The AI/ML FDA Plan

The “Proposed Framework”?

  • A change that introduces a new risk or modifies an existing risk that could result in significant harm
  • A change to risk controls to prevent significant harm
  • A change that significantly affects clinical functionality or performance specifications of the device
  1. Performance (e.g. retraining, change in AI architecture…)
  2. Inputs used by the algorithm (e.g. adding new input data types and sources)
  3. Intended use (e.g. from a confidence score that is ‘an aid in diagnosis’ (drive clinical management) to a ‘definitive diagnosis’ (diagnose) or inclusion of pediatric population where the SaMD was initially intended for adults ages 18 years or older)
  1. Quality Systems and Good Machine Learning Practices (GMLP)
  • SaMD Pre-Specifications (SPS): manufacturer’s anticipated modifications to “performance”, “inputs,” or “intended use”
  • Algorithm Change Protocol (ACP): Specific methods in place to achieve and control the risks of the anticipated modifications in the SPS. See table below for more details on what is expected in the ACP.
  • Intensive Care Unit (ICU) SaMD
  • Skin Lesion Mobile Medical App (MMA)
  • X-Ray Feeding Tube Misplacement SaMD

Feedback

Pilot: AI-guided Cardiac Ultrasound

Digital Health Center of Excellence

  • Connect and build partnerships
  • Share knowledge
  • Innovate regulatory approaches

The Action Plan

  1. Tailored Regulatory Framework for AI/ML-based SaMD
  • What they heard: Many suggestions including re:Predetermined Change Control Plan
  • What they’ll do: Update the proposed framework, including issuance of Draft Guidance on the Predetermined Change Control Plan for public comment
  • What they heard: Strong general support for the GMLP + a call for FDA to encourage harmonization through consensus standards efforts.
  • What they’ll do: Encourage harmonization of Good Machine Learning Practice development through institutions such as IEEE and AAMI
  • What they heard: Call for further discussion on how AI/ML-based technologies interact with people, including their transparency to users and to patients
  • What they’ll do: Following up on recent Patient Engagement Advisory Committee meeting, next step: hold public workshop on how to support transparency to users and enhance trust in AI/ML-based devices
  • What they heard: : Stakeholders described the need for improved methods to evaluate and address algorithmic bias and to promote algorithm robustness.
  • What they’ll do: Support regulatory science efforts to develop methodology for the evaluation and improvement of algorithms, including for the identification and elimination of bias, and promotion of algorithm robustness (e.g. Centers for Excellence in Regulatory Science and Innovation (CERSIs) at UCSF, Stanford, and Johns Hopkins University)
  • What they heard: Stakeholders described the need for clarity on Real-World Performance (RWP) monitoring for AI/ML software
  • What they’ll do: Work with stakeholders who are piloting the RWP process for AI/ML-based SaMD
  • Update the proposed regulatory framework in the AI/ML-based SaMD discussion paper, including issuance of a Draft Guidance on the Predetermined Change Control Plan.
  • Harmonize development of GMLP through additional FDA participation in collaborative communities & standards development efforts.
  • Continue to host discussions on the role of transparency to users, including holding a public workshop on medical device labeling to support transparency
  • Support regulatory efforts on methodology for evaluation & improvement of algorithms, including for the identification & elimination of bias, & robustness and resilience of algorithms.
  • Advance real-world performance pilots in coordination with stakeholders and other FDA programs, to provide additional clarity.

A last-minute hurdle?

--

--

--

Cofounder/CTO at Curai (AI for healthcare). Former Quora VP, Netflix Director. Software, Machine Learning, Data, Recsys... From Barcelona, in the Valley

Love podcasts or audiobooks? Learn on the go with our new app.

Recommended from Medium

Risks of Artificial Intelligence, its impacts and why you should care.

OVERVIEW ON ROBOTICS – THE MAIN FUTURISTIC FORM OF ARTIFICIAL INTELLIGENCE.

Patient Perceptions Of AI-Driven Systems

WeBlock AMA Session With CONNECTOME: Where is the margin of imagination:when artificial…

Read These 4 Books to Get into AI — Especially If You Can’t Code

How Artificial Intelligence can change medicine forever if implemented strategically

Doctor working with futuristic technology

How AI Can Help Fashion Retailers Fight Amazon

AI and Ethics

Get the Medium app

A button that says 'Download on the App Store', and if clicked it will lead you to the iOS App store
A button that says 'Get it on, Google Play', and if clicked it will lead you to the Google Play store
Xavier Amatriain

Xavier Amatriain

Cofounder/CTO at Curai (AI for healthcare). Former Quora VP, Netflix Director. Software, Machine Learning, Data, Recsys... From Barcelona, in the Valley

More from Medium

AI and synthetic data generation

AI can write better than humans now?

To Build Trust, Answer The Why Question — IDEATE + EXECUTE

The IOT: Where We Are, And Where We’re Heading Next