Bradley Merrill Thompson Quoted in “Device Experts Say FDA Is Too Heavily Invested in Pre-Cert Program”


Bradley Merrill Thompson, Member of the Firm in the Health Care & Life Sciences practice, in the firm’s Washington, DC, office, was quoted in InsideHealthPolicy, in “Device Experts Say FDA Is Too Heavily Invested in Pre-Cert Program,” by Kelly Lienhard. (Read the full version – subscription required.)

Following is an excerpt:

FDA’s latest report on its software precertification program shows the agency is investing substantial time on the new project, despite not having statutory authority or answering key questions from the industry, device software industry lawyers say. …

Brad Thompson, a member of the firm at Epstein Becker & Green and who started the AI Startups in Health Coalition, thinks FDA is spending too much time on the program.

He told Inside Health Policy that FDA’s large investment in the new software program is concerning because, without authorization from Congress, the effort could end up being a waste of the agency’s time and resources.

FDA is banking on Congress not just authorizing the program, but also authorizing FDA’s specific approach to precertification, according to Thompson.

“Usually the process is the other way around. Congress specifies the new approach in specific legislative language, and then FDA implements it. FDA has no idea whether Congress will authorize a precertification program, and if Congress does, FDA has no idea what parameters Congress might put around the design of that precertification program,” Thompson wrote in an email to IHP. …

FDA could be wasting time on a program that not only lacks congressional signup but is also unacceptable to industry stakeholders, Thompson said.

The pre-cert program aims to evaluate developers first, rather than the product, which has some digital health developers concerned that the process will become more about “who you are” rather than what a company can do to develop a device. Developers want assurance from the agency that the program will not subjectively award precertification to companies that are “friends of the FDA,” according to Thompson.

The software-based device industry is also calling for FDA to give specific details about how the pre-cert program would help companies.

Specifically, Thompson asserts that FDA has said it will save time, but developers want to know where in the process time is saved.

“I have honestly never seen any suggestion that this program will actually save time in the first instance,” Thompson said. “FDA seems to be arguing that it will save time in the long run. I am always suspicious of that kind of argument from a federal agency." …

“[FDA needs to] spend more time working through the issues that may actually determine whether the program fails. It is not the logistics that will likely cause it to fail. There are potentially major -- big picture -- design decisions that have to be made that balance FDA's regulatory interests with the interests of patients/provider and the industry,” Thompson said.

The pre-cert program also could pose additional challenges for AI software, Thompson said. The program update does not clarify whether the data FDA wants to collect and potentially share with the public for the program would include data that AI firms consider proprietary. …

That’s time that could be better invested in other reforms that have more industry support, both Thompson and Mickelson said.

Thompson pointed to the artificial intelligence/machine learning framework from 2019 that has generated substantial industry excitement, especially within AI startups.

Under that framework, AI-based devices would follow pre-specified performance objectives and use a validation process that ensures improvements to performance, safety and effectiveness of the software. It would also allow for device performance to be monitored in the real world once the device is on the market, which could allow AI-based medical devices to update autonomously while still letting FDA provide a reasonable assurance of safety and effectiveness.