Experts offer advice on preparing PCCPs

Developing pre-defined change control plans (PCCPS) for devices that leverage artificial intelligence and machine learning (AI/ML) requires significant effort and attention to detail, and companies should not underestimate the time needed to create these plans, according to legal experts who spoke at the Food and Drug Law Institute’s (FDLI) annual meeting on 15 May.

They also noted that the number of applications with AI and ML elements have “dramatically increased” in recent years.

In December 2024, the FDA issued final guidance on PCCPs for AI/ML software products. The guidance recommends how sponsors can proactively plan product updates without filing new premarket applications for certain product changes. (RELATED: Final FDA guidance on PCCP includes clarification on version control, Regulatory Focus 3 December 2024).

Cassie Scherer, senior director for global digital health policy at Medtronic, stated, “We at Medtronic are huge fans of PCCPs.” She noted that the company currently has three approved PCCPs in place.

Preparing PCCPs is quite labor-intensive; developing them requires significant time and effort, she said.

Scherer also advised sponsors to “get in front” of their submissions and to notify FDA if they are planning a presubmission meeting involving a PCCP.

Marta Villarraga, a principal with Exponent concurred. “It is complicated, and it takes a team to put things together on the sponsor's side for the PCCPs,” she said.

Ariel Seeley an attorney with Morgan Lewis, said, “If you do feel like you’re running into an issue that a review is not going the way you expected make sure you look who is on your team, and look at the FDA digital health center of excellence” for additional information or questions.

In 2023, the FDA approved approximately 220 AI/ML devices. While the number of devices approved in 2024 is not yet final, it stands at about 170 so far. In contrast, only three devices with AI features were approved by the FDA in 2012, according to the agency’s database.

Most approved AI/ML devices were in the radiology areas, followed by cardiovascular and neurological devices.

The number of PCCPs in approved AI/ML devices has also increased significantly, said Villarraga. In 2024, there were 13 AI/ML decisions with PCCPs in the fields of radiology, cardiovascular medicine, endocrinology, and neurology. In contrast, there were only two devices with PCCPs in 2020.

Experts offer advice on preparing PCCPs

Related topics

Recommended content

Welcome to the new RAPS Digital Experience