Connect with us

International Circuit

FDA, Health Canada, MHRA release additional guiding principles for MLMDs

The US Food and Drug Administration (FDA), Health Canada, and the UK Medicines and Healthcare products Regulatory Agency (MHRA) have released additional guiding principles for machine learning-enabled medical devices (MLMDs) that focus on transparency of information for stakeholders that interact with the device.

“The underlying objectives of these guiding principles are to foster international harmonization and to underscore the importance of considering transparency throughout the life cycle of MLMDs,” Troy Tazbaz, director of the Digital Health Center of Excellence at FDA’s Center for Devices and Radiological Health, stated in a news release.

Transparency in the guiding principles refers to information that contains context of use for the device’s end-user in addition to communication strategies and mediums. The guiding principles build on a separate document that was created and jointly published by FDA, Health Canada, and MHRA in 2021 that focus on developing Good Machine Learning Practice for MLMDs.

“This information holds the potential to influence the trust of healthcare professionals and patients toward a medical device and inform decisions regarding its use,” Tazbaz said. “The comprehensive integration of these guiding principles of transparency across the entirety of the product life cycle serves to ensure that informational requirements are adequately addressed, thereby promoting the safe and effective utilization of MLMDs.”

In a document outlining the guiding principles, the agencies note that transparency should include explanations of the relevant audience for the device, motivation for using the device, relevant information, where information will be placed, timing of communication, and the use of human-centered design principles.

The authors of the guiding principles said descriptions of who is using a device should mention everyone that cares for a patient, and particularly individuals using the device for health care purposes and who impact patient outcomes surrounding use of the device.

There is a wide range of relevant information that can be shared across MLMDs, and what is appropriate for a device will depend on its type and intended users. When describing a MLMD, good practices include providing information about the intended use and intended users, the purpose and function of the device, and what diseases or conditions it is meant to treat. Describing how the device functions in a healthcare workflow, such as how it impacts decisions or judgments from healthcare professionals, is also good practice. The lifetime benefits and risks of the device should also be described as well as its clinical limitations, such as gaps in information or contraindications.

How device information is accessed is also important, according to the document. “A good practice is to optimize use of the software user interface so that the information it conveys is responsive to the user. This software user interface can allow information to be personalized, adaptive and reciprocal,” it notes.

When communication occurs matters, and information needs may change at various stages of a device’s total product lifecycle, according to the document. Other considerations include use of notifications for device updates, and targeted information as needed during a workflow. “Considering the information needs throughout each stage of the total product lifecycle can support successful transparency,” it notes. “Detailed device information may be needed when considering whether to acquire or implement a device, and whether and how to use it.”

Being transparent in why a device is being used can impact the safety and effectiveness of an MLMD, and can also help identify errors and biases in systems and outputs. “The transparent and consistent presentation of information, including known gaps in information, can have many benefits. It builds fluency and efficiency in the use of MLMDs,” the authors said. “It can also foster trust and confidence in the technology and encourages the adoption and access to beneficial technologies.”

Describing how an MLMD is used “requires a holistic understanding of users, environments and workflows,” and transparency in this area can be aided by human-centered design principles that consider responsive and iterative design, validation, monitoring and communications, the authors said.

“As outlined in the guiding principles document, a comprehensive understanding of users, environments, and workflow is paramount in addressing transparency for MLMDs,” Tazbaz said in the news release. “Employing human-centered design methods can provide an approach for developing MLMDs with a high degree of transparency.”

FDA noted it is continuing to solicit feedback from the public on its proposed regulatory framework for modifications to medical devices that use artificial intelligence and machine learning. RAPS.org

Copyright © 2024 Medical Buyer

error: Content is protected !!