Guiding Principles for Predetermined Change Control Plans in ML-Enabled Medical Devices

1. Background

In 2021, the U.S. Food and Drug Administration (FDA), Health Canada, and the U.K.’s Medicines and Healthcare products Regulatory Agency (MHRA) jointly identified 10 guiding principles that can inform the development of Good Machine Learning Practice (GMLP). https://www.gov.uk/government/publications/predetermined-change-control-plans-for-machine-learning-enabled-medical-devices-guiding-principles/predetermined-change-control-plans-for-machine-learning-enabled-medical-devices-guiding-principles GMLP supports the development of safe, effective, and high-quality artificial intelligence/machine learning technologies that can learn from real-world use and, in some cases, improve device performance. These principles are detailed at Good Machine Learning Practice for Medical Device Development: Guiding Principles. https://www.fda.gov/medical-devices/software-medical-device-samd/good-machine-learning-practice-medical-device-development-guiding-principles Advancements in digital health technologies include artificial intelligence/machine learning-enabled medical devices (MLMD), with key terms and definitions available at Machine Learning-Enabled Medical Devices: Key Terms and Definitions. https://www.imdrf.org/documents/machine-learning-enabled-medical-devices-key-terms-and-definitions Regulatory expectations aligned with best practices for development and change management, such as those in the GMLP Guiding Principles, can support the quality of such devices, leading to patient benefits like earlier access to innovative technologies or more accurate diagnoses. https://www.gov.uk/government/publications/predetermined-change-control-plans-for-machine-learning-enabled-medical-devices-guiding-principles/predetermined-change-control-plans-for-machine-learning-enabled-medical-devices-guiding-principles The change management process ensures ongoing safety and effectiveness of devices throughout the total product lifecycle (TPLC). However, changes to MLMDs, such as modifications to a model or algorithm, may be substantive and require regulatory oversight, including additional premarket review, which may not align with the rapid pace of MLMD development. https://www.gov.uk/government/publications/predetermined-change-control-plans-for-machine-learning-enabled-medical-devices-guiding-principles/predetermined-change-control-plans-for-machine-learning-enabled-medical-devices-guiding-principles Internationally, the medical device community is discussing predetermined change control plans (PCCPs) to manage device changes requiring regulatory authorization. PCCPs help align regulatory processes with rapid change management in MLMDs, manage risks timely through monitoring, maintenance, and performance improvement, and uphold high standards for device safety and effectiveness. https://www.gov.uk/government/publications/predetermined-change-control-plans-for-machine-learning-enabled-medical-devices-guiding-principles/predetermined-change-control-plans-for-machine-learning-enabled-medical-devices-guiding-principles For this document, a PCCP is a plan proposed by a manufacturer that specifies certain planned modifications to a device, the protocol for implementing and controlling those modifications, and the assessment of impacts from modifications. PCCPs may vary by regulatory jurisdiction. https://www.gov.uk/government/publications/predetermined-change-control-plans-for-machine-learning-enabled-medical-devices-guiding-principles/predetermined-change-control-plans-for-machine-learning-enabled-medical-devices-guiding-principles The objectives of the 5 Guiding Principles for PCCPs for MLMD are to provide foundational considerations for robust PCCPs and facilitate stakeholder engagement and collaboration. This document lays a foundation for PCCPs and encourages international harmonization to support responsible innovations in digital health. https://www.gov.uk/government/publications/predetermined-change-control-plans-for-machine-learning-enabled-medical-devices-guiding-principles/predetermined-change-control-plans-for-machine-learning-enabled-medical-devices-guiding-principles Feedback is welcomed via the FDA public docket at FDA-2019-N-1185. https://www.regulations.gov/docket/FDA-2019-N-1185 Contact the Digital Health Center of Excellence (FDA) at Digitalhealth@fda.hhs.gov, the Medical Devices Directorate Digital Health Division (Health Canada) at mddpolicypolitiquesdim@hc-sc.gc.ca, or the software and AI team (MHRA) at software@mhra.gov.uk. https://www.gov.uk/government/publications/predetermined-change-control-plans-for-machine-learning-enabled-medical-devices-guiding-principles/predetermined-change-control-plans-for-machine-learning-enabled-medical-devices-guiding-principles

2. Guiding Principles

2.1 Focused and Bounded

A PCCP describes specific changes a manufacturer intends to implement, limited to modifications within the intended use or purpose of the original MLMD. This includes:

• The extent of planned changes and scope of the MLMD with changes implemented.
• Plans to safely modify the device within PCCP bounds, including methods for verifying and validating changes, and mechanisms to detect and revert or stop implementation if changes fail performance criteria.
• Impacts of planned changes. https://www.gov.uk/government/publications/predetermined-change-control-plans-for-machine-learning-enabled-medical-devices-guiding-principles/predetermined-change-control-plans-for-machine-learning-enabled-medical-devices-guiding-principles

2.2 Risk Based

The value and reliability of a PCCP are strengthened by a risk-based approach adhering to risk management principles, relevant throughout the TPLC from inception through implementation and use. This ensures individual and cumulative changes remain appropriate over time for the device and its use environment. https://www.gov.uk/government/publications/predetermined-change-control-plans-for-machine-learning-enabled-medical-devices-guiding-principles/predetermined-change-control-plans-for-machine-learning-enabled-medical-devices-guiding-principles

2.3 Evidence Based

Evidence generated throughout the TPLC ensures ongoing safety and effectiveness of the device with a PCCP, demonstrates benefits outweigh risks, and establishes risks are managed. Considerations include:

• Methods and metrics to measure device performance that are scientifically and clinically justified, proportional to risk, and consistent with other TPLC evidence.
• Specified methods to generate evidence demonstrating benefits and risks before and after PCCP changes. https://www.gov.uk/government/publications/predetermined-change-control-plans-for-machine-learning-enabled-medical-devices-guiding-principles/predetermined-change-control-plans-for-machine-learning-enabled-medical-devices-guiding-principles

2.4 Transparent

Best practice for PCCPs involves providing clear information and detailed plans for ongoing transparency to users and stakeholders, ensuring awareness of device performance and use before and after changes. Examples include:

• Characterization of data used in development and modifications, reflecting the intended population.
• Comprehensive testing for planned changes.
• Characterization of the device before and after changes.
• Monitoring, detection, and response to deviations in performance. https://www.gov.uk/government/publications/predetermined-change-control-plans-for-machine-learning-enabled-medical-devices-guiding-principles/predetermined-change-control-plans-for-machine-learning-enabled-medical-devices-guiding-principles

2.5 Total Product Lifecycle (TPLC) Perspective

Creating and using a PCCP from a TPLC perspective elevates quality by considering stakeholder perspectives and risk management throughout the TPLC. It supports existing regulatory, quality, and risk measures to ensure safety through monitoring, reporting, and response to concerns. https://www.gov.uk/government/publications/predetermined-change-control-plans-for-machine-learning-enabled-medical-devices-guiding-principles/predetermined-change-control-plans-for-machine-learning-enabled-medical-devices-guiding-principles