fbpx Skip to main content

Artificial Intelligence/Machine Learning (AI/ML)-Based Software as a Medical Device (SaMD) Action Plan

Artificial Intelligence/Machine Learning (AI/ML)-Based Software as a Medical Device (SaMD) Action Plan

 

FDA recently released the AI/ML Action Plan. In April 2019, FDA put out a draft which outlined a premarket review for artificial intelligence and machine learning-driven software modifications and requested feedback. This final plan addresses that feedback and aligns closely with the mission of the recently launched Digital Health Center of Excellence.

  1. Tailored Regulatory Framework for AI/ML-based SaMD

The draft plan proposed a framework for modifications to AI/ML-based SaMD that relies on the principle of a “Predetermined Change Control Plan.” The SaMD Pre-Specifications (SPS) describe “what” aspects the manufacturer intends to change, and the Algorithm Change Protocol (ACP) explains “how” the device will change while remaining safe and effective. FDA intends to issue a draft guidance for public comment in this area.

  1. Good Machine Learning Practice (GMLP)

The draft plan used the term Good Machine Learning Practice to describe a set of AI/ML best practices (e.g., data management, feature extraction, training, interpretability, evaluation and documentation) that are similar to good software engineering practices or quality system practices. 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/ML-based technologies. These GMLP efforts will be pursued in close collaboration with the Agency’s Medical Device Cybersecurity Program.

  1. Patient-Centered Approach Incorporating Transparency to Users

FDA acknowledges that AI/ML-based devices have unique considerations that necessitate a proactive patient-centered approach to their development and utilization that takes into account issues including usability, equity, trust, and accountability. The Agency is currently compiling input and aims to hold a public workshop to share learnings and to elicit input from the broader community on how device labeling supports transparency to users.

  1. Regulatory Science Methods Related to Algorithm Bias & Robustness

FDA notes that bias and generalizability are not issues exclusively to AI/ML-based devices, though it is especially important to consider with such devices. FDA is supporting numerous regulatory science research efforts to develop these methods to evaluate AI/ML-based medical software.

  1. Real-World Performance (RWP)

The draft plan described the notion that to fully adopt a total product lifecycle (TPLC) approach to the oversight of these devices, collection and monitoring of real-world data may be used in this process. 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.