FDA Releases Action Plan for AI, Machine Learning Medical Software

The FDA has released its first artificial intelligence and machine learning action plan, a multi-step approach designed to advance the agency’s management of advanced medical software.

The Artificial Intelligence/Machine Learning (AI/ML)-Based Software as a Medical Device (SaMD) Action Plan is a response to stakeholder feedback on the FDA’s 2019 regulatory framework for AI and ML-based medical products.

“This action plan outlines the FDA’s next steps towards furthering oversight for AI/ML-based SaMD,” said Bakul Patel, director of the Digital Health Center of Excellence in the Center for Devices and Radiological Health (CDRH).

“The plan outlines a holistic approach based on total product lifecycle oversight to further the enormous potential that these technologies have to improve patient care while delivering safe and effective software functionality that improves the quality of care that patients receive. To stay current and address patient safety and improve access to these promising technologies, we anticipate that this action plan will continue to evolve over time.”

In the new action plan, the FDA noted that it will aim to further develop the proposed regulatory framework for AI/ML-based SaMD. This step will include issuance of draft guidance on a predetermined change control plan, which pertains to software that learns over time.

The agency plans to publish this draft guidance in 2021. Other areas of development will include refinement of the identification of types of modifications appropriate under the framework, as well as specifics on the focused review, such as the process for submission/review and the content of a submission.

The agency will also aim to support the development of good machine learning practices to evaluate and improve machine learning algorithms. The FDA noted that the development and adoption of AI/ML best practices is important not only to guide product design, but also to facilitate the oversight of these advanced devices.

The FDA stated that it has actively participated in several efforts related to the development of AI/ML best practices, including standardization projects and collaborative communities.

“As part of this Action Plan, FDA is committing to deepening its work in these communities in order to encourage consensus outcomes that will be most useful for the development and oversight of AI/M-based technologies,” the agency said.

“In keeping with FDA’s longstanding commitment to a robust approach to cybersecurity for medical devices, these good machine learning practice efforts will be pursued in close collaboration with the Agency’s Medical Device Cybersecurity Program.”

The agency also stated that it would foster a patient-centered approach to AI/ML-based SaMD, including device transparency to users.

The FDA noted that transparency is especially important for AI and ML devices, which may learn and change over time and incorporate algorithms that exhibit a degree of opacity.

To ensure transparency in AI and ML medical device software, the FDA held a Patient Engagement Advisory Committee (PEAC) meeting in October 2020. Patients offered input on what factors impact their trust in these technologies.

“We intend to consider this input for identifying types of information that FDA would recommend a manufacturer include in the labeling of AI/ML-based medical devices to support transparency to users,” the FDA said.

“These activities to support the transparency of and trust in AI/ML-based technologies will be informed by FDA’s participation in community efforts, such as standards development and patient-focused programs. They will be part of a broader effort to promote a patient-centered approach to AI/ML-based technologies based on transparency to users.”

Additionally, the agency plans to develop methods to evaluate and improve machine learning algorithms. The FDA will focus on issues of bias and generalizability in AI and ML algorithms, developing methods that identify and eliminate bias as well as evaluating the ability for AI to withstand changing inputs and conditions.

Finally, the action plan outlines the FDA’s goal of advancing real-world performance monitoring pilots. Gathering performance data on the real-world use of SaMD could allow manufacturers to understand how their products are being used, identify opportunities for improvement, and respond proactively to any concerns.

“As part of this Action Plan, the Agency will support the piloting of real-world performance monitoring by working with stakeholders on a voluntary basis. This will be accomplished in coordination with other ongoing FDA programs focused on the use of real-world data,” the FDA said.

The agency expects that this new action plan, in conjunction with feedback from industry leaders, will help advance the use of AI and machine learning in healthcare.

“The AI/ML Software as a Medical Device Action Plan described in this document was developed in direct response to stakeholder feedback and is intended as a multipronged approach to further advance the Agency’s oversight of these technologies. Continued stakeholder engagement is crucial for the success of this work,” the FDA concluded.