funding away from even successful medical device companies, while depressing valuations and making companies that would otherwise go public stay private longer, Makower said. He highlighted FDA and CMS’s RAPID proposal as a necessary step forward in accelerating reimbursements for innovative tech.

When it comes to generative AI programs that can learn and update themselves, they may not fit neatly within the current FDA frameworks, said Makower, even its predetermined change control plans that outline potential future changes from the start.

“At the moment, I probably believe that it makes sense for the FDA to have some role in authorizing generative AI — let’s say agents or bots, or what have you — that are performing functions and evolving,” he said. “It’s hard to imagine any other entity placing some degree of value, or a halo of acceptability, other than the FDA.”

In August, the FDA published a discussion paper on approaches for regulating generative AI programs, including a “competency-based evaluation” akin to how physicians are credentialed after going through medical school.

“There’s no other entity that has that respect and can bring the communities together like FDA,” said Ritter, “But I think we’re also asking FDA to lead in an area where they haven’t necessarily led before, because of the convergence of these technologies all at one time.”

“I think that’s a really critical point to think about, in the future of this whole area, which is that all stakeholders need to have trust in FDA’s decisions because they’re confident that FDA had the right tools and the expertise to make those judgments in the first place,” said MDIC President and CEO Andrew Fish.

The FDA is taking public comments and feedback on the meeting’s discussions and the agency’s future regulatory science agenda, with a docket open through Nov. 24.